Thanks Derek, I have to say its turned out pretty accurate I am still surprised how well things can be made in a home workshop with relatively simple equipment. I have made a wireless optical version circuit for this but I am holding off at the moment as I may have another way to make a sensor that could be even more accurate with less precision to make it. Thanks for the comments its always good to hear what people think. Cheers.
Thanks for taking the time to make this video series -- very informative and well presented. The steps you went through to characterize your probe (and tips) is thorough and well described. I'm in the process of CNCing my mill and just ordered one of the $60 chinese variants of the probe you made - once I get it setup I plan to run through your procedure for characterization. It will be interesting to see how well those probes compare. Thanks again.
I’m so glad that Planet CNC is making the measurement/calibration of the probe automatically. You only need a 20€ 30mm calibration ring and a script will do all calculations.
That’s interesting I have had a look at planet CNC’s website and can’t find any information about it. Sounds like a really great idea. I hope to revisit this project I have some ideas I would like to try out. I am tied up with my new workshop probably until the new year. Cheers
@@CazualHaze yeah you have to turn the probe 180* one time and it figures out how off center the probe is too. It’s a great controller software. Gantry squaring, a lot of probe scripts, tool wear compensation, backlash compensation etc.
Really interesting stuff. I love how thoroughly you quanitifed the error, it's really cool that the lobe shape is visible in the plot. I wonder if you could get higher precision out of the probe by marking a dot on the stylus and manually rotating the probe so that it always contacts the work piece from the same direction? I.e. rotate the dot to face the thing you're measuring. Probably not worth it if the probe precision already exceeds the mill's repeatability. Thanks for making these videos, your channel is rapidly becoming one of my favourites, I wish you great success with the whole UA-cam thing.
Thanks James I think if you could rotate the probe so it only measured in one direction you could get some really accurate measurements. If you did this you could also I guess make a simpler probe (one set of ball contacts). But as you say the probe looks to be more accurate then the mill anyway! Thanks for the kind words I enjoy making these videos. Cheers
Working in a metrology department I can tell you that we datum our probes 2 or 3 times a day. If a probe hasn't been used, or had been swapped out of the mount it is datumed before use. The idea that you don't have to calibrate your system is fine if you're not looking for micron accuracy. Yes I have forced the company to scrap thousands of posts for being 9 microns out of spec, that was fun explaining that.
thanks #grapes008 I get that its clearly the difference between a bloke with a hobby and someone needing to hit a precise tolerance. The more I get into the higher spec stuff the more I get bogged down in the rabbit hole. It is fascinating the lengths you have to go to to get things made really accurately. Cheers
This may be a stupid question, but would an unsprung metal probe act as a probe, but without the error of the spring and contacts. Assuming you are machining metal, the probe could just touch off the part itself, and become electrical continuous between the part and probe.
Thanks @The Orange Baron, Yep in theory you dont need a sprung probe but your machine will bend it. I made one and that's exactly what happened! I think the best ones use a strain gauges instead of the kinematic ball arrangement. So a very light touch can be picked up. Maybe you could also use a microphone arrangement to listen to the probe touching the part, it would be sensitive but prone to picking up the machine moving. Cheers
Would it have been better to clamp on the inner race of the bearing rather than the outer race, thus eliminating any inner race looseness in the bearing?
probably! I really ought to get a proper standard for this testing something known good and accurate. I still find it amazing you can get down to the microns in a home shop! Cheers
That's great stuff and a subject close to my own heart. I didn't see any info on what you use for CAM, but if it's Fusion360, would you be interested in macros and a post-processor to integrate probing with the CAM? I've done all that stuff both for Mach3 and for PathPilot for the Tormach users.
Touch probe with three contact points by 120 degree separations construction inherently have errors differently in four sides measurements. Using four different compensation values will solve it. I wrote program to find them out and make use of them by locating them in routine probe touching program procedures (may be my probe is not so accurate too, with adjusted the three setting screws, accurate result is not got every time). For the last part of your video mentions, l prefer to have a reference protruded block surface on the machine table or on toolchanger box, every time the probe(or new tip) is brought to use, there is program automatically bring it touch it first to set z zero. You got the Mach3(or Mach4)probing interface and probing program, Is it a free download one?
Blum do a gauge ring rest. Then spin the probe 180 degrees and re run it. This allows you to estimate the "runout" of the probe. Also the offset of the spindle from zero Their calibration macros are interesting in that they do a single touch and this gives you a combination estimate of trigger distance and probe signalling delay. As long as you probe at the same speed every time it remains micron accurate (eg I probe at 1,000mm/min and see micron repeatability) Finally note that you measured your table top to spindle, so you can find the length of the probe by just probing the table? Seems faster than even measuring stuff with calipers? No?
It does look like there are several ways to calibrate one of these probes. Keeping the speed constant is a really good way to keep accuracy. I did buy a digital height gauge to measure the tool lengths but I got one with only0.05mm accuracy so it was a waste of cash!! Cheers
@@CazualHaze I recently bought one of the micron Resolution 0.001mm mechanical toolsetters for adjusting heights of inserts on a facemill. I can't speak for how accurate it is, but all three numbers move and repeat (more or less) as I spin the facemill. So it's good enough for relative measurements.
If you put your rough stock dimensions into the Vers.by wizard then press the corresponding button (inside circle. Outside circle, length etc) then it will measure it for you. Cheers
Most interesting videos I have watched in a while, thanks for sharing both the video and your thoughts.
Thanks Derek,
I have to say its turned out pretty accurate I am still surprised how well things can be made in a home workshop with relatively simple equipment. I have made a wireless optical version circuit for this but I am holding off at the moment as I may have another way to make a sensor that could be even more accurate with less precision to make it.
Thanks for the comments its always good to hear what people think.
Cheers.
This is pure gold!
In a shed, with reasonable equipment. “Only” takes a methodical operator and time. 😉 Clever!👌
Glad you liked it!
Cheers
Thanks for taking the time to make this video series -- very informative and well presented. The steps you went through to characterize your probe (and tips) is thorough and well described. I'm in the process of CNCing my mill and just ordered one of the $60 chinese variants of the probe you made - once I get it setup I plan to run through your procedure for characterization. It will be interesting to see how well those probes compare. Thanks again.
Glad you enjoyed it! let us know what you get in terms of repeatability and accuracy it would be interesting to know.
Cheers
Super FKN entertaining video, work well documented and the whole process shown. Love you mofo!
Glad you enjoyed!
That was a good addition to making video, like that systematic approach
Cheers
I’m so glad that Planet CNC is making the measurement/calibration of the probe automatically. You only need a 20€ 30mm calibration ring and a script will do all calculations.
That’s interesting I have had a look at planet CNC’s website and can’t find any information about it. Sounds like a really great idea. I hope to revisit this project I have some ideas I would like to try out. I am tied up with my new workshop probably until the new year.
Cheers
@@CazualHaze yeah you have to turn the probe 180* one time and it figures out how off center the probe is too. It’s a great controller software. Gantry squaring, a lot of probe scripts, tool wear compensation, backlash compensation etc.
I’ll take a closer look at it I need a new controller for my old laser cutter
Cheers
Good work. We shared this one on our forum alongside your previous related vids 😎
Awesome, thank you!
With that probe and a DIY rectifying machine I believe you could make your Mill so much accurate.
Cheers keep watching I think I can do better!
Really interesting stuff. I love how thoroughly you quanitifed the error, it's really cool that the lobe shape is visible in the plot. I wonder if you could get higher precision out of the probe by marking a dot on the stylus and manually rotating the probe so that it always contacts the work piece from the same direction? I.e. rotate the dot to face the thing you're measuring. Probably not worth it if the probe precision already exceeds the mill's repeatability. Thanks for making these videos, your channel is rapidly becoming one of my favourites, I wish you great success with the whole UA-cam thing.
Thanks James
I think if you could rotate the probe so it only measured in one direction you could get some really accurate measurements. If you did this you could also I guess make a simpler probe (one set of ball contacts). But as you say the probe looks to be more accurate then the mill anyway!
Thanks for the kind words I enjoy making these videos.
Cheers
Working in a metrology department I can tell you that we datum our probes 2 or 3 times a day. If a probe hasn't been used, or had been swapped out of the mount it is datumed before use. The idea that you don't have to calibrate your system is fine if you're not looking for micron accuracy. Yes I have forced the company to scrap thousands of posts for being 9 microns out of spec, that was fun explaining that.
thanks #grapes008 I get that its clearly the difference between a bloke with a hobby and someone needing to hit a precise tolerance. The more I get into the higher spec stuff the more I get bogged down in the rabbit hole. It is fascinating the lengths you have to go to to get things made really accurately.
Cheers
Thanks for sharing. I have a Renishaw head that I picked up for $100 a few years ago that I need to get wired up to Mach3.
Thanks Keith, you did well to get a proper probe for so little!
Wiring up is very easy tis worth the effort.
Cheers
This may be a stupid question, but would an unsprung metal probe act as a probe, but without the error of the spring and contacts. Assuming you are machining metal, the probe could just touch off the part itself, and become electrical continuous between the part and probe.
Thanks @The Orange Baron,
Yep in theory you dont need a sprung probe but your machine will bend it. I made one and that's exactly what happened! I think the best ones use a strain gauges instead of the kinematic ball arrangement. So a very light touch can be picked up. Maybe you could also use a microphone arrangement to listen to the probe touching the part, it would be sensitive but prone to picking up the machine moving.
Cheers
@@CazualHaze Neat, I understand, thanks for they reply!
Would it have been better to clamp on the inner race of the bearing rather than the outer race, thus eliminating any inner race looseness in the bearing?
probably! I really ought to get a proper standard for this testing something known good and accurate.
I still find it amazing you can get down to the microns in a home shop!
Cheers
That's great stuff and a subject close to my own heart.
I didn't see any info on what you use for CAM, but if it's Fusion360, would you be interested in macros and a post-processor to integrate probing with the CAM? I've done all that stuff both for Mach3 and for PathPilot for the Tormach users.
Thanks David,
That sounds interesting to have a look at.
My mail is cazual.haze@gmail.com
Cheers
Touch probe with three contact points by 120 degree separations construction inherently have errors differently in four sides measurements. Using four different compensation values will solve it. I wrote program to find them out and make use of them by locating them in routine probe touching program procedures (may be my probe is not so accurate too, with adjusted the three setting screws, accurate result is not got every time).
For the last part of your video mentions, l prefer to have a reference protruded block surface on the machine table or on toolchanger box, every time the probe(or new tip) is brought to use, there is program automatically bring it touch it first to set z zero.
You got the Mach3(or Mach4)probing interface and probing program, Is it a free download one?
Thanks @ray-charc3131
I did not know there was an auto tool height routine available I will try and get it loaded into my mill.
Cheers
Blum do a gauge ring rest. Then spin the probe 180 degrees and re run it. This allows you to estimate the "runout" of the probe. Also the offset of the spindle from zero
Their calibration macros are interesting in that they do a single touch and this gives you a combination estimate of trigger distance and probe signalling delay. As long as you probe at the same speed every time it remains micron accurate (eg I probe at 1,000mm/min and see micron repeatability)
Finally note that you measured your table top to spindle, so you can find the length of the probe by just probing the table? Seems faster than even measuring stuff with calipers? No?
It does look like there are several ways to calibrate one of these probes. Keeping the speed constant is a really good way to keep accuracy. I did buy a digital height gauge to measure the tool lengths but I got one with only0.05mm accuracy so it was a waste of cash!!
Cheers
@@CazualHaze I recently bought one of the micron Resolution 0.001mm mechanical toolsetters for adjusting heights of inserts on a facemill. I can't speak for how accurate it is, but all three numbers move and repeat (more or less) as I spin the facemill. So it's good enough for relative measurements.
Sounds good repeatability is key for me!
how i get that tool to size the material ?
If you put your rough stock dimensions into the Vers.by wizard then press the corresponding button (inside circle. Outside circle, length etc) then it will measure it for you.
Cheers