I just retrofitted an S866 display with a Brainpower controller onto wifes bike which previously had only PAS on/off, i.e. did not have selectable PAS levels. Got it all working, BUT the PAS does not behave as I think it should. It seems to begin powering the motor soon after you start pedaling, but rather than adding power as your cadence begins to slow for an uphill it reduces the amperage. This is backwards to what I'd expect. On my other bike, the assist (and amperage to motor) increases as cadence slows in an attempt to help maintain the speed of the selected power level. Instead, in this current setup, the assist falls away as help is needed. Basically, it increases the assist as I pedal faster and reduces the power as cadence slows. Ideas would be appreciated.
PAS troubles S866 with Brainpower controller
- Thread starter cjxj
- Start date
Brainpower are cheap and a bit crappy on power delivery as they use speed control, mostly the power is all or nothing. Each assist level having a max assist speed so one 8-9mph, two 12 /13mph and three 15.5mph or so. Once you near the optimum speed for the assist level the current will ramp down so one will feel the power reducing.
Last edited:
Thanks neilh. Your description is what I was expecting, but I get the opposite. As cadence increases, so does the current to the motor util the assist level speed is reached. Then it falls off as one would expect. As cadence decreases, as when you begin an uphill, the current also decreases rather than increase, which is opposite what I believe should happen.
An update: I'm suspicious of my pas sensor. I *sort of* tested it when I was putting this together in order to count magnets, and saw enough response for count them. I just read a few posts where the author had pas issues with this controller and solved his problem by making transistor "not gate" to switch the polarity of his pas signal. I made a similar circuit, but in the process I noted that the signal from my pas is irregular and insuficient to even trigger this circuit. While I don't understand why the bike behaves as it does, I see no reason to continue the head scratching until I know I have a good pas sensor. On order now. Hopefully will be able to post a good resolution.
An update: I'm suspicious of my pas sensor. I *sort of* tested it when I was putting this together in order to count magnets, and saw enough response for count them. I just read a few posts where the author had pas issues with this controller and solved his problem by making transistor "not gate" to switch the polarity of his pas signal. I made a similar circuit, but in the process I noted that the signal from my pas is irregular and insuficient to even trigger this circuit. While I don't understand why the bike behaves as it does, I see no reason to continue the head scratching until I know I have a good pas sensor. On order now. Hopefully will be able to post a good resolution.
The hub might not be very powerful ir the winding speed might be fast so struggles on hills , hence the lack of power.
So, a happy update (for anyone interested)... Here's what I found
1) Turns out the original pas sensor was bad. With 5v across the power connections, I got very little response from the signal output. So... I purchased a new pas sensor from Amazon. The one I bought was a KT12L.
2) Although the new pas tested good using a multimeter, I got no pas response at all from the bike.
3) So, I plugged in the "not gate" I had made and described earlier, between the pas and controller......... and wonder of wonders, it's alive and now seems to work fine.
4) I still have no idea why it behaved as it did originally, adding power when pedaling fast but dropping to no assist as candence slowed. I also have no idea why the new pas didn't work, nor why inverting the square wave signal made things work. I have some suspicions, but at this point, I don't really care to investigate. I'm just happy it works.
5) It seems that more than a few people have had trouble with the pas sensor on this controller/display combination. The "not gate" idea apparently worked for persons in a couple of posts I saw and I had some workable parts laying around. I used a 2n3904 general purpose npn transistor and 1 each 500 ohm and 1k ohm resistors. I had the 3pin jst male and female connectors to make it plug in between the pas connector and controller connector, but one could build it into the pas lead directly. I used heat shrink to tie it all together neatly. Google "transistor not gate" and you'll see plenty of examples of this circuit, but I've attached what I did..
1) Turns out the original pas sensor was bad. With 5v across the power connections, I got very little response from the signal output. So... I purchased a new pas sensor from Amazon. The one I bought was a KT12L.
2) Although the new pas tested good using a multimeter, I got no pas response at all from the bike.
3) So, I plugged in the "not gate" I had made and described earlier, between the pas and controller......... and wonder of wonders, it's alive and now seems to work fine.
4) I still have no idea why it behaved as it did originally, adding power when pedaling fast but dropping to no assist as candence slowed. I also have no idea why the new pas didn't work, nor why inverting the square wave signal made things work. I have some suspicions, but at this point, I don't really care to investigate. I'm just happy it works.
5) It seems that more than a few people have had trouble with the pas sensor on this controller/display combination. The "not gate" idea apparently worked for persons in a couple of posts I saw and I had some workable parts laying around. I used a 2n3904 general purpose npn transistor and 1 each 500 ohm and 1k ohm resistors. I had the 3pin jst male and female connectors to make it plug in between the pas connector and controller connector, but one could build it into the pas lead directly. I used heat shrink to tie it all together neatly. Google "transistor not gate" and you'll see plenty of examples of this circuit, but I've attached what I did..
Attachments
The KT 12 magnet sensors have lots of compatibility issues. The 8 and 10 magnet ones are much better. You should have asked first.
You probably could have solved that problem with a simple pull up or pull down 10k resistor. You got a solution, which is all that matters and people can copy it if stuck. Thanks for the info.
Hmm. I'm new to all this. The KT12L was $11 delivered overnight, so I figured there was not much to lose.
saneagle, I suspect you might be right about using (in this case) a pull down resistor. The signal line when not connected, from the controller sits logic high. From the KT12, it is low between magnets and drives high when passing a magnet. I didn't look at it when I first connected it, but guess it probably stayed at least somewhat elevated all the time, hence a poor logic level square wave. A resistor selected to pull the signal low between magnets might have done the trick in a simpler fashion. I'd try that first next time. However, I had already made the little circuit with connectors and it worked. The circuit is also somewhat of a signal buffer, is super simple, and costs almost nothing. Shared because other discriptions I saw were not very clear. Perhaps the discussion might help someone....
saneagle, I suspect you might be right about using (in this case) a pull down resistor. The signal line when not connected, from the controller sits logic high. From the KT12, it is low between magnets and drives high when passing a magnet. I didn't look at it when I first connected it, but guess it probably stayed at least somewhat elevated all the time, hence a poor logic level square wave. A resistor selected to pull the signal low between magnets might have done the trick in a simpler fashion. I'd try that first next time. However, I had already made the little circuit with connectors and it worked. The circuit is also somewhat of a signal buffer, is super simple, and costs almost nothing. Shared because other discriptions I saw were not very clear. Perhaps the discussion might help someone....