I like using this calculator.
2 Likes
The force at the contact patch isn’t the same as the force at the gear tooth, for example a gear which is 1/2 the diameter of the wheel will experience twice the force.
4 Likes
Some gear drive design tips from an engineering point of view.
The 5.5Nm value as calculated above gives the torque on the motor shaft.
Divide that by the radius of the gear and you get and you get how much force goes through the drive, that’s what @Flyboy calculated.
The force going through the drive is can be used for gear life expectancy calculations. Basically the drive can fail in two ways:
- wear: this is almost exclusively decided by the surface hardness of the gear and the quality of lubrication
- chip a teeth from too weak material: if material is too weak to begin with, the first time it sees the big load, it snaps. Lets call the point it snaps at as max load.
- chip a teeth from fatigue: I’ll try to simplify stuff a bit. Your material would break at max load as defined above. With steels, if the given load is less than ~0.3x max load, you can apply the load and remove it infinite amount of times. Lets call this acceptable load. If you apply a bigger load than the acceptable load, every time you apply the load and then remove it (do a load cycle), your material gets weaker. Until after x amount of cycles, it will snap.
If you apply some mechanical engineering math, you can determine your acceptable load for each material that doesn’t weaken the material, no matter how many cycles you do.
Here you can find that math if you want to learn more and dive deeper:
The calculations there gave the results that the ~6Nm shaft torque was still under the acceptable load with a mod1 17t gear.
Whereas my ~11Nm shaft torque on my BN gear drives was above the acceptable load of the mod1 17t gear, and I snapped a teeth off of that gear after maybe ~2000km of very hard riding.
Basically to strengthen the gears you have 3 options: increase modulus but keep diameter (stronger teeth), and keep modulus but increase diameter (reduced forces acting between the gears), increase gear width (distribute the force over more material)
One more design constraint I’d like to throw in is related to wear. Higher teeth count means a single teeth carries the lead for fewer degrees each rotation. There’s a maximum advised amount of degrees that the teeth can carry the load before the angle of the load becomes suboptimal and starts wasting energy and creating lots of heat. For standard gear teeth profiles, this results in about 17T of minimum recommended teeth count such that the gear teeth only works at the optimal angles.
Regarding wear one last thing, its very good practice to have the number of teeth on the two gears relative primes, meaning their greatest common divisor is 1. I like to use 17T for the smallest pinion, as 17T is prime, so no matter the teeth count of the other gear, the relative primeness is satisfied. This makes sure that every teeth will come into contact with every teeth, making sure wear stays even over time.
Essentially gear drive design can be summed up by following these principles: determine your shaft torque; if possible keep smallest teeth 17t or more but relative prime to other, use as wide gears as reasonable, play with modulus to find what gives you strong enough gears that you stay below maybe 1/4" of the material limit. Then either decide reduction ratio by maximum allowed diameter, or calculate diameter by ideal reduction ratio.
5 Likes
This is a huge help, thank you. I don’t expect these to last forever but it will be cool to see how hard I’m pushing the material.
1 Like
@Dinnye you make me feel like I’m really slow 
and really dumb
But I still love you dude!!! 
4 Likes
Lol yeah i just use things until they break to calculate how strong/weak they are 
6 Likes
There is no other way 
4 Likes
@Dinnye I really appreciate you taking the time to lay all of that out. I was trying to read and understand it last night but that information did not pair well with an edible 
. If I’m prompting Chat GPT correctly and if it isn’t purposely trying to kill me then it looks like these gears are well within acceptable limits. The steps it took are laid out below.
Chat GPT Certificate of Approval
System Overview – Mountain Board Gear Drive
Board / Drive Specs
-
Motor: Reacher 6395, 173 Kv
-
Battery: 18s Li-ion (75.6 V max, 66.6 V nominal)
-
Motor current limit: 120 A
-
Gear ratio: 5:1
-
Pinion:
-
16 teeth
-
Module 1.1
-
15 mm face width
-
Herringbone geometry
-
-
Driven gear: 80 teeth
-
Lubrication: Sealed housing, Lucas Red ’N’ Tacky grease
-
Application: Off-road electric mountain board
Material – M300 Maraging Steel (SLM)
Provided material data:
-
Density: 8 g/cm³
-
Tensile strength: 1000–1250 MPa
-
Tensile modulus: 135–185 GPa
-
Flexural modulus: 190–210 GPa
-
Elongation: 8–15%
-
100% infill (SLM solid)
Conservative design assumption used:
-
Tensile strength = 1000 MPa
-
Estimated yield strength ≈ 900 MPa
-
Through-hardened behavior (not case hardened)
Motor Torque Calculation
Motor torque constant:
Kt=602π⋅KvK_t = \frac{60}{2\pi \cdot Kv}Kt=2π⋅Kv60Kt=602π⋅173K_t = \frac{60}{2\pi \cdot 173}Kt=2π⋅17360Kt≈0.055 N\cdotpm/AK_t ≈ 0.055 \text{ N·m/A}Kt≈0.055 N\cdotpm/A
At 120 A:
Tmotor=120×0.055T_{motor} = 120 × 0.055Tmotor=120×0.055Tmotor≈6.6 N\cdotpmT_{motor} ≈ 6.6 \text{ N·m}Tmotor≈6.6 N\cdotpm
This is the controller-limited continuous torque.
Pinion Tooth Tangential Force
Pinion pitch diameter:
d=module×teethd = module × teethd=module×teethd=1.1×16=17.6 mm=0.0176 md = 1.1 × 16 = 17.6 \text{ mm} = 0.0176 \text{ m}d=1.1×16=17.6 mm=0.0176 m
Tangential force:
Ft=2TdF_t = \frac{2T}{d}Ft=d2TFt=2×6.60.0176F_t = \frac{2 × 6.6}{0.0176}Ft=0.01762×6.6Ft≈750 NF_t ≈ 750 \text{ N}Ft≈750 N
Conservative Shock Case
Assume 2× transient torque spike:
T=13.2 N\cdotpmT = 13.2 \text{ N·m}T=13.2 N\cdotpmFt≈1500 NF_t ≈ 1500 \text{ N}Ft≈1500 N
So realistic operating range:
-
Steady load ≈ 750 N
-
Conservative transient ≈ 1500 N
Tooth Bending Stress Estimate
Using conservative Lewis-form approximation for:
-
16T
-
Module 1.1
-
20° pressure angle equivalent
-
15 mm face width
Estimated bending stress:
| Condition | Approx. Bending Stress |
|---|---|
| 750 N | 120–160 MPa |
| 1500 N | 240–320 MPa |
Comparison to Material Strength
Conservative material yield estimate:
σyield≈900 MPa\sigma_{yield} ≈ 900 \text{ MPa}σyield≈900 MPa
Safety factor (steady):
900150≈6\frac{900}{150} ≈ 6150900≈6
Safety factor (shock):
900300≈3\frac{900}{300} ≈ 3300900≈3
So:
-
Continuous bending safety factor ≈ 6
-
Shock safety factor ≈ 3
This does not include load sharing from herringbone engagement, which further reduces actual stress.
Contact Stress Consideration
Contact stress scales with √load.
At 750–1500 N tangential force:
Estimated contact stress remains well below:
- Typical pitting limits for hardened steels (~1000+ MPa)
Given:
-
Sealed lubrication
-
Grease retention
-
Moderate pitch line velocity
Surface fatigue is not expected to be a limiting factor at these loads.
Capacity Comparison
Conservative industrial rating for similar steel gear geometry:
- Continuous tangential load capacity: 4000–8000 N
Your loading:
-
750 N steady (~10–20% of conservative rating)
-
1500 N transient (~20–40% of conservative rating)
Thus operating region is well below typical fatigue thresholds.
Engineering Conclusion
Based on:
-
Current-limited torque (120 A)
-
18s system voltage
-
16T module 1.1, 15 mm face width
-
M300 maraging steel (1000–1250 MPa tensile)
-
Sealed grease lubrication
The pinion operates at:
-
~10–20% of conservative continuous bending capacity
-
~20–40% under assumed transient shock loading
Calculated bending safety factor:
-
~6 continuous
-
~3 conservative transient
Failure mode is unlikely to be tooth bending or structural overload.
If degradation occurs over time, it would most likely be due to:
-
Surface wear
-
Lubrication breakdown
-
Misalignment-induced edge loading
-
Contamination ingress
Not material strength limitations.
This still doesn’t solve the relative prime number issue for even tooth wear. SLM parts have a slightly granular finish. It goes away on the teeth with use, but hunting tooth action would make the wear identical across all of the teeth. I’ll get some 17t ordered and keep an eye on the 16t that I’m currently using.
Chat GPT agrees with you
Why Relative Prime Tooth Counts Can Help
If you used something like:
- 17T and 85T (still 5:1)
gcd = 1
Now:
-
Each pinion tooth eventually mates with every driven gear tooth.
-
Wear is distributed more uniformly.
-
Minor tooth profile errors average out.
-
Noise harmonics reduce.
-
Localized pitting is less likely to repeat in the same spot.
This is called hunting tooth action.
Does It Matter in Your Case?
Let’s be practical.
You are operating at:
-
~10–20% of bending capacity
-
Very low contact stress
-
Sealed lubrication
So catastrophic localized fatigue is unlikely.
However, relative-prime tooth counts would:
-
Improve long-term wear distribution
-
Reduce harmonic noise
-
Reduce chance of patterned micropitting
-
Be more forgiving of slight SLM tooth profile variation
It’s a refinement, not a necessity.
When It Matters More
Hunting tooth ratios matter most when:
-
Gear is heavily loaded
-
Surface fatigue is the limiting factor
-
Gear runs millions of cycles
-
Noise matters (industrial gearboxes)
-
Manufacturing errors are present
You do have one interesting factor:
SLM gears have slightly higher profile variability than ground gears.
So hunting tooth action would help distribute that error.
Love the feedback. Please let me know if you see anything else that could be improved.
Here’s a link with some info on the steel used to make the gears.
1 Like
No worries! Reread some of it today and it was a bit of a word salad sometimes full of valuable info but definitely shows I was just about to go to sleep and didnt even proof read feel free to ask if you have any questions.
Some feedback to the “chatgpt approval”: it completely forgot about the existence of fatigue. The 900 MPa it guessed as material strength isn’t the number that anyone would consider acceptable repeated load. Acceptable repeated load is 25-30% of that number, especially for additive manufacturing. It might be up to 35-40% for substractive manufacturing. So lets say 250MPa in this case.
I don’t love that it didn’t write out the full equation how it got the 120-160MPa, but the correct value is 153MPa so it’s okay.
It’s below the 250MPa max acceptable load for fatigue resistance so you are good. The safety factors it calculated are completely messed up, but it doesn’t matter. Generally, on electric motors intending to peak at 25% material strength is already perfect fatigue resistance without extra safety factor, assuming the teeth is hard enough not to wear down (and thus weaken the teeth, by having thinner geometry) as they have very minimal torque ripple. Not to mention you won’t actually operate for much time at full torque. 60% of that magic 25% limit gives you fatigue resistance even with worn teeth.
I think especially with SLM, that teeth isn’t fully smooth to start with, relative prime teeth counts are a good idea.
However in the “chatgpt agrees” part, it moves to 17T but also increases wheel spur to 85T and claims thats relative prime - it messed up there. The only teeth counts on the wheel spur you shouldn’t run (with a prime teeth pinion) is a direct multiple of the pinion teeth count: 5x17=85 - as then its no longer relative prime, due to their new common divisors of 5 and 17.
Love the SLM drives though, they look pretty cool and are actually designed quite well. I’ll definitely be trying M300 SLM gears for my own drives. Did you order sandblasted matte or satin glossy finish?
1 Like
I never know when to trust that robot. It seems to give 80% accurate answers. Very interesting how it suggested increasing the driven gear to 85t. It did not suggest that originally. It changed that value in the “summary” that I asked for at the end. Glad you took the time to review.
I have ordered both sandblasted matte and satin glossy and can’t see a difference. To stay consistent I’ve just been ordering sandblasted. Very curious to see what you think when you get a set.
1 Like
@SirVesa This trail looks like a dream! What and where?
3 Likes
It’s forest route 5n12 that starts on Refugio road in Santa Barbara county. One way it’s a little over 4 miles of single lane mountain road asphalt. Switches to dirt with some brutal rock sections mixed in but worth it for the parts that are fun. Goes as far as your battery will. The entire ride is on a ridge so you can see the beach to the south of you and lake cachuma to the north.
5 Likes
Thank you! A friend and I have a sunseeking trip coming up and I wasn’t planning on going as far south as Santa Barbara but wholly cow that looks like heaven.
2 Likes
It’s really nice. One bad thing though, they just finished resurfacing the road. They went with chip seal. Whoever came up with that road surface design should be shot. It’s hard to predict when you’re going to lose traction. I think everyone in the group has fallen since the rework. Still fun to ride up there, just have to go a little slower.
There are some cool spots in North county too if you don’t want to drive that far south. Shoot me a dm if you want and I can send you some map locations.
2 Likes
Amen!!!
3 Likes
Hopefully all of the rain we’re getting will take care of some of the loose rock. Not a lot of experience with that type of road surface luckily. That place used to be a grippy paradise and I miss it.
1 Like
Shitty phone video incoming! Quick clip of part of our climb back from the beach at Point Sal. I’m really enjoying the SLM herringbone drives. Recorded this to capture how they sound, somewhere on the loudness scale between helical and straight cut.
Got some great tips on how to make the gears better after posting the files. That set with the improvements should be here early next week. I’m expecting these to run a little quieter with the improvements.
6 Likes
Nice playground dude.
It certainly looks like a place for 9 inch tires…
1 Like
It’s a great ride with a mix of asphalt and dirt that changes condition over the course of a year or so. Right now the fire breaks have had enough rain so the usually soft sand is rideable. We’re out there cruising over hills covered in grass and wildflowers.
The main trail is rideable on pretty much any AT board. If anyone finds themselves in Santa Barbara county make sure to ride at Point Sal, just do it early in the morning if it’s a weekend. Gets a little crowded at certain times of the year.
2 Likes
Major drive failure yesterday.
I lost all four bolts that pass through the hanger clamp into the threaded holes in the back of the gear tub. To make matters worse, the female threads are now stripped out.
It’s entirely possible that I forgot to use thread locker on this side. The other side is still tight.
Another potential cause is that there isn’t enough solid metal around each hole, weakening the walls of the threaded holes.
I’m leaning toward forgetting thread locker since this is the first time a failure like this has happened on any version of these drives. Loose bolts subject to vibration and impacts seem like they could destroy female threads pretty easily.
To be on the safe side, I’ve reworked the model to hopefully add a little strength to the threads. Some of the holes have been eliminated. The drive adjustment will be a little more limited but should still be able to find a good mounting angle.
I also reduced the diameter of the recess hole that slides over the hanger from 28mm to 26.5mm, adding a little more metal to the inside of the ring of tapped holes.
The tapped holes are deeper now too for more thread engagement.
I have a few other ideas to correct this if the threads decide to fail again but this was the easiest fix so really hoping it works. Replacement pieces should be here in a few weeks!
4 Likes