In the process of upgrading my 3D printer and preparing to build my own, I decided that I also needed to design the ultimate extruder. Before I picked an overall style for the design, I wanted to determine which gear had the best grip and could apply the greatest amount of torque on the filament. So, I ordered 10 different gears, designed a printable test rig, and got to work! Here is the chronicle of my quest.

Step 1: The Test Rig

I thought for quite a while about how to test the different gears on an even playing field. The easy answer would be to install each of them in the same extruder and to measure the amount of pull that each one generated before slipping or skipping... The problem with that plan is that each gear has slightly different dimensions and wouldn't all fit in the same extruder. That type of test would also result with a lot of skipped motor steps instead of actually testing the limits of each gear.

The solution? A test rig that can apply the same pressure for each test, one that keeps the gears from rotating, and one that is adjustable enough to reasonably accommodate the varied dimensions of the gears that I was testing.

My design uses a 3D printed fixture with a v-grooved filament path and a pivoting gear axle made out of clevis pins. The pins are ground down to fit 5mm and 8mm drive gears. The cross-drilled holes through both pins have the same spacing, so using one hole as a pivot point and the opposite hole to hang a weight from provides a consistent force against the filament regardless of which axle is being used.

gear test rig.stl
Download View in 3D

Step 2: The Gears

I procured 10 different drive gears from a handful of vendors. Interestingly, they were all a little different from each other. The exception was a duplicate Ultimaker drive gear that I found from two different suppliers (making it 11 gears in total, actually). The gears I tested were:

SeeMeCNC Hobbed Drive Roller 5mm drive

E3D Hobb-Goblin 5mm

E3D Hobb-Goblin 8mm

BangGood MK7 Drive Gear 8mm Drive

BangGood Ultimaker Knurled Gear

RobotDigg Ultimaker Drive Gear 5mm Drive

BangGood Brass Extruder Gear 5mm Drive

RobotDigg MK7 Drive Gear 5mm Drive

RobotDigg MK8 Drive Gear 5mm Drive

RobotDigg Drive Gear 5mm

RobotDigg Drive Gear 8mm

Step 3: The Test

To perform the test, I loaded the rig with a length of filament, hung a 6.2lb weight from the end of the axle (a large mic6 aluminum plate), and tightened the drive gear so that it was ideally aligned with the filament path. After everything was set, I gradually pulled on the filament until something happened. Some gears slipped smoothly, others sort of chattered down the filament, and a few really dug in and grabbed hard.

When I pulled the filament through each gear, I took crude measurements with a mechanical pull-scale to compare the amount of grip each gear had. I originally intended to use a digital luggage scale, but its peak hold function didn't work in a way that gave me results that I was happy with. For these initial tests, I actually used a fish scale from the hunting and camping aisle. Not a scientific instrument, but good enough.

After getting a fairly consistent measurement 3 or 4 pulls in a row, I'd loosen the grub screw and rotate the gear to expose fresh teeth and to run the test again. Some gears would build up plastic in the teeth, so turning the gear between tests was important for consistency.

I performed this type of test 3 separate times on each gear, and a fourth time if I had a measurement that seemed like an outlier.

*I fully understand that this method is not very scientific. My measurements can only be counted as approximations, and they're only good enough to rank the drive gears form best to worst. To obtain "real" measurements I would want to use a metal testing fixture and a real load-cell to test the pull of each gear.

Step 4: The Results

After testing all 11 gears, there was a wider spread in the results than I expected. A few of the gears surprised me, and a few of the gears seemed like they should have performed better. In no particular order, here are the measurements that I recorded. (Also summarized, in order, in the next step.)

Ultimaker Drive Gears

RobotDigg - 11.3 pounds

BangGood - 13.3 pounds

The two Ultimaker style gears were visually identical - I had to mark them to tell them apart. They're interesting to look at, and using a knurled surface instead of cut gears is certainly a unique design that provides some interesting possibilities like accommodating multiple filament diameters, positions, etc.

E3D HobGoblin 5mm

16 pounds

E3D cuts a very clean looking gear. Compared to some of the others, the teeth looked very precise and consistent, and the machining is clean. Out of all of the gears, this one and the MK8 have the smallest "drive diameter", which I think is advantageous to the torque that they can apply and the finer resolution that you can achieve with them (more steps per mm of extrusion).

E3D HobGoblin 8mm

17 pounds

The 8mm drive HobGoblin is just as pretty and precise as the 5mm variant. The teeth on both gears were deep, and wide enough to accommodate 1.75mm filament or 3mm. The drive diameter on this gear was smaller than the rest of the 8mm drive competition, so the same torque and resolution advantages would apply.

MK8 Drive Gear 5mm

9 pounds

The MK8 has a very small drive diameter, which is great. The teeth aren't as deep as some of the other gears, but that means that it has more of them which can help with certain types of filament. It was interesting to see the MK8 next to the 7's that I had because you can clearly see the differences between them and how they've evolved.

MK7 Drive Gear 5mm

11 pounds

The MK7 drive gear is widely used in the 3D world, and for good reason. It has lots of sharp little teeth, and it also has a deeper cut groove for the filament to pass through, which theoretically would give you more surface contact on the filament.

MK7 Drive Gear 8mm

9.6 pounds

The 8mm drive version of the MK7 seemed to be machined just a touch cleaner than the 5mm one. The teeth were cut a little more cleanly, and the surface finish of the machining looked a little more controlled.

RobotDigg 5mm Gear

20+ pounds

The RobotDigg gears feel sharp, and oh man do they bite. Both gears bit down aggressively, and I would reach the tensile strength of the filament by snapping it before getting the gears to slip. The filament was breaking near the middle of the pull, so I wasn't seeing failure due to it cutting the plastic. The 20# measurement is the maximum that I reached before the filament snapped, so I don't actually know what these gears would hold if that hadn't happened. I suspect that soon after this pull strength they would have cut the filament in half instead of slipping.

RobotDigg 8mm Drive Gear

20+ pounds

Both gears (5 and 8mm) are the same outside diameter, so it's feasible to use them interchangeably when swapping between motors. The gears appear to be EDM cut instead of machined with a mill or hobbing fixture. They also appear to be hardened, so I suspect that they'd stay sharp over time. The 8mm drive version of the gear has a very small drive diameter compared to some of the other gears, and because of that there's not a lot of meat for the set screw to bite into. That's my only criticism of the 8mm gear, as it seems like it wouldn't be impossible to strip out its set screw.

BangGood Brass Gear

17.3 pounds

The brass gear appears to be repurposed from something else because the tooth profile looks like it's designed to mesh with a second gear. This one is also different than the others because the teeth extend beyond the mounting sleeve diameter, the other gears have teeth cut into them. The amount of bite that this gear had surprised me, it pulled harder than I expected it to. It's also useful for a few different purposes because of the straight cut teeth, you're not married to a specific filament diameter or position.

SeeMeCNC Drive Gear

12.3 pounds

SeeMeCNC's drive gear is my old standby, the one in this picture has about 10km of filament under its belt. The diameter is relatively compact, the teeth are wide, and they have a sharp wire edge from the machining process that bites the filament nicely. This gear feels like a tank more than the other gears, and they're durable for the long haul.

Step 5: Conclusion

To summarize, here are the measurements from all of the tests, in order. An asterisk indicates the *8mm drive gears.

MK8 - 9#

MK7* - 9.7#

MK7 - 11#

Ultimaker - 11.3# (*&^%*O)

SeeMeCNC - 12.3#

Ultimaker - 13.3# (BangGood)

HobGoblin - 16#

HobGoblin* - 17#

Brass Gear - 17.3#

RobotDigg - 20+#

RobotDigg* - 20+#

One of the biggest things that I learned by doing this shootout is that the pull strength isn't the end all be all of performance. It's extremely important and may trump some other considerations, but it's not everything. Some other thoughts about selecting the right gear for you:

-Since most of the gears aren't available in two different motor shaft diameters, you need to decide which kind of motor you want to work with up front. The exceptions to this rule are the MK7 gears and the RobotDigg gears, one extruder could run either motor, if the motor mount holes allowed it.

-I think that smaller, more frequent teeth would provide a better grip on flexible filaments. I've printed NinjaFlex with the SeeMeCNC gear, but haven't tried some of the grittier ones to test my theory.

-A smaller drive diameter (the diameter of the active teeth) will give you a slightly greater torque before the motor would skip steps. Smaller diameter also raises your steps per mm on the extruder slightly and would give you a higher resolution.

I hope you've enjoyed my gear testing journey. If I add more to the collection, I'll make sure to post about it here. Let me know what your experience has been and what gears you like, too!