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

The Grin 10k is sealed in plastic but looks thick. My KTY experimental isn’t coated in plastic. I don’t have or know what epoxy etc. I should be using? I hope it’s simple, it doesn’t make me happy.

I'm not sure what the temperature limit of the plastic coated 10k Grin thermistor is. I might not use it inside the motor can, however, you could use it on the outside. Bury it in the bottom of a cooling fin, on the back side so it sees as little wind as possible, and cover it with a dab of silicone gasket maker to secure it and shield more wind. The silicone gasket material is pretty easy to pull off also. The current settings in your CA for temp protection assume the thermistor is against the motor windings so use a 160/170C soft/hard limit. With the sensor on the outside of the unit, you'll want to lower these limits, I would say 70/80C should be pretty conservative- this would safely allow the windings to be twice the temp as the casing. Sorry for the inconvenience, put the 10k on the outside and adjust the temp setpoints.

If you want to ship back the motor, I'll pay and install a new sensor against the motor windings. Tangent gearbox disassembly instructions- these show a gen2 unit, obviously if there's a separate 3220 engine, you'd remove the end cap with the 6 allen screws instead of the snapring, but otherwise instructions are the same.

1. Remove the snapring behind the freewheel using a flat blade screwdriver.



2. Remove the M6x14mm screw in the center of the freewheel (this holds the FW against the GBX output). Thread in a long M6 screw, I've begun shipping these with the kits, an M6x50mm should do it. These M6 threads are in the GBX output piece; the long M6 will thread far into the unit and push against the motor shaft. Without the snapring in place, the entire gearbox assembly will pull out of the housing. Don't drop it, don't get it dirty. Once the output cover and large bearing are free, the assembly should pull out easily by hand.


3. Remove the motor key so the rotor will pass through the seal.



4. Remove the motor cover snapring same as the FW snapring.




5. Remove the motor cover (it's easiest to push the rotor shaft towards the windings a bit, it'll push the cover out), push the rotor shaft towards the windings and pull out the rotor. Make note of the spacers and wave washers on the shaft and their arrangement, don't lose any when the rotor pops free of the stator (magnets).



6. Epoxy the thermistor to the motor windings. Route the wires through the windings so as the thermistor can be glued securely to the windings. Ensure the wires are not interfering with the rotor air gap, ensure no epoxy is left in the air gap. Use about any high temp, thermally conductive/electrically insulative epoxy such as JB Weld (no kidding, works great). Most epoxies are electrically insulators due to the much higher content of plastic than metal additives, but read the label to make sure.



7. Reinstall in reverse order. Careful with the insertion of the rotor, don't let it slam into place.

Mugenski and other Tangent users,
I have a first generation 3K tangent drive with a thun torque sensing B.B. (amp limited pedal assist adjusted with 3 grin position switch as well as a domino throttle).
I believe I am having the same issue as you with unexpected throttle surge. I often use my bike with a 40 percent (of max.80 amps) assist level and the throttle when needed. This is great for the steep & rocky technical climbs but going from assist only and then back to throttle I can get surge on initial throttle re-application that can be anywhere from scary to downright dangerous. Thanks for the insight that it occurs when switching durring non zero throttle input. I will test that next time I ride. Anyone out there riding tangent pedelec having this issue?
When setting Thrl->MinInput and Thrl->MaxInput to match actual throttle voltage range I can get different “live throttle in” readings in the SETUP THROT IN screen (CA3) from one time to another. Could this be a factor?

Mark, what is your 'up rate' set to? Not the 'fast up rate' (this is freewheel engagement). Only the newest 3.1 firmware on the CA allows for different throttle up-rates for PAS and throttle input. PAS usually feels better with a slow up-rate (<0.5msec/sec) whereas the throttle uses a higher up-rate to be responsive (4.0msec/sec is what I ship the kits with). A slow up-rate will dampen throttle input. The CA will use whichever input is higher between the throttle and PAS. If it takes too long for the throttle input to respond, you may be causing the surge by opening the throttle too far because the motor hasn't responded yet. If your up-rate is low, try increasing it but watch the PAS, it'll get more aggressive.

Dave,
Thanks for super quick reply. That was my first post!
I made suggested adjustments to throttle and PAS up-rates.
Throttle up rate was 4.0Msec/sec as you recommend for quick throttle response is now set at 3.0Msec/sec.
The screen on CA shown as PAS rate (I assume meaning PAS up-rate) was 0.05 is now 0.45Msec/sec
These adjustments appear to have solved the problem. One observation from test drive. If I am in assist mode (Thun signal) and then accelerate via throttle overriding Thun signal to max revs then coast down to a slower speed( human cadence) and then pedle enough to re engage Thun signal for assist sometimes slight surge or maybe just the more aggressive assist threshold you mentioned.
This drive continues to amaze me. My wife has a $4K Haibike that is pretty seamless and pleasant to ride but my All-Mountain bike with the Astro 3210 and your amazing reduction is infinitely more capable, powerful, fast and fun. I can’t imagine what the 6K version Is like.

mark saunto said:

I can’t imagine what the 6K version Is like.

Click to expand...

Pretty much is frocking crazy. I dont even have 6kw available (due to battery) currently and its just crazy what the motor/controller combo can do.

Tangent 3K users with PAS - Do you get runaway throttle?
It does not appear there is too many of us 3K drive users out there with the Thun torque sensing B.B. so I am hoping someone with a similar experience will see this post.
I have been making small throttle adjustments for many weeks trying to eliminate an intemitant runaway throttle. I made the (uprate) adjustments as described in Tangent Daves last post which helped and after a short test ride I thought I had it solved. (By the way Dave has been really fantastic with support but I hate to bug him with this intermittent problem and love his drive system in every other way). After an extensive multi hour ride yesterday it reared its ugly head again and I flipped the bike which is scary in the technical and very rocky terrain I ride in. “Mugenski” mentioned it in an earlier post and he thought it may have to do with switching when throttle was in a greater than zero voltage position but I have trouble monitoring exactly what is happened when it decides to go nuts. Might it have to do with how CA3 processes throttle signal, anyone with any insight?
I ride with 40 percent assist when climbing and in rock gardens which is a fair amount of power (40 percent of 60 amp maximum) as well as using throttle on and off. The CA must switch between whichever signal is dominant and sometimes when I come back onto throttle input I get a surge or runaway throttle (flip).
It takes a lot of finess in rough terrain but having proportional assist AND throttle is fantastic when it works. I am starting to questioning if I should continue with this configuration after I almost rode off the side of a cliff a few weeks ago. Any input on this would be much appreciated. Cheers.

This isn't a runaway throttle condition, right? Power cuts as it should when you close the throttle? Is the surge from more torque than you requested or is the amount of torque correct but applied abruptly?

Go ahead and alter the settings inside the CA, write down the values as they are now for a baseline. I like to change the values by 50% or more so the effect is drastic and obvious. First I might try increasing the 'stop threshold' to 45rpm or higher (thinking something wonky happening if the CA is flipping between PAS and throttle, so ensure the PAS signal is zero as soon as you stop pedaling- can also rock the pedals reverse to stop PAS immediately). Second, I might try increasing the 'assist average' to smooth out any PAS spikes. Third, I'd cut the throttle 'up rate' in half (after verifying in the SETUP THROT OUT the 'fast up rate' is 0.05msec/sec and the 'fast threshold' is 2.0amps---actually, check these two values first). You can also watch the value of the throttle input on the SETUP THROT IN main screen, check that the physical throttle device isn't sending a bad signal somehow (can also set the throttle fault voltage to be .05V higher than the max input to be safe). Fourth, I might update the CA to the new 3.1 firmware and give 'er another try.

-dave

Crank /spindle problem

Looks like I have bent the crank or spindle. The chain ring shows side to side movement when I rotate the crank. I experienced lots of this back in my bmx days having bent cranks, back when cranks were mostly single piece due to my extreme preference for one for forward over the other. I'm guessing that is the problem here too. Might just be the crank arms too. Will have to disassemble to be certain.

Dave is the crank spindle provided from sick bike parts? What about the crank arms? What are other options for the spindle?

I imagine some of the threaded trials crank arms would be much stronger.

Video of the wobble to come...

Mark, I also use Dave's 3210 with PAS and have experienced the same issues as you. Trying to tune them out with the CA. I feel this could be a lengthy process. I learn something new every time I ride however.

This little guy looks hungry. Let's feed him a gearbox...

Ooh I hope one of those things is heading for Sweden!

I really appreciate the thick thermal sensor wiring. I've accidentally torn the tiny one that Astro fits the regular motor once or twice. They use headphone wiring in it, thin as a hair. Easy to fix but annoying.

Is there any special service on the sealed gearbox? Should I stock up on some special grease and clean it out every 1000km?

Took the time to make and install our own temp sensors instead of risk Astro's tiny wires (which I think have caused a couple issues). I used the same KTY84/130 but with 22ga teflon wire plus soldering the crappy JST connection to the wire for security. I try to deal with all the issues a customer may see.

The grease is Nabtesco Vigogrease RE0, super cheap stuff, availabe online (ebay). I am unsure of the maintenance interval, but yes, it will consists of pulling the gearset out, wiping the old grease, applying new, re-installing gearset. 1000kms sounds like a good time to peek inside and check the condition of the lube. There's 20mL of supergrease sealed in each unit now- the only way out is past the o-ring behind the motor freewheel, so if anybody sees significant leakage, let me know. Pulling the gearset is easy and takes a couple minutes, will post procedure later.

Hello everybody!

I did my second home-to-office travel today, I love the tangent!!!

No particular issue during the installation except the installation of the crankset (who looked veiled but it's settled, I had squeezed a little too, it must be allowed time to align well during break-in).

I did no changes on the CA settings. That works like a charm.

However, my battery can only deliver 3kw (Triangle 14s6p from Em3ev). It's a bit frustrating because I reach very often this limitation.

Again, thanks a lot Dave, now I hope that will be work for a long time.

You're only half way to the fun zone. Loads of torque can be had at low RPM even with a small battery but big torque and high speeds need a big power supply. Careful though, once you feed Tangent big amps, you may not go back.

@tangentdave

a short "side-question" --> wahts teh overall width of the Drive (from outer motor-case-lid to the Freewheel ? )

Greets

Notger

The mounting face of the drive unit isn't the same as the outside. The mounting face sits flush with the BB face, in this arrangement, the motor FW chainline is about 13mm outboard of the BB face. The BB face sets the chainline, independent of the frame centerline.

The overall width of the unit, outside edge of FW to outside edge of motor case is 139mm. Drive unit housing width is 123mm.

Hello Dave,

I changed the current limit of the CA from 60A (max burst amp of BMS) to 70A: no issue except during long run at max speed. So I setted 90A (max for 6x30Q): no problem and... OMG so good this tangent

90A is 50% more torque than 60A; 120A is 30% more torque than 90A; 150A is 25% more torque than 120A...

Just a note about the build, the parts and the experience.

It began with a donor bike, a 2002 Stumpjumper. I found one with upgraded fork, shock brakes and wheels. To that I added a new drive kit, 3 x 6 amp/hour Luna batteries, a Luna power switch, Grin tech brake switch, and lots of wire and connectors.

Wiring the batteries took a while with 10 gauge wire. Wiring connections were crimped. One battery is in an under the seat bag, 2 in the triangle. Extra paths for the wire in and out of the bags were opened with a hot screw driver. Very neat. Using a Luna 52 v charger each pack was charged separately, voltage matched (using a light bulb for a load) to 1/10 volt and plugged into a parallel pack. I have since charged as a combined pack several times. I am aware of the potential problems – but none so far.

Note that the Luna switch appears to be only 12 gauge wire, a minor bottleneck for current. It has not heated up significantly. The switch was mounded through the side of the battery bag with a pair of washers. Discreet, easy to reach and so far, reliable.

Installing the drive was not difficult overall. With the proper bottom bracket spacers, the chain line was spot on. The chain line from bottom bracket to the 9 speed cluster was pretty well centered. After the first ride I blocked out the first gear (32) as unnecessary and requiring too much chain diagonal. Rather than using the provided stainless strap to secure the drive to the downtube, I used a monster wire tie. Easier to install and tighten. Has not been an issue.

I did the break in on the work stand. I was surprised at how much the drive loosened up with just a few minutes of operation. I was almost impossible to roll the bike backwards pre break in. Afterwards it moved readily. Drive operation has become smoother and quieter as the first 50 miles accumulated.

I’ve had no issues with the drive to crank chain after the first 10 miles. I did manage to throw the chain once early on. Since then, nothing. I changed nothing and have no idea how I threw it, or why it has been better. If anything, my riding has been rougher and less cautions as miles have accumulated. The chain tensioner seems adequate, though I wonder how long the thing will last at the speeds it sees. Has anyone worn one out yet?

I was surprised with the performance once I got out on the road. Throttle response is big! Once rolling it is probably like a light weight 100 cc bike with moderate tune. Off the line it is like that bike with the throttle wound up and the clutch dumped. I was fortunate not to have an audience when I landed on my ass in the middle of the street after a moments inattention. It went over very quickly!

There have been a few issues and challenges.

I continue to have skipping (chain on cog) on the larger cogs. This is also where there is the most diagonal chain run. Modest throttle – no problem. Full throttle it skips. The chain and cogs are virtually new with no significant wear. My next step will probably be a different brand cluster. If no luck there, I will custom build a cluster to move all used cogs closer to the ideal chain line.

I have begun to install PAS. The first issue was that the off-the-shelf 12 magnet sensors were too large to fit between the crank and the controller heat sink. I made a plexiglass adapter of smaller diameter using the same magnets which functions just fine. I am working on tuning the CA3 so that PAS is mellow enough so my wife isn’t spooked by the bike.

I noted a question some comments ago wondering if the contributor had bent his bottom bracket as the sprocket wobbles. I note that both freewheels – crank and drive unit – have some wobble. Wobbly freewheels are not that uncommon. The runout is approx. 1.5 mm lateral on a 42-tooth chainring on the bottom bracket.

I have two areas of less than ideal wire management. One is at the drive unit. I currently have a coil of wire tied cable in a very vulnerable position under the drive unit. No clear solution as yet – but I will find one. The other rat’s nest is under the CA3. Lots of wires meeting there. Lots of excess length. Again, wire tied. I will neaten that up when I have the final connections made for the PAS, computer links, phone charger and coffee maker.

Finally, a question about efficiency. As yet I have not matured sufficiently to do careful efficiency testing. Still tearing around on the road or in the woods with little regard for efficiency. Has anyone done careful experiments to see how efficient the unit can be at 15 mph on the level? I did remove the knobby tires which helped a lot, but I’d like to know what I can reasonably expect if I behave myself.

I remain awestruck by the quality and elegance of the unit. Just what I’d hoped for.

You going Dave?

For sure I'll be there. The ranch is beautiful with tons of private acreage for riding. All ebikes are welcome so come out and nerd out. There's lots, lots, lots of great riding. It's a 12 hour drive for me, it is a good time and gorgeous country. This is year 3, come out to an ebike playground for a day or two.

Chris, the host, is an ebike nut with a fleet of Santa Cruz/cyclone builds and now a fleet of the Fantic pedelecs as well. Last I heard, the format for the ebike events were still being finalized, I'll push for a gnarly option if anybody's like me and wants to push themselves and bikes hard and fast. I'm thinking an 'open' class race for the big bore throttle bikes, 12-15 mile trail race prob split into a couple laps, 30-45min race time, time trials format, functioning crankset required, no power/battery limitations. Say there's a 2 hour block of time to start your laps, you can start as many times as you want/need but your race time must consist of the consecutive 12-15miles. This gives people time to fix technical issues and restart or get a second try to reduce your time. Thoughts? Interest?

I'll have a new build, keep up if you can.

Free support/shade/beers (till they're gone)/good times for Tangent riders! (actually, anybody cool is always welcome). Come on out, let's go ride!

Bummer it’s not closer....this sounds right down my alley! I’m in the middle of the country and about 1600 miles away!

Take pictures for the rest of us.

Tom

If you really want to get gnarly Dave lets ditch the stopwatch and do a mass start. The old format doesn't give you enough time to fix any real issues or recharge your battery anyway

thunderdome style.