Lower layer height is weaker because you were increasing the number of layer bonds available and layer bonds are always your weak point in 3D printing so if anything you need to increase your layer height not decrease it
craazyy22 this is only partially true lower layer height is always weaker no exceptions the week is part of a 3D print is the bonding between layers higher temperatures to a point and the kneeling to a point will reduce how much more we can take but it will still be weaker than the same exact process with a thicker layer height
I believe that now, but I really didn't understand how that was possible before yesterday, i'm running with the assumption that a thicker layer has more thermal mass and can warm the previous layer more and bond to it better. along with there being more bonds with thinner layers, it can only be as strong as it's weakest layer, right? Otherwise it just doesn't make sense to me that thinner layers are weaker, I kinda thought that a thinner layer would be mashing the nozzle into the part, helping to fill any tiny little cracks and possibly radiate some heat onto the previous layer from the nozzle, apparently that isn't the whole story though.
@@GearDownForWhat layer bonds are the weak points but layers only go one direction on a prints part so it is possible to rotate something in this case 90 degrees and print it vertically that way the force on the gear teeth isn't opposing the layer lines, would should increase the strength of the gears. But even if you do the disc would probably only last a couple minutes still.
Most metal 3D printers are more convolution than they are worth. It's still a multi-step process. In which case just use standard FDM to produce a mold positive, and sand cast the part that you need.
I wonder how much longer it would last if you changed the orientation of the part with respect to layer lines (the teeth will be their weakest printed like that) , or if you annelied it (or both) ? Rotating it 90 degrees would be overkill but id be interested to see if tilting it 30 degrees or so would make much difference.
Gear Down For What? It depends on how it's held together inside the pumpkin imagine if you were to go inside your pumpkin with a laser and slice it in half without hitting any bolts imagine that theoretical scenario would it pull apart or the bolts hold it where it is? If the latter then it should work fine If I had the technical means to do this well I guess more of the time to do this I would put a .8 or even a 1 mm nozzle on the printer print out a 75% of that layer height slow and hot and then don't do it in a truck get a cheap 80s rear wheel drive 1500 pound car :-) think Geo Metro but something with a pumpkin It would be neat to see if he can hold together obviously if you punched it it wouldn't hold together but if you gingered it and we're careful could you go say 20 miles :-)
Thicker extrusion and bigger layers = more lamination. The more surface area which connects each extrusion line, the better the lines will adhere to each other. It's a trade-off between surface quality and strength. Increase your Extrusion Multiplier in Simplify3d (It's called 'Flow' in other slicers, iirc) and your layer thickness to insure layer bonding.
I love how you can see the teeth actual dropping out one by one when you are backing up. Thanks for revisiting, it was fun, even though you got your teeth broken! Did you just thread directly into the Nylon X for the circle of bolts?
Gear Down For What? The only way this can work is by printing it vertically since force will be applied perpendicular in relation to the printed layers and that will make able to stand more torque and force, the way you are printing the part will always break in the layer lines and strip them apart but anyway great video.
Look if you want a strong part you have to make sure the layers adhere very well to each other. to accomplish this you should increase the nozzle temp to the highest you can go before getting major defects from the heat and have a heated enclose. Also slowing down the speed helps giving the plastic time to melt the layer beneath and on the sides. Using overlap constantly also helps but then again need to slow down the print to compensate for the extra friction. I am not sure what the glass transition temperature is for the nylonX but if you can make a enclosure that keeps that exact temperature. so the layers bond instantly when heated a little bit. And when the part have cooled you shall put it in the oven to annealing it for all the internal stress that might cause shattering.
Looking at the cross section of 0.1 vs 0.2 layer height, the adhesion surface area increases by 15% for 0.1 mm. Also the gap volume increases by 4x for 0.2 mm. The lower your nozzle is to the previous layer, the more it re-heats, creating a better adhesion surface. Assuming perfect conditions (no cooling, heated chamber, etc), lower layer height always wins. I would bet a 0.01 mm layer height would be comparable to injection molded strength.
I would guess wider nozzle diameter was a big factor in failure. I picture short stick vs a long stick, small bricks vs 1 big brick. Many small cables or 1 solid cable. I think the wider diameter made it break in chunks. smaller lines layered while warm probably helped seal and adhere layers better.
Would changing the build direction help ? Forces would be applied on a perpendicular plane to the layers instead of parallel. Would be stronger and the teeth would hold better.
You need to retrofit a 5 - axis mill with a printer head to print this for maximum strength, might even have to be a 6 - axis one so the table and head move independently of each other, ideally printing each tooth as an outside growth...might work on a mill-lathe, growing it from the inside of the tooth to the outside.
If you've seen a worn out steel ring gear from an old pickup, you'd know that those differentials need all the strength they can get! Looks like it was a fun project though!
One thing you could try is printing it at a 45° angle (x or y axis rotation). You'll get layers that go vertically into the teeth. It takes a lot longer to print and uses a lot of support material, but it's much stronger.
Benjamin King I think it would be just better to add tree holes below each tooth so that it crosses the base of each tooth and then just screw some 5/32 screws from the other side directly into the plastic. That way the screws could take some of the tensile force of the teeth though it would take like 3 screws for each tooth.
That Nylon X differential gear you put on might just be passable for a temporary holdover part until a proper replacement is found. In the near future, you'll be able to replace that gear made from steel with one 3D printed from an even stronger version of Nylon X that potentially incorporates carbon nanotubes into its construction.
Back in the 1970's I was a mechanic at a Volvo shop (yes, I'm older than rocks). At that time the 4 cylinder engines had a "fiber" timing gear on the camshaft (the big gear) and a steel gear on the crankshaft (the small gear). Meanwhile the 6 cylinder engines used steel for both gears. The gears were exactly the same size, so you could put a steel camshaft gear on a 4 cylinder engine if you so desired. We found that the fiber gears lasted around 60K - 80K miles. The steel gears on the 6 cylinder engines (and 4 cylinder engines when we replaced fiber with steel) never broke. My point is that there are lots of gears on a car, and 3D printing might be appropriate for some of them. That ability would be a great thing for restoring antique cars where you can't buy the parts anymore. You may be trying to 3D print the most highly stressed gear on the whole vehicle. How about trying to print something like the gear that drives the oil pump or the distributor?
Try printing piece in an orientation where layer boundaries don't align with the base of the gear teeth (i.e. print ring vertically instead if horizontally on bed, or maybe at an angle with support rafts.)
Your Part warped while printing - this is also the reason for the broken glasstray. When you put it on the lathe, you've given it his deadsentence right away: doing the horizontal(? - the facing plate side) - pass didn't make it straight at the teethside. So the teeth had uneven pressure and fit, the first few broke right away and made it look work, the others couldn't hold up then... Still it's impressive the car moved at all, I wonder what would happen, if you'll try a straight part again.
Trying anyway: It would take some fancy gcode wrangling, but what about printing the teeth so the layers/sheer is at an angle to the applied force. Basically 45-degree angle printing (if not this, maybe something to try for other gear projects)
could try changing the print orientation or using your printer in non planer mode so it doesn't have layers it takes some work to slice it like that but in this application i thing it might help
Huge thanks for making a second version!! Very interesting!! Would be interesting with a little calculation on the difference between nylonX and hardened steel and then what tensile strength would be required at a minimum to make a ring gear that works long term in a car. Thanks!!
give it a higher layer height, fix the extrusion width to the original value, bake the part in your oven for 30 minutes at the glass transition temperature of the filament type. then try again. also maybe consider a carbon fiber filament as well.
So the first one's problem was that you had stopped the print to insert the nuts; but its teeth were perfectly fine. It looks that this time you actually tapped the plastic to get the threads... Which is perfect. The layer height changes though...? Guess that didn't work out well. Combine the two and it should last longer? >_> Also, maybe try giving it a quick flash bake to ensure it's as fused together as possible. Just an idea for the future...
If at first you don't succeed, try, try, try again. You should look into Non-Planar Slicing. You'll need to compile a new slicer though. Also, you should probably upgrade your 3D Printer so that you can build at higher temperatures, and then construct a plaster cast around the gear and anneal it (assuming that the material you're using can handle tempering and annealing). The plaster cast will prevent excess warping. There's also probably some kind or glaze or coating that you could treat the finished part with to increase its durability.
Redesign the part and incorporate a triple joggle interlock joint-also consider using a toughened carbon reinforced ABS and the Toughened Epoxy can actually achieve a molecular bond to the ABS
I suspect the carbon fibre mixed in, is actually making the layer adhesion weaker. Properly bonded plain nylon will not make that ripping sound when you were pulling the teeth off. Nylon normally will bend and stretch before breaking. I know you dont want to try again, but i think a plain nylon filament, printed real slow and hot would be a better choice for a gear than Nylon X.
Alot of people have said they were disappointed with how brittle Nylon X is. There are better carbon fiber blends out there. I would look at 3DXTechs carbon fiber blend lol
nylonx is crap . but thats his sponsor . those teeth fell of even before load , that a joke . with good filament from 3dxtech or treed it would hold much better
i wonder if both gears were plastic if it would work better? im not sure how to explain why but i feel like that would move some of the force away from the teeth a little bit and into the gear since both teeth are the same hardness
Carson W234 Actually having a bit higher width gives you better layer adhesion because the molten plastic gets pushed more against the last layer and so sticks better with less air gaps, but you shouldnt exceed your nozzle's point diameter or the excess plastic will just come out of the sides where it doesnt push against the lower layer and so you get an ugly dragged finish and not any better layer adhesion.
You might try making all of your layers and infills in a circular pattern so that the fibers are always perpendicular to the pinion gear. I noticed that the infill on your first gear used rectilinear infill. Rectilinear will make the pattern square w.r.t. the build plate. This causes the fiber direction to be slightly different in each gear tooth. As soon as the pinion cracks the weakest tooth it accelerates and applies a weed whacker action to all of the gears. The layers should essentially have their fibers in concentric rings so that they will be perpendicular to the pinion gear. The fibers are much stronger along their length. The little bits of carbon fiber will align parallel to each other as the filament is extruded along its length. This is its strongest orientation. You might also increase the number of perimeters until the entire gear is almost made of nothing but perimeters. Just a few thoughts. Don't let the negative comments from the nay sayers get to you . Even though you explain that you are not repairing your truck for daily use with the printed gear and that it's just a test to see how far you can push the material strength (and entertainment purposes), many people seem to miss that point... Breakthoughs come from insane problems. Think about where the greatest forces are and how the fibers should point in those places. Great video!
Something i just thought off. It would require you to print two gears and properly modify the differentiel, but if you could print the teeth with a less step angle then the force would be better connected through all the layers and greatly reduce the risk of teeth shearing at the weakest point. Basicly if you could change the angle of the bevel gears. Just incase you ever consider looking at this again. Maybe buy a cheap lawn tractor to play with if you can afford it. Would be perfectly fine to get a used one.
Cool, maybe you could leave tiny slits in each teeth and push a small piece of flat stock so each teeth has a steel core with purchase into the base ring below the shear line
Alternatively, you could print a gear, dump it in high temperature plaster, cure and evaporate the plastic in a kiln, then fill the cavity with bronze powder and finally sinter a real metal gear
Is heat treatment possible? With high strength metal parts you'd typically use a very thick build plate to control distortion, and then a stress relieving heat treatment before removal
Did you properly dry out the nylon before printing? Nylon filament absorbs water from the air like crazy and printing with nylon that's full of moisture will leave tiny gaps all over your print and weaken it significantly.
I was like.....DAMN the underside of his pickup is rusty....then I saw the Minnesota plate....nevermind XD...I lived in Bemidji for 12 years. I miss real winter, ice fishing, snowmobile drags on the lakes, lutefisk, and lefse.
Okay, I won't talk you into trying this again... but I would love to see you design, engineer, and print a gear box small enough to fit under the seat or somewhere similar that is capable of winching your truck out of a ditch or back onto the road on a snowy day. Ideally powered by a drill or impact wrench. Something capable of a "real world work" similar to a drill powered "come-a-long" if that is something that exists.
Spectt84 when he gets to the "engineer " part, he will realize that's not remotely possible. The mechanics of materials calculations will lead him to the conclusion of "don't bother". I have a phrase I use at my job for ideas that won't work: Divide by Zero.
What would happen if you cover the gear with fiber glass resin, I mean, just fully sumerge the gear with the liquid resin and using the finest fiber?? Do you think it will make it stronger???
@3dprintingnerd What if you printed it vertically or at an angle? By printing horizontally you create natural “fault” lines from where the teeth cut into the gear. Printing vertically with a larger diameter nozzle (maybe even a dual diameter printer if that is even possible) would mean you would be making the teeth “against the grain” instead of with it.
really enjoying your videos as I venture into 3d printing gears. I gotta ask, what is going on with that lathe "tool"? Grind one properly(big relief& rake angles), or buy one and you'd have a glass like finish without all the stringing.
Heyyyy, that shattered like my 3D printer did when I tried to add loctite to the little captive nut screw things... And that lube you sprayed in there looked like it was just soaked up by the nylon. Maybe next try, douse it in molylube or graphite powder?
Like you said, there is a reason that part is made of hardened steel. I would love to see you use that filament again but don't try to destroy it immediately.
You should have printed the gear Vertically Standing up So the layers would be against the grain where the force is being applied it would be Much less prone to teeth breaking off.
Thinner layers improve the surface finish but decrease the strength. Think of each adhesion point as a possible weak spot and it makes more sense. The thinner layers don’t add more material than they do weak points so the part gets weaker.
Not gonna ask you to retry, but.. Having watched through your tests I'm quite confident that it'd be possible to use plastic in automotive power transfer. You simply can't replace a metal tooth with a plastic tooth, but when the whole gearing was redesigned around plastic it would probably work. Eventually. Prints are becoming stronger each day!
If you are going to try again, I will suggest Instead of embedding the nuts, try insert steel thread inserts with epoxy after the ring is completed, since it is going to just dealing with rotational force not pulling...
Maybe someday you can make an axle instead as that materials low shear strength won't be an issue. The printing size requirements will be an issue though.
Did you guys check / adjust the backlash? I incorrectly set my backlash on my factory made steel ring and pinion and shatterd the ring when taking a turn and hitting the gas too hard in a numerically lower gear (higher) - with more tq. a carbon fiber gear may hold up properly but there is a lot that goes into it than just installing the gear and sticking it in the car. Frankly - a 3d printed non-CCF gear would be weaker than a steel one or a reinforced one. But the backlash is important. Dial gauge and all that (Source: i do all my own car repairs, and have destroyed a ring gear before)
The broken gears at the start would have caused the rest to fail, as at one point instead of 3 gears holding all the force you would drop down to 2 and then becoming the weakest link. starting a domino effect as they shared off. i think this is why the lats print was ''stronger''
would love to see somebody try this with the parts printed so that the pinion was not putting shear force against the layer boundaries. turn it vertical and print with supports or print in pieces?
Keep in mind that the extrusion width - at least noted within Slic3r - is calculated based on the layer height, not the nozzle diameter. If your nozzle diameter is 0.4mm printing at 0.2mm height, setting the extrusion width to (say) 125% would set the extrusion width to only 0.25mm (0.2mm height X 125%), not the 0.5mm you might expect (0.4mm diameter X 125%).
Maybe you would just use the printer to make CAST parts, and seeing as you have some machine tools at your disposal, just clean the cast parts up so they have good surface finish and last a nice long time.
Try next time to print your model perpendicular to the hot bed you will gain more strength the forces will be acting perpendicular as well think in a piece of wood and the wall is laminated a perpendicular cut is more difficult than a parallel
The Wider the line extrusion width the weaker your layers become! For truly invincible nylon carbon fiber parts, you make it 0% infill and 99999999 perimeters or walls and with about 5 to 8 top and bottom layers depending on the tolerance required for such part. But it would work if you just lowered the layer width exponentially
I know it's annoying to end a video on a negative but I think this was still really valuable, especially the part about layer heights. Do you know how NylonX would compare to Polycarbonate? Another issue is the layer direction - you're applying a shearing force directly across the layer at its weakest point, if it were possible to print the geometry vertically somehow this force would be applied 'through' the layers instead.
I know this was probably defeating on a materialistic, emotional and financial level, but man it was good to see that you preserved to do it at all. About the only way I could see this working with "plastic" bits would be to print out your ring gear with precise dimensional tolerance, mold it in something real sturdy (top shelf casting resin) and use a forged carbon treatment with the mold under high vacuum. I'm currently experimenting with turbine blades and a few other treatments with chopped basalt fibers to see how long they'll last in a high heat, high pressure environment. Not quite the same as mechanical torque with direct shearing forces, but if one was determined to have a "plastic" drivetrain component, that'd be my solution. Just don't do a burnout on it. =D
i dont know what engine and tranny your s10 has, but assuming its got a 2.2 with a Borg Warner, it puts out 140 lb ft of torque at the crank. 1st gear ratio is 3.35 to 3.97, taking it to 469-555 lb ft of torque at the pinion (input of the differential). The ratio between the pinion and ring is anywhere between 3.08 and 4.56, taking the 469 to 555 ib ft and increasing to 1444 to 2530 lb ft at the wheels. If you look at your factory codes in your glovebox you can get exact figures, very good video though.
Didnt look like you set up any backlash on the ring you can have the same results with the harden steel gear as what you had here if the diff is not setup properly..
Interesting idea. Something to consider.....when a machine/structure/part is initially built in 1 material then made lighter by being built from another material ie a trailer built of steel versus 1 built of aluminum....often times the lighter redesign is built of material thicknesses/shapes that are very different.... 1 of the most obvious things that come to mind are airplane wings.... consider how a wood/aluminum/composite wing are built differently (at times).... My point in all this is, it may be possible to make a functional unit if the entire differential was completely redesigned, perhaps finer/wider teeth and a much larger overall diameter of ring and pinion. I would be surprised if it couldn't be done by some outside the box means on a light duty application. Cool attempt at least
Actually, what would have been more interesting is replacing with carbon fiber infused filament. It surely would break in the end, but I assume, it would hold much longer than nylon. Probably even longer that peek or pekk filaments.
It's a shame that there's no feasible way to print that ring gear on its side. The teeth won't shear off at one of the layers if the layers could be in the direction of the height if the teeth if that description was clear enough
I think the problem is that there wasn't enough contact. You'll notice at the base, there's a decent amount of meat that wasn't being used. and didn't shear.
Good try, but bigger layers are stronger, not smaller. You're adding more shear points. The wider than nozzle width will just add smooshed stuff to the sides, not guaranteed to be laid down properly, so weaker again. If you're worried about the teeth shearing, I'd try printing it vertically, with as large a layer as you can manage, staying at no more than 80% of your nozzle width.
I can't wait for the third version you promised at the end of the video....
Yeah
I’m new
Lower layer height is weaker because you were increasing the number of layer bonds available and layer bonds are always your weak point in 3D printing so if anything you need to increase your layer height not decrease it
they don't have to be weak if you have a insanely high nozzle temp and annealing the part afterwards
craazyy22 this is only partially true lower layer height is always weaker no exceptions the week is part of a 3D print is the bonding between layers higher temperatures to a point and the kneeling to a point will reduce how much more we can take but it will still be weaker than the same exact process with a thicker layer height
I believe that now, but I really didn't understand how that was possible before yesterday, i'm running with the assumption that a thicker layer has more thermal mass and can warm the previous layer more and bond to it better. along with there being more bonds with thinner layers, it can only be as strong as it's weakest layer, right?
Otherwise it just doesn't make sense to me that thinner layers are weaker, I kinda thought that a thinner layer would be mashing the nozzle into the part, helping to fill any tiny little cracks and possibly radiate some heat onto the previous layer from the nozzle, apparently that isn't the whole story though.
Nerys it cools down too quickly to help the next later bond.
@@GearDownForWhat layer bonds are the weak points but layers only go one direction on a prints part so it is possible to rotate something in this case 90 degrees and print it vertically that way the force on the gear teeth isn't opposing the layer lines, would should increase the strength of the gears. But even if you do the disc would probably only last a couple minutes still.
well now we HAVE to try to convince you to try again
Seems like you need a metal 3d printer...
Most metal 3D printers are more convolution than they are worth. It's still a multi-step process.
In which case just use standard FDM to produce a mold positive, and sand cast the part that you need.
I wonder how much longer it would last if you changed the orientation of the part with respect to layer lines (the teeth will be their weakest printed like that) , or if you annelied it (or both) ?
Rotating it 90 degrees would be overkill but id be interested to see if tilting it 30 degrees or so would make much difference.
MiggyManMike I would slice it in half and print both parts 90 degrees
This
I was considering that, its possible that that could work if the bolts could hold the two halves together enough
Gear Down For What? It depends on how it's held together inside the pumpkin imagine if you were to go inside your pumpkin with a laser and slice it in half without hitting any bolts imagine that theoretical scenario would it pull apart or the bolts hold it where it is? If the latter then it should work fine
If I had the technical means to do this well I guess more of the time to do this I would put a .8 or even a 1 mm nozzle on the printer print out a 75% of that layer height slow and hot and then don't do it in a truck get a cheap 80s rear wheel drive 1500 pound car :-) think Geo Metro but something with a pumpkin
It would be neat to see if he can hold together obviously if you punched it it wouldn't hold together but if you gingered it and we're careful could you go say 20 miles :-)
A Chevette or "AE86" Corolla would fit the bill I think. My Chevette needs a new ring gear anyway...
Thicker extrusion and bigger layers = more lamination. The more surface area which connects each extrusion line, the better the lines will adhere to each other. It's a trade-off between surface quality and strength. Increase your Extrusion Multiplier in Simplify3d (It's called 'Flow' in other slicers, iirc) and your layer thickness to insure layer bonding.
I wish more people where willing to try more real world use case scenerios for 3d printed parts. Thanks man!
Try annealing the print to increase layer adhesion. You can pack it in salt to reduce shrinkage.
Also use grease. WD-40 is not a lubricant.
I love how you can see the teeth actual dropping out one by one when you are backing up. Thanks for revisiting, it was fun, even though you got your teeth broken! Did you just thread directly into the Nylon X for the circle of bolts?
yeah I just threaded it this time, I guess I didn't need to implant nuts into it last time haha
Gear Down For What? The only way this can work is by printing it vertically since force will be applied perpendicular in relation to the printed layers and that will make able to stand more torque and force, the way you are printing the part will always break in the layer lines and strip them apart but anyway great video.
Look if you want a strong part you have to make sure the layers adhere very well to each other. to accomplish this you should increase the nozzle temp to the highest you can go before getting major defects from the heat and have a heated enclose. Also slowing down the speed helps giving the plastic time to melt the layer beneath and on the sides. Using overlap constantly also helps but then again need to slow down the print to compensate for the extra friction. I am not sure what the glass transition temperature is for the nylonX but if you can make a enclosure that keeps that exact temperature. so the layers bond instantly when heated a little bit. And when the part have cooled you shall put it in the oven to annealing it for all the internal stress that might cause shattering.
this would work much better if you printed it standing, or at a sharp angle 45º or more.
Looking at the cross section of 0.1 vs 0.2 layer height, the adhesion surface area increases by 15% for 0.1 mm. Also the gap volume increases by 4x for 0.2 mm.
The lower your nozzle is to the previous layer, the more it re-heats, creating a better adhesion surface. Assuming perfect conditions (no cooling, heated chamber, etc), lower layer height always wins. I would bet a 0.01 mm layer height would be comparable to injection molded strength.
Amazed it moved at all. I've shredded half the teeth off a hardened steel ring gear, tons of stress on that, especially on a fullsize truck
I would guess wider nozzle diameter was a big factor in failure. I picture short stick vs a long stick, small bricks vs 1 big brick. Many small cables or 1 solid cable. I think the wider diameter made it break in chunks. smaller lines layered while warm probably helped seal and adhere layers better.
You should try printing it at a 45° angel up instead of flat on the glass
or just flat 90'
Would changing the build direction help ? Forces would be applied on a perpendicular plane to the layers instead of parallel. Would be stronger and the teeth would hold better.
You need to retrofit a 5 - axis mill with a printer head to print this for maximum strength, might even have to be a 6 - axis one so the table and head move independently of each other, ideally printing each tooth as an outside growth...might work on a mill-lathe, growing it from the inside of the tooth to the outside.
If you've seen a worn out steel ring gear from an old pickup, you'd know that those differentials need all the strength they can get! Looks like it was a fun project though!
One thing you could try is printing it at a 45° angle (x or y axis rotation). You'll get layers that go vertically into the teeth. It takes a lot longer to print and uses a lot of support material, but it's much stronger.
Benjamin King
I think it would be just better to add tree holes below each tooth so that it crosses the base of each tooth and then just screw some 5/32 screws from the other side directly into the plastic. That way the screws could take some of the tensile force of the teeth though it would take like 3 screws for each tooth.
That Nylon X differential gear you put on might just be passable for a temporary holdover part until a proper replacement is found. In the near future, you'll be able to replace that gear made from steel with one 3D printed from an even stronger version of Nylon X that potentially incorporates carbon nanotubes into its construction.
Back in the 1970's I was a mechanic at a Volvo shop (yes, I'm older than rocks). At that time the 4 cylinder engines had a "fiber" timing gear on the camshaft (the big gear) and a steel gear on the crankshaft (the small gear). Meanwhile the 6 cylinder engines used steel for both gears. The gears were exactly the same size, so you could put a steel camshaft gear on a 4 cylinder engine if you so desired. We found that the fiber gears lasted around 60K - 80K miles. The steel gears on the 6 cylinder engines (and 4 cylinder engines when we replaced fiber with steel) never broke. My point is that there are lots of gears on a car, and 3D printing might be appropriate for some of them. That ability would be a great thing for restoring antique cars where you can't buy the parts anymore. You may be trying to 3D print the most highly stressed gear on the whole vehicle. How about trying to print something like the gear that drives the oil pump or the distributor?
Try printing piece in an orientation where layer boundaries don't align with the base of the gear teeth (i.e. print ring vertically instead if horizontally on bed, or maybe at an angle with support rafts.)
Your Part warped while printing - this is also the reason for the broken glasstray. When you put it on the lathe, you've given it his deadsentence right away: doing the horizontal(? - the facing plate side) - pass didn't make it straight at the teethside. So the teeth had uneven pressure and fit, the first few broke right away and made it look work, the others couldn't hold up then...
Still it's impressive the car moved at all, I wonder what would happen, if you'll try a straight part again.
Trying anyway: It would take some fancy gcode wrangling, but what about printing the teeth so the layers/sheer is at an angle to the applied force. Basically 45-degree angle printing (if not this, maybe something to try for other gear projects)
Holly shit i am glad i live in texas and domt have crazy rust like that
thank you for trying this a second time EDIT: I was not going to ask you to try it again, however... thank you for trying it a second time.
Now use Taulman Alloy 910, I made a whole gear set for my craftsman 109 lathe, and I haven't had a single issue. Really tough material. Bonds strong.
You should do it one more time with a larger nozzle and do larger layer Heights I bet it would hold up
Higher layer height, higher temp. And if you are getting free material and have some time, print it vertically rather than flat.
I’m impressed that you even tried. Almost like Don Quixote given the incredible forces involved.
I would offer up to print on a Markforged Onyx Pro, but the build volume isn't big enough and it hasn't been delivered yet.
could try changing the print orientation or using your printer in non planer mode so it doesn't have layers it takes some work to slice it like that but in this application i thing it might help
Huge thanks for making a second version!! Very interesting!! Would be interesting with a little calculation on the difference between nylonX and hardened steel and then what tensile strength would be required at a minimum to make a ring gear that works long term in a car. Thanks!!
give it a higher layer height, fix the extrusion width to the original value, bake the part in your oven for 30 minutes at the glass transition temperature of the filament type. then try again. also maybe consider a carbon fiber filament as well.
I would expect that shorter layers only introduce more points of failure. Probably higher temps in general would be more effective
Hello, I'd like some info on how to print the ring gear like you have..
"How is this worse than it was before?" I'm betting that the other plastic was more giving, it would bend without actually breaking.
So the first one's problem was that you had stopped the print to insert the nuts; but its teeth were perfectly fine. It looks that this time you actually tapped the plastic to get the threads... Which is perfect. The layer height changes though...? Guess that didn't work out well. Combine the two and it should last longer? >_>
Also, maybe try giving it a quick flash bake to ensure it's as fused together as possible. Just an idea for the future...
That was a very cool project. You should try it on a much smaller car, maybe a Go-cart or so
If at first you don't succeed, try, try, try again.
You should look into Non-Planar Slicing. You'll need to compile a new slicer though.
Also, you should probably upgrade your 3D Printer so that you can build at higher temperatures, and then construct a plaster cast around the gear and anneal it (assuming that the material you're using can handle tempering and annealing). The plaster cast will prevent excess warping.
There's also probably some kind or glaze or coating that you could treat the finished part with to increase its durability.
Redesign the part and incorporate a triple joggle interlock joint-also consider using a toughened carbon reinforced ABS and the Toughened Epoxy can actually achieve a molecular bond to the ABS
I suspect the carbon fibre mixed in, is actually making the layer adhesion weaker. Properly bonded plain nylon will not make that ripping sound when you were pulling the teeth off. Nylon normally will bend and stretch before breaking. I know you dont want to try again, but i think a plain nylon filament, printed real slow and hot would be a better choice for a gear than Nylon X.
Alot of people have said they were disappointed with how brittle Nylon X is. There are better carbon fiber blends out there. I would look at 3DXTechs carbon fiber blend lol
nylonx is crap . but thats his sponsor . those teeth fell of even before load , that a joke .
with good filament from 3dxtech or treed it would hold much better
Larger layer height actually makes it stronger, as well as printing at a slightly higher temperature to get better bonding between layers.
i wonder if both gears were plastic if it would work better? im not sure how to explain why but i feel like that would move some of the force away from the teeth a little bit and into the gear since both teeth are the same hardness
Your extrusion width should be the same as (or close to, up to 125%) your nozzle diameter.
Carson W234
Actually having a bit higher width gives you better layer adhesion because the molten plastic gets pushed more against the last layer and so sticks better with less air gaps, but you shouldnt exceed your nozzle's point diameter or the excess plastic will just come out of the sides where it doesnt push against the lower layer and so you get an ugly dragged finish and not any better layer adhesion.
I agree with Carson W234. I figure that the over extrusion caused the parts to not fit together so well and that is probably why it failed faster
Have you tried using fillament made to cast molds and then casting the part.
You might try making all of your layers and infills in a circular pattern so that the fibers are always perpendicular to the pinion gear. I noticed that the infill on your first gear used rectilinear infill. Rectilinear will make the pattern square w.r.t. the build plate. This causes the fiber direction to be slightly different in each gear tooth. As soon as the pinion cracks the weakest tooth it accelerates and applies a weed whacker action to all of the gears. The layers should essentially have their fibers in concentric rings so that they will be perpendicular to the pinion gear. The fibers are much stronger along their length. The little bits of carbon fiber will align parallel to each other as the filament is extruded along its length. This is its strongest orientation. You might also increase the number of perimeters until the entire gear is almost made of nothing but perimeters. Just a few thoughts. Don't let the negative comments from the nay sayers get to you . Even though you explain that you are not repairing your truck for daily use with the printed gear and that it's just a test to see how far you can push the material strength (and entertainment purposes), many people seem to miss that point... Breakthoughs come from insane problems. Think about where the greatest forces are and how the fibers should point in those places. Great video!
If it's nylon, does that make it a limited slip? lol...
Very limited.
Something i just thought off. It would require you to print two gears and properly modify the differentiel, but if you could print the teeth with a less step angle then the force would be better connected through all the layers and greatly reduce the risk of teeth shearing at the weakest point.
Basicly if you could change the angle of the bevel gears. Just incase you ever consider looking at this again. Maybe buy a cheap lawn tractor to play with if you can afford it. Would be perfectly fine to get a used one.
Cool, maybe you could leave tiny slits in each teeth and push a small piece of flat stock so each teeth has a steel core with purchase into the base ring below the shear line
Alternatively, you could print a gear, dump it in high temperature plaster, cure and evaporate the plastic in a kiln, then fill the cavity with bronze powder and finally sinter a real metal gear
sounds like a threat "if you print that diff ring gear you will lose your teeth"
Is heat treatment possible? With high strength metal parts you'd typically use a very thick build plate to control distortion, and then a stress relieving heat treatment before removal
Did you properly dry out the nylon before printing? Nylon filament absorbs water from the air like crazy and printing with nylon that's full of moisture will leave tiny gaps all over your print and weaken it significantly.
you should retry but with more modern filament like Petg, and use a nice thick layer height
Just the fact you were able to get 10' is ridiculous. Have you tried a carbon firbe infused pla?
I'd love to try this with a MarkForged printed part reinforced with continuous fiber strands!
I was like.....DAMN the underside of his pickup is rusty....then I saw the Minnesota plate....nevermind XD...I lived in Bemidji for 12 years. I miss real winter, ice fishing, snowmobile drags on the lakes, lutefisk, and lefse.
Maybe i'm late, but your gear need some adjustments after installing to have an optimal work conditions. look for some info in internet.
Okay, I won't talk you into trying this again... but I would love to see you design, engineer, and print a gear box small enough to fit under the seat or somewhere similar that is capable of winching your truck out of a ditch or back onto the road on a snowy day. Ideally powered by a drill or impact wrench. Something capable of a "real world work" similar to a drill powered "come-a-long" if that is something that exists.
Spectt84 when he gets to the "engineer " part, he will realize that's not remotely possible. The mechanics of materials calculations will lead him to the conclusion of "don't bother". I have a phrase I use at my job for ideas that won't work: Divide by Zero.
What would happen if you cover the gear with fiber glass resin, I mean, just fully sumerge the gear with the liquid resin and using the finest fiber?? Do you think it will make it stronger???
Try the salt baking method - fusing all the layers together?
@3dprintingnerd What if you printed it vertically or at an angle? By printing horizontally you create natural “fault” lines from where the teeth cut into the gear. Printing vertically with a larger diameter nozzle (maybe even a dual diameter printer if that is even possible) would mean you would be making the teeth “against the grain” instead of with it.
really enjoying your videos as I venture into 3d printing gears. I gotta ask, what is going on with that lathe "tool"? Grind one properly(big relief& rake angles), or buy one and you'd have a glass like finish without all the stringing.
Heyyyy, that shattered like my 3D printer did when I tried to add loctite to the little captive nut screw things... And that lube you sprayed in there looked like it was just soaked up by the nylon. Maybe next try, douse it in molylube or graphite powder?
I want to see it tried with annealed nylon. Pushing it to truly the limit. looking forward to a followup
Like you said, there is a reason that part is made of hardened steel. I would love to see you use that filament again but don't try to destroy it immediately.
Just a thought but could the arousal propellant (frequently a solvent) from the grease be deteriorating the nylon?
The key would be orientation of fibers. For me looks like delamination of layers. But how to print this? :-)
You should have printed the gear Vertically Standing up So the layers would be against the grain where the force is being applied it would be Much less prone to teeth breaking off.
Thinner layers improve the surface finish but decrease the strength. Think of each adhesion point as a possible weak spot and it makes more sense. The thinner layers don’t add more material than they do weak points so the part gets weaker.
Not gonna ask you to retry, but..
Having watched through your tests I'm quite confident that it'd be possible to use plastic in automotive power transfer. You simply can't replace a metal tooth with a plastic tooth, but when the whole gearing was redesigned around plastic it would probably work. Eventually. Prints are becoming stronger each day!
If you are going to try again, I will suggest Instead of embedding the nuts, try insert steel thread inserts with epoxy after the ring is completed, since it is going to just dealing with rotational force not pulling...
Maybe someday you can make an axle instead as that materials low shear strength won't be an issue. The printing size requirements will be an issue though.
I've blown up the gears in my truck three times and it only has a four cylinder. (and 46in tires) It's amazing 3D printed gears worked at all.
Did you guys check / adjust the backlash? I incorrectly set my backlash on my factory made steel ring and pinion and shatterd the ring when taking a turn and hitting the gas too hard in a numerically lower gear (higher) - with more tq. a carbon fiber gear may hold up properly but there is a lot that goes into it than just installing the gear and sticking it in the car.
Frankly - a 3d printed non-CCF gear would be weaker than a steel one or a reinforced one. But the backlash is important. Dial gauge and all that (Source: i do all my own car repairs, and have destroyed a ring gear before)
The teeth were breaking with almost no load because your pinion engagement and lash were not set properly.
The broken gears at the start would have caused the rest to fail, as at one point instead of 3 gears holding all the force you would drop down to 2 and then becoming the weakest link. starting a domino effect as they shared off. i think this is why the lats print was ''stronger''
would love to see somebody try this with the parts printed so that the pinion was not putting shear force against the layer boundaries. turn it vertical and print with supports or print in pieces?
Keep in mind that the extrusion width - at least noted within Slic3r - is calculated based on the layer height, not the nozzle diameter. If your nozzle diameter is 0.4mm printing at 0.2mm height, setting the extrusion width to (say) 125% would set the extrusion width to only 0.25mm (0.2mm height X 125%), not the 0.5mm you might expect (0.4mm diameter X 125%).
Maybe you would just use the printer to make CAST parts, and seeing as you have some machine tools at your disposal, just clean the cast parts up so they have good surface finish and last a nice long time.
Why was the build plate busting ?? Was it so hot or heavy it warped the glass and broke it or what.???
Someone sponsor this man a sintering 3d printer!
Nicely done though.
You need a heated chamber and print as hot as possible. cool slowly, then anneal.
Try next time to print your model perpendicular to the hot bed you will gain more strength the forces will be acting perpendicular as well think in a piece of wood and the wall is laminated a perpendicular cut is more difficult than a parallel
The Wider the line extrusion width the weaker your layers become! For truly invincible nylon carbon fiber parts, you make it 0% infill and 99999999 perimeters or walls and with about 5 to 8 top and bottom layers depending on the tolerance required for such part. But it would work if you just lowered the layer width exponentially
I know it's annoying to end a video on a negative but I think this was still really valuable, especially the part about layer heights. Do you know how NylonX would compare to Polycarbonate? Another issue is the layer direction - you're applying a shearing force directly across the layer at its weakest point, if it were possible to print the geometry vertically somehow this force would be applied 'through' the layers instead.
I know this was probably defeating on a materialistic, emotional and financial level, but man it was good to see that you preserved to do it at all.
About the only way I could see this working with "plastic" bits would be to print out your ring gear with precise dimensional tolerance, mold it in something real sturdy (top shelf casting resin) and use a forged carbon treatment with the mold under high vacuum. I'm currently experimenting with turbine blades and a few other treatments with chopped basalt fibers to see how long they'll last in a high heat, high pressure environment.
Not quite the same as mechanical torque with direct shearing forces, but if one was determined to have a "plastic" drivetrain component, that'd be my solution. Just don't do a burnout on it. =D
i dont know what engine and tranny your s10 has, but assuming its got a 2.2 with a Borg Warner, it puts out 140 lb ft of torque at the crank. 1st gear ratio is 3.35 to 3.97, taking it to 469-555 lb ft of torque at the pinion (input of the differential). The ratio between the pinion and ring is anywhere between 3.08 and 4.56, taking the 469 to 555 ib ft and increasing to 1444 to 2530 lb ft at the wheels. If you look at your factory codes in your glovebox you can get exact figures, very good video though.
Awww, I was hoping, like you, for a better outcome. Sorry it didn't work out. Cheers, JAYTEE
Hey you
Hey 😂
Didnt look like you set up any backlash on the ring you can have the same results with the harden steel gear as what you had here if the diff is not setup properly..
Interesting idea. Something to consider.....when a machine/structure/part is initially built in 1 material then made lighter by being built from another material ie a trailer built of steel versus 1 built of aluminum....often times the lighter redesign is built of material thicknesses/shapes that are very different.... 1 of the most obvious things that come to mind are airplane wings.... consider how a wood/aluminum/composite wing are built differently (at times)....
My point in all this is, it may be possible to make a functional unit if the entire differential was completely redesigned, perhaps finer/wider teeth and a much larger overall diameter of ring and pinion.
I would be surprised if it couldn't be done by some outside the box means on a light duty application.
Cool attempt at least
fun watch! I wonder how it would change if you annealed the nylon ;)
Actually, what would have been more interesting is replacing with carbon fiber infused filament. It surely would break in the end, but I assume, it would hold much longer than nylon.
Probably even longer that peek or pekk filaments.
You should get shapeways to sponsor you to do this with their various metal 3d prints
It's a shame that there's no feasible way to print that ring gear on its side. The teeth won't shear off at one of the layers if the layers could be in the direction of the height if the teeth if that description was clear enough
Thanks for trying again! I finally have some closure!
Seems like you already knew this had no chance, but extremely cool experiment.
I think the problem is that there wasn't enough contact. You'll notice at the base, there's a decent amount of meat that wasn't being used. and didn't shear.
Good try, but bigger layers are stronger, not smaller. You're adding more shear points. The wider than nozzle width will just add smooshed stuff to the sides, not guaranteed to be laid down properly, so weaker again.
If you're worried about the teeth shearing, I'd try printing it vertically, with as large a layer as you can manage, staying at no more than 80% of your nozzle width.