Erider torque voltage output consistency issue

[colsaunders2]

I picked up the T9 (36-pole version) in June and was having similar issues to those described above. Basically, a sufficiently-hard stop with regen (brake switch + throttle) would result in no Hwatts output. I had other issues that I was focused on fixing, so I used basic/cadence mode for a while. However, when my Baserunner died last month, I tried again with the "Infineon" controller (no dice) and then with the replacement Baserunner. Lo and behold, torque assist worked! 95% of the time, anyway. I've had a couple of instances where a sharp bump or carrying the kids trailer resulted in null output, but that's only been 2-3 times in 2 weeks (not every single ride, like before).

BR #2 - and this may be the key detail - does not have working speed passthru. My MAC (not GMAC) has regen but no internal speed sensor, so I'm running an external pickup. I also re-routed the CA3-WP and PAS cables below the downtube, perhaps keeping the gremlins away for the most post.

 

[matmat]

BR #2 - and this may be the key detail - does not have working speed passthru. My MAC (not GMAC) has regen but no internal speed sensor, so I'm running an external pickup. I also re-routed the CA3-WP and PAS cables below the downtube, perhaps keeping the gremlins away for the most post.

Interesting detail! Maybe the problem is in the speed passthru wire?

About 10 days ago, I made a simple change: I grounded my battery's negative terminal to the frame. Since then, I've not had one instance of the CA freaking out and no longer reading the ERider correctly. I've done all kinds of aggressive braking (as much as 1.4 kW regen!), adjusting regen using the throttle while pedaling, and using the throttle while pedaling. I still have some RF chokes in place, but I've removed the separate power supply that I had going to the ERider, i.e., the ERider is again powered by the CA. So, my setup is basically stock + some RF chokes on the GMAC/CA/ERider cables + battery negative terminal grounded to frame.

I will try to add some pictures in the next couple of days, but for me, it appears I can close this issue for the time being.

FWIW, IMO, the issue is squarely with the CA and is probably caused by EMI from the GMAC in the ground and/or power lines. The issue continued for me even when the ERider was powered by a separate power supply but could be fixed by simply power cycling the CA. However, power cycling the CA did not power cycle the ERider (its power supply was powered directly from the battery), which leads me to think that the problem was fundamentally in the CA.

I wish everyone dealing with this issue the best of luck and hope you can get back out there to ride safely and happily as soon as possible.

 

[matmat]

Here is a picture of the quick-and-dirty grounding loop I'm using to ground the battery negative to the frame:



FWIW, I still also have RF chokes in a few places:

• Between the PR and GMAC: Immediately out of the PR and just before the L10 connector bridge to the GMAC
• Between the PR and the CA: Immediately out of the PR and just before the CA
• Between the CA and the ERider: Immediately out of the ERider and just before the CA
• Between the CA and the throttle: Just before the CA
• Between the CA and the brakes: Just before the CA

But, as I noted before, having just the RF chokes and/or having the ERider powered separately did not fix the problem.

 

[SolarFreak]

@matmat:

Thank you for posting your cure info - I also have an eRider T9 & have been suffering with the same problem (DuckDuckGo pointed me to this thread). Ferrite cores helped somewhat, but your simple -ive grounding to the frame sounds promising. I'll try it tomorrow & report back with my results.

Thanks again.

 

[kuz]

I had the same issue. Some 10-15 minutes into into a trip torque signal from Erider drops. Resetting motor controller fixes the problem for some 5 minutes. I don't even have regen, no E-brakes either. I can't say a this time that I could correlate the onset of the problem with hitting a pothole.

Grounded battery negative wire - the problem's gone. Thank you, matmat!

My setup: G02 from stoke monkey rigged as front hub. CA3_DPS with external speed pickup (my G02 came without internal speed sensor.) Erider-T9. CA3 firmware CA3-15b3. As I said no regen no E-brakes. Yuba Mundo Classic is the platform.

 

[dreyco]

What gauge wire did you use for the grounding? I'm having a similar issue. I have a Baserunner installed in my battery housing so I'm not exactly sure how I'd ground that.

 

[dreyco]

Update on my side. I ground one of the two negative terminals that goes between my Baserunner and battery. I noticed that my human watt numbers now no longer go negative and seem able to stay constant but they do seem like they maybe go too high (250-350w). I'm no greek god on a bike so likely they are wrong. I'll see if I need to rezero the torque sensor.

 

[dreyco]

Update on my side again, looks like grounding the battery did indeed fix the issue. I did also halve the Torque scale down to 35N/m so that my Human Watts were in the realm of sanity but I'm still not convinced it's at all accurate, it's just somewhat more proportional to reality. At least now my Torque multipler won't send me to outer space when I step on the pedals.

 

[matmat]

Hi @dreyco, sorry about the delayed response. I don't often get to check ES.

I'm happy to hear it looks like grounding the -'ve terminal has resolved the issue for you, too.

Regarding the wire gauge, if your grounding loop also carries battery current, then I think you've got to go with something pretty sizable. To hazard a guess, I'd say maybe 1.5-2 mm of stranded copper in diameter (sorry, don't know what gauge that would be.) It sounds like in your case the grounding wire simple splits off from the -'ve terminal, so, I think even something light duty should be acceptable.

Actually, Grin asked me to test this fix by creating a grounding wire that has a fuse between the -'ve terminal and the frame. I just got around to building this tonight, and will report on back on 1. If the fix still works, and 2. If the fix doesn't work if I remove the fuse. Importantly, Grin thought the fuse would be a good idea in case another component shorted to ground via the frame (e.g., if a phase wire inside the motor became chafed), and that way, the fuse would blow instead of there being sparks (in this case, the current would travel from the motor to the frame and from the frame to the battery -'ve terminal through the grounding loop.) For now, I've set this up with a 1 A fuse. Again, I'll report back on if this works and if removing the fuse does indeed stop it from working.

 

[dreyco]

Thanks for the update! Appreciate it. Did they suggest 1A fuse or did you just guessimate it?

 

[dreyco]

Grin got back to me and suggested a 100mA fuse

 

[qwerkus]

Actually, Grin asked me to test this fix by creating a grounding wire that has a fuse between the -'ve terminal and the frame. I just got around to building this tonight, and will report on back on 1. If the fix still works, and 2. If the fix doesn't work if I remove the fuse. Importantly, Grin thought the fuse would be a good idea in case another component shorted to ground via the frame (e.g., if a phase wire inside the motor became chafed), and that way, the fuse would blow instead of there being sparks (in this case, the current would travel from the motor to the frame and from the frame to the battery -'ve terminal through the grounding loop.) For now, I've set this up with a 1 A fuse. Again, I'll report back on if this works and if removing the fuse does indeed stop it from working.

Absolutely! When I read about your "grounding the negative terminal" solution, it seemed way to dangerous. Using a small wire or a fuse is a must here.

 

[matmat]

Hi guys, interim update. The fix still works with the fuse between -'ve terminal and the frame.

@qwerkus, indeed, you were right and Grin agree with you. Thanks for pointing it out!

I now need to verify that it stops working if I remove the fuse.

Happy to hear that the fix has worked for others, too. I will try stepping down to a 100 mA fuse.

 

[matmat]

Hi Everyone,

I've verified that the fix stops working if I remove the fuse. Furthermore, if I remove the fuse and the CA gets in a bad state, I can fix the situation merely by reinserting the fuse, i.e., without power cycling the CA or the ERider.

Here're some pictures of the fuse and the wiring. In essence, it's just a Y-splitter on the ground line, with one leg connected to the frame via a fuse. It currently has a 1 A fuse, but I will probably switch it out to a 100 mA fuse as was recommended above by Grin.

Here're the fuses I bought: https://www.amazon.com/dp/B07F8RLMPB

And, the APP connectors and a crimper: https://www.amazon.com/dp/B07QPZWFHJ

It could probably be made a lot safer by orienting the fuse vertically and securing the wires a bit better.

 

[velijukka]

Update on my side. I ground one of the two negative terminals that goes between my Baserunner and battery. I noticed that my human watt numbers now no longer go negative and seem able to stay constant but they do seem like they maybe go too high (250-350w). I'm no greek god on a bike so likely they are wrong. I'll see if I need to rezero the torque sensor.

Can you put a picture of the grounding? I also have a Baserunner L10 and a similar problem with the sensor

 

[matmat]

Hi Everyone, I've verified that the fix still works with a 100 mA fuse.

There is maybe a little bit more wobbliness in the CA's estimate of the motor power than before, but it's probably just my initial impression. In any case, the issue with the CA's measurement of the ERider's torque output flatlining has not happened again.

I will update if the issue returns and/or I need to make any adjustments to the fuse rating.

 

[dreyco]

Update on my side. I ground one of the two negative terminals that goes between my Baserunner and battery. I noticed that my human watt numbers now no longer go negative and seem able to stay constant but they do seem like they maybe go too high (250-350w). I'm no greek god on a bike so likely they are wrong. I'll see if I need to rezero the torque sensor.

Can you put a picture of the grounding? I also have a Baserunner L10 and a similar problem with the sensor

I'm almost too embarrassed to take a photo. It's just an 18ga wire "tied" around one of the negative prongs and then a loop of bare wire wrapped around a water bottlecage screw that goes into the frame.

 

[SolarFreak]

So, after a bit of an illness & a lot of constant horizontal rain it finally cleared up & I felt well enough to test out the fix for the first time yesterday. I kept the ferrite cores in situ & added a -ive earth from the battery to the frame with an inline 100mA fuse fitted for safety & it seems that all my issues have gone - fantastic! Everything now works as it should - no cut-outs, no irregular pulses & a smooth as silk torque action at last - what a difference!

As a FYI - I was a little concerned about weather it would play nicely with my AC hub dynamo or if it would cause my lights to fail, but I needn't have worried as the hub dynamo & lights/cache battery work & charge fine.

Thank you to everyone who helped test this simple but perfect solution - this issue had been baffling me for months, I'm so happy - thanks again to all involved.

 

[cnrd]

Crossing my fingers, I just ordered a fuse holder and 100mA fuse.

I have been battling various noise related issues since I got my GMAC 2 years ago, including a resetting ERider T9.

Are you just connecting the GND wire to the frame? And should it be done close to any specific component?

 

[SolarFreak]

Crossing my fingers, I just ordered a fuse holder and 100mA fuse.

I have been battling various noise related issues since I got my GMAC 2 years ago, including a resetting ERider T9.

Are you just connecting the GND wire to the frame? And should it be done close to any specific component?

As long as it's a "good" GND to the frame & that it continues throughout the frame you should be good. I even checked that it continued to the pedal crank itself, seeing as that's where the sensors involved are, just to be sure.

I don't know how or why this fix works so well or why it's only needed on certain eRider T9 setups & why that is, but I would be very interested to find out why. If anyone could enlighten me I'd be grateful.

Edit: I've now been able to test it thoroughly in both wet & dry weather & can confirm it works perfectly.

 

[cnrd]

As long as it's a "good" GND to the frame & that it continues throughout the frame you should be good. I even checked that it continued to the pedal crank itself, seeing as that's where the sensors involved are, just to be sure.

I don't know how or why this fix works so well or why it's only needed on certain eRider T9 setups & why that is, but I would be very interested to find out why. If anyone could enlighten me I'd be grateful.

Edit: I've now been able to test it thoroughly in both wet & dry weather & can confirm it works perfectly.

Thanks

I ended up just crimping a ring terminal onto a piece of wire and crimped that to the fuse holder wire.

Luckily I already have a central terminal block for both the positive and GND wires, so I just connected into that and then screwed the ring terminal to the closest screw.

Unfortunately it is currently snowing too much for me to comfortably test it, I did just do a small test run up and down the street to make sure that everything still worked, will report back when I get to take a longer run!

 

[velijukka]

Already two ERider-T9N torque sensors returned to Grin. The other gave only about 20-30 HW of torque signal. Another work somehow, sometimes gave more than 300HW, sometimes 25-50 HW, sometimes fell asleep completely, no logic in action.

Driven about 1300 km of test run with torque sensors, not a single distance (20-40km) went smoothly. Battery grounding did not help.
A normal pedal sensor had to be installed until Grin figured out the cause.

 

[matmat]

@velijukka, out of curiosity, did you verify that your grounding wire made good contact with the metal of the frame (and not just the surface?)

 

[velijukka]

@velijukka, out of curiosity, did you verify that your grounding wire made good contact with the metal of the frame (and not just the surface?)

Grounding to the frame is OK. It looks like even an ordinary pedal sensor requires that ground. I had to switch to one because I didn't get the torque sensors to work properly.

 

[DerekWillmott]

Update on my side. I ground one of the two negative terminals that goes between my Baserunner and battery. I noticed that my human watt numbers now no longer go negative and seem able to stay constant but they do seem like they maybe go too high (250-350w). I'm no greek god on a bike so likely they are wrong. I'll see if I need to rezero the torque sensor.

Dreyco, can you describe or post a picture of how you tapped into the negative lead? Is your baserunner installed in the battery mounting rail?