Cycle Analyst V3 preview and first beta release

Get real world experience and user feedback on the electric bicycle products.
Iambuilderman   1 mW

1 mW
Posts: 15
Joined: Sep 01 2020 3:08pm

Re: Cycle Analyst V3 preview and first beta release

Post by Iambuilderman » Apr 06 2021 12:06am

justin_le wrote:
Sep 29 2020 3:23am
HrKlev wrote:
Sep 29 2020 2:03am
It is a mode for tecnical terrain where the output relies on the torque input from the BB torque sensor, and not torque and rpm combined.
Yes, actually more to the point I'm strongly considering in the next update that we just drop the watts/Hwatt scaling factor and have it instead be Watts/Nm. [...]
Has there been any progress in the implementation of being able to choose a torque multiplier instead of a power multiplier for pas using a torque sensor?

And a related question for @justin: can you implement a customizable timeout for the pas support in ca v3? Because I think the current implementation has a far too long timeout leading to dangerous spontaneous power surges on a high power ebike. Let me try to explain:

I have built a powerful direct drive ebike with sabvoton controller and ca v3 + sempu v4 torque sensor. It can output enough torque to do a power wheelie (150 amp Max). To tame the beast, I have set throttle ramp rather slow (5v/s), pas ramp even slower (1v/s) and basically disabled fast ramp (0 amp). I have ramp down really fast (15v/s). While biking, the bike is very well behaved, with controllable throttle and torque based pas. However, when I stop pedaling for a moment, e.g. 0,5s (?) for example for a corner, and then continue pedaling, the cycle analyst seems to still think I was pedaling and spikes the throttle to the Max to continue its previous power level. This leads to very dangerous situations, where I have already had a hard crash with lots of damage because the bike suddenly did an unintentional wheelie while trying to cross a street.

You could argue that the bike is too powerful but that is not the point here. I also have a modded kt controller that I put custom firmware on myself. In that firmware, you can enter a pas timeout to prevent pas engaging unintentionally.

Maybe my settings in the ca v3 are not optimal, but I think they are rather soft and safe actually. That is, when I am pedaling, not when I pause for a moment. I use 4 pole torque averaging but for the power surges it does not help to change this to 2 or 12. I have pas start at 0,15s and stop at 0,10s between poles. Changing this setting does not seem to help either; it just influences the required cadence but once you are going there seems to be another invisible pas timeout waiting for the next pole?

I have the sabvoton at minimum response time (0,010s ramp up and down) to make sure that there are no oscillations between the ca control logic and the sabvoton logic.

So it boils down to the question: is there any way to deny the ca v3 to create power surges when you briefly interrupt your pedaling motion? I think torque mode + a customizable pas timeout (“listening time”) would help. Also, I think a global setting for Max voltage ramp rate for pas, even when in power based assist mode, is desperately needed.

Please let me know if I am doing something wrong and there is a simple solution to prevent these ca v3 pas related voltage/power surges! Because right now pas is just downright dangerous...

Ingo   1 W

1 W
Posts: 55
Joined: Nov 03 2016 6:22am
Location: Germany

Re: Cycle Analyst V3 preview and first beta release

Post by Ingo » Apr 13 2021 2:01pm

hi all,

i noticed some bugs at the new firmware 3.15b1

the present name does not match with the name in software. (showes different between CA and software) e.g. pdl ast = legal
some bugs in calculated %Regen. show me 987% (should be about 1%)

someone can help me here?

regards ingo

HrKlev   100 W

100 W
Posts: 227
Joined: Dec 06 2019 10:24am

Re: Cycle Analyst V3 preview and first beta release

Post by HrKlev » Apr 13 2021 4:00pm

What is the news in this 3.15b version? I cant find release notes anywhere.
DIY lightweight mid drive with torque sensing build thread: viewtopic.php?f=6&t=107622

Limbs   10 mW

10 mW
Posts: 34
Joined: Sep 03 2020 4:11pm

Re: Cycle Analyst V3 preview and first beta release

Post by Limbs » Apr 13 2021 6:24pm

Iambuilderman wrote:
Apr 06 2021 12:06am
justin_le wrote:
Sep 29 2020 3:23am
HrKlev wrote:
Sep 29 2020 2:03am
It is a mode for tecnical terrain where the output relies on the torque input from the BB torque sensor, and not torque and rpm combined.
Yes, actually more to the point I'm strongly considering in the next update that we just drop the watts/Hwatt scaling factor and have it instead be Watts/Nm. [...]
Has there been any progress in the implementation of being able to choose a torque multiplier instead of a power multiplier for pas using a torque sensor?

And a related question for @justin: can you implement a customizable timeout for the pas support in ca v3? Because I think the current implementation has a far too long timeout leading to dangerous spontaneous power surges on a high power ebike. Let me try to explain:

I have built a powerful direct drive ebike with sabvoton controller and ca v3 + sempu v4 torque sensor. It can output enough torque to do a power wheelie (150 amp Max). To tame the beast, I have set throttle ramp rather slow (5v/s), pas ramp even slower (1v/s) and basically disabled fast ramp (0 amp). I have ramp down really fast (15v/s). While biking, the bike is very well behaved, with controllable throttle and torque based pas. However, when I stop pedaling for a moment, e.g. 0,5s (?) for example for a corner, and then continue pedaling, the cycle analyst seems to still think I was pedaling and spikes the throttle to the Max to continue its previous power level. This leads to very dangerous situations, where I have already had a hard crash with lots of damage because the bike suddenly did an unintentional wheelie while trying to cross a street.

You could argue that the bike is too powerful but that is not the point here. I also have a modded kt controller that I put custom firmware on myself. In that firmware, you can enter a pas timeout to prevent pas engaging unintentionally.

Maybe my settings in the ca v3 are not optimal, but I think they are rather soft and safe actually. That is, when I am pedaling, not when I pause for a moment. I use 4 pole torque averaging but for the power surges it does not help to change this to 2 or 12. I have pas start at 0,15s and stop at 0,10s between poles. Changing this setting does not seem to help either; it just influences the required cadence but once you are going there seems to be another invisible pas timeout waiting for the next pole?

I have the sabvoton at minimum response time (0,010s ramp up and down) to make sure that there are no oscillations between the ca control logic and the sabvoton logic.

So it boils down to the question: is there any way to deny the ca v3 to create power surges when you briefly interrupt your pedaling motion? I think torque mode + a customizable pas timeout (“listening time”) would help. Also, I think a global setting for Max voltage ramp rate for pas, even when in power based assist mode, is desperately needed.

Please let me know if I am doing something wrong and there is a simple solution to prevent these ca v3 pas related voltage/power surges! Because right now pas is just downright dangerous...
I had similar, though less powerful, surges on initial PAS pull-away & PAS re-engagement on my recently built twin motor/CA V3/PhaseRunner project, regardless of HWatts multiplier. I tamed it by reducing PAS ramp-up to 2v per second and increasing WGain to a setting well in excess of the “recommended range” of 8-25 @ 90.

Post Reply