Hey John, good subject matter here. This is something I have often wanted to test myself. One thing I feel may be causing error is the way the part was being held and indicated in the 4 jaw. If you only indicated one spot, you may not being running absolute concentric. If so once you stick a tool in there and read a different position you will get runout. Getting parts indicated to within tenths in a 4 jaw is tricky, but I like how your not scared to give it a try. I see in a comment below your going to give it another try. 👍🏻👍🏻
I was also wondering why after he indicated at the Chuck and switched to the end of the holder why he wasn't bumping it around with a mallet and then going back to the Chuck and checking run-out there. He instead went to the end and started adjusting the chuck again. I usually run two at the same time indicators one at the Chuck and one on the end. Tightening the chuck fucks up the end Everytime. I do the same thing in the mill to check my reamers. Why would he put it in the lathe. Just load it in the mill put it in neutral and spin the damn thing.
They are not that fragile. You don't throw them on the bench when not in the box, but they could take it. I use the same indicator multiple times every day. Mine is 4 years old and has been bumped and knocked around a number of times. Still accurate.
About the larger toolholder. I didn't see you check the bore at multiple depths. By checking at one place you dial it to be concentric at that depth, but the holder is likely not co-axial with the spindle, as the chuck grabs it by the edges. Try measuring at the edge, and then go as far into the bore as you can. This toolholder doesn't appear to have machined faces, but if a part have a face, you would normally dial the face to not have a deviation, and thus making the part co-axial with the spindle, and then dial the bore to have no runout, making the bore be concentric.
Tuugem I was thinking the same thing. check the holder at deeper depth. also try checking the tool as close as you can to the holder where you centred it.
exactly. i instantly thought when he put the tool in and suddenly mesured that it has runout that it has to be off axis or bend :p i would have mesured the test gage pin first further away to try to track down where the runout comes from mostly, also mesured the gage pin 180° turned and subtracted those results to eliminate anythin in the testpin. i would love to see a "do-over" or whatever its called in english :)
All these "what-if's" and excuses for the results could be totally obviated by simply installing the toolholder into the machine spindle and measuring the runout in-situ. Whatever the runout reads there is what the metal sees, and that's really the only measurement that matters. Stick the toolholder into the machine taper, measure the bore, then install the tool and measure the tool. Would probably have been faster than the way he did it originally, but I doubt the results will change drastically.
i think the cat40 tool holder axis was not lined up with the spindle axis, so having the endmill in the bore would throw the runout more (the farther away from the holder the more runout). You should try the same test with the holder in the VMC spindle and see what happens (so it is lined up by the taper and not the face of the lathe chuck), i bet you'll probably see the same results as with the tormach holder.
I suspect the added runout from the 3/4" holder is because the holder wasn't parallel with the spindle, thereby causing angular runout. Just like when a rifle barrel is dialed in it is checked at two different locations 4-6" apart.
I'm doubtful. The angle would have with be 0.23 degrees (0.002" of runout about 0.5" from the end of the collet (arctan(0.002/0.5)), which is highly unlikely. There would have had to be debris in the holder, which wouldn't have been possible due to the extremely tight fit of the tool in the holder. What it likely was, was different ambient conditions than what the tool was turned at, as well as anisotropic properties, meaning the tool didn't expand/contract outwards at the same rate in every radial direction.
"Apron" is the front part, with the knobs and handles on it, that hangs down in front of the lathe ways like a cook's apron. "Carriage" is the name for the entire moving assembly, including the part that rides on the ways and holds the crossfeed screw, as well as the apron itself. Also, there's likely far more slop between the ways and the carriage (on an old used lathe like this) than the runout numbers he's measuring in the endmill holders, so it wouldn't work to measure it off the carriage.
Some remarks: You have a system with a mechanical indicator with gears, shafts, and bearings mounted on a tall flexible mount. A tiny amount of energy is stored in the mag base mast going +and released going -: it takes a little more force to deflect the indicator + direction than it returns going - . This is the dreaded stiction. Your Mitutoyo indicator is a superb tool, new and free running, but it's still afflicted with stiction and this is manifested in the repeatability of its readings. The longer the mag base mast, the worse the stiction affects the reading. A small amount of vibration or a light tap with a pencil is usually enough to settle an indicator/mag base system to its lowest energy state. This is a long way around to suggest you run a motor idle and disengaged so it's slight vibration dithers (actual technical term) the indcator set up settling out the stiction making the readings more reliable. I realize the motor/drive noise poses problems when shooting a video but I thought I'd mention it for your viewers. Another point is wear in the spindle tooling. A coralliary of Murphy's Law dictates that wear never cancels error. In time yous end mill adaptor will wear. Collets wear bell-mouthed but stay roughly concentric but set screw (Weldon style) adaptors wear quicker than collets. The cutting forces in a Weldon adapTer produce a stirring action in the gripping bore. The serscrew produces a large load on the opposite wall of the bore and it is here where the wear will be greatest. Further, the tooling bore will wear bell-mouthed. Figure on replacing them oftener than collet adapters. How often is a matter of intermittant monitoring the state of wear and taking into account the degree of cutter performance required by your work.
John, in a weldon-screw in type of holder you have a single point and a very small almost a line like surface contact on your tool and holder. It may seem rigid but if you run any real cuts it will fret. For grinding apps it will work since you're removing material in very small amounts. You can obviously use it on endmills for roughing, if the endmill has a land for weldon. I recently had a long extension issue and purchased a ShearHog SH787 just for that purpose, holding it in any type of holder other than a high performance milling chuck didn't work. The SH worked but the quality was pathetic due to vibration. That was really educational to see that a milling chuck (they are mechanical, usually more expensive than either shrink fit or hydraulic chucks) eliminated the vibrations. Your setup may seem to work great with tools that have small extension. But if you truly want quality, you need to up your tool holding to shrink fit or hydraulic or high perf. Also instead of dialing in your lathe chuck, you need to look at the entire chain, for both displacement and stiffness, spindle-toolholder-tool-tooltip, and measure the tool tip runout. Add a solid carbide extension if there's vibration upto 5D extension.
The run out you are seeing on the 3/4 diameter is most likely largely influenced by the centerline of this diameter not in line with the centerline of the lathe spindle. Case in point when Abom dials a diameter in his lathe he checks it for runout on the diameter and a face or machined shoulder several times till both directions are as close as the workpiece will allow. In your situation the best thing you can do is put the taper into your machine spindle then check runout. This will not be perfect either but will greatly reduce the error shown now. Once this is achieved you can then proceed to the task at hand checking the push of the set screw. After all we are talking about milling and not grinding so where the rubber hits the road is the holder in the spindle. The push of the set screw will certainly be a different value on the tip as compared to the shank next to the holder. Bottom line alot of variables are influencing the rotational accuracy of an end mill. One more thing to think about. All standard carbide end mills are a minus tolerance on the cutting diameter so any runout/push or combination of the two will in effect use up some of that negative tolerance and maybe bring it closer to nominal. BTW you should never toss an indicator around and expect it to maintain accuracy. Metrology is a very intense subject not easily covered here. Just some of my .02
Gonna put this "throwing" the indicator thing to rest. They get built in Japan, packaged, thrown in a truck, bounced around till they reach port, get tossed in a container, lifted onto a freighter, tossed around at sea for a few days, gets unloaded, tossed on a truck, container gets emptied, thrown on more trucks, bounced around for a few hundred miles, thrown on a shelf, thrown BACK on a truck, bounced around some more before it FINALY gets to you. If it survive the journey to your hands, a 3" slight toss isn't going to do jack. Simmer down.
Just because you can doesn't mean you should... My truck has a few dings from the previous owner, it doesn't mean I'm going to intentionally do anything that could cause more.
Joshua Stinson True, but I'd think a precise mechanical instrument like the one in the video has some vibration tolerance engineered in. It would be interesting to find out what the design requirements for the test indicator was. Of course if you baby it, the instrument will last longer, but is that worth someone's time? Some would say yes, others no. As an engineer though, I'd say this type of abuse should be accounted for. If not, somebody should be ashamed. :D
John, you have to make sure you are concentric and square to the spindle . you have to check multiple depths on the bore. you only checked one area. any run out would be amplified. also get some thin copper to put on your jaws bid
with the tormach holder you may have had more of a cylindrical surface to orientate to the spindle so it was less apparent. if you were only grabbing the outer ring of a cat 40 even a 1 deg kick would deceive you. to remedy try putting the pin back in and get to concentric readings or reindicate the bore to match in two places. also when doing fussy indicating on machine only spin one direction . it takes the slop out and gives a little preload, so the jumps are consistent. happy holidays. keep up the good work.
John The initial movement on the DTI when you rotate the chuck is the headstock bearing rolling up on the lubrication film that's why it stays a tad off ( never returns to zero ) while you are rotating it ,plus the influence of you hand. In the steam turbines I have worked on you can see the shaft rise when they are barred up prior to putting on the main steam . It's oil film ramp
Stuart Hardy shouldn't that ramp in the opposite direction when reversing the spindle? Even by hand? It's been awhile since the class that included hydrodynamic bearings/effects, and we didn't go into effects of hydrodynamic forces on roller bearings. Not doubting you at all, just looking to expand my knowledge base.
Wooo! Thanks for taking the time to measure, sir. OK, brass tacks. First, you're using excellent USA made and brand new toolholders that were manufactured recently, with USA made cutting tools that were also manufactured recently. That is the best possible set of circumstances for a test. It only goes downhill from there. Depending on the shank tolerance on your tool and how old your setscrew holder is, it could go way downhill. Second, like I said on Instagram, half a thou will not cause the machine to explode or the tool to snap off at the holder. What it WILL do is impact surface finish when side milling. No effect when end milling (more or less), which is what you were doing in the injection molding video. If you have a two flute tool with one flute a half thou higher than the other, that change in scallop height when sidemilling makes a visible pattern. The more runout the worse the pattern gets. The more flutes, the more complex the pattern gets, but it is still there. You can't feel it with your fingernails, but you can see it, and for some things that pattern can be a real problem. For instance, I used to make paintball gun parts, which had to be polished to a mirror finish before anodizing, and this visible pattern caused me no end of irritation during polishing. It IS real, because the deeper marks were still there long after the shallower ones had gone. I fought and fought my process trying to improve the surface finish, and in the end I came up with two solutions that pretty much worked equally well. 1) I bought a milling chuck that was guaranteed to have less than two tenths runout at the tool shank, and they weren't lying. TIR on the tool was just under two tenths installed in the machine spindle. THAT toolholder combined with judicious replacement of tools used for finishing (they would do X parts as a finishing tool, then got demoted to roughing for the rest of their lives) solved the surface finish problem. I had been doing the tool demotion thing before the milling chuck too, but after switching to the milling chuck the parts per finishing tool (that number X) went up about 20%. It wasn't huge, but their life did increase with reduced runout. 2) An oldtimer I respected over at the CNCZone forums suggested that I try a single flute tool, either by simply removing one cutting edge from a two flute, or buying a single flute made for high speed routing for aluminum or plastic. That tool would also leave an acceptable finish, even in a setscrew holder because to it runout was irrelevant. Wherever that one flute ended up in relation to the spindle centerline didn't matter (except as a matter of total cutting OD, but these parts were never splitting a thou tolerances to start with, it was all about surface finish), because where it touched the work, it was always the same distance from center. That tool too solved the surface finish issue. Cost wise it was about a wash. The milling chuck was about $250 (in 2004) and the good Atrax single flute carbide router mills were about $95 a pop. The router mills were cheaper up front, but they wear out, whereas I still have and use the milling chuck. These were all 1/2" tools, solid carbide, USA made. I ended up sticking with the milling chuck method because I'd already sunk the money into it, and in the long run it was much cheaper (those were parts that I made for close to five years, so it more than undercut the router mills cost wise over that span), but I kept the single flute trick in my back pocket, and it saved my ass on several occasions later on. Like I said on Instagram, it DOES matter, but not for everything. You just have to be aware of what's going on with your tooling and use it accordingly. That's why I suggested that you measure to see for yourself what your toolholders look like. The setscrew holders I own are terrible runout wise, so I don't use them for things where that will be a problem. I use them for things like big roughing endmills, or for other single flute tooling where runout doesn't matter (like form tools or half round engravers). The things I use in setscrew holders most often are indexable endmills that I use for roughing (like that ShearHog you're so proud of). The tolerances on most inserts are so large that whatever the toolholder itself is contributing to the high/low of the flutes radially is just making a poor situation slightly worse, and since I'm roughing, I don't care what the sidemilling finish looks like anyway. I'd be interested to see what the TIR looks like with that MariTool holder actually in the spindle of your new HAAS machine. That's how I measure all of my tooling, because ultimately that's the only place it matters. Anyway, thanks again for taking the time to measure, I'm glad you took video of it.
John, Nice demonstration. My takeaway. Measure the tolerances in your tooling if your work requires it, and adjust accordingly (as you demonstrated in the injection mold video). Otherwise I would probably look for error elsewhere. If you were to repeat this test I would look at doing it on a precision ground v-block on your surface plate. Attach the indicator to your height gauge. The lathe and chuck add a lot of potential error sources. Thanks for the great video, your insight, and continued contributions to the CNC/machining community.
John, you should have used a dowel pin and measured the runout at the TIP of where the tool end would be, and not at the base of the tool. If the tool is tipped to an angle by the setscrew, the runout at the end of the tool tip would be multiplied. I have normally only heard this runout argument made on small conical engraving tools where the cutting tip is super small, and even a thou or so could make a difference on fine engraving.
AndTheCorrectAnswerIs Not quite, what you have is a perfectly machined true tube, holding a perfectly machined true shaft. Therefore the set screw won't make the tool tilt, it will push it against one side of the tool holder, but still parallel (or as close as possible) to the tool holder. So, the runout will be equal (or damn close to) regardless of where the indicator is reading from.
AndTheCorrectAnswerIs I am, because they are reasonably safe assumptions to make, considering they are precision parts. And if the tool is worn, it will go skiwhiff in a Collet, Jacobs or Set screw chuck. And if the tool holder is off by enough to be significant, it needs binning.
John, I like the concept of checking runout but I wonder if the tape on the jaws isn't allowing the cat40 holder to be skewed to the spindle centerline. you have the nose on centerline but I bet the back of the toolholder is off center resulting in a cylinder that wobbles the farther you get from the true running nose. It may be easier to test this in your mill spindle.
The larger the hole and shaft, the larger the tolerance on the hole and shaft diameters. It doesn't matter whether we're talking tooling or bearings. The larger the part, the more thermal expansion becomes an issue both in manufacturing the tools and in use. For example, a 2" diameter part heated from 70 degrees to 250 degrees expands twice as much as a 1" diameter part. So when manufacturing a part that is heating up as you cut it, it becomes harder to hold tight tolerances on larger part. Use of coolant and CNC machinery has made that easier, but it still affects the final part. Thanks, John
I also cringed at the "toss". However, there were not any other employees around and that would be important as they follow what the boss does, not what he says. Also, I think that the "other" factors that could have mitigated the deviations could, just as likely, made the results worse! Also, tape can not be used when measuring in tenths. Also, when measuring in tenths, you can only do it when the actual cutting setup is used (CAT40 in the spindle of the machine to be used). NYC CNC videos have been of a great benefit to me as well as the comments. It is generally true that the comments are more helpful to me when a mistake is made. Great video!
John, When ever you indicate something in a 4-Jaw chuck you also have to indicate the face run out. It might be that you have a close run out at one place in the bore but not in every place in the bore (because of worn chuck jaws or the chuck jaws springing open when you tighten them). So to really get something running true you would have to indicate at the front of the bore than at the back of the bore (You may have to tap the part in the chuck than readjust the jaws and repeat as many time as necessary). You can also indicate the face if it runs true to the bore. So with your end mill test you may be running true in the bore in one place but you move the indicator out on the shank of the end mill and its not running true because the bore not running true in the back due to face run out.Also how tight were you tightening the set screw in your test? As tight as you do to really hold a tool? Usually it been my experience that you (I mean every one not just you) tighten them jest about as tight as you can get it. This does set the tool off center. Some end mill holder manufactures intentionally grind the bore off center to compensate for this.
I work in a shop that also grinds tools. run out can be significant in older, well worn and cheap import holders. Compounded by Chinese mills with shanks that are sometimes undersized by a few tenths. We know our customers well, and the ones that use old or cheap holders get their tools ground in cheap import holders. If the tools don't have weldon flats, we add them to keep the slop to the same side of the tool when set. This way, the shank can be pushed to the side of a sloppy holder by the set screw, but we grind the cutting edges true in that position. Less effective run out. You get what you pay for. A high quality holder in good condition with a quality tool set in it will have negligible run out and can be pushed a lot harder without making my machines "sing".
The Command brand ones I get to use at work are usually well within .0005 Measured on a Zoller, HSK100a holders. If you want little runout as possible: use shrink, hydraulic or powRgrip. Usually all that ends up in an endmill holder is roughing tools.
For your test indicator look up specs on hysteresis. IIRC the compac (tesa) indicator I have is speced at less than 2 tenths (might be 1), but that's a multi rev. Metrology is a wicked business!! You're also dealing with a not so parallel tool to bore axis if the set screw doesn't split the engagement length of the tool. Not much just off the end of the holder, but it can still be there, and get larger further out.
GOOD NEWS! ---> John, You have a couple of compounding tolerances in what you are measuring. Take the actual tool and/or pin diameter in consideration of the runout that you have measured, plus still a little iffy on the tape, you forgot to cut the mesurement in half for diameter. So your runout appears to be actually 0.00015" on the tormach holder. Might be fun to mic everything too!
Retry this with the holder in the VM-3 spindle. Check the spindle runout (1 tenth or less likely), then the tool holder bore runout and finally the clamped tool runout. This should give you A) your total runout to compare with the ER and B) if you subtract the first 2, the runout attributable to the setscrew. Also, you're used to 1.5hp. I absolutely would not use a setscrew holder without a Weldon flat on the tool in a 30hp machine. I've had 1/2" endmills pull out of a perfectly tightened ER32 mid cut well below full power and a setscrew holder without a flat to grab onto has far less pullout resistance. You'll also find that the machine has enough torque to overcome the holding force of the set screw and spin the tool in the holder...ruining the bore in addition to snapping the tool. Completely unrelated tip: I'm not sure if they changed this with the Next Gen Control, but be VERY careful with the tool offset probe. I had a close call when I forgot to mark in the offset page that a 2in shell mill was not a spot drill. The probing cycle changes based on the type of tool you tell the machine you're using. It plunged straight down for what it thought was a spot drill and the inserts were hovering over the main body of the probe when I hit feed hold. Every time I put in a new tool I would immediately change the offset/probing page and run a probe cycle to prevent any mishaps like that. Also, with the work offset probe, it's not active if the lights aren't flashing. That means you can crash it during hand jog. The tip is designed to break to protect the super expensive body, but it still costs $100+ to replace the tip.
I think John should retest the cat40 in the VMC I don't think it was cylindricaly aligned or parallel to the lathe there wasn't much to grab onto with the 4 jaw chuck. I think that a thou of runout is quite allot for a 3/4 end mill even if they are not import end mill holders. I thought the set screw run out would be more than a few ten thousands of an inch. Nice video John.
My big lathe has about 3 tenths of error in the spindle bearings, no amount of dialing in will get the error to less than 3 tenths. I only know this because I have used sub-tenth indicators to try and dial in parts to DNO. Making sure that you have at least twice the resolution is important. That means using a 50 millionths indicator to read single tenths, and a tenths indicator to read 2 tenths, and 5 tenths to read 1 thou. A 1 thou indicator gets you within 2 thou.
Ok, here's the bottom line. You have some clearance tolerance built into the tool holder so you can get the tool into it. Unless you are using shrink fit holders, a solid holder will always have a very slight amount of runout due to the clearance. The better the holder and the better the tools you use, the lower the error will be, but it will never be zero. So, yes, the endmill will be offset slightly from the centerline of the holder, unless the maker ground the hole eccentrically to account for the clearance.
I kinda like the description "set screw holder". I knew immediately what you were talking about, although over here we call them Weldon tool holders. I never used them because the tool supplier did not recommend them at the time I first bought tool holders. I have been using ER32 and ER25 collets and they work just fine. Seeing this video I probably never will use Weldon holders. What strikes me is that you do not use weldon tool bits in this demonstration. Weldon tool holders require btis with a notch in the shaft to prevent the tool from ever pulling itself out of the holder.
I am with you John, the tool holders will have little effect on your finished product. For instance in cutting a slot. If it need to be .5000" Then you would not use a 1/2" cutter to do that anyway. You would use maybe a 3/8". So all you really have going on, is a flute on one side of the cutter taking a slight bit more. Maybe that will result in a tool wearing out sooner... who knows! To me the finished part xize, and what it measures; after it is cut with any tool... is what is important. Thus if you have a cutter, that is cutting to large a slot in this instance; possibly form run out. Or lack of any machine tolerance. You simply adjust the size of your tool to reflect this in you CAM software. Tell the system the cutter is .5005" in diameter, or tell the machine the slot should be .4995". There are many ways to "fool" the system, so that the finished part is what you get. I highly doubt, you will ever get a 1/2 slot, being cut with a .5000" cutter. Even on a brand new machine. Because of all the things you are bringing up. However, with a few brains upstairs; all that can for sure be overcome. My machine shop teacher back in 1979, said this: "It is not the machine, no matter how old; it is the machinist!". Thus you can put a person, with little to no experience on the new Hass, and they will not be able to make good or fitted parts. They just don't understand machining. However, you put as experience Machinist, on a lathe made in the 1870's; and somehow, by some absolute magic, that person can pump out good accurate parts all day long. So again I agree, the accuracy of the tool; and its tool holder; on a slow machine. With larger cutters, don't mean much, except in the final part size. However on a new High Speed spindle, will most likely reduce vibration and such... thus being more important, to have tight tolerances. As when using small cutters, .060" or smaller; real easy to break on sloppy or old machines, I have experienced that. Enjoyed the video! Thanks.
The tool holder is centered with respect to the ground spindle mating surface. So you would have to insert this in a spindle or something that can hold on to the taper. That outside part of the tool holder doesn't need to be machined with very tight tolerances compared to the spindle mating surfaces and the where the tool attaches. That might account for the big swing in dial indicator.
I will agree with most of the other comments, that this should be done with vee blocks on the surface plate. Also even 20 years ago, I remember the best tool holder makers talking about grinding the bore offset from the axis to compensate for the set screw. I measured about 3 tenths of intentional offset in their brand new 3/8" holder at the time. So you can't dial in the bore, you need to dial in the spindle end, not the bore. I guess these days the real winner is hydraulic holders or shrink fit. I've never had the chance to use them, but I've done other shrink fits like piston pins into rods, and I can imagine how perfect and rigid that would be in a tool holder.
I'm pretty sure, if John did this video after he had his haas he would have not only talked about it, but would have used it to do this test. he uploaded a back up video for times of busy, clearly the man is busy, which is great! ps. I'm really looking forward to seeing you use your haas!
You need to make sure not only that your bore is on center but that it's also in there straight along the the length of the bore because you may not be in there straight along the axis of the tool holder. Somebody mentioned putting it into the spindle and that is actually probly the easiest way to do that since the spindle is precision ground and new so you will most likely have very little if any play in the spindle. Interested to see if that straightens it out and give better readings.
you were trying to use a "very used" 4 jaw Chuck as a perfectly flat surface and we're adding run-out to the part. you should have used a longer gauge pin tightened into the holder and measured at both ends to make sure the holder was perfectly aligned to the centerline of the lathe.
At 10:43 if you mark the high and low positions of setscrew chuck, then you can know whether to add or subtract that from your measurement of the deflection of the tool when clamped with the setscrew.
The problem may be be runout when you're actually measuring and need to measure concentricity. Also if the video is about using a collet vs a set screw holder you should do the same test with a collet. Maybe the collet would will be even worse. Who knows? I've heard that the real problem is when the tool is under pressure. Being held by one screw versus the whole circumference effects how it flexes during a cut. The collet and endmill will both probably wear more evenly then the setscrew holder or it's endmill. The setscrew holder might be more rigid though. Does it matter if it's a weldon shank? I think you're probably right however. People have all sorts of superstitions when it comes to machining. I think the time you save with setscrew holders probably nudges out the tiny possible downsides for almost all jobs.
The tolerance of AccuHolders is .0002" on the ID, measured with ten-thousandths carbide gage pins, and the "hand feel" is everything ;) Add the usual .0002" of clearance on tool shank OD, plus or minus a bit of chance, and there go your three tenths
Here in Seattle the aerospace shops don't like ER collects for milling on CNC machining centers they like the set screw end mill holders or for 0 zero run out they use milling chucks they say the ER collects fail sometimes on long runs with heavy cuts
The second tool holder was likely slightly cocked in the 4 jaw making the tool magnify that error. Basically the bore wasn't straight with the axis if the machine. But who knows when dealing with such a tiny thing. Great videos. Thx.
It is very possible that the tool is orbiting so you have to check multiple places to insure the bore is perpendicular to the bore of the lathe. That’s where cmm’s rock. Happy checking.
For doing general purpose machining, not mold making or hard milling, I haven't really had any runout issues with set screw holders in VMCs. Grind a set screw flat on you end mill and lock her down and she probably won't pull out while you're roughing. I've had more run out issues with beat up collets and cracked collet holders than anything. Good when they're but tend to go to hell after a few years.
just an idea. The manufacturer must give a certain amount of clearance in the chuck bore on purpose for ease of tool insertion. since it is safe to assume this clearance is quantifiable, maybe they offset the center of the bore towards the set screw by half of the given clearance value. this would allow the tool to be concentric to the axes of rotation when screw is tight. I would love to see this test done by zeroing in on the toolholder shaft instead of the bore.
12:20 OK, you were surprised at +6 -1 But the point is that when you tightened the set screw, the +6 -1 did not change. So wherever your error was coming from, setscrew tension wasn't affecting it. For the actual tool, tighening the setscrew reduced the total runout.
So the setscrew collet-held tool had runout roughly similar to the collet alone. Would be interesting to mark where the high point was for the collet alone to see how it compares to the high point of the held tool. Really neat.
You should take in consideration the heat from the machining process. It could reduce the play between the tool and the holder. And use a end mill that is designd to be used with that holder (weldon ) the shank could be grinded at a specific tolerance for the holder.
Just my thoughts after opening my eyes when you tossed the indicator ;( Why not just insert the holders into the mill spindle to check runout? That would also include any tool holder to spindle runout possibilities. After my 20+ years as a manual machinist I never found an endmill less than a thou out unless it was heat shranked.
When the Haas is up and running measure the 0.75" solid tool holder in the spindle. I bet you'll get better results. I use 1/8" solid holders and I only get a few tenths of run-out.
Not sure I'm buying the other sources of runout error theory. Sure, in principle, that sounds reasonable, but your test showed that the Tormach tool holder runout was exactly where it would be if it was entirely attributed to the set screw biasing the tool. Given the results of this one test, I'd be inclined to say that the error is almost entirely the result of the set screw holder. It would be just as valid to say that the lathe chuck and other factors produced .0005" of error in the other direction, and the set screw runout error was actually .0008", but that error was reduced by .0005" by the opposing runout error of the chuck. IMO, dialing in the ID of the tool holder would null any other setup errors. Of course, in real world use, those other sources of error would not be nulled, so spindle runout, runout between the tool holder shank and the tool holder ID, etc. would all be in play. I'm a pud knockin' home gamer, and this looks like fairly significant runout to me, particularly on a smaller end mill. Not end of the world, but not insignificant. I will admit that the rigidity of the set screw tool holder is typically better than an ER20 collet tool holder, particularly for aggressive cuts, and that's certainly an important point. I bet Maritool hand picks all the stuff they send you in their metrology department from now on! Comparing a 1/2" import tool holder to the vurra nice looking 3/4" Maritool tool holder isn't exactly an apples to apples comparison, but 1/2" to 3/4" isn't so much of a difference that the Maritool should have had .0016" of runout when the Tormach had .0003" of runout.
It might be possible that while you dialed in the front of the hole, the rest of the toolholder's bore might not be in line with it (ground crooked), and that the tool sits at an angle into the toolholder. It would be interesting to see how the measurements are affected further down the gauge pin's length, say at 1" distance from the chuck, if the measurements get amplified down the shaft, that would be your answer. Ciao, Marco.
To all the people commenting on the indicator that he "threw" on the table... that indicator didn't even care about that very slight drop from a few inches high. It gets way more beat up in the mail on its way to the distributor, then again from the distributor to you. It will be perfectly fine. (also he didn't throw it)
Hi the type of tool holder you have is Weldon holder . A god weldon holder can not be to cold for mounting a cutter in it's not shrink chuck but it's where narrow in tolerances. I normally heated white a hot air gun to 50celcius
Hey... great video as always. I reckon your toolholder is not parallel withe the spindle centerline. So while the runout may be 0 at a SPECIFIC point on the holder, that WILL change down the length. You should probably run the indicator up and down the gauge pin in 2 axes to get it parralal first but then I guess you are skewing results again. Needs to be in a proper spindle I guess... try it in your brand spanking new Haas 😀
the bigger the tool the larger tolerances, if you're truly worried about run out and this is for all your viewers swap to shrink fit if you can mortgage your house. we use them almost exclusively especially in the mag 3 at 33k spindle .00005-.0001 tir
I would suggest clamping the holder in a v block with the set screw at 90 degrees to the orientation of the v groove, and then "flip flop" the v block on a surface plate and measure with a height gauge. This would show you exactly how much error there is in the plane of the set screw
Another thing would be to manually probe your tool radius at each flute and see if it matches with the test here. Any good probing subprogram will run the tool backwards and measure from the highest point anyways.
The real test is to chuck that holder in the Tormach, and very lightly tighten the setscrew on a gage pin or tool shank. Measure the runout on the tool shank. Note the relation of the runout to the setscrew location. Now crank down on the setscrew to normal torque and measure the runout again. Now you know the inherent runout in the holder, and you know any distortion from tightening the setscrew.
What if they grind the center bore slightly off center to compensate for the tool being pushed over I did not see you measure for that thanks for the videos. Mason
The error in the reading was most likely due to the tool not being coaxial. Also the reason I do not like Weldon holders is the fact that you are pressing a carbide cylinder against a slightly larger bore; so the contact is a single line the length of the tool inside the holder and one or two contacts from the screws. Another major concern with Weldon holders is they should only be used with tools that have flats milled/ground into them; otherwise the gripping force is insufficient for a 30hp spindle and the tool will spin in the holder, possibly pulling out, pushing in, or welding itself in place.
all the older guys tap the indicator base and the arms after dialing onto zero. It gets that initial bogus reading out so you don't have to go around 3 - 5 times until the indicator settles in.
The issue isn't really so much that the tool will push it out. The surface area takes the pressure fine you shouldn't get any run out. Any deformation would take place in the screw since it's taking the same amount of pressure but on a lot smaller surface. Beyond that the main issue is they begin the run out relatively quickly compared to a milling chuck that lasts quite a while just replacing the sleeves. Same with ER collets but I don't run much bigger than 3/8' endmills in an er collet.
That up-down-up-down that doesn't follow a nice orbit, that's your spindle bearings. You see a high spot, a low spot, then a mid-high spot, then low, then high. According to your indicator, your spindle bearings are 1-2 tenths out.
I use these Set Screw Holders (BT40) all the time, also for finishing in inox steel, and can mill in 0.02mm! easily and reapetedly :D Then for de ER collets (especially on SK40) you shouldn't be concerned with the rigidity. For example I use ER16 150mm long chucks on a 5 axis machine and really didnt have any issues. Regards Sam (CH)
Ive done these tests years ago, and came to the same conclusion. Every time, there was a difference. Just remember that even two tens means that one tooth is cutting more deeply than the other, and can lead to four tens larger diameter cut. Realistically, we can’t really get it perfect, though, if we’re lucky, errors cancel out. Most likely they cancel out partially. But, cheap holders can have close to 0.001” runout. The best; 0.0002”. If you want to drive yourself crazy, you can rotate the tool in the holder while measuring runout, and marking holder and tool, always put the tool in on the marks. I do that with some things. With the holders in the four jaw, you’re squeezing the holders from four directions. Even slight pressure can distort that thick holder. Remember we’re talking about just a few tens.
The answer is yes the tool will move some when tightened as the slack in the bore is taken up. Some holders are ground with a bias in the bore to allow for that. Like Briney tru holders www.brineytooling.com/files/flyers/brineytrupos.png The issue is how much does it really matter. If your feed rates are .002 tooth load you will end up with actual .001 to .003 chip loading depending on what side of the tool you are on. I am sure this effects tool life but I can't tell doing what I do with my tools in a job shop but what it does do is effect finish. The greater your tool run out the slower feed you will have to run to get a good finish. Hydraulic holders like my Schunks are the easiest and best solution to this for most applications as they are more rigid and repeatable than collets but for maximum rigidity and run out you have to go with a shrink fit tooling setup. You need to check run out with the tool in the spindle where it matters. I have had tooling with run outs up to .003 on the big stuff and it still cuts great and will hog off material as good as a true running tool but again when you need the finish you will be running half speed.
the offset comes from the screw flexing the body, not the diametrical difference. Thats also why they make facemills where each adjustable insert can be dialed in.
I would like to see a comparison with the set screw holder and the clamping tool holder (ist this the right name for it?)! If you do a remake with the mill spindle that would be awesome! I know this discussion of which tool holder to use very well. Best regards from Germany! :-)
Wasn’t there an update where he talked to frank from maritool about how they measure their runout and he put the cat40 holder in his VM3 and re-measured the runout? I was re-watching this because I couldn’t remember what the runout was when it was tested properly...
That around 1 tenth equals 0,01mm pretty precisely. A runout like that does matter when doing e.g. a 0,03mm cut on a 2-flute end mill. The other one is taking 0,04 and the other one 0,02 chips. Also the spiral flute will change it's cutting when it rotates around the end mill and the center is not rotating around itself.
Just catching up on all your videos now. I'm a bit behind at the moment. The CAT 40 would have been hard to hold. I liked your note at the top of the screen. Should keep the keyboard warriors at bay. Keep up the great work John. cheers, Aaron.
OK, brass tacks: First, you're using excellent USA made and brand new toolholders that were manufactured recently, with USA made cutting tools that were also manufactured recently. That is the best possible set of circumstances for a test. It only goes downhill from there. Depending on the shank tolerance on your tool and how old your setscrew holder is, it could go way downhill. Second, like I said on Instagram, half a thou will not cause the machine to explode or the tool to snap off at the holder. What it WILL do is impact surface finish when side milling. No effect when end milling (more or less), which is what you were doing in the injection molding video. If you have a two flute tool with one flute a half thou higher than the other, that change in scallop height when sidemilling makes a visible pattern. The more runout the worse the pattern gets. The more flutes, the more complex the pattern gets, but it is still there. You can't feel it with your fingernails, but you can see it, and for some things that pattern can be a real problem. For instance, I used to make paintball gun parts, which had to be polished to a mirror finish before anodizing, and this visible pattern caused me no end of irritation during polishing. It IS real, because the deeper marks were still there long after the shallower ones had gone. I fought and fought my process trying to improve the surface finish, and in the end I came up with two solutions that pretty much worked equally well: 1) I bought a milling chuck that was guaranteed to have less than two tenths runout at the tool shank, and they weren't lying. TIR on the tool was just under two tenths installed in the machine spindle. THAT toolholder combined with judicious replacement of tools used for finishing (they would do X parts as a finishing tool, then got demoted to roughing for the rest of their lives) solved the surface finish problem. I had been doing the tool demotion thing before the milling chuck too, but after switching to the milling chuck the parts per finishing tool (that number X) went up about 20%. It wasn't huge, but their life did increase with reduced runout. 2) An oldtimer I respected over at the CNCZone forums suggested that I try a single flute tool, either by simply removing one cutting edge from a two flute, or buying a single flute made for high speed routing for aluminum or plastic. That tool would also leave an acceptable finish, even in a setscrew holder because to it runout was irrelevant. Wherever that one flute ended up in relation to the spindle centerline didn't matter (except as a matter of total cutting OD, but these parts were never splitting a thou tolerances to start with, it was all about surface finish), because where it touched the work, it was always the same distance from center. That tool too solved the surface finish issue. Cost wise it was about a wash. The milling chuck was about $250 (in 2004) and the good Atrax single flute carbide router mills were about $95 a pop. The router mills were cheaper up front, but they wear out, whereas I still have and use the milling chuck. These were all 1/2" tools, solid carbide, USA made. I ended up sticking with the milling chuck method because I'd already sunk the money into it, and in the long run it was much cheaper (those were parts that I made for close to five years, so it more than undercut the router mills cost wise over that span), but I kept the single flute trick in my back pocket, and it saved my ass on several occasions later on. Like I said on Instagram, it DOES matter, but not for everything. You just have to be aware of what's going on with your tooling and use it accordingly. That's why I suggested that you measure to see for yourself what your toolholders look like. The setscrew holders I own are terrible runout wise, so I don't use them for things where that will be a problem. I use them for things like big roughing endmills, or for other single flute tooling where runout doesn't matter (like form tools or half round engravers). The tools I put in setscrew holders most often are indexable endmills that I use for roughing (like that ShearHog you're so proud of, and rightly so). The tolerances on most inserts are so large that whatever the toolholder itself is contributing to the high/low of the flutes radially is just making an already poor situation slightly worse, and since I'm roughing, I don't care what the sidemilling finish looks like anyway. I'd be interested to see what the TIR looks like with that MariTool holder actually in the spindle of your new HAAS machine. That's how I measure all of my tooling, because ultimately that's the only place it matters. Anyway, thanks again for taking the time to measure, I'm glad you took video of it.
I'm not sure what the deal is here. I tried to edit my original comment, but instead ended up posting a new comment, and now I can't edit THAT comment either. lol WTF UA-cam?
Hi John. interesting topic, great video. A couple things that I was wondering about: Did you test the cylindrical error of your test gages or milling cutters? IIRC the +/- .0002" for over or under is the size (diameter) deviation from nominal on the gauge pins, not the tolerance for cylindrical error which depends on the class (grade) of the pin. What about end shake without the grub screw tightened? a good slip fit (that pops) can be anywhere from a couple tenths to almost a thou in my experience. What about a sweep of the hole in the holder to see if it's out of round? What about tolerance stacking? inquiring minds want to know!
longrangehunter is on it I think. being parallel to the spindle is key. You could easily dial in to .0003 at one set point without notice a 1 or 2 degree kant of the tool holder in relation to the lathe spindle. End result being that when you meassure the tool it will swing on both side of zero as the median cross over is where you zeroed it.
when u are trying to tighten something precise in 4-jaw u need to dial in in multiple places across the hole. soc the hole might be dialed perfectly on one spot but further down it might have run out.
I think the reason you got more runout with the .750 endmill is because it had to be a little bit crooked, and the further your the tool is stuck out, the more runout you will have. Even if you were able to get it very close with just the tool holder, I think that it was slightly crooked not in terms of concentricity but with the angle of the holder. nothing you could have improved on, but it had to have not been perfect just given the setup. I think if you could get the holder exactly straight with the lathe, the results would be different
I'd be curious to see if the runout on your mill changes before and after you turn it 90 degrees in the holder. I'm not sure which error that would resolve but perhaps lower runout is possible with a little tweaking.
It will push it off center, but only the clearence between the end mill outside diameter and the tool holder's inside diameter, so just a few tenths or less. Be careful on a lathe turning between centers and using a bent tail, drive dog. Use a straight tail, lathe dog if possible. I have seen the leverage of the bent tail ones push the shaft off center and cause runout.
Size for size doesn't fit. So technically set screw pushes it over a few microns. Set screws are nice, cheap, and cant pull out. I use them, but any tools over 5-6" its better in a collet. Collets are better cause you can make the tool any length. At my job i mostly use Nikkon holders and put a shrink fit extension in it. That is really nice, expensive but worth it. I can't stand the people saying a reamer should be in a collet. Drill chuck for me. Video of that would be nice. I do use collet over1" reamer though
My issue with set screw holders is not pushing it over to one side but when name not mentioned inserts an endmill without a flat ground on the shank can cause pull out, especially in hard materials and titanium.
I've never had any issues with all my years of programing and running Haas machines. I am really impressed with them. It's all about setup and selecting the right tools for the job. I was referring to a manager at a job shop who incorrectly used the wrong tool holder and even pointing that out to him, said run the f**en thing! Needles to say, his ignorance destroyed a set of hard jaws.
Mounting a tapered shaft holder with tape around the short distance of the ring at the base of the taper could result in the holder being slightly out of alignment of the headstock axis and holder axis. While it may not be much, the variation you are measuring are not big. Wouldn't it be more accurate to mount the tapered holder in a matching tapered machine spindle. That would give a real indicator of both machine accuracy and potential holder accuracies. Lop
Probably a reason he didn't show the holders in a mill and that's because Tormachs have a bit of runout. I picked up a 96 Brigeport VMC for $3,500 and added Shars ER32 collets & CT40 holders. Mill ran under 0.0001" runout @ taper and the ER32 setup would maintain 2 or 3 tenths tops @ end mill shank. Backlash on all 3 axis was under 2 tenths on a 20 year old mill that can step @ 0.00006" over 20 years ago. Expecting a Tormach to hold 1 or 2 tenths on the holders isn't exactly unreasonable in this day and age.
The runout = the difference between the Arbor bore diameter and the shank diameter of the cutter. A bore Mic and an external Mic could have given you a difinitive answer without all that messing about. The real point here, is that if you get a slightly undersized cutter shank, on a Collet chuck it will always spin true. The other thing, is that Grub screw puts some unwanted pressure on the shank of the cutter, I have seen them snap before if they are not heat treated correctly..
If the Cat 40 tool holder isn't inline with the 4-jaw along its center axis then the error will compound the further out you measure from the tool holder.
when you center a bore; measure in 2 places. one close to the edge and 1 about an inch deep. this wil make sure you have no runout over longer distances.
As long as that end mill comes around with the same runout every rotation it will cut straight. Unless you are cutting a slot the exact same size as the end mill you’re using, it will be fine.
this is fine on 2 flt and 3 flt under say 6000 rpm, jump to 4 - 6 flt and some higher speeds and feeds and that difference may start to show problems. get up to 10+ and beyond the collet system should be used, once u get into 12000 and on shrink fit tooling and some type of harmonic system should be used like blue swarf which will give you the optimum rpm ranges for chatter free cutting at the appropriate speeds and feeds
out of curiosity, whens the last ime you checked your lathe spindle itself? all errors 'stack' and for an accurate baseline you need to know the error your starting with, if any.
I haven’t made it all the way through the video so I’m just questioning as i go but, even if there is a measurable precise amount of run out, couldnt you fix that in a four jaw?
Always tram your tool at tip of tool, longer tool means more runout, checking at tool holder isn't knowing what that tool is doing at tip. Used to use cellphone on taper to compensate, the closer you get it running true at end of tool the better the finish, and you'll be able to run faster with a trammed tool.
Hey John, good subject matter here. This is something I have often wanted to test myself. One thing I feel may be causing error is the way the part was being held and indicated in the 4 jaw. If you only indicated one spot, you may not being running absolute concentric. If so once you stick a tool in there and read a different position you will get runout. Getting parts indicated to within tenths in a 4 jaw is tricky, but I like how your not scared to give it a try. I see in a comment below your going to give it another try. 👍🏻👍🏻
Abom79 thinking the same thing
I was also wondering why after he indicated at the Chuck and switched to the end of the holder why he wasn't bumping it around with a mallet and then going back to the Chuck and checking run-out there. He instead went to the end and started adjusting the chuck again. I usually run two at the same time indicators one at the Chuck and one on the end. Tightening the chuck fucks up the end Everytime. I do the same thing in the mill to check my reamers. Why would he put it in the lathe. Just load it in the mill put it in neutral and spin the damn thing.
Am I the only one that jumped in my chair....just a little when that dial indicator box hit the table top....?
No, but that was a stunt indicator, a simple trick of the camera.
hahaha yup metoo!!!
Well it “was” a 10ths indicator
They are not that fragile. You don't throw them on the bench when not in the box, but they could take it. I use the same indicator multiple times every day. Mine is 4 years old and has been bumped and knocked around a number of times. Still accurate.
I’m sure it saw tougher handling when it was shipped, that’s why the case has foam and supports in it
About the larger toolholder. I didn't see you check the bore at multiple depths. By checking at one place you dial it to be concentric at that depth, but the holder is likely not co-axial with the spindle, as the chuck grabs it by the edges. Try measuring at the edge, and then go as far into the bore as you can. This toolholder doesn't appear to have machined faces, but if a part have a face, you would normally dial the face to not have a deviation, and thus making the part co-axial with the spindle, and then dial the bore to have no runout, making the bore be concentric.
Tuugem I was thinking the same thing. check the holder at deeper depth. also try checking the tool as close as you can to the holder where you centred it.
Tuugem I was about to write a very similar comment.
exactly. i instantly thought when he put the tool in and suddenly mesured that it has runout that it has to be off axis or bend :p
i would have mesured the test gage pin first further away to try to track down where the runout comes from mostly, also mesured the gage pin 180° turned and subtracted those results to eliminate anythin in the testpin.
i would love to see a "do-over" or whatever its called in english :)
All these "what-if's" and excuses for the results could be totally obviated by simply installing the toolholder into the machine spindle and measuring the runout in-situ.
Whatever the runout reads there is what the metal sees, and that's really the only measurement that matters.
Stick the toolholder into the machine taper, measure the bore, then install the tool and measure the tool. Would probably have been faster than the way he did it originally, but I doubt the results will change drastically.
i think the cat40 tool holder axis was not lined up with the spindle axis, so having the endmill in the bore would throw the runout more (the farther away from the holder the more runout). You should try the same test with the holder in the VMC spindle and see what happens (so it is lined up by the taper and not the face of the lathe chuck), i bet you'll probably see the same results as with the tormach holder.
I suspect the added runout from the 3/4" holder is because the holder wasn't parallel with the spindle, thereby causing angular runout. Just like when a rifle barrel is dialed in it is checked at two different locations 4-6" apart.
I agree with you. Its easy to verify. You could mesure the angular runout with the moving part of the lathe.
i think its the carriage in English. I mean with the magnetic base on the carriage.
I'm doubtful. The angle would have with be 0.23 degrees (0.002" of runout about 0.5" from the end of the collet (arctan(0.002/0.5)), which is highly unlikely. There would have had to be debris in the holder, which wouldn't have been possible due to the extremely tight fit of the tool in the holder. What it likely was, was different ambient conditions than what the tool was turned at, as well as anisotropic properties, meaning the tool didn't expand/contract outwards at the same rate in every radial direction.
Apron is what I've always heard it called.
"Apron" is the front part, with the knobs and handles on it, that hangs down in front of the lathe ways like a cook's apron. "Carriage" is the name for the entire moving assembly, including the part that rides on the ways and holds the crossfeed screw, as well as the apron itself. Also, there's likely far more slop between the ways and the carriage (on an old used lathe like this) than the runout numbers he's measuring in the endmill holders, so it wouldn't work to measure it off the carriage.
Some remarks:
You have a system with a mechanical indicator with gears, shafts, and bearings mounted on a tall flexible mount. A tiny amount of energy is stored in the mag base mast going +and released going -: it takes a little more force to deflect the indicator + direction than it returns going - . This is the dreaded stiction. Your Mitutoyo indicator is a superb tool, new and free running, but it's still afflicted with stiction and this is manifested in the repeatability of its readings. The longer the mag base mast, the worse the stiction affects the reading. A small amount of vibration or a light tap with a pencil is usually enough to settle an indicator/mag base system to its lowest energy state.
This is a long way around to suggest you run a motor idle and disengaged so it's slight vibration dithers (actual technical term) the indcator set up settling out the stiction making the readings more reliable. I realize the motor/drive noise poses problems when shooting a video but I thought I'd mention it for your viewers.
Another point is wear in the spindle tooling. A coralliary of Murphy's Law dictates that wear never cancels error. In time yous end mill adaptor will wear. Collets wear bell-mouthed but stay roughly concentric but set screw (Weldon style) adaptors wear quicker than collets. The cutting forces in a Weldon adapTer produce a stirring action in the gripping bore. The serscrew produces a large load on the opposite wall of the bore and it is here where the wear will be greatest. Further, the tooling bore will wear bell-mouthed. Figure on replacing them oftener than collet adapters. How often is a matter of intermittant monitoring the state of wear and taking into account the degree of cutter performance required by your work.
John, in a weldon-screw in type of holder you have a single point and a very small almost a line like surface contact on your tool and holder. It may seem rigid but if you run any real cuts it will fret. For grinding apps it will work since you're removing material in very small amounts. You can obviously use it on endmills for roughing, if the endmill has a land for weldon. I recently had a long extension issue and purchased a ShearHog SH787 just for that purpose, holding it in any type of holder other than a high performance milling chuck didn't work. The SH worked but the quality was pathetic due to vibration. That was really educational to see that a milling chuck (they are mechanical, usually more expensive than either shrink fit or hydraulic chucks) eliminated the vibrations. Your setup may seem to work great with tools that have small extension. But if you truly want quality, you need to up your tool holding to shrink fit or hydraulic or high perf. Also instead of dialing in your lathe chuck, you need to look at the entire chain, for both displacement and stiffness, spindle-toolholder-tool-tooltip, and measure the tool tip runout. Add a solid carbide extension if there's vibration upto 5D extension.
The run out you are seeing on the 3/4 diameter is most likely largely influenced by the centerline of this diameter not in line with the centerline of the lathe spindle. Case in point when Abom dials a diameter in his lathe he checks it for runout on the diameter and a face or machined shoulder several times till both directions are as close as the workpiece will allow. In your situation the best thing you can do is put the taper into your machine spindle then check runout. This will not be perfect either but will greatly reduce the error shown now. Once this is achieved you can then proceed to the task at hand checking the push of the set screw. After all we are talking about milling and not grinding so where the rubber hits the road is the holder in the spindle. The push of the set screw will certainly be a different value on the tip as compared to the shank next to the holder. Bottom line alot of variables are influencing the rotational accuracy of an end mill. One more thing to think about. All standard carbide end mills are a minus tolerance on the cutting diameter so any runout/push or combination of the two will in effect use up some of that negative tolerance and maybe bring it closer to nominal. BTW you should never toss an indicator around and expect it to maintain accuracy. Metrology is a very intense subject not easily covered here. Just some of my .02
Gonna put this "throwing" the indicator thing to rest.
They get built in Japan, packaged, thrown in a truck, bounced around till they reach port, get tossed in a container, lifted onto a freighter, tossed around at sea for a few days, gets unloaded, tossed on a truck, container gets emptied, thrown on more trucks, bounced around for a few hundred miles, thrown on a shelf, thrown BACK on a truck, bounced around some more before it FINALY gets to you.
If it survive the journey to your hands, a 3" slight toss isn't going to do jack. Simmer down.
''Mitutoyo'' might be a Japanese name, it would by far not always ''made in Japan'' in my home Country they have a plant for decades;-)
Arend Wolbers
If it's the Switzerland, Germany area than that's even further away.
Japan isn't far that far from the CA coast 😊
Just because you can doesn't mean you should... My truck has a few dings from the previous owner, it doesn't mean I'm going to intentionally do anything that could cause more.
Joshua
Of course if you can it means you should... we'd be stuck in the Stone Age still if everyone adhered to that 😝
Joshua Stinson True, but I'd think a precise mechanical instrument like the one in the video has some vibration tolerance engineered in. It would be interesting to find out what the design requirements for the test indicator was. Of course if you baby it, the instrument will last longer, but is that worth someone's time? Some would say yes, others no. As an engineer though, I'd say this type of abuse should be accounted for. If not, somebody should be ashamed. :D
I cringed when you threw that indicator lol.
jbbchinook haha me too
jbbchinook yep me too. I was about to make the same comment.
Same lol, but we have one and they seem to hold up pretty well to bumps. Mitutoyo makes pretty good stuff
John, you have to make sure you are concentric and square to the spindle . you have to check multiple depths on the bore. you only checked one area. any run out would be amplified. also get some thin copper to put on your jaws bid
with the tormach holder you may have had more of a cylindrical surface to orientate to the spindle so it was less apparent. if you were only grabbing the outer ring of a cat 40 even a 1 deg kick would deceive you. to remedy try putting the pin back in and get to concentric readings or reindicate the bore to match in two places. also when doing fussy indicating on machine only spin one direction . it takes the slop out and gives a little preload, so the jumps are consistent. happy holidays. keep up the good work.
"This micrometer is amazing and super accurate!" Then tosses it on the bench...
"Dial test indicator"
Right... I keep forgetting. They looks similar to the paper mics we use at work.
Eco Mouse I couldn't believe my eyes
That's how much he trusts mitutoyo lol. They really do make solid measuring equipment.
Eco Mouse that wasn't a micrometer it was a dial indicator
John
The initial movement on the DTI when you rotate the chuck is the headstock bearing rolling up on the lubrication film that's why it stays a tad off ( never returns to zero ) while you are rotating it ,plus the influence of you hand. In the steam turbines I have worked on you can see the shaft rise when they are barred up prior to putting on the main steam . It's oil film ramp
Stuart Hardy shouldn't that ramp in the opposite direction when reversing the spindle? Even by hand?
It's been awhile since the class that included hydrodynamic bearings/effects, and we didn't go into effects of hydrodynamic forces on roller bearings.
Not doubting you at all, just looking to expand my knowledge base.
Wooo!
Thanks for taking the time to measure, sir.
OK, brass tacks.
First, you're using excellent USA made and brand new toolholders that were manufactured recently, with USA made cutting tools that were also manufactured recently. That is the best possible set of circumstances for a test. It only goes downhill from there. Depending on the shank tolerance on your tool and how old your setscrew holder is, it could go way downhill.
Second, like I said on Instagram, half a thou will not cause the machine to explode or the tool to snap off at the holder.
What it WILL do is impact surface finish when side milling. No effect when end milling (more or less), which is what you were doing in the injection molding video.
If you have a two flute tool with one flute a half thou higher than the other, that change in scallop height when sidemilling makes a visible pattern. The more runout the worse the pattern gets. The more flutes, the more complex the pattern gets, but it is still there.
You can't feel it with your fingernails, but you can see it, and for some things that pattern can be a real problem. For instance, I used to make paintball gun parts, which had to be polished to a mirror finish before anodizing, and this visible pattern caused me no end of irritation during polishing. It IS real, because the deeper marks were still there long after the shallower ones had gone.
I fought and fought my process trying to improve the surface finish, and in the end I came up with two solutions that pretty much worked equally well.
1) I bought a milling chuck that was guaranteed to have less than two tenths runout at the tool shank, and they weren't lying. TIR on the tool was just under two tenths installed in the machine spindle. THAT toolholder combined with judicious replacement of tools used for finishing (they would do X parts as a finishing tool, then got demoted to roughing for the rest of their lives) solved the surface finish problem. I had been doing the tool demotion thing before the milling chuck too, but after switching to the milling chuck the parts per finishing tool (that number X) went up about 20%. It wasn't huge, but their life did increase with reduced runout.
2) An oldtimer I respected over at the CNCZone forums suggested that I try a single flute tool, either by simply removing one cutting edge from a two flute, or buying a single flute made for high speed routing for aluminum or plastic. That tool would also leave an acceptable finish, even in a setscrew holder because to it runout was irrelevant. Wherever that one flute ended up in relation to the spindle centerline didn't matter (except as a matter of total cutting OD, but these parts were never splitting a thou tolerances to start with, it was all about surface finish), because where it touched the work, it was always the same distance from center. That tool too solved the surface finish issue.
Cost wise it was about a wash. The milling chuck was about $250 (in 2004) and the good Atrax single flute carbide router mills were about $95 a pop. The router mills were cheaper up front, but they wear out, whereas I still have and use the milling chuck. These were all 1/2" tools, solid carbide, USA made.
I ended up sticking with the milling chuck method because I'd already sunk the money into it, and in the long run it was much cheaper (those were parts that I made for close to five years, so it more than undercut the router mills cost wise over that span), but I kept the single flute trick in my back pocket, and it saved my ass on several occasions later on.
Like I said on Instagram, it DOES matter, but not for everything. You just have to be aware of what's going on with your tooling and use it accordingly.
That's why I suggested that you measure to see for yourself what your toolholders look like.
The setscrew holders I own are terrible runout wise, so I don't use them for things where that will be a problem.
I use them for things like big roughing endmills, or for other single flute tooling where runout doesn't matter (like form tools or half round engravers).
The things I use in setscrew holders most often are indexable endmills that I use for roughing (like that ShearHog you're so proud of). The tolerances on most inserts are so large that whatever the toolholder itself is contributing to the high/low of the flutes radially is just making a poor situation slightly worse, and since I'm roughing, I don't care what the sidemilling finish looks like anyway.
I'd be interested to see what the TIR looks like with that MariTool holder actually in the spindle of your new HAAS machine. That's how I measure all of my tooling, because ultimately that's the only place it matters.
Anyway, thanks again for taking the time to measure, I'm glad you took video of it.
"Amazing that we can have stuff this stuff this day and age for so little money" *SLAM*
John,
Nice demonstration. My takeaway. Measure the tolerances in your tooling if your work requires it, and adjust accordingly (as you demonstrated in the injection mold video). Otherwise I would probably look for error elsewhere.
If you were to repeat this test I would look at doing it on a precision ground v-block on your surface plate. Attach the indicator to your height gauge. The lathe and chuck add a lot of potential error sources.
Thanks for the great video, your insight, and continued contributions to the CNC/machining community.
John, you should have used a dowel pin and measured the runout at the TIP of where the tool end would be, and not at the base of the tool. If the tool is tipped to an angle by the setscrew, the runout at the end of the tool tip would be multiplied. I have normally only heard this runout argument made on small conical engraving tools where the cutting tip is super small, and even a thou or so could make a difference on fine engraving.
Excellent point.
AndTheCorrectAnswerIs Not quite, what you have is a perfectly machined true tube, holding a perfectly machined true shaft. Therefore the set screw won't make the tool tilt, it will push it against one side of the tool holder, but still parallel (or as close as possible) to the tool holder. So, the runout will be equal (or damn close to) regardless of where the indicator is reading from.
+Tom Jackson You're ASSUMING the holder is perfectly machined, AND that the tool is perfectly machined and has no wear.
AndTheCorrectAnswerIs I am, because they are reasonably safe assumptions to make, considering they are precision parts. And if the tool is worn, it will go skiwhiff in a Collet, Jacobs or Set screw chuck. And if the tool holder is off by enough to be significant, it needs binning.
John, I like the concept of checking runout but I wonder if the tape on the jaws isn't allowing the cat40 holder to be skewed to the spindle centerline. you have the nose on centerline but I bet the back of the toolholder is off center resulting in a cylinder that wobbles the farther you get from the true running nose. It may be easier to test this in your mill spindle.
The larger the hole and shaft, the larger the tolerance on the hole and shaft diameters. It doesn't matter whether we're talking tooling or bearings.
The larger the part, the more thermal expansion becomes an issue both in manufacturing the tools and in use. For example, a 2" diameter part heated from 70 degrees to 250 degrees expands twice as much as a 1" diameter part. So when manufacturing a part that is heating up as you cut it, it becomes harder to hold tight tolerances on larger part. Use of coolant and CNC machinery has made that easier, but it still affects the final part.
Thanks,
John
I also cringed at the "toss". However, there were not any other employees around and that would be important as they follow what the boss does, not what he says. Also, I think that the "other" factors that could have mitigated the deviations could, just as likely, made the results worse! Also, tape can not be used when measuring in tenths. Also, when measuring in tenths, you can only do it when the actual cutting setup is used (CAT40 in the spindle of the machine to be used).
NYC CNC videos have been of a great benefit to me as well as the comments. It is generally true that the comments are more helpful to me when a mistake is made.
Great video!
John, When ever you indicate something in a 4-Jaw chuck you also have to indicate the face run out. It might be that you have a close run out at one place in the bore but not in every place in the bore (because of worn chuck jaws or the chuck jaws springing open when you tighten them). So to really get something running true you would have to indicate at the front of the bore than at the back of the bore (You may have to tap the part in the chuck than readjust the jaws and repeat as many time as necessary). You can also indicate the face if it runs true to the bore. So with your end mill test you may be running true in the bore in one place but you move the indicator out on the shank of the end mill and its not running true because the bore not running true in the back due to face run out.Also how tight were you tightening the set screw in your test? As tight as you do to really hold a tool? Usually it been my experience that you (I mean every one not just you) tighten them jest about as tight as you can get it. This does set the tool off center. Some end mill holder manufactures intentionally grind the bore off center to compensate for this.
I work in a shop that also grinds tools. run out can be significant in older, well worn and cheap import holders. Compounded by Chinese mills with shanks that are sometimes undersized by a few tenths.
We know our customers well, and the ones that use old or cheap holders get their tools ground in cheap import holders. If the tools don't have weldon flats, we add them to keep the slop to the same side of the tool when set. This way, the shank can be pushed to the side of a sloppy holder by the set screw, but we grind the cutting edges true in that position. Less effective run out.
You get what you pay for. A high quality holder in good condition with a quality tool set in it will have negligible run out and can be pushed a lot harder without making my machines "sing".
The Command brand ones I get to use at work are usually well within .0005
Measured on a Zoller, HSK100a holders.
If you want little runout as possible: use shrink, hydraulic or powRgrip. Usually all that ends up in an endmill holder is roughing tools.
For your test indicator look up specs on hysteresis. IIRC the compac (tesa) indicator I have is speced at less than 2 tenths (might be 1), but that's a multi rev. Metrology is a wicked business!!
You're also dealing with a not so parallel tool to bore axis if the set screw doesn't split the engagement length of the tool. Not much just off the end of the holder, but it can still be there, and get larger further out.
GOOD NEWS! ---> John, You have a couple of compounding tolerances in what you are measuring. Take the actual tool and/or pin diameter in consideration of the runout that you have measured, plus still a little iffy on the tape, you forgot to cut the mesurement in half for diameter. So your runout appears to be actually 0.00015" on the tormach holder.
Might be fun to mic everything too!
Retry this with the holder in the VM-3 spindle. Check the spindle runout (1 tenth or less likely), then the tool holder bore runout and finally the clamped tool runout. This should give you A) your total runout to compare with the ER and B) if you subtract the first 2, the runout attributable to the setscrew.
Also, you're used to 1.5hp. I absolutely would not use a setscrew holder without a Weldon flat on the tool in a 30hp machine. I've had 1/2" endmills pull out of a perfectly tightened ER32 mid cut well below full power and a setscrew holder without a flat to grab onto has far less pullout resistance. You'll also find that the machine has enough torque to overcome the holding force of the set screw and spin the tool in the holder...ruining the bore in addition to snapping the tool.
Completely unrelated tip: I'm not sure if they changed this with the Next Gen Control, but be VERY careful with the tool offset probe. I had a close call when I forgot to mark in the offset page that a 2in shell mill was not a spot drill. The probing cycle changes based on the type of tool you tell the machine you're using. It plunged straight down for what it thought was a spot drill and the inserts were hovering over the main body of the probe when I hit feed hold. Every time I put in a new tool I would immediately change the offset/probing page and run a probe cycle to prevent any mishaps like that.
Also, with the work offset probe, it's not active if the lights aren't flashing. That means you can crash it during hand jog. The tip is designed to break to protect the super expensive body, but it still costs $100+ to replace the tip.
I think John should retest the cat40 in the VMC I don't think it was cylindricaly aligned or parallel to the lathe there wasn't much to grab onto with the 4 jaw chuck. I think that a thou of runout is quite allot for a 3/4 end mill even if they are not import end mill holders. I thought the set screw run out would be more than a few ten thousands of an inch. Nice video John.
My big lathe has about 3 tenths of error in the spindle bearings, no amount of dialing in will get the error to less than 3 tenths. I only know this because I have used sub-tenth indicators to try and dial in parts to DNO. Making sure that you have at least twice the resolution is important. That means using a 50 millionths indicator to read single tenths, and a tenths indicator to read 2 tenths, and 5 tenths to read 1 thou. A 1 thou indicator gets you within 2 thou.
Ok, here's the bottom line. You have some clearance tolerance built into the tool holder so you can get the tool into it. Unless you are using shrink fit holders, a solid holder will always have a very slight amount of runout due to the clearance. The better the holder and the better the tools you use, the lower the error will be, but it will never be zero. So, yes, the endmill will be offset slightly from the centerline of the holder, unless the maker ground the hole eccentrically to account for the clearance.
I kinda like the description "set screw holder". I knew immediately what you were talking about, although over here we call them Weldon tool holders. I never used them because the tool supplier did not recommend them at the time I first bought tool holders. I have been using ER32 and ER25 collets and they work just fine. Seeing this video I probably never will use Weldon holders.
What strikes me is that you do not use weldon tool bits in this demonstration. Weldon tool holders require btis with a notch in the shaft to prevent the tool from ever pulling itself out of the holder.
I am with you John, the tool holders will have little effect on your finished product. For instance in cutting a slot. If it need to be .5000" Then you would not use a 1/2" cutter to do that anyway. You would use maybe a 3/8". So all you really have going on, is a flute on one side of the cutter taking a slight bit more. Maybe that will result in a tool wearing out sooner... who knows!
To me the finished part xize, and what it measures; after it is cut with any tool... is what is important. Thus if you have a cutter, that is cutting to large a slot in this instance; possibly form run out. Or lack of any machine tolerance. You simply adjust the size of your tool to reflect this in you CAM software. Tell the system the cutter is .5005" in diameter, or tell the machine the slot should be .4995". There are many ways to "fool" the system, so that the finished part is what you get.
I highly doubt, you will ever get a 1/2 slot, being cut with a .5000" cutter. Even on a brand new machine. Because of all the things you are bringing up. However, with a few brains upstairs; all that can for sure be overcome.
My machine shop teacher back in 1979, said this: "It is not the machine, no matter how old; it is the machinist!".
Thus you can put a person, with little to no experience on the new Hass, and they will not be able to make good or fitted parts. They just don't understand machining.
However, you put as experience Machinist, on a lathe made in the 1870's; and somehow, by some absolute magic, that person can pump out good accurate parts all day long.
So again I agree, the accuracy of the tool; and its tool holder; on a slow machine. With larger cutters, don't mean much, except in the final part size. However on a new High Speed spindle, will most likely reduce vibration and such... thus being more important, to have tight tolerances.
As when using small cutters, .060" or smaller; real easy to break on sloppy or old machines, I have experienced that.
Enjoyed the video! Thanks.
The tool holder is centered with respect to the ground spindle mating surface. So you would have to insert this in a spindle or something that can hold on to the taper. That outside part of the tool holder doesn't need to be machined with very tight tolerances compared to the spindle mating surfaces and the where the tool attaches. That might account for the big swing in dial indicator.
I will agree with most of the other comments, that this should be done with vee blocks on the surface plate. Also even 20 years ago, I remember the best tool holder makers talking about grinding the bore offset from the axis to compensate for the set screw. I measured about 3 tenths of intentional offset in their brand new 3/8" holder at the time. So you can't dial in the bore, you need to dial in the spindle end, not the bore. I guess these days the real winner is hydraulic holders or shrink fit. I've never had the chance to use them, but I've done other shrink fits like piston pins into rods, and I can imagine how perfect and rigid that would be in a tool holder.
I'm pretty sure, if John did this video after he had his haas he would have not only talked about it, but would have used it to do this test. he uploaded a back up video for times of busy, clearly the man is busy, which is great!
ps. I'm really looking forward to seeing you use your haas!
You need to make sure not only that your bore is on center but that it's also in there straight along the the length of the bore because you may not be in there straight along the axis of the tool holder. Somebody mentioned putting it into the spindle and that is actually probly the easiest way to do that since the spindle is precision ground and new so you will most likely have very little if any play in the spindle. Interested to see if that straightens it out and give better readings.
you were trying to use a "very used" 4 jaw Chuck as a perfectly flat surface and we're adding run-out to the part. you should have used a longer gauge pin tightened into the holder and measured at both ends to make sure the holder was perfectly aligned to the centerline of the lathe.
At 10:43 if you mark the high and low positions of setscrew chuck, then you can know whether to add or subtract that from your measurement of the deflection of the tool when clamped with the setscrew.
The problem may be be runout when you're actually measuring and need to measure concentricity. Also if the video is about using a collet vs a set screw holder you should do the same test with a collet. Maybe the collet would will be even worse. Who knows?
I've heard that the real problem is when the tool is under pressure. Being held by one screw versus the whole circumference effects how it flexes during a cut. The collet and endmill will both probably wear more evenly then the setscrew holder or it's endmill. The setscrew holder might be more rigid though.
Does it matter if it's a weldon shank?
I think you're probably right however. People have all sorts of superstitions when it comes to machining. I think the time you save with setscrew holders probably nudges out the tiny possible downsides for almost all jobs.
The tolerance of AccuHolders is .0002" on the ID, measured with ten-thousandths carbide gage pins, and the "hand feel" is everything ;) Add the usual .0002" of clearance on tool shank OD, plus or minus a bit of chance, and there go your three tenths
Here in Seattle the aerospace shops don't like ER collects for milling on CNC machining centers they like the set screw end mill holders or for 0 zero run out they use milling chucks they say the ER collects fail sometimes on long runs with heavy cuts
The second tool holder was likely slightly cocked in the 4 jaw making the tool magnify that error. Basically the bore wasn't straight with the axis if the machine. But who knows when dealing with such a tiny thing. Great videos. Thx.
It is very possible that the tool is orbiting so you have to check multiple places to insure the bore is perpendicular to the bore of the lathe. That’s where cmm’s rock. Happy checking.
For doing general purpose machining, not mold making or hard milling, I haven't really had any runout issues with set screw holders in VMCs. Grind a set screw flat on you end mill and lock her down and she probably won't pull out while you're roughing.
I've had more run out issues with beat up collets and cracked collet holders than anything. Good when they're but tend to go to hell after a few years.
just an idea. The manufacturer must give a certain amount of clearance in the chuck bore on purpose for ease of tool insertion. since it is safe to assume this clearance is quantifiable, maybe they offset the center of the bore towards the set screw by half of the given clearance value. this would allow the tool to be concentric to the axes of rotation when screw is tight. I would love to see this test done by zeroing in on the toolholder shaft instead of the bore.
12:20 OK, you were surprised at +6 -1 But the point is that when you tightened the set screw, the +6 -1 did not change. So wherever your error was coming from, setscrew tension wasn't affecting it. For the actual tool, tighening the setscrew reduced the total runout.
So the setscrew collet-held tool had runout roughly similar to the collet alone. Would be interesting to mark where the high point was for the collet alone to see how it compares to the high point of the held tool. Really neat.
You should take in consideration the heat from the machining process. It could reduce the play between the tool and the holder.
And use a end mill that is designd to be used with that holder (weldon ) the shank could be grinded at a specific tolerance for the holder.
Just my thoughts after opening my eyes when you tossed the indicator ;(
Why not just insert the holders into the mill spindle to check runout? That would also include any tool holder to spindle runout possibilities.
After my 20+ years as a manual machinist I never found an endmill less than a thou out unless it was heat shranked.
When the Haas is up and running measure the 0.75" solid tool holder in the spindle. I bet you'll get better results. I use 1/8" solid holders and I only get a few tenths of run-out.
Test it in your Haas VM3 when you get cozy with that
My Maritool holders typically show 1 to 2 tenths max when chucked up. Many show 0 tenths, the needle quivers and that's it.
Not sure I'm buying the other sources of runout error theory. Sure, in principle, that sounds reasonable, but your test showed that the Tormach tool holder runout was exactly where it would be if it was entirely attributed to the set screw biasing the tool. Given the results of this one test, I'd be inclined to say that the error is almost entirely the result of the set screw holder. It would be just as valid to say that the lathe chuck and other factors produced .0005" of error in the other direction, and the set screw runout error was actually .0008", but that error was reduced by .0005" by the opposing runout error of the chuck.
IMO, dialing in the ID of the tool holder would null any other setup errors. Of course, in real world use, those other sources of error would not be nulled, so spindle runout, runout between the tool holder shank and the tool holder ID, etc. would all be in play.
I'm a pud knockin' home gamer, and this looks like fairly significant runout to me, particularly on a smaller end mill. Not end of the world, but not insignificant.
I will admit that the rigidity of the set screw tool holder is typically better than an ER20 collet tool holder, particularly for aggressive cuts, and that's certainly an important point.
I bet Maritool hand picks all the stuff they send you in their metrology department from now on! Comparing a 1/2" import tool holder to the vurra nice looking 3/4" Maritool tool holder isn't exactly an apples to apples comparison, but 1/2" to 3/4" isn't so much of a difference that the Maritool should have had .0016" of runout when the Tormach had .0003" of runout.
It might be possible that while you dialed in the front of the hole, the rest of the toolholder's bore might not be in line with it (ground crooked), and that the tool sits at an angle into the toolholder. It would be interesting to see how the measurements are affected further down the gauge pin's length, say at 1" distance from the chuck, if the measurements get amplified down the shaft, that would be your answer. Ciao, Marco.
To all the people commenting on the indicator that he "threw" on the table... that indicator didn't even care about that very slight drop from a few inches high. It gets way more beat up in the mail on its way to the distributor, then again from the distributor to you. It will be perfectly fine. (also he didn't throw it)
Hi the type of tool holder you have is Weldon holder .
A god weldon holder can not be to cold for mounting a cutter in it's not shrink chuck but it's where narrow in tolerances.
I normally heated white a hot air gun to 50celcius
Hey... great video as always. I reckon your toolholder is not parallel withe the spindle centerline. So while the runout may be 0 at a SPECIFIC point on the holder, that WILL change down the length. You should probably run the indicator up and down the gauge pin in 2 axes to get it parralal first but then I guess you are skewing results again. Needs to be in a proper spindle I guess... try it in your brand spanking new Haas 😀
the bigger the tool the larger tolerances, if you're truly worried about run out and this is for all your viewers swap to shrink fit if you can mortgage your house. we use them almost exclusively especially in the mag 3 at 33k spindle .00005-.0001 tir
Great content!! We are all awaiting that chip stream from the mighty VMC!! 👍🏻
I believe the tolerance for the ID of Kennametal "shrinker" tool holders is + -.00015. I don't remember exactly. It could be even tighter than that.
I would suggest clamping the holder in a v block with the set screw at 90 degrees to the orientation of the v groove, and then "flip flop" the v block on a surface plate and measure with a height gauge. This would show you exactly how much error there is in the plane of the set screw
Another thing would be to manually probe your tool radius at each flute and see if it matches with the test here. Any good probing subprogram will run the tool backwards and measure from the highest point anyways.
The real test is to chuck that holder in the Tormach, and very lightly tighten the setscrew on a gage pin or tool shank. Measure the runout on the tool shank. Note the relation of the runout to the setscrew location. Now crank down on the setscrew to normal torque and measure the runout again. Now you know the inherent runout in the holder, and you know any distortion from tightening the setscrew.
What if they grind the center bore slightly off center to compensate for the tool being pushed over I did not see you measure for that thanks for the videos.
Mason
The error in the reading was most likely due to the tool not being coaxial. Also the reason I do not like Weldon holders is the fact that you are pressing a carbide cylinder against a slightly larger bore; so the contact is a single line the length of the tool inside the holder and one or two contacts from the screws. Another major concern with Weldon holders is they should only be used with tools that have flats milled/ground into them; otherwise the gripping force is insufficient for a 30hp spindle and the tool will spin in the holder, possibly pulling out, pushing in, or welding itself in place.
all the older guys tap the indicator base and the arms after dialing onto zero. It gets that initial bogus reading out so you don't have to go around 3 - 5 times until the indicator settles in.
The issue isn't really so much that the tool will push it out. The surface area takes the pressure fine you shouldn't get any run out. Any deformation would take place in the screw since it's taking the same amount of pressure but on a lot smaller surface. Beyond that the main issue is they begin the run out relatively quickly compared to a milling chuck that lasts quite a while just replacing the sleeves. Same with ER collets but I don't run much bigger than 3/8' endmills in an er collet.
That up-down-up-down that doesn't follow a nice orbit, that's your spindle bearings. You see a high spot, a low spot, then a mid-high spot, then low, then high. According to your indicator, your spindle bearings are 1-2 tenths out.
I use these Set Screw Holders (BT40) all the time, also for finishing in inox steel, and can mill in 0.02mm! easily and reapetedly :D
Then for de ER collets (especially on SK40) you shouldn't be concerned with the rigidity. For example I use ER16 150mm long chucks on a 5 axis machine and really didnt have any issues.
Regards Sam (CH)
Use aluminum 1/8 thick by .75 by .5 between part and each jaw. Hope you read this and try it. It will protect your parts from jaws better than tape.
You should use aluminum shims so you don't damage existing parts..
Ive done these tests years ago, and came to the same conclusion. Every time, there was a difference. Just remember that even two tens means that one tooth is cutting more deeply than the other, and can lead to four tens larger diameter cut.
Realistically, we can’t really get it perfect, though, if we’re lucky, errors cancel out. Most likely they cancel out partially. But, cheap holders can have close to 0.001” runout. The best; 0.0002”.
If you want to drive yourself crazy, you can rotate the tool in the holder while measuring runout, and marking holder and tool, always put the tool in on the marks. I do that with some things.
With the holders in the four jaw, you’re squeezing the holders from four directions. Even slight pressure can distort that thick holder. Remember we’re talking about just a few tens.
The answer is yes the tool will move some when tightened as the slack in the bore is taken up. Some holders are ground with a bias in the bore to allow for that. Like Briney tru holders www.brineytooling.com/files/flyers/brineytrupos.png The issue is how much does it really matter. If your feed rates are .002 tooth load you will end up with actual .001 to .003 chip loading depending on what side of the tool you are on. I am sure this effects tool life but I can't tell doing what I do with my tools in a job shop but what it does do is effect finish. The greater your tool run out the slower feed you will have to run to get a good finish. Hydraulic holders like my Schunks are the easiest and best solution to this for most applications as they are more rigid and repeatable than collets but for maximum rigidity and run out you have to go with a shrink fit tooling setup. You need to check run out with the tool in the spindle where it matters. I have had tooling with run outs up to .003 on the big stuff and it still cuts great and will hog off material as good as a true running tool but again when you need the finish you will be running half speed.
the offset comes from the screw flexing the body, not the diametrical difference. Thats also why they make facemills where each adjustable insert can be dialed in.
I would like to see a comparison with the set screw holder and the clamping tool holder (ist this the right name for it?)! If you do a remake with the mill spindle that would be awesome! I know this discussion of which tool holder to use very well.
Best regards from Germany! :-)
Wasn’t there an update where he talked to frank from maritool about how they measure their runout and he put the cat40 holder in his VM3 and re-measured the runout? I was re-watching this because I couldn’t remember what the runout was when it was tested properly...
great video John.
I'll have to check the runout on the set screw holders I have for my Atlas mill.
remember your Atlas mill?
That around 1 tenth equals 0,01mm pretty precisely. A runout like that does matter when doing e.g. a 0,03mm cut on a 2-flute end mill. The other one is taking 0,04 and the other one 0,02 chips. Also the spiral flute will change it's cutting when it rotates around the end mill and the center is not rotating around itself.
Just catching up on all your videos now. I'm a bit behind at the moment. The CAT 40 would have been hard to hold. I liked your note at the top of the screen. Should keep the keyboard warriors at bay. Keep up the great work John. cheers, Aaron.
OK, brass tacks:
First, you're using excellent USA made and brand new toolholders that were manufactured recently, with USA made cutting tools that were also manufactured recently. That is the best possible set of circumstances for a test. It only goes downhill from there. Depending on the shank tolerance on your tool and how old your setscrew holder is, it could go way downhill.
Second, like I said on Instagram, half a thou will not cause the machine to explode or the tool to snap off at the holder.
What it WILL do is impact surface finish when side milling. No effect when end milling (more or less), which is what you were doing in the injection molding video.
If you have a two flute tool with one flute a half thou higher than the other, that change in scallop height when sidemilling makes a visible pattern. The more runout the worse the pattern gets. The more flutes, the more complex the pattern gets, but it is still there.
You can't feel it with your fingernails, but you can see it, and for some things that pattern can be a real problem. For instance, I used to make paintball gun parts, which had to be polished to a mirror finish before anodizing, and this visible pattern caused me no end of irritation during polishing. It IS real, because the deeper marks were still there long after the shallower ones had gone.
I fought and fought my process trying to improve the surface finish, and in the end I came up with two solutions that pretty much worked equally well:
1) I bought a milling chuck that was guaranteed to have less than two tenths runout at the tool shank, and they weren't lying. TIR on the tool was just under two tenths installed in the machine spindle. THAT toolholder combined with judicious replacement of tools used for finishing (they would do X parts as a finishing tool, then got demoted
to roughing for the rest of their lives) solved the surface finish problem. I had been doing the tool demotion thing before the milling chuck too, but after switching to the milling chuck the parts per finishing tool (that number X) went up about 20%. It wasn't huge, but their life did increase with reduced runout.
2) An oldtimer I respected over at the CNCZone forums suggested that I try a single flute tool, either by simply removing one cutting edge from a two flute, or buying a single flute made for high speed routing for aluminum or plastic. That tool would also leave an acceptable finish, even in a setscrew holder because to it runout was irrelevant. Wherever that one flute ended up in relation to the spindle centerline didn't matter (except as a matter of total cutting OD, but these parts were never splitting a thou tolerances to start with, it was all about surface finish), because where it touched the work, it was always the same distance from center. That tool too solved the surface finish issue.
Cost wise it was about a wash. The milling chuck was about $250 (in 2004) and the good Atrax single flute carbide router mills were about $95 a pop. The router mills were cheaper up front, but they wear out, whereas I still have and use the milling chuck. These were all 1/2" tools, solid carbide, USA made.
I ended up sticking with the milling chuck method because I'd already sunk the money into it, and in the long run it was much cheaper (those were parts that I made for close to five years, so it more than undercut the router mills cost wise over that span), but I kept the single flute trick in my back pocket, and it saved my ass on several occasions later on.
Like I said on Instagram, it DOES matter, but not for everything. You just have to be aware of what's going on with your tooling and use it accordingly.
That's why I suggested that you measure to see for yourself what your toolholders look like.
The setscrew holders I own are terrible runout wise, so I don't use them for things where that will be a problem.
I use them for things like big roughing endmills, or for other single flute tooling where runout doesn't matter (like form tools or half round engravers).
The tools I put in setscrew holders most often are indexable endmills that I use for roughing (like that ShearHog you're so proud of, and rightly so). The tolerances on most inserts are so large that whatever the toolholder itself is contributing to the high/low of the flutes radially is just making an already poor situation slightly worse, and since I'm roughing, I don't care what the sidemilling finish looks like anyway.
I'd be interested to see what the TIR looks like with that MariTool holder actually in the spindle of your new HAAS machine. That's how I measure all of my tooling, because ultimately that's the only place it matters.
Anyway, thanks again for taking the time to measure, I'm glad you took video of it.
I'm not sure what the deal is here. I tried to edit my original comment, but instead ended up posting a new comment, and now I can't edit THAT comment either. lol WTF UA-cam?
Hi John.
interesting topic, great video. A couple things that I was wondering about: Did you test the cylindrical error of your test gages or milling cutters? IIRC the +/- .0002" for over or under is the size (diameter) deviation from nominal on the gauge pins, not the tolerance for cylindrical error which depends on the class (grade) of the pin. What about end shake without the grub screw tightened? a good slip fit (that pops) can be anywhere from a couple tenths to almost a thou in my experience.
What about a sweep of the hole in the holder to see if it's out of round? What about tolerance stacking?
inquiring minds want to know!
The Amount of Deflection should equal the amount of the Clearance between the tool and the tool holder.
longrangehunter is on it I think. being parallel to the spindle is key. You could easily dial in to .0003 at one set point without notice a 1 or 2 degree kant of the tool holder in relation to the lathe spindle. End result being that when you meassure the tool it will swing on both side of zero as the median cross over is where you zeroed it.
when u are trying to tighten something precise in 4-jaw u need to dial in in multiple places across the hole. soc the hole might be dialed perfectly on one spot but further down it might have run out.
I think the reason you got more runout with the .750 endmill is because it had to be a little bit crooked, and the further your the tool is stuck out, the more runout you will have. Even if you were able to get it very close with just the tool holder, I think that it was slightly crooked not in terms of concentricity but with the angle of the holder. nothing you could have improved on, but it had to have not been perfect just given the setup. I think if you could get the holder exactly straight with the lathe, the results would be different
Hi John
would be interesting to see the test in the Tormach and in the Haas
cheers
I'd be curious to see if the runout on your mill changes before and after you turn it 90 degrees in the holder. I'm not sure which error that would resolve but perhaps lower runout is possible with a little tweaking.
Good stuff. I'm sure someone with $50k in metrology equipment is going to pipe up, and hopefully they'll make a video too :)
It will push it off center, but only the clearence between the end mill outside diameter and the tool holder's inside diameter, so just a few tenths or less.
Be careful on a lathe turning between centers and using a bent tail, drive dog. Use a straight tail, lathe dog if possible. I have seen the leverage of the bent tail ones push the shaft off center and cause runout.
Size for size doesn't fit. So technically set screw pushes it over a few microns. Set screws are nice, cheap, and cant pull out. I use them, but any tools over 5-6" its better in a collet. Collets are better cause you can make the tool any length. At my job i mostly use Nikkon holders and put a shrink fit extension in it. That is really nice, expensive but worth it. I can't stand the people saying a reamer should be in a collet. Drill chuck for me. Video of that would be nice. I do use collet over1" reamer though
My issue with set screw holders is not pushing it over to one side but when name not mentioned inserts an endmill without a flat ground on the shank can cause pull out, especially in hard materials and titanium.
makun16 He will learn when he starts running an end mill larger than 3/16 dia. and can't hold tolerances with the HAAS.
Never had problems holdin tolerances on any Haas machine. It's all in the machinist.
I've never had any issues with all my years of programing and running Haas machines. I am really impressed with them. It's all about setup and selecting the right tools for the job. I was referring to a manager at a job shop who incorrectly used the wrong tool holder and even pointing that out to him, said run the f**en thing! Needles to say, his ignorance destroyed a set of hard jaws.
Mounting a tapered shaft holder with tape around the short distance of the ring at the base of the taper could result in the holder being slightly out of alignment of the headstock axis and holder axis. While it may not be much, the variation you are measuring are not big.
Wouldn't it be more accurate to mount the tapered holder in a matching tapered machine spindle. That would give a real indicator of both machine accuracy and potential holder accuracies. Lop
Probably a reason he didn't show the holders in a mill and that's because Tormachs have a bit of runout. I picked up a 96 Brigeport VMC for $3,500 and added Shars ER32 collets & CT40 holders. Mill ran under 0.0001" runout @ taper and the ER32 setup would maintain 2 or 3 tenths tops @ end mill shank. Backlash on all 3 axis was under 2 tenths on a 20 year old mill that can step @ 0.00006" over 20 years ago. Expecting a Tormach to hold 1 or 2 tenths on the holders isn't exactly unreasonable in this day and age.
The runout = the difference between the Arbor bore diameter and the shank diameter of the cutter.
A bore Mic and an external Mic could have given you a difinitive answer without all that messing about.
The real point here, is that if you get a slightly undersized cutter shank, on a Collet chuck it will always spin true. The other thing, is that Grub screw puts some unwanted pressure on the shank of the cutter, I have seen them snap before if they are not heat treated correctly..
If the Cat 40 tool holder isn't inline with the 4-jaw along its center axis then the error will compound the further out you measure from the tool holder.
when you center a bore; measure in 2 places. one close to the edge and 1 about an inch deep. this wil make sure you have no runout over longer distances.
i allways wondered about this! Glad i found this video now.
As long as that end mill comes around with the same runout every rotation it will cut straight. Unless you are cutting a slot the exact same size as the end mill you’re using, it will be fine.
this is fine on 2 flt and 3 flt under say 6000 rpm, jump to 4 - 6 flt and some higher speeds and feeds and that difference may start to show problems. get up to 10+ and beyond the collet system should be used, once u get into 12000 and on shrink fit tooling and some type of harmonic system should be used like blue swarf which will give you the optimum rpm ranges for chatter free cutting at the appropriate speeds and feeds
out of curiosity, whens the last ime you checked your lathe spindle itself? all errors 'stack' and for an accurate baseline you need to know the error your starting with, if any.
Just set the R8 adaptor on a Vblock on a granite surface plate.
I haven’t made it all the way through the video so I’m just questioning as i go but, even if there is a measurable precise amount of run out, couldnt you fix that in a four jaw?
Always tram your tool at tip of tool, longer tool means more runout, checking at tool holder isn't knowing what that tool is doing at tip. Used to use cellphone on taper to compensate, the closer you get it running true at end of tool the better the finish, and you'll be able to run faster with a trammed tool.
Love my Noga double arm