Strange regen and PAS behavior after new T9 install

[bombadero]

Hello, I had an arc fault a few months ago that has taken a while to repair. I've recently replaced the torque sensor, an ERider T9N, and while my bike is mostly restored now, the regen, human watts and other metrics are diplaying outlandish numbers and I have to set the starting watts of my PAS to -30 HW in the CA3 PAS screen to get it to work at all. That's the TL;DR. More on the arc fault, which is now fully repaired, can be found on this separate thread.

If you've read this far, here's a bulleted summary in between the TL;DR and reading that other thread in full:

0. I have a dual-motor, dual-battery, 52V e-bike wired in parallel. His name is Bruce the Moose. He is a bike tractor.
1. I had an accident that tacoed my rear wheel; I did not realize at the time, but it also mechanically damaged an XT-90 connector coming out of the rear controller.
2. The negative wire on the XT-90 came loose from the solder cup during a ride, causing an arc fault that fried my CA3, wire harness and torque sensor.
3. I soldered a new, custom wire harness with XT-90 and XT-90S connectors that is shorter and better protected by the frame, and repalced my CA3. So far, was so good.
4. I replaced my T9N sensor with a new unit from Grin, but had weird issues with it. It would only provide assistance if the bike was pedaled in reverse. There are some additional details to this I can provide here or you can read about them towards the end of the other thread.
5. I replaced the T9N again with a sensor from Aliexpress, and the bike now works but I have these weird outputs and have to set the start threshhold to -30 HW to gain any assistance, and even at that level I have to pedal pretty hard to get assistance.

I should also mention the assistance is a little inconsistent. Every once in a while it cuts out altogether at speed (but well below top speed). In these cases, If I hit the throttle, it becomes available again after I release the throttle. At other times, it will kick in with very light pedaling, which one would expect with a negative 30 threshhold, but when starting from zero I almost always have to use the throttle no matter how hard I pedal to start.

I have ordered yet another sensor from Grin that I haven't picked up yet. One thing still to try is to install that sensor and compare to the other two. Summary of the sensors/history for clarity:

0. Original sensor. Burnt out in arc fault.
1. 1st replacement (Grin). Damaged, provided no assistance (magnet spindle somehow bent).
2. Hybrid sensor (new Grin sensor + old sensor's un-bent magnet spindle). Only provided asistance in reverse. When setting 5V gain to reverse (which is the opposite of what this sensor requires), rather than flipping the behavior as expected, this hybird sensor instead provides no assistance.
3. 2nd replacement (Aliexpress). Currently installed, working with described wonky behavior.
4. 3rd replacement (Grin). Not yet picked up or installed for comparison. TBD.

Photos of the strange outputs from the CA3:


I used -47 Ah on my 14 mile ride!!!

My bike is a power generator apparently!!!

 

[amberwolf]

assistance.
I should also mention the assistance is a little inconsistent. Every once in a while it cuts out altogether at speed (but well below top speed). In these cases, If I hit the throttle, it becomes available again after I release the throttle. At other times, it will kick in with very light pedaling, which one would expect with a negative 30 threshhold, but when starting from zero I almost always have to use the throttle no matter how hard I pedal to start.

Inconsistent operation of a device usually means there is an intermittent connection, or electrical noise on a signal line.

Other people have had noise issues with the Erider; some solutions proposed and some tested here:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=112368&hilit=erider&start=50
and here
https://endless-sphere.com/forums/viewtopic.php?f=2&t=107750&hilit=erider
and probably elsewhere in one or more of these.
https://endless-sphere.com/forums/search.php?keywords=erider&terms=all&author=&sc=1&sf=titleonly&sr=topics&sk=t&sd=d&st=0&ch=300&t=0&submit=Search

Photos of the strange outputs from the CA3:
20220402_161711.jpg
I used -47 Ah on my 14 mile ride!!!
If that always happens, negative current, then the shunt is wired backwards (blue and white wires swapped to the CA).

If it is inconsistent or intermittent, there is a poor connection between the shunt and the CA.

If it's some other result, would need more specific details to work out the cause.

 

[bombadero]

Okay, been a while since I posted here as I've been going back and forth with Grin info on this issue. I've made a small amount of progress. Justin suggested I upgrade to the new 3-2b3 firmware, thinking that might fix my issues. Unfortunately, it made things slightly worse: all of the old symptoms, plus the new symptom of having to tap the throttle every time after coasting to get PA to kick back in. After a fruther suggestion from Justin, still under the new firmware, I changed from Active Hi/0.5 mOhm in the calibration display to Active Lo/1.000 mOhm/Dual Motors=Yes. This, thankfully, resolved all of my PAS issues immediately, so that's progress :thumb:. The bike runs great now--smooth and responsive.

Unfortunately, there are still some annoying little issues remaining with the CA3 logging negative Ah and bizarrely high regen numbers. I examined my new wire harness thoroughly and there are no crossed wires anywhere and everything still probes out to ~58.8V. With PAS issues resolved, I think it's safe to rule out the T9, the PAS cable and EMF noise now. I examined both the Baserunner and Phaserunner in the Phaserunner suite and according to it, they are running perfectly. I thought maybe it could be the brake wire, but that just tells the regen whether or not to kick in. It seems doubtful the motors themselves would cause any issues.

To me that leaves two candidates: either wires damaged on the Baserunner by the arc fault perhaps, which wouldn't show up in the Phaserunner suite but hypothetically corrupt the data downstream, although I don't know if that's even a feasible possibility; or your earlier suggestion about a swapped shunt wire on the new CA3. Since the regen behavior was fine with the old CA3, this swapping wouldn't be in the Baserunner or the CA-WP extension cable. It would have to be in the CA3 itself, either in the cable or at the solder points on the PCB.

There are two ways I have available to tackle this: either I repair the old CA3, which appears to have a burnt out Q1; or I open up the brand new CA3. Since the bike is otherwise running fine and I would like to continue using it, I would prefer to try repairing the old CA3 to see if it exhibits different behavior. Unfortunately Grin has not been forthcoming with how to source that component, or the display, which is another candidate for being damaged. Q1 has 'SIDN 2450 194R00H' printed on it, but that doesn't yield any results on Google. Unfortunately I don't have lot of experience hacking around with electronics other than soldering a display or some GPIO devices purchased from Adafruit onto a raspi, so I don't know how to obtain those components. I could probably find a display easily enough just by searching for one with the same dimensions, but the Q1 component has me stumped. Presumably it would be a pretty easy solder job and I could pretty quickly test whether the new CA3 is at fault or the problem lies elsewhere.

On the side, another thing I wonder is where the regen calculations take place, in the main motor controller or the CA3. I assume it happens in the CA3 rather than in the motor controller with the CA3 as a dumb terminal, but I don't really know.

If it is feasible that damaged wires coming out of the Baserunner could be corrupting the data coming out of it, the only thing I could really do is replace the controller since it's potted. Not the cheapest solution; I might do so anyway at some point and replace it with a Phaserunner/Frankenrunner just because 1) it's kind of a pain getting at the Baserunner on my bike to effect repairs or examine it and 2) since I had to re-wire it for a dual-battery parallel setup, the wires have to take two immediate, and very tight, 90 degree turns out of the Baserunner, which seems less than ideal for longevity for those wires. If I do that, I may just leave the Baserunner down in that rear Reention housing, unplugged, as a back-up controller.

Side note...

...I finally got my sidecar frame. It took forever because the guy who makes them is just a Swedish dude in his garage. He has one NA distributor, a small chain of e-bike stores out here in California called Mad Dogs & Englishmen, after the song. They shipped the frame directly to me, but unfortunately sent me the wrong frame. There is an 18" frame--referring to the wheel size--that will work on smaller bikes with 24" and smaller wheels, and a 20" frame for bikes with 26" tires and above. I ordered the large but they shipped me the small. Since they have a shop in Mill Valley and the shop manager lives in SF, he was able to swap frames with me at my garage door, but for some reason it took like two months to arrange that.

But anyway, it has arrived, and I've started to prototype with planks of wood and crates to try to get the dimensions right for my own sidecar fab:



It works great, but needs some work, both in terms of the product itself and my own usage of it. This is what I have learned so far:

• Unlike on a trailer, hitting the curb with the tire is pretty dangerous. On a trailer, there is enough range/axes of motion, that the trailer will rebound or the tire will roll up on the curb. On the sidecar, it's attached physically to the bike, which is quite heavy, so you don't rebound enough and just kind of grind into the curb. This can obviously be quite dangerous at high speeds. In my own build, I plan to add skate or rollerblade wheels vertically on the outside of the enclosure, and build it to overhang the wheel. That way, just those wheels will hit things like curbs, and roll off of or along them.

• The mounting hardware for the chainstay bracket just consists of a few M6 SS u-bolts, which I found tend to get loose after hitting a number of potholes and speed bumps following just a couple of rides. There is an M6 threaded hole near the front of this bracket in the center along the chainstay however, so I put a Grin Bottle Bob behind that hole clamped with a sufficiently small SS hose clamp that I happened to have on hand from a recent garden hose repair. Not ideal, as the Bottle Bob takes an M5, so I'm just using that M6 threaded hole as a pass through, but it locks the bracket firmly in place. Speaking of which, on my bike I can't get the u-bolt that is supposed to attach to the seat stay into a position where it won't block my chain shifting from like 3rd to 4th on my high range. I have a beefy rack on the back, so I attached it to the rack stays instead, which is less than ideal, but with that extra Bottle Bob hardware on the front, it seems to work just fine.

• Security is an issue. It uses a custom 240mm quick release axle to attach to the bracket bolted to the stay. I thought of using a Hexlox axle, but they only make wheel axles in standard sizes, and there is no standard wheel axle as big as 240mm, they only go up to 183. I thought of using a sleeve nut on the end to secure it, which has an allen socket, so I could put a Hexlox magnet on there, but they only come in M12x1.75 for some reason, and bike axles are M12x1.5. Same issue with round allen nuts. What I settled on is ordering a 12mm set screw collar with an M6 set screw that I will place around the axle, tightening the set screw and placing a Hexlox in the set screw. I also placed a Hexlox in the Bottle Bob's M5 bolt's cap, and if need be will lock the wheel with a folding lock when parking the bike.

• Since it has no suspension of its own, it makes the bike a little wobbly.

• The generic Mitas tire it ships with is decent, but I ordered some Schwalbe 20" Pick-up tires for it, which have a load capacity of 150 kg.

• If what you have on the sidecar frame is too far to the left, you cannot pedal on righthand turns; only coast or use the throttle. I plan to build it far enough to the right so this is not an issue. I stacked 3 crates on top of each other today to do a test and found that with something this tall on the frame, your right hip runs into it on righthand turns, preventing you from banking further. ~18" seems to be about as high as you can go, which as it turns out is about how tall his own sidecar bodies are. I might go taller than that actually, but simply slope it away from the rider at a 45 degree angle at that point, then level off horizontally. I will prototype that first, of course. It would kind of complicate having a top hatch though.

• You can temporarily use the sidecar as a kickstand, but the bike will still fall eventually, so you still need to use a kickstand. A cool alternative in the future however could be to find a way to attach the bike rigidly to the sidecar with a handlebar brace.

If I post more progress about the sidecar I'll start a new thread for that.

 

[amberwolf]

I changed from Active Hi/0.5 mOhm in the calibration display to Active Lo/1.000 mOhm/Dual Motors=Yes.

Do you mean "high range" and "low range"? (I don't know of anything relating to calibration using "active" and hi or lo).

Unfortunately, there are still some annoying little issues remaining with the CA3 logging negative Ah and bizarrely high regen numbers. I examined my new wire harness thoroughly and there are no crossed wires anywhere and everything still probes out to ~58.8V. With PAS issues resolved, I think it's safe to rule out the T9, the PAS cable and EMF noise now. I examined both the Baserunner and Phaserunner in the Phaserunner suite and according to it, they are running perfectly. I thought maybe it could be the brake wire, but that just tells the regen whether or not to kick in. It seems doubtful the motors themselves would cause any issues.

If during normal operation you see negative current flow, and during regen braking you see positive current flow, then somewhere between the CA PCB and the shunt the two wires (blue and white) are swapped.

Wire damage wouldn't cause that; they'd have to be physically reversed to cause this to happen.

If current flow is reversed, regen will show actual power usage instead of regen, and actual power usage will be the regen instead. (or rather, it will be a negative number representing the difference between the actual power and the actual regen).

If you see negative current flow for both regen and traction motor usage, I can't imagine anything other than a software issue inside the CA itself that could possibly cause that.

To me that leaves two candidates: either wires damaged on the Baserunner by the arc fault perhaps, which wouldn't show up in the Phaserunner suite but hypothetically corrupt the data downstream, although I don't know if that's even a feasible possibility; or your earlier suggestion about a swapped shunt wire on the new CA3.
<snip>
On the side, another thing I wonder is where the regen calculations take place, in the main motor controller or the CA3. I assume it happens in the CA3 rather than in the motor controller with the CA3 as a dumb terminal, but I don't really know.

There is no data to corrupt; there is no communication between the CA and any other device, except when you connect a computer to the serial cable and run the setup program. (if you use the Analogger, there is serial *output* from the CA to it, but there is nothing going into the CA except when you are running the setup program from a computer).

All calculations for data the CA uses and displays happen inside the CA; there is nowhere else for them to happen.

There are two ways I have available to tackle this: either I repair the old CA3, which appears to have a burnt out Q1; or I open up the brand new CA3. Since the bike is otherwise running fine and I would like to continue using it, I would prefer to try repairing the old CA3 to see if it exhibits different behavior. Unfortunately Grin has not been forthcoming with how to source that component, or the display, which is another candidate for being damaged. Q1 has 'SIDN 2450 194R00H' printed on it, but that doesn't yield any results on Google. Unfortunately I don't have lot of experience hacking around with electronics other than soldering a display or some GPIO devices purchased from Adafruit onto a raspi, so I don't know how to obtain those components. I could probably find a display easily enough just by searching for one with the same dimensions, but the Q1 component has me stumped. Presumably it would be a pretty easy solder job and I could pretty quickly test whether the new CA3 is at fault or the problem lies elsewhere.

There is some info about this kind of repair in my CA Repair Thread. I don't remember the part number, but I do recall that unless they have changed the design, it uses a Depletion Mode FET, rather than the typical FET, as a pass-thru regulator.
https://endless-sphere.com/forums/viewtopic.php?f=2&t=19540&start=25#p678269


Note that you can also bypass the entire regulator and run the CA on a low voltage instead; the same post says how to test that way, and there is a header on the CA to run using an external power source that's low voltage, and use a separate pin on the header to monitor the battery voltage. There should be info on how to do that in the CA manual for the high-voltage setup, I think. I may even describe it in that thread, since I ended up running that first CAv3 I got that way after I blew up the Q1/etc.

 

[bombadero]

Do you mean "high range" and "low range"? (I don't know of anything relating to calibration using "active" and hi or lo).

Sorry, yes, I meant hi and lo range.

There is some info about this kind of repair in my CA Repair Thread.

Thanks, although the product number of depletion mode FET referenced there is sold out everywhere. It seems like there are a lot of other depletion mode FETs available, but I don't know what voltage and amperage I'm supposed to be looking for, and they come in a lot of flavors.

Looking at the barrel plug on the CA3-WP cable, it looks like it is a preassembled component. I'm not sure how to take it apart even if I cut the cable, as it's overmolded plastic with the leads seemingly glued in place inside it. So it seems unlikely the cable would be the culprit. It seems like if there is a QC issue then with the blue and white leads reversed on S+/S- it would be inside the new CA3. So I guess I need to open up my brand new replacement CA3 to see if white is soldered to S+ and blue to S-.

 

[bombadero]

And Grin info just confirmed that as the best approach.

 

[bombadero]

So I opened up the CA3 and the white wire is soldered to S+ and the blue wire is soldered to S-, which is correct. So the problem appears to lie elsewhere.

 

[amberwolf]

It's possible that the wires are swapped inside the overmolded connector. I've had this kind of error happen with all sorts of molded connectors over the decades, including more than one standard IEC-ended AC power cord.

If you measure continuity (ohms) between the pins they should be on at that connector end and the PCB pads inside the CA, are they correct?


If none of the wiring or parts that contains shunt wires or shunts, other than what is attached to the new CA, has changed from the time it worked, then the problem has to be within the new CA or it's attached wiring.

If anything else with shunt or shunt wiring has changed, then it's possible for the problem to be within those things as well.



Regarding the Q1 FET, this is the datasheet for the p/n IXTY08N50D2-ND posted by one user in that thread
https://ixapps.ixys.com/DataSheet/DS100178A(IXTY-TA-TP08N50D2).pdf
with basic specs of "N-Channel 500 V 800mA (Tc) 60W (Tc) Surface Mount TO-252AA"
I don't know what p/n is used in any of the CA itself that I have; thought I ahd posted that but apparently not.

Justin's image of the schematic shows DN2470K4 for Q1,
N-Channel 700 V 170mA (Tj) 2.5W (Ta) Surface Mount TO-252, (D-Pak)
https://www.microchip.com/mymicrochip/filehandler.aspx?ddocname=en570581

So presumably an N-channel Depletion-mode FET that can handle 500V (drain-source), with at least 170mA D-S current, and at least 2.5W power handling, in the TO-252 format, would work. Given Justin's note about turn-on voltage, it probably has to have a gate turn on voltage of <10v (whatever the zener is in the CA) to work without changing that zener.

This is the best filtering I could get from Digikey for stuff they actually have in stock
https://www.digikey.com/en/products/filter/transistors-fets-mosfets-single/278?s=N4IgjCBcoGwJxVAYygMwIYBsDOBTANCAPZQDaIALAAxwDMdIAuoQA4AuUIAymwE4CWAOwDmIAL6EYVGIhApIGHAWJkQtAOwAOODqasOkbnyGiJIKZtnzFeQiUjkYmzRQohCLqgFZ178GGkwCEIwagoAJhkQr284cL8wTXDw5wS4uBg3QnCwWkjLbM0YcLy-FOLaWgSwOHV1Lz8fV2CQJoiEmHUYMF9sqjAYSrL%2BwYbsmEGqLMo8uCK-dXCYgvAqOukEiJoqPRB2Th4BEXFCOC8EaDk0LFsVBzVaTSpwt2Y9gyMj0xC5i%2BRrpR2VQ7N77QyHEwnEDhaSWS7WG7KezkGJUAAEADU-FJ0VjJDFMQsqLi-E8SSFieTwOpiYTQR8AKqCfhsADyqAAsrh0NgAK68XBQgC08XhUD4vKRqgajDEcqAA
what I would recommend is using their messaging / help feature to ask them for the right p/n based on the info we have above. If they can't help, Mouser should be able to. Or Farnell, etc.

 

[bombadero]

I told Grin info that the shunt wiring appeared to be correct, but they insisted that I re-wire the CA3 shunt backwards so that white goes to S- and blue goes to S+. I did that, and it fixed the regen and wh/mi reporting, but reversed HWhrs and avg HW to negative numbers and completely killed PA. I informed Grin and am waiting for a response.

Nothing has changed shunt-wise apart from the new CA3, which was burned out by the arc fault, and the fact that the arc fault occurred on the Baserunner. I still maintain that could be an issue; perhaps the heat from the arc fault fused the shunt leads inside the Baserunner so that current is flowing directly between the leads.

Correct me if I'm wrong, but having switched the leads in the CA3 so they are backwards, would that not have corrected swapped wires in the connector? So it seems like either the issue is not in the CA3 at all, indicating the Baserunner, as everything else has been ruled out; or that the CA3 is even more defective. I feel like Grin should replace it, but we'll see what they say.

Thanks for the info on the FET, I may still attempt that.

When reversing the shunt, it was like playing Operation. I tried removing the display from the GPIO pins, but wasn't successful. Do you have any advice on that? I've only ever placed displays onto GPIO headers, I've never had to remove them.

 

[amberwolf]

I told Grin info that the shunt wiring appeared to be correct, but they insisted that I re-wire the CA3 shunt backwards so that white goes to S- and blue goes to S+. I did that, and it fixed the regen and wh/mi reporting, but reversed HWhrs and avg HW to negative numbers and completely killed PA.

That's bizzare. The shunt shouldn't have any thing to do with the human watts, as that comes from the torque sensor via calculations the CA does. I can't think offhand of a way they could interact this way--hopefully Grin can. :?

Nothing has changed shunt-wise apart from the new CA3, which was burned out by the arc fault, and the fact that the arc fault occurred on the Baserunner. I still maintain that could be an issue; perhaps the heat from the arc fault fused the shunt leads inside the Baserunner so that current is flowing directly between the leads.

If that happened, you'd probalby read no current at all, because it would simply short out the voltage that should be sent to the CA on those shunt leads, and the CA wouldn't read any current without any voltage on them.

Additionally, current wouldn't flow thru the shunt measuring leads (if something was wrong that caused this (which would take quite some doing), they'd just melt and open the circuit because they're too small for the controller operational current).

Additionally...the heat from the arc fault couldn't have reached the shunt wires and damaged them, without first reaching the BR's internals and damaging those (like it's shunt, PCB, etc); to actually melt the wires together would require enough heat to noticeably damage the PCB itself first. And even to do that, it would have had to occur at or nearly at the BR PCB itself, and be sustained long enough.

Correct me if I'm wrong, but having switched the leads in the CA3 so they are backwards, would that not have corrected swapped wires in the connector?

Yes, it would.

So it seems like either the issue is not in the CA3 at all, indicating the Baserunner, as everything else has been ruled out; or that the CA3 is even more defective. I feel like Grin should replace it, but we'll see what they say.

There's nothing that could *change* in the Baserunner that could cause backwards current flow readings; to do that requires swapping the wires. (they could be installed that way in the first place, but they couldnt' change).

All the wires do is connect to each end of a shunt that is in the battery negative wire to the controller (usually on the actual controller board, except in the case of an external shunt in the battery wiring).

Current flows thru the shunt (not the blue and white shunt measuring wires), and creates a tiny voltage across teh tiny resistance of teh shunt.

When reversing the shunt, it was like playing Operation. I tried removing the display from the GPIO pins, but wasn't successful. Do you have any advice on that? I've only ever placed displays onto GPIO headers, I've never had to remove them.

It's been a while since I modded or fixed one of my CAs, so I don't remember exactly what I did...but for edge wiring like the shunt wires, I don't think I did anything with the display.

However, in general, whenever I desolder anything I care about, I use solderwick and/or a soldersucker, and a temperature-controlled Weller soldering station with whichever tip is best fit for teh application (usually a fat chisel-tip for desoldering work, to quickly heat the desoldering points and wick to get teh solder off before damaging pads, parts, etc).

(desoldering bad parts I may have less care for, by cutting them off their pins, etc., then pulling the pins out with needlenose while heating from the other side of the pad, etc)

 

[ERP]

So I opened up the CA3 and the white wire is soldered to S+ and the blue wire is soldered to S-, which is correct. So the problem appears to lie elsewhere.

hi, I just bought the erider and installed it and I am having problems with it but I still need to update the firmware on my v3.1 ca to make things right. My question is, when you check to make sure everything is okay from the get go by making sure the P&D on the display are moving up and down is it normal for just the P to move up and down and not the D? I emailed grin but haven't gotten a response yet. Any help would be appreciated.

 

[offGridDownUnder]

Hello,

Can you say if you have further information or a solution?

I am just at the beginning of diagnosing my problem, but see some (not all) similarities in what you report.

In my case, the symptom is that when I activate the brake switch and use the throttle to modulate the braking, the motor ceases to have any apparent effect and the momentary power shows '0'.

Grin All-Axle on a Q45 Cruzbike (so the back wheel is a 100mm front wheel with the All-Axle in it), Baserunner & CA. I have spares and I'm going to try swapping in to narrow down the possibilities, but any information you might have may be helpful now or in the future.

Cheers,
sadhak

 

[offGridDownUnder]

I replaced the Baserunner with a Phaserunner, and on the stand the behavior is that I push the throttle and the wheel spins but then slows a bit. The power on the CA shows ~160W and then drops to with no change in throttle position. On several attempts it dropped to ~20W and once it dropped to 0.

I plugged in a different throttle, and the behavior remains (that would have been too easy, sigh). The indicated power on the CA bounces back and forth between ~17 & ~34, and it also went to 0 once as well.

 

[amberwolf]

BR and PR are current-controlling throttles, so setting a throttle position will allow (not force) that amount of phase current to the motor. If it has no load, like on a stand, it won't draw much of that current, but it may reach max speed really quick, depending on how everything is setup.

If the CA is setup to control speed or current or power, it may create odd behavior on the test stand that won't occur during a ride because of the above.

To ensure the PR itself has the correct behavior you want for the system, first set the CA to Bypass for throttle mode, so that none of it's settings will affect the throttle operation (no limiting, ramping, etc). Throttle will go directly as-is from your throttle to the controller's signal input. (PAS won't work, though, but you can at least make sure throttle does what it should.)

 

[offGridDownUnder]

To ensure the PR itself has the correct behavior you want for the system, first set the CA to Bypass for throttle mode, so that none of it's settings will affect the throttle operation

Thanks for the suggestion. I'm a bit pressed for time, as I'm planning to sell the bike, and it's replacement is imminent, and I don't have space for both. So, I felt hurried about the matter.

I implemented your suggestion as a trial, and it made no obvious difference apart from no PAS, but that got me poking into the settings. I was not diligent about recording everything before starting, so I'm not completely certain, but in one of the modes at least PAS was set to 0. I set it on the road and after that all worked as I expected, with regenerative braking working according to throttle input.

I consider I may have simply 'lost' part of my configuration due to the movements of the heavens or something.

I'll test some more (hot in the afternoons in Perth, and I'm supposed to be working), but it's nice to have some feeling of control.

Regards,
sadhak

 

[amberwolf]

I consider I may have simply 'lost' part of my configuration due to the movements of the heavens or something.

That might be closer than you'd think to a possiblity:

Cosmic rays (or other radiation) can cause memories of various kinds to lose data (or even permanently damage them). It's one reason for random computer glitches and crashes.

While the chances of any particular device being affected at any particular time are vanishingly small, there *is* still a chance....

Sometimes it isn't a user-accessible setting that's affected, but rather some data or code in a flashable device, so when you check the settings, they're all good, but it still doesn't work. If the device has a factory reset function that resets those data, it can fix a problem like this (though that is unlikely to fix code).

Modern computers and complex portable devices can have error-checking code to deal with this sort of thing, but I doubt the tiny MCU in things like the CA have room for this.