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

[SlowCo]

state of ebike controllers

Although bigger than an ESC, the Nucular controllers seem very promising:
https://endless-sphere.com/forums/viewtopic.php?f=31&t=75494

 

[knurf]

It's just a bit annoying, I mean this definitely looks like a phase runner on the top tube, around the 1:10 mark, and the bike seems to move decent enough (even though this is Bruno and he's always excited about stuff).

Ah well, I have a fun plan B too, build up something around a ATMega8 to simulate the RC pulse and just use a HV80. Use a simple ohm meter thing for the thermistor to get a feel for the motor temp. But would be cool to try to push 72v into the Astro...

[youtube]6ojifdokMKw[/youtube]

 

[tangentdave]

The drive in this video is from the first ever batch I made. If it's not the first customer unit, it's definitely first five. Those 40:1 gearboxes were insanely torquey.

 

[district9prawn]

Ahah oh crap Bruno's YT video of my bike is still out there.

However I can't put ANY sort of real load on it or it will cut out with phase over current error. I can sort of get it to stop doing that by raising the "current regulator bandwidth" setting but the best I've reached was not cutting out with error, instead losing any sort of power (its less than what a 250w bike puts out). I've also fiddled with the "PLL bandwidth" to no avail. Anyone have tried this setup?

If you are getting over current faults on the PR at high throttle try decreasing the current regulator Ki. You need to set current regulator bandwidth to 0 first to allow the Ki change to take. This problem mostly went away when I switched to a lower kv astro motor. (Higher inductance).

The other problem I had tuning the asi controller was the freewheel engagement. This was one area where I really liked the vescs. If you look at the source code it calculates the rotor position without needing to know the rotor speed. So it is mostly immune to sudden changes in speed (like when the freewheel engages).

I haven't had luck with the BAC800 (phaserunner) running either a 3210 or 3220, phase overcurrent errors abound when the motor is loaded. ASi didn't design these controllers with high speed inrunners in mind. I can get the BAC2000 to play nicely with a 3220, but that controller is sort of unwieldy on a MTB, could be an option for Qulbix-like bike. According to ASi, the control scheme of the BAC800 is different from the BAC2000 and they weren't surprised when I had issues with the smaller controller on RC engines.

I've had success with the larger asi controllers too on my neu bike. Despite the inductance being less than half their recommended minimum, it never trips the current fault and the bike goes from zero to death wobble in like 5 seconds.

I didn't know Flipsky were working on a 300A vesc. Good to hear. Their "6.6" vesc is excellent. If they release a 300A I assume it will be 80v max current sensors just like the official 75/300 one.

state of ebike controllers

If you are game you could always try one of these
http://teamtriforceuk.com/a200s-v1-3-pcb-kit/

 

[knurf]

Ahah oh crap Bruno's YT video of my bike is still out there.

Ah cool, was hoping to reach who ever put that bike together Looks totally usable from here.

If you are getting over current faults on the PR at high throttle try decreasing the current regulator Ki. You need to set current regulator bandwidth to 0 first to allow the Ki change to take. This problem mostly went away when I switched to a lower kv astro motor. (Higher inductance).

The other problem I had tuning the asi controller was the freewheel engagement. This was one area where I really liked the vescs. If you look at the source code it calculates the rotor position without needing to know the rotor speed. So it is mostly immune to sudden changes in speed (like when the freewheel engages).

Yeah I got it almost usable. It propels me, but far from excellent. Worst part is a horrible oscillation once it gets going a little faster.
I noticed it's a different beast compared to the simpler CC controllers. When it detects the torque it puts out going to no use because of low gearing, it eases down the rpm which confused me. I also think the speed limiter came into play and disturbed some of my first attempts.
I was sort of hoping you'd remember some values or even have some old xml dump of the settings..

 

[tangentdave]

We'll have a report on that big A200 ESC in the future, E1Allen is in possession of his kit and we're getting it set up on his GT Fury (the moving BB needs to use the ISCG tabs to hold rotation, his new mounting plate is on the way). We'll be able to compare how the Talon/CA combo rides compared to a big VESC. His engine is the Neutronics 3215/1Y, 220kV, 0.018mOhm.

-dave

 

[E1Allen]

I was hoping the side plate arrived today. I'll be busy all day tomorrow, but I have Friday off. So my plan is to test ride Friday with everything. As far as the A200, I ordered that back in July. At least I got a email saying it's awaiting shipment. So I'll at least get to test the talon prior to the A200 to see a difference.

 

[aCeMadMod]

hi I made my vesc water cooled made a real water block for it on my mill

it works well what am waiting for dual vesc . to use 2 of Dave gearbox . in a. Gti frame.


once I work how I can upload pic I show you all.≥ 8)

 

[aCeMadMod]

Hi Dave any others
I am asking for the best stock gearing for a 3220 And 40.1 Dave gearbox

What are you all using thank.

 

[cr0m08]

Hi Ace,
For referance, I’m running a 3220 3 turn, 20:1 gearbox on 12S LiPo. CA set for 120A max and most I’ve pulled so far is 110A with my gearing. As you are using 40:1 gearbox, and the 3220 has loads of torque, I would suggest you run 20T on motor and 24T receiving on the crank. That will give you a total reduction at crank output of 48:1, whereas mine is 38:1 (20T motor and 38T on crank 265rpm). Now at the other end, I’m running 32T on crank going to the lowest 6 cogs from an XT 9 speed cluster, 17-34T. Which is on a single wide block spider, so there are no gouges on my freewheel. That gives me a gear range of 200%, or roughly 25 - 50 Kph.
Given the torque of the 3220, I think Dave said the sweet spot for crank rpm is around 300 - 350rpm. I found when I had the early system, 40:1 gearbox and then a 2:1 to the crank (80:1 total or 125rpm at crank) it never used torque, the motor just screamed high rpm, and it was like a truck, just forever going through gears. But now my current system just pulls so hard, it winds out in each of the gears like a dirt bike, I even shift two gears at a time, so 3 shifts for my whole cluster!
Darryl in Perth.

 

[aCeMadMod]

Hi thank for math..
We're the hell do you find a 24teeth front ebike chianring?? In au

Yes will able to custom made some thing I get sick of this. If needed.

I like buy parts off the shelf

Any help would be good
thank you all 8)

 

[cr0m08]

My 24T came from Dave in one of the iterations we went through. He made a single sided spider to carry the 104 BCD chainring, and also used one of the cyclone kits 24T steel, which has the 5 holes BCD pattern of the flanged freewheels. Not sure if anyone else came up with an alternative?

 

[aCeMadMod]

Hi great Dave made a spider with 24teeth and 104bcd ring


Is they any way of. Take a pic and post here

Thank you

 

[cr0m08]

Just had a quick look through Dave's Tangent web site. Damn it, he's too efficient and got rid of all the older images , where there was a couple pics of it with 24T. I will go through my spares box tomorrow and load up a couple shots of it for you. Its basically the same as the standard spider, except the outer ends of the four arms were ground down to 3mm thick instead of 6mm I think it was. Sorry I don't want to get rid of mine as you never know when stuff like that comes in handy!

 

[cr0m08]

Here is the spider I spoke of. I used it with 24T on opposite side to its intended use, because the version 1.5 gearbox has drive chain on the outside rather than the inside as the spider was designed for on the first version gearboxes. Main body of spider is 6.4mm thick and the arms are 3.8mm thick.
Also might be worth to contact Dave as he may still have some of the 20:1 gearsets kicking around for the version one gearboxes. That would make it much easier for you to use off the shelf parts.

 

[knurf]

On the topic of (cheap) chainrings, I just mounted a couple of https://www.aliexpress.com/item/New...13-4c4d-b5ae-5ed1638e4469&transAbTest=ae803_3 Ekfan chainrings. Narrow wide, supposedly made of 7075-T6 and available in pretty colors. They're like 1/5 of the price of the ones I've been using, going to be interesting to see how well they last. At least they didn't implode on my first session.

 

[tangentdave]

Sickbikeparts and Cyclone-tw.com have 24t sprockets, these have the same 2.43"x5 bolt pattern as the flanged freewheels (2.43...2.53, can't remember of the top of my head). They're sprockets are soft and sometimes we had trouble with chain drops on the small 24t sprocket. And I think a 32t/32t on the freewheel crankset looks the best and is simplest to set up.

You'll probably want to gear your unit for 350RPM at the crankset, this makes a 32t front ring into a 32t cog the 30mph gear (27.5" tires). If you were to run a single speed, this is the gearing I would suggest. Hill climbing wheelie torque that still runs to a useful vehicle speed. It doesn't matter if you have a 3turn or 4turn 3220, 20t or 14t motor freewheel, 32t or 24t at the crankset, 20:1 or 40:1 gearbox. Gear it for 350RPM to start with and adjust it from there. 6kW @ 350RPM is 160Nm torque. The Talon can keep up with phase current demands at lower RPM, so watch out for bending bicycle components if your motor cadence is less than this.

Amazon and Aliexpress have a whole raft of cheap chainrings, I've used 'em without issue. Only problem is they're usually not countersunk for the chainring bolts.

cr0m08, try turning the 'max amps' in the CA up to 140A....

-dave

 

[cr0m08]

Dave,
I will look at playing a bit more with the amps, I’ve been running around 260rpm at cranks on fat 24” wheels. Already folded over cogs on the cluster when I powered on before shift completed!
But I have to wait still longer yet for medical clearance before I can ride, so frustrating. It’s been 2months already since a ride, a new wheelset waiting to try out, and I’m foaming at the mouth to get a ride. Fracture back takes some time!
Darryl

 

[cr0m08]

So... I finally got my two version 1.5 kits sorted and I think completed for now (Nukeproof 3220 for me, and KeeWee 3210 for my wife). Both got new wheelsets I built myself.
Alas fate has been a fickle mistress with me, just as I'm about to go on a 1st ride on new wheels, I got injured again (last 12months with shoulder injury from work). This time I slipped taking photos on a wet granite rock and ended up in a back brace for a few months with a compression fracture in L1 vertebrae. Groan!!!! So now I have to wait to see what the wheels have done for my 3220.
Don't it look pretty and ready to ride!

 

[knurf]

Ok maybe I spoke too soon. Those ekfan rings seem to be made of some kind of sturdy cheese rather than 7075. I never seen failures like this with the expensive stuff, and I've tried a lot bigger rings before (hope 44t @ 3kw+ just fine).

I'm also trying various brands of ebike rated chains on this rig and so far so good. 450km offroad with loads 2-7kw, no drama. Only a bent rear cassette (it has a weak design on cog #3).

Water and road salt does bad stuff, beware. I'm still running fine but only because I've now encased all wires+connectors in inner tubes and reassembled the ESC with silicon conformal coating and new thermal paste. Blasting ACF-50 in bare connectors only works for minutes at a time in this weather.

I did incinerate a Talon by accident and stupidity, I think (not 100%) I might had a weak/bad connection to one of the phases and didn't realise. It just stutters and won't turn in this case. But do not push more current at the talon in anger, like I did.. Big yellow fireball under the bike (Could have been caused by some dormant problem from rescuing the ESC from earlier water damage, but it worked fine for a good while after being dried out and coated.. hm).

 

[aCeMadMod]

Hi all I can you best use 5 hole 110Bcd
Chianrings they alot stonger
I Found years ago. With cyclone

All full steel $5 chain ring you don't need a spider

Dave seem to not reply his pm much days has fast he use too

You a lucky man cr0m08

You got 2 Get 1.5 get gearboxs. You can't ride them must drive you nuts
Dave stop my order for 2 X get 1.5 sill don't know why I was happy to wait..

Thank you uploads pic for spider 8)

 

[tangentdave]

There's discussion about whether a gearbox is necessary in an EV. I would love to have less moving parts on my bike but the gearbox accomplishes a critical function for our weight/size-are-a-priority application. The size of the motor decreases. It's obvious, everybody knows this, but when you can fit 8kW into your coffee mug, it drives the point home. If we were to move the weight of the gearset into more copper, what would the dimensions of an engine need to be in order to produce 100ft/lbs with 600RPM? Sort of like a giant hub motor... Sure, a high torque mid-drive pushes on the bike components pretty hard and a hub wouldn't, but mass centralization is key. I rode a new Zerode Taniwha last week with a Pinion gearbox, and I can feel the difference not having a cassette/derailleur hanging on the back wheel. The bike is so planted, you can feel how light both ends are. Let's not sacrifice the ride for the sake of easy drivetrain design.

-dave

 

[tangentdave]

I'm digging those 2.8" Minions and the Pulse. Not digging the cheddar Ekfan ring. Perhaps Ekfan decided to forgo stress analysis? Probably since all those Chinese chainrings on Amazon and Aliexpress are identical. We have almost reached full simulacrum, only a little farther.

 

[SlowCo]

...mass centralization is key. I rode a new Zerode Taniwha last week with a Pinion gearbox, and I can feel the difference not having a cassette/derailleur hanging on the back wheel. The bike is so planted, you can feel how light both ends are. Let's not sacrifice the ride for the sake of easy drivetrain design.

-dave

You know that would be the ultimate combo: a Tangent drive on a bike with a Pinion gearbox

 

[tangentdave]

I think the Effigear would lend itself better to the addition of a motor. Looks like they use 3 parallel shafts, one for pedal spindle then two for the gearing. If we could add power to the second shaft we could probably measure human watts moving between pedal spindle and second shaft....