Dave's 80:1 RC mid drive kit build log

Sorry guys, life. Took a job at an aerospace company and it's slowly depleting my soul, early morning by early morning. Design progress has been made on two versions of gen3 but no parts or testing yet. I'll be approaching this version a bit slower than previously. More copper, more metal, bigger bearings. Also, way less copper, way less metal, optimized bearings.

Lost Sierra is coming up...I hear there's real stakes in this year's race...

-dave
 
He lives!

I think I pinged you back in January or so about new Tangent kits - a couple months after your reply I figured that life had probably intruded, heh.

Good to hear that the spark is still alive!
 
tangentdave said:
Lost Sierra is coming up...I hear there's real money at stake in this year's race...

-dave

You're coming right? It might be real luna vs HPC chest puffing battle fest. It would be awesome if you swooped in and won the cash
 
Beautiful work, subbed :)

Mind if I pick your brains on something? I was thinking of this type of reduction for in-wheel drive on a very light 4 wheeler (trials car, space saver rims with motocross tyres). Seemed like it could be mounted on multiple cranks/eccentrics to leave the centre open for off-the-shelf hubs and drum brakes.

The aim was to try 3d printing the gear train, see what kind of abuse it could take but I didn't go beyond back of the envelope planning because it would only be single stage and it appeared to be equivalent to driving the large, coarse final drive gears at full motor speed with a regular multi stage reduction, ie. that the gears are essentially turning at full motor speed and would need to be extremely accurate to get any kind of efficiency, far more accurate than would be possible with consumer 3d printing.

Is that right or was I missing something? Hard to visualise how the loads work in this kind of gear train.
 
Calculate the amount of load on each gearbox output (torque), then make some assumptions regarding how much contact area you have, which gives you the pressure at the contact areas. It's a simple lookup in material data tables to see how much bending or deformation you will have. Then probably double the pressure due to inaccuracies and check again.

Surface finish is huge. Accuracy is huge. Amount of eccentricity is important. There is sliding in this geartrain. There will be higher pressures than you think. Cycloidals don't backspin well so you'd need to freewheel the gearbox to allow the vehicle to coast. Plastic anything will not last more than 2 seconds. There is no substitute for harder steel. Cycloidals are incredibly tough but not a magic bullet.

Gearboxes are for shrinking the size of the electric engine. Given enough space, they're not necessarily needed. A car for instance, has lots of room for battery and controller and copper. Power is power, there is no cheating science.

Torque (lb.in) = 63,025 x Power (HP) / Speed (RPM)

Power (HP) = Torque (lb.in) x Speed (RPM) / 63,025

Torque (N.m) = 9.5488 x Power (kW) / Speed (RPM)

Power (kW) = Torque (N.m) x Speed (RPM) / 9.5488

[courtesy http://wentec.com/unipower/calculators/power_torque.asp]

These equations are equivalent everywhere in the universe, minus the singularity. On the motor shaft. On the gearbox output. On the wheels. Gearboxes only convert speed to torque, they don't create anything, allow less current for more torque or make up for a mismatched electric system. The do give the vehicle designer options. Power is power.

I bet 3d printed cycloidals won't be great, even printed metal. Ideally the positional tolerances for the ring gear and cycloid are under 0.001" each, 0.005" is marginal and anything more 0.01" will bang itself to death. Several tolerances stack up in the case of cycloidals, more so than an involute planetary, so keeping everything tight without binding is important.

-dave
 
Thanks. Did some playing around with that afterwards, bearings as roller teeth would probably work but adds loads of weight so not much point. Some hydraulic pumps/motors use the same kind of thing and many have hardened rollers as the outer teeth. Tolerances are extremely tight on the hydraulic motors, almost a press fit but they probably depend on fluid pressure to balance things out.

You're still selling these units? A friend is on the fence about going electric, if he comes around I'd recommend them as they seem pretty close to the ideal for power to size and weight.
 
Would really love to know if tangentdave will sell me a 3210 tangent controller-I nuked mine. I had one of the very first few units-and it's been great. Barring that-has anyone successfully hooked up a Phaserunner to one of these?
 
No but just get a standard castle creations hv80, minor programming changes, that’s it. The phase runner is hard to get to behave well with a 3210, had no real great success with that.
Also consider opening the hv80 and glaze it with silicone conformal coating since they are highly susceptible for water damage otherwise. Replace the thermal paste with some non conductive cpu paste. Easily pushes up to 100a bursts.
 
It's probably more expensive, but the Talon HV 120 that Dave uses is a much better unit than HV80 in terms of thermal capacity, because of its bilt in BEC and greater actual mass. I believe this means that the unit has more aluminium around it and I think is better able to dissipate the heat build up. The actual amps can then be limited back to 60A (Dave's original setting, or more if wanted) by the CA V3. I'm running HV80 on my 3210 and HV160 on my 3220, but I have increased the cooling fins on both of them and I don't push them as hard as some by only running 12S LiPo not 14S which is possibly getting too close to the limits of the units.
 
Recently I bought a early Tangent set second hand because I wanted a MX feeling on a MTB. Drove a year with the Bafang but the lack of ground clearence was killing me. Got a 3 Kw and have it build beatiful inside a Kona Entourage Deluxe and was verry pleased with it, except when I was having the amps at 50 and suddenly the freewheel didn't gave the torque to the bike anymore. So after a closer inspection I saw that the threat for the freewheel was gone on the aluminium gearbox site, but the motor and gearbox was running fine. Anybody has the same experience an how do I get the bike back running. Sent Dave a mail but did not have anny responce.
 
Xtr6 said:
... suddenly the freewheel didn't gave the torque to the bike anymore. So after a closer inspection I saw that the threat for the freewheel was gone on the aluminium gearbox site, but the motor and gearbox was running fine. Anybody has the same experience an how do I get the bike back running. Sent Dave a mail but did not have anny responce.

Have a machine shop make/copy the adapter out of steel.
 
Have a machine shop make/copy the adapter out of steel.
Yeah this is actually something we should look into if we want to keep these things spinning for a little while longer. Anyone gone this route yet? I think Dave used a lot of different dimensions for the internals, so I wonder how much a machine shop would charge to clone one of these parts.

That output freewheel adapter piece is a weak spot (as is the freewheel itself, if it doesn't unthread itself like this, the freewheel will die in a few thousand km and it will then be cold welded to the threads of this piece). A replacement would keep one of mine units running for a good while.
 
I think it is one off the first generation. Going next week to a machineshop to ask for a solution. The drive is only 3 Kw. The torque is high when the bike is accelerated. Try to weld it with JB cold weld and it lasted 10 km with only 20 amps. on the controller. Se what te Guys at the machine shop are going to come with.
 
I think it is one off the first generation. Going next week to a machineshop to ask for a solution. The drive is only 3 Kw. The torque is high when the bike is accelerated. Try to weld it with JB cold weld and it lasted 10 km with only 20 amps. on the controller. Se what te Guys at the machine shop are going to come with.
JB Weld or other commercial epoxy won't cut it I'm afraid (I've tried and failed a few times). Maybe that more expensive belzona stuff would do, who knows. Even in combination with a wedge, it's very hard to get it to keep together for any real extent. One problem is that these epoxies get softer with increasing temperature, and the whole gearbox gets kinda heated, even out to the freewheel. But do tell what the machine shop guys says!
 
The fix with JB weld is holding 60 km now. I make 2 wedges and weld these inside the freewheel then I cut 2 small pieces out the alloy that where the same dimensions as the wedge where. Did a few rides with this and it is stil holding. It will probally not last but in the meanwhile I can ride the bike and have some fun. At this moment I have 50 Amps. and will keep it for a while like this. Will make a picture.
 
Yeah I did something exactly like that on one of my old units.

The JB epoxy is heat resistant but the bond is relatively weak and it is down to the two wedges to take most of the load. Mine broke after a few rides at 2500w, but I also had a problem with the pedal crank freewheel breaking up and pulling the chain all over the place, probably didn't help.

I might try a final time with better wedging and other more expensive industrial epoxy, but the only real solution is a new transfer piece for these units. Hopefully these parts can be synthesised in 7075 or even steel without being more expensive than say a new X1 pro.
 
Maybe if you guys ask Dave, he will send a CAD file and drawing for the part so you can get a quote from a machine shop for a run of 5 or 10 (not sure how many drives have been sold) replacements parts. Unit cost for a handful will be a fraction of the cost for one.
Seems like the right thing to do if he doesn't have time to make more parts.
 
PaulD said:
Maybe if you guys ask Dave, he will send a CAD file and drawing for the part so you can get a quote from a machine shop for a run of 5 or 10 (not sure how many drives have been sold) replacements parts. Unit cost for a handful will be a fraction of the cost for one.
Seems like the right thing to do if he doesn't have time to make more parts.
That's an idea, but Dave seems very busy nowadays. I've found it hard to reach him even for the smallest of favours. I know he's also used multiple variations of the same design, so matching a particular part to the correct CAD might be a problem. I can also imagine he'd rather not give away parts of his intellectual property for free just like that.
 
OK guys, I can make a new part and most likely more if needed.
For the first user in the US who will send me a complete kit with broken part, I will do it for $100.
I will evaluate the problem and let you know what I think is the best solution.
Maybe this post will get Dave going and he will spring to action.
If not, at least you have another option.
 
Ecyclist said:
OK guys, I can make a new part and most likely more if needed.
For the first user in the US who will send me a complete kit with broken part, I will do it for $100.
I will evaluate the problem and let you know what I think is the best solution.
Maybe this post will get Dave going and he will spring to action.
If not, at least you have another option.
No takers?
Never mind. I'm in my final preparation for my trip to Japan with my e-bike and I'm running out of time.
I will be there for one mouth. After that I will be playing catch-up game.
Good luck to you all.
 
Visited the tangentmotors Webpage.
Can I still buy a drive kit from them ?

Sent a message to Dave but no reply.
 
Back
Top