Hi all,
I've been building my ebike over the course of 18 months after seeking advice from these forums and lurking/reading a lot, and it's pretty much done. There are just a couple of outstanding bugs that I haven't been able to resolve by lurking here, Googling, or by reading datasheets.
The bugs are primarily related to the relationship between my Cycle Analyst 3 and my Sempu BB PAS/torque sensor, which probably isn't surprising as a few people have reported issues here before. Here are the details:
A note on the PAS - I bought directly from China, so mine outputs a 48pulse/rev square wave on one wire, and a direction logic signal on another, rather than a quadrature output. The CA3 config doesn't seem to allow a 48 pulse PAS input, so I've set it to 24 pulse/rev. That should result in the CA3 overestimating my human watts by a factor of two, unless I halve the torque value. I don't see how that could cause my issue, but I'm open to suggestions
I've been building my ebike over the course of 18 months after seeking advice from these forums and lurking/reading a lot, and it's pretty much done. There are just a couple of outstanding bugs that I haven't been able to resolve by lurking here, Googling, or by reading datasheets.
The bugs are primarily related to the relationship between my Cycle Analyst 3 and my Sempu BB PAS/torque sensor, which probably isn't surprising as a few people have reported issues here before. Here are the details:
- 1.) If I apply sufficient torque through the pedals, the output voltage from the CA3 to my Phaserunner will go to fault voltage (3.64V) and wait a few seconds before I can apply power by either peddaling or using the throttle input. This is seeply annoying if I'm climbing a steep hill. The throttle never causes the CA3 to output a fault voltage, just peddaling.
2.) The power delivered by the motor tends to surge unpredictably when peddaling. I believe this is the result of the CA3 getting an incorrect calucaulation of human watts and sending a corresponding speed signal to the Phaserunner. I assume the human watts figure is calucated using boththe torque signal and the pulse rate from the PAS, and I am planning to use a portable DSO to figure out which signal is to blame. It's plausible that the orientation of my Sempu is to blame, but it's the latest version with no orientation marks and no directions from the manufacturer to orient it in any particualr way. But if anyone has suggestions based on past experience, I'd like to hear it.
A note on the PAS - I bought directly from China, so mine outputs a 48pulse/rev square wave on one wire, and a direction logic signal on another, rather than a quadrature output. The CA3 config doesn't seem to allow a 48 pulse PAS input, so I've set it to 24 pulse/rev. That should result in the CA3 overestimating my human watts by a factor of two, unless I halve the torque value. I don't see how that could cause my issue, but I'm open to suggestions
