PAS does not fit

ymd

100 mW
Joined
Aug 3, 2017
Messages
40
Location
Bay Area, California
I finally received my ebike conversion kit from em3ev. Thank you all for helping me decide on the kit. Here is the link to the post where you all helped me. https://endless-sphere.com/forums/viewtopic.php?f=3&t=89856.

It took very long for me to get the kit as I was waiting for the new triangle battery packing em3ev was making. I will post pictures after I finish the installation. Right now I need help with installing the PAS on my bike.

The PAS sensor I have got is the one in the photo I have attached.

The problem I have is that this sensor does not fit behind the crank as there is very little gap on my bike. I even removed the crank and put it directly on the Bottom bracket. But, when I install the crank, the PAS sensor simply gets jammed and does not rotate.

I will be thankful if anyone helps me to install the sensor? I don't want to be riding long distances without a pedal assist.

Thanks for all your help.
 

Attachments

  • PAS-Display.jpg
    PAS-Display.jpg
    58.2 KB · Views: 5,609
Two options. Either may also require creative mounting of the sensor.
1) Mount disk/sensor on the other side of there is room.
2) Remove the center of the disk and center the magnet ring to the chain wheel with glue if that allows clearance for the disk to spin.
 
Thank you for your suggestions. There is not enough room on the chain side too to fit the PAS in the gap from the crank. I need to take out the chain ring and see if I can glue the magnet disc and have space to install the sensor.

Would getting a longer bottom bracket solve the problem? If so, any suggestions?

docw009 said:
Two options. Either may also require creative mounting of the sensor.
1) Mount disk/sensor on the other side of there is room.
2) Remove the center of the disk and center the magnet ring to the chain wheel with glue if that allows clearance for the disk to spin.
 
A longer bottom bracket will give you more space to mount the magnet ring, but it will also change the chainline-- probably for the worse.

If you spend most of your time in the higher gears because of the electric assist, then moving the chainline outward might not be a problem.

For a square taper bottom bracket, I recommend Shimano BB-UN26 as a very inexpensive and reliable unit. It's widely available in sizes up to 127.5mm.
 
You can cut out the middle, drill some holes and use zip-ties to fix it to your inner chain-ring, or you can epoxy it.

Be aware that there are four permutations of disc direction and disc orientation (flip). For some sensors like that, only one works, for others, there's two possibilities, so test it first to see which ways it works before gluing or assembling anything, otherwise you have to undo everything you did when you got it wrong.
 
The sensor you are trying to use is a KT-D12R, made by KT Kunteng in China. You can also purchase the similar unit, KT-D12L which is specifically designed to fit on the Left Hand crank. The right hand version you have will only work on the Right side crank.
I am using the KT-D12L Left side sensor on my bike and to make it fit into a crank set with tight clearance, you can trim the plastic fingers in the centre of the magnet disc with a sharp modelling knife to reduce their thickness, and open up the internal diameter, if it is too tight on the crank. I have reduced the thickness of my disc by half, in order to fit on my bike and it works perfectly.
it is much easier to fit the left side version IMO, and produces a very neat looking and serviceable system. A few vendors on Aliexpress offer the KT-D12L version, and maybe Em3ev stock it as well. I recommend you check it out.

IMG_20170305_150745.jpg
 
Thanks for your suggestion. I had referenced a wrong sensor in my OP. The actual sensor I have is CA3 Cycle Analyst PAS Speed Sensor (see attached stock picture) which can be mounted on either right or left crank. I will try to trim the thickness of the sensor and see if it fits as you suggest.

What is your opinion about "TDCM Torque Sensing BB For CA3 Cycle Analyst"? Would it be better than a speed sensor? Is it too complicated to install it?

Johne-bike said:
The sensor you are trying to use is a KT-D12R, made by KT Kunteng in China. You can also purchase the similar unit, KT-D12L which is specifically designed to fit on the Left Hand crank. The right hand version you have will only work on the Right side crank.
I am using the KT-D12L Left side sensor on my bike and to make it fit into a crank set with tight clearance, you can trim the plastic fingers in the centre of the magnet disc with a sharp modelling knife to reduce their thickness, and open up the internal diameter, if it is too tight on the crank. I have reduced the thickness of my disc by half, in order to fit on my bike and it works perfectly.
it is much easier to fit the left side version IMO, and produces a very neat looking and serviceable system. A few vendors on Aliexpress offer the KT-D12L version, and maybe Em3ev stock it as well. I recommend you check it out.

View attachment 1
 
The PAS kit in your new photo is a standard KT-D12R or KT-D12L modified by EM3ev to take the CA plug. I can only tell you that this PAS kit, when used with matching KT controller and KT display, has to be fitted on the designated Right or Left side crank to work. I'm not sure whether this is the same with a CA display and different controller.
Another point to bear in mind is the sensor unit must be a maximum of 5mm from the magnet disc, otherwise it will not work.
I have no experience of using the Torque sensing bottom bracket, so can't help you there.
Good luck with your installation!
 
That is a Grin Technologies product that is a actually modified sensor from King Meter. This version operates in simple quadrature mode which means the two hall sensors just feed the two outputs directly. The CA determines pedaling direction by the phase of the two signals. Putting the device on the other side reverses this phase relationship because the magnets activate the halls in reverse order as you pedal and so makes it look like you are pedaling in reverse (no assist).

To make the unit work on the other side you need to get the hall orientation and crank motion in line with what the CA thinks is 'FWD pedaling' by either mounting the sensor backwards (REV crank motion + REV sensor = FWD signals) or, easier still, just change what the CA thinks is FWD. Go into CA settings and reverse the PAS->DirPlrty setting. Done. In any case, set PAS->SignlType to '2-wire' (quadrature).

In line with the suggestion above, just wave the sensor over the PAS wheel of your proposed installation to see if it will work before you do any fab work. Reverse the CA setting if required to get the direction part squared away. It actually sounds like you would have been better off with their chainring-side unit (PAS_12P_Chr = http://www.ebikes.ca/shop/electric-bicycle-parts/torque-sensors/pas-12p-chr.html), but you should be able to get what you have working ok.

You can also look at the CA PAS Settings Preview Screen and watch the two little arrows which show the two hall signals. If they are alternating UP/DOWN when you move the PAS wheel then your sensor is picking up the magnets okay. If that is happening but you're not getting assist, the the hardware is installed okay but you have a CA PAS configuration glitch.

PasPreviewScreen.png
 
I was able to install the PAS magnet disc by cutting off the "fingers" and gluing it to the left crank. Now I am able to get pedal assistance as I pedal. However, the motor is very jerky. It feels like the motor is turning on and off every rotation of the crank. I have checked the LED on the PAS sensor and it blinks every time the magnets on the disc pass by it when I pedal backwards by hand. Looks like the sensor is getting the signal correctly. I don't know why I am getting jerky response from the motor. If I use the throttle there is no jerky movement. Also, I have noticed that when I am pedaling fast or when I am in the first level of pedal assist the jerkiness is lesser.

Is this behavior due to a faulty installation of the PAS or is there a setting I need to use to on my Cycle Analyst 3 to make it better?

Thank you all for all your help.
 
ymd said:
Is this behavior due to a faulty installation of the PAS or is there a setting I need to use to on my Cycle Analyst 3 to make it better?

A described in the above post:
Please verify the (P) and (D) arrows change direction when you pedal FWD and make the two CA settings.
 
Thank you for your suggestion.

I just checked the two arrows change direction when I pedal FWD (or even REVERSE).

I have set PAS->SignlType to '2-wire'. The magnet disc is mounted on the left crank (not on the chain side). So, I did not reverse the PAS->DirPlrty setting.

Is there anything else I can check?

A described in the above post:
Please verify the (P) and (D) arrows change direction when you pedal FWD and make the two CA settings.[/quote]
 
Firmware Version is 3.0. Start and Stop thresholds are each 5 RPM. Do I need a firmware update?

teklektik said:
Should be working...
what firmware are you running? (see splash screen)
What are you start and stop thresholds in the PAS Setup submenu?
 
ymd said:
Firmware Version is 3.0. Start and Stop thresholds are each 5 RPM. Do I need a firmware update?
Not sure why you altered the default start/stop thresholds - I would recommend restoring the defaults of 10 and 17 respectively. This isn't the problem, but it will make the PAS work better - when it finally does.

I cannot really think of a reason for the behavior you describe except that the sensor is not picking up the magnets for part of the crank revolution. The sensor should be no more than about 5mm away from the magnet disk - closer is better. That said, if you are getting proper Up/Down motion of the little arrows for the full revolution then things should be okay. If the arrows don't toggle as expected, then the CA will think you stopped pedaling and kill the assist only to restart it again when the arrows behave. This is the ON/OFF symptom in one crank revolution that you describe.

There are lots of new features in the later 3.1 firmware, but you don't need to upgrade to get this to work. Best not to go changing things until you get this sorted.
 
I changed the start/stop thresholds to get a quicker response when I have stopped at a red light. I will restore to the defaults as you suggest and make small changes to see if that helps.

I looked at the magnet disc and the sensor more closely. The two halves of the magnet disc are at about 1-2 mm distance difference from the sensor. It may be because, before I glued the magnet discs to the crank, I used a dremel tool to smooth out the rough edges after cutting the "fingers" off. I may have shaved off more plastic on one half than the other in the magnet disc. I have now moved the sensor as close as I could without touching the magnet disc.

Even after these changes, the issue is still there. Would it be possible that if the magnet discs are not at the same distance from the sensor even though both halves of the disc are less than 5mm from the sensor, the Cycle Analyst ignores the weaker signal coming from the disc that is slightly farther? If that is a possibility, I will try to remove the magnet discs from the crank and re-glue them ensuring both halves of the disc are at the same distance from the sensor.

I was again checking to confirm the arrows are getting proper Up/Down motion over a full revolution. They do move up/down while I pedal over a full revolution. However, I noticed that the motor was not turning on while I was in "setup" to check the PAS arrows though the arrows were moving up/down. Is this normal? I had to come out of setup to get the pedal assist working.

Is there a way to connect the PAS sensor directly to the controller bypassing Cycle Analyst?

teklektik said:
Not sure why you altered the default start/stop thresholds - I would recommend restoring the defaults of 10 and 17 respectively. This isn't the problem, but it will make the PAS work better - when it finally does.

I cannot really think of a reason for the behavior you describe except that the sensor is not picking up the magnets for part of the crank revolution. The sensor should be no more than about 5mm away from the magnet disk - closer is better. That said, if you are getting proper Up/Down motion of the little arrows for the full revolution then things should be okay. If the arrows don't toggle as expected, then the CA will think you stopped pedaling and kill the assist only to restart it again when the arrows behave. This is the ON/OFF symptom in one crank revolution that you describe.

There are lots of new features in the later 3.1 firmware, but you don't need to upgrade to get this to work. Best not to go changing things until you get this sorted.
 
When in Setup, the motor will not run to avoid unpleasant mishaps during configuration.

The PAS sensor is digital - if the arrows are toggling then the disk is okay. The CA is unaware of the sensor distance - it either works or it doesn't. It sounds like you have things set up correctly in the CA - I can think of a thing of two to look at out of desperation, but frankly nothing jumps out as a problem area and I can't think of a way to misconfigure the CA to make it behave as you report (I actually tried a few on the bench). I think it's time for an email to ebikes.ca to see if they have some insights - there's no substitute for experience with the product...

Although some controllers support a PAS wheel directly, that particular one is sort of special and it's unlikely to work with anything but a CA. The CA does pretty nice PAS - this problem is just puzzling...
 
I was sure the PAS sensor would be digital. Just wanted to confirm if I am missing something. Thanks for the clarification. I just emailed to ebikes.ca. I hope they will be able to help.

Thank you again for all your help. I very much appreciate it.

teklektik said:
When in Setup, the motor will not run to avoid unpleasant mishaps during configuration.

The PAS sensor is digital - if the arrows are toggling then the disk is okay. The CA is unaware of the sensor distance - it either works or it doesn't. It sounds like you have things set up correctly in the CA - I can think of a thing of two to look at out of desperation, but frankly nothing jumps out as a problem area and I can't think of a way to misconfigure the CA to make it behave as you report (I actually tried a few on the bench). I think it's time for an email to ebikes.ca to see if they have some insights - there's no substitute for experience with the product...

Although some controllers support a PAS wheel directly, that particular one is sort of special and it's unlikely to work with anything but a CA. The CA does pretty nice PAS - this problem is just puzzling...
 
I wanted to come back and post how my issue was resolved by Justin from Grin Technologies. This youtube video he sent me helped resolve the issue: https://www.youtube.com/watch?v=W_hnUzM5ZuQ.

My PAS controller was working correctly. Only the watts limit feedback gain was tuned incorrectly. The WGain was set to 50 in my CA3 when I got it. In the video Justin suggests a value around 15. After I changed the WGain to 15, the response of my bike has improved greatly. I think this was the issue and I am very happy that it is resolved.

I hope this post will help someone in the future.
 
Excellent news!
Thanks for posting back - this was looking like a puzzler...

Things seem to have gone off the track (for me) with this remark which seemed to indicate power fluctuations in synchrony with pedal motion:

ymd said:
It feels like the motor is turning on and off every rotation of the crank.
This seemed to suggest a sensor failure since WGain issues are generally unrelated to crank position.

My Bad. Tunnel vision....

Anyhow, for other folks in reading here to resolve this or other CA PAS problems, try looking at page 45 "17.Tuning Hints" of section "5.7 Pedal Assist" of the [strike]Un[/strike]official Guide. This discusses the WGain 'power surging' adjustment and other stuff.

Also - now that you have this basically working, the 3.1 beta firmware also includes a new ThrO->PASRate adjustment not present in 3.0; this allows you to control the rate of power application in PAS mode distinct from power application under throttle control. This can very much reduce the need to back down the WGain setting which controls both throttle and PAS power. This remedies an issue in the 3.0 firmware where throttle response often became sluggish due to adjustments of WGain or UpRate needed to make PAS behave.
 
Back
Top