Torque sensor problem with v3.03 Cycle Analyst

[jloe]

Hi all,

I have a cargo bike that I electrified with ebikes.ca parts a couple of years ago. I recently had my local bike shop upgrade the bottom bracket with an ebikes.ca ERider-T9N torque sensor bb. That setup worked great for a few months. Then, about two weeks ago, it stopped working for no apparent reason. I initially suspected a faulty connection, but when I go in to the Cycle Analyst (firmware v3.03) in the "Setup TRQ Sensor" screen, I see that the torque sensor is putting out "1.90V 26.7Nm" when the torque sensor is connected, and "4.55V 212 Nm" when disconnected. I don't know if these values are appropriate, but the difference tells me that the torque sensor is sending some sort of signal.

This has led me to the conclusion that I have messed up the settings in the Cycle Analyst somehow - either initially, and the problem has just now manifested itself, or possibly more recently without even realizing I had messed up the settings.

A complicating factor to all of this is that I have not kept the firmware on the Cycle Analyst up to date. I am still running v3.03. This is a problem because there is not a "restore to factory defaults" option in v3.03, and all of the instructions for configuring torque sensors on ebikes.ca are written for the v3.1x firmwares, which have significantly different setup menu options. As a result, I'm having a hard time figuring out which values I might have set incorrectly in the Cycle Analyst.

So, it seems like my next step ought to be to update the CA firmware to v3.14. When I try to do that, I run in to another problem. I have a Grin Technologies USB-TTL cable, but when I hook everything up to my Mac (High Sierra 10.13.6), and click on the "READ CA" button, the Cycle Analyst displays the text: “…PC CONNECTION”, and after a few moments, this error message is displayed:”Unable to communicate with the Cycle Analyst in a timely manner. Please make sure the device is properly connected. Please cycle power to the device and try again.” Eventually, this message is displayed: “Failed to read setup from Cycle Analyst! Please make sure you have selected the correct COM port and that your Cycle Analyst is turned on and properly connected, then try again.”

So, I searched here, and found this thread: https://endless-sphere.com/forums/viewtopic.php?t=37964&start=4175 and it made me wonder if perhaps my v3.03 of the firmware had a similar timing issue on my laptop, so I tried to download the settings from the CA, by clicking on the “READ CA” button first, and then turning on the power to the CA. This didn’t work even though I tried it multiple times at different points in time in the bootup process.

After that, I decided to attempt to update the CA’s firmware even though I was worried about bricking the CA. I attempted this by clicking the “UPDATE Firmware” button, selecting the 3-14 firmware that I downloaded from https://www.ebikes.ca/downloads/cafw/release/, selecting the proper serial port, clicking the “Update Firmware” button, and then turning on the CA. I attempted this probably 6 times, each time trying to get slightly different timing of turning on the CA. Each time, I got this red error screen:


Based on this information, is it possible to tell if I have a faulty CA? Or possibly a faulty USB-TTL cable? Is there some additional testing that I should perform that I haven't done?

Any advice you could provide would be much appreciated as this is one of our primary sources of transportation. Thank you!

 

[amberwolf]

Hi all,

I have a cargo bike that I electrified with ebikes.ca parts a couple of years ago. I recently had my local bike shop upgrade the bottom bracket with an ebikes.ca ERider-T9N torque sensor bb. That setup worked great for a few months. Then, about two weeks ago, it stopped working for no apparent reason. I initially suspected a faulty connection, but when I go in to the Cycle Analyst (firmware v3.03) in the "Setup TRQ Sensor" screen, I see that the torque sensor is putting out "1.90V 26.7Nm" when the torque sensor is connected, and "4.55V 212 Nm" when disconnected. I don't know if these values are appropriate, but the difference tells me that the torque sensor is sending some sort of signal.

This has led me to the conclusion that I have messed up the settings in the Cycle Analyst somehow - either initially, and the problem has just now manifested itself, or possibly more recently without even realizing I had messed up the settings.

It is much more likely that there is a fault in the sensor itself, or in it's wiring or connections. I would start there, and see if something is loose, disconnected, wiring damage, etc.

If there was water of any kind involved at any time between intial install and now, I would investigate every possible place it could have gotten into and checking for corrosion / oxidation, etc. Especially JST connectors, as they are very vulnerable to this, not only at the contact surfaces, but where the wires crimp into the back of teh contacts.

If you were altering settings when the problem occured, then you should change the setting back to what it was and see if it works. If you did not change any settings then it is unlikely to be a settings issue, and is probably hardware (sensor or connections.

I am still running v3.03. This is a problem because there is not a "restore to factory defaults" option in v3.03, and all of the instructions for configuring torque sensors on ebikes.ca are written for the v3.1x firmwares, which have significantly different setup menu options. As a result, I'm having a hard time figuring out which values I might have set incorrectly in the Cycle Analyst.

I was going to refer you to the historical archives of Grin documentation but I don't see it on their site now, and I don't remember if they did have this available there or not. You could try looking for a version of it here
https://web.archive.org/web/20140401000000*/https://www.ebikes.ca/learn/documents.html
picking a year and then a month and day and time for the available archives, and seeing if what you're after is in there. Alternately, Grin Tech may have this informaiton more easily available via direct contact.

Based on this information, is it possible to tell if I have a faulty CA? Or possibly a faulty USB-TTL cable? Is there some additional testing that I should perform that I haven't done?

I would refer you to another recent Apple-to-CA thread where he solved the problem,
https://endless-sphere.com/forums/viewtopic.php?f=1&t=105760
but he didn't bother to say what he did.

After much trial and error I have succeeded in getting the CA Software Utility to work on my MacBook Pro. Any one out there with Mac issues should contact me as I've looked every which way and experimented a lot to find a fix.\

You could try contacting him directly and see if he will say, and then post the exact steps that he tells you as you test them, so others that have this problem won't be stuck doing the same dance.

Other than that, contacting Grin Tech directly is the only other good option I know of.

 

[amosthedog]

I am actually ok. Everything connects as it should. I do plan to write up steps for the Mac so anyone else having a connection problem will know what to do.

 

[amberwolf]

We appreciate that.

 

[jloe]

amberwolf - Thank you for your well-considered reply. I especially appreciate the suggestion to use the Wayback Machine to get the older documentation - it helped me go through the CA's settings with a fine-toothed comb. I ended up using this version.

I agree with you, a problem in the sensor, or wiring, or connections does seem most probable. I've just examined all of the wiring and connectors again. I don't see anything that appears damaged in any way. I asked tech support at ebikes.ca last week if there is any way to test the voltage output of the torque sensor directly, but they did not provide me with that information - I assume because the voltage output of the strain gauge in the torque sensor is quite small, but I don't know for sure. Instead they instructed me to test the voltage on pins 1 and 2 of the CA's Torque/PAS input. ebikes.ca said that the 11.23V that I observed with my multimeter would be considered within typical range.

One piece of data that seems to indicate that the wiring and connections are working as intended is that while in the "SETUP TRQ SENSOR" screen, with the torque sensor disconnected, the preview line reads "4.98V 212 Nm". When I connect the torque sensor, the preview line changes to read "1.94V -0.1Nm". This leads me to believe that the wiring and connections from the torque sensor to the CA are in proper working order. Is that a reasonable assumption to make based on this data?

As an aside, and just to be sure that I cover all of the bases, does that disconnected torque sensor value of "4.98V 212 Nm" sound correct?

The following information seems support to the argument that there is a problem with the sensor inside of the bottom bracket: The v3.0 CA documentation states that in the "SETUP TRQ SENSOR" screen the "Preview line shows the measured voltage from the torque sensor signal. If the torque sensor is enabled, then it also shows the corresponding calculated N-m of torque on the cranks." When I stand on the pedals while in the "SETUP TRQ SENSOR" screen, I don't see a change in the preview line, just a steady state of "1.94V -0.1Nm". This seems to me to indicate a problem with the sensor itself. What do you all think?

Lastly, in response to this question:

If you were altering settings when the problem occured, then you should change the setting back to what it was and see if it works. If you did not change any settings then it is unlikely to be a settings issue, and is probably hardware (sensor or connections.

I had not changed any settings since the initial setup back in about February. I've just gone over the settings in detail, and the only obvious potential issue that I saw was in Trq Offset. When the torque sensor is connected, it outputs 1.94V, and I had had this value set at 1.52V for an unknown reason. I changed this value to the correct 1.94V, but it did not help when I took the bike out for a test ride.

Thank you again for all of your help amberwolf! (I've messaged amosthedog privately, and they indicated that they might be able to post more information about the Mac->CA connection issue on Monday.)

 

[amberwolf]

I agree with you, a problem in the sensor, or wiring, or connections does seem most probable. I've just examined all of the wiring and connectors again. I don't see anything that appears damaged in any way.

The problem with wires and connectors is just that--you may not see anything wrong, but it still could be. Wires can break inside their insulation, usually this happens where they bend or where they enter or leave a connector, or wherever they are strained around something (tight).

But they could all be fine, and it could be the actual TS itself that's failed. It does happen. :/ If it is the TS, it might be repairable depending on what's wrong, but if it came from Grin they usually have good warranty coverage, so you'd just send it back to them.

I asked tech support at ebikes.ca last week if there is any way to test the voltage output of the torque sensor directly, but they did not provide me with that information - I assume because the voltage output of the strain gauge in the torque sensor is quite small, but I don't know for sure. Instead they instructed me to test the voltage on pins 1 and 2 of the CA's Torque/PAS input. ebikes.ca said that the 11.23V that I observed with my multimeter would be considered within typical range.

Yes, that's important because it's what powers the TS--without that, the rest of it doesnt' work.

The output of the TS itself is the voltage you are seeing on the CA. If you measure the rest of the pins on that TS/PAS connector, relative to ground (black meter lead on ground) you will see one that matches whatever you see on the CA's display at that moment. I don't remember which pin it is, but the CA documentation should have that info if you need it.

The straingauge inside the TS will have very very small changes in properties. If it's resistive then it won't have a voltage output at all, unless it is still hooked up to the wheatstone bridge or other instrument amplifier system; whatever they used in that TS.

One piece of data that seems to indicate that the wiring and connections are working as intended is that while in the "SETUP TRQ SENSOR" screen, with the torque sensor disconnected, the preview line reads "4.98V 212 Nm". When I connect the torque sensor, the preview line changes to read "1.94V -0.1Nm". This leads me to believe that the wiring and connections from the torque sensor to the CA are in proper working order. Is that a reasonable assumption to make based on this data?

Probably. But...a poor connection, or an imperfect but almost good one, can give a voltage different from what it is supposed to be.

As an aside, and just to be sure that I cover all of the bases, does that disconnected torque sensor value of "4.98V 212 Nm" sound correct?

I don't know. I have a spare CA not on the trike that I can power up and check what it reads with no TS attached. (I use the TDCM, and have a spare of that too). On my TDCM, IIRC it is always around 2.5v-ish when no force is applied to the pedals.

When I stand on the pedals while in the "SETUP TRQ SENSOR" screen, I don't see a change in the preview line, just a steady state of "1.94V -0.1Nm". This seems to me to indicate a problem with the sensor itself. What do you all think?

Probably, if there's no change, no response, the sensor has a problem.

When the torque sensor is connected, it outputs 1.94V, and I had had this value set at 1.52V for an unknown reason. I changed this value to the correct 1.94V, but it did not help when I took the bike out for a test ride.

Probably it originally was 1.52v when at zero force input, would be the reasonable explanation for that initial value.

 

[donn]

As an aside, and just to be sure that I cover all of the bases, does that disconnected torque sensor value of "4.98V 212 Nm" sound correct?

I have no idea about any of this, but just happened to have read this passage in the 3.1 manual recently, so I pass it on since it appears to confirm your thinking on this:

The preview screen also shows the voltage coming from the torque sensor. On most torque devices this will sit somewhere between 1-2.5V, and then either increase or decrease as you apply force on the pedals. With no torque sensor attached the voltage will rise to nearly 5V.

 

[amosthedog]

Jloe,

This is how I fixed my Cycle Analyst (CA) connection with my Macbook.

Be sure the downloaded file from Grin, (located in the bottom part of this document:https://tinyurl.com/y89ndn4g) is in the Mac Applications folder, not wherever the unzipped file resides.

The trick here is making sure the file is in the application folder. Most Mac developers add a graphical screen instructing the user to drag a file into the app. folder. Grin does not do that.

So, when you unzip the file it stays in the location you designated. For me, it was my desktop.

Don’t be fooled by the CA setup utility that opens when the file is clicked. The Setup Utility will open even if the files are not in the applications folder, but the program will not work!! That’s where you get the error messages when the CA is corded to your computer and you attempt to upload your CA activity.

Go to your application file and be sure that ca_Setup_Utilty_v1.54 is there. If not you will need to copy and/or drag the file into the Mac Application folder.

Be sure that Ca3-14 is in the data file or your changes will not transfer back to the CA. If you don’t have it you can load CA3-14 firmware file directly (just unzip the archive and move it to your installation folder in Applications) from here:
https://www.ebikes.ca/downloads/cafw/release/  

I hope this helps. Feel free to let me know if you have questions or don’t understand any instructions above.

 

[jloe]

donn - thanks for the confirmation on that voltage value.

amosthedog - thanks for the info about what caused your connection problem. I've been launching the app from the Applications folder, so I don't think that's the issue that I've been encountering. I've been reading up on the FTDI drivers, which I initially installed before even trying to connect to the CA. After reading some of the FTDI documentation, it sounds like my Mac should be able to communicate automatically to the USB-TTL cable without the drivers, so perhaps my mistake was installing the drivers without verifying that I actually needed them. Do you recall if you installed the FTDI drivers?

amberwolf - thank you again for your help, your suggestion that the wires / connectors might be damaged in a non-apparent way led to research how to inspect JST-SM crimps, and I found out that the male connectors are very difficult to get out of their housing. (And I suspected the male connectors as the probable culprit.) I really wanted to avoid dealing with the male connectors, so I gave the rest of the system another inspection, and I discovered that the lockring on the non-drive side of the ERider bottom bracket had come loose. I was able to hand tighten the lockring, which was enough to re-enable the torque sensor, though not enough to re-enable the pedal motion detection functionality. I'll pull off the crank arm tomorrow and try to further tighten the lockring then. I'm surprised that this was the culprit as I had the bottom bracket installed at a super reputable shop!

Thank you all for your help in diagnosing this problem. It was not at all obvious what the problem was, and all signs pointed in other directions!

 

[amberwolf]

I discovered that the lockring on the non-drive side of the ERider bottom bracket had come loose. I was able to hand tighten the lockring, which was enough to re-enable the torque sensor, though not enough to re-enable the pedal motion detection functionality. I'll pull off the crank arm tomorrow and try to further tighten the lockring then. I'm surprised that this was the culprit as I had the bottom bracket installed at a super reputable shop!

Thank you all for your help in diagnosing this problem. It was not at all obvious what the problem was, and all signs pointed in other directions!

Wiring/connections are almost always the problem, when something works and then doesn't. Or if something is "wierd". But sometimes it's something else mechanical that's loose. Sometimes it's worse, because there can be multiple independent causes for the "same" problem, all happening at the same time.

FWIW, I would bet that the underlying problem is where the cable comes out of the TS-BB unit, or whereever it passes out of the BB shell, it's damaged from the TS-BB rotating inside the BB shell due to the locknut not being tight.

So the cadence sensor may have an actual broken wire, while the TS wire might only be damaged enough to either have poor connection under some conditions, or the insulation is damaged enough to allow touching the frame.

 

[amosthedog]

I did download the drivers but I don't think that was the problem. Or, maybe it was a compound issue...I'll never know if the drivers made a difference since I had them installed before I found a fix for my issues. Glad you've solved your problem!

 

[jloe]

I just wanted to close the loop on this topic. I was finally able to update the CA's firmware using a borrowed Mac computer. The difference this time was that I did not install the FTDI drivers on the borrowed Mac. After updating the firmware on the CA, I was able to properly set all of the parameters and get the torque sensor working again. I can't explain why a loose lockring would cause a problem with the CA, but Robbie from ebikes.ca hypothesized this: "The only thing that would make sense to me is if the lockring loosening caused a high current draw from the 12V bus on the CA, which may have caused it to not power down correctly."

Also, I went back and tried to uninstall the FTDI drivers from my laptop using FTDI's instructions, but my laptop was still unable to communicate with the CA. So the moral of that story is - don't install the FTDI drivers unless you run in to a problem communicating with your CA!