Cycle Analyst V3 preview and first beta release

Get real world experience and user feedback on the electric bicycle products.
User avatar
amberwolf   100 GW

100 GW
Posts: 27915
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by amberwolf » May 05 2019 3:37am

ok, that opens. not sure if it's the cause but you have pas config settins -- assist scale factor set to 0.00 w/rpm. maybe change that to something significantly higher to see if it does anything, then tune it down to where you actually want it once you know it's working?

Lowtech_velorizontal   1 µW

1 µW
Posts: 3
Joined: May 03 2019 3:16pm
Location: Guérande

Re: Cycle Analyst V3 preview and first beta release

Post by Lowtech_velorizontal » May 05 2019 4:58am

Thanks for the answer.

Just tried to modify the W/rpm, with several values until the max of 40 W/rpm but still the same problem.

ebike11   100 kW

100 kW
Posts: 1163
Joined: Nov 16 2013 8:01pm
Location: Korea

Re: Cycle Analyst V3 preview and first beta release

Post by ebike11 » May 05 2019 2:20pm

Hi guys
2 part question

A) I have a couple of large controllers that im testing with my MXUS motor.
Im using the external shunt here

https://www.ebikes.ca/shop/electric-bic ... t-ca3.html

And was wondering, if the controller specs 96V 150A are higher than the shunt specs, then must i upgrade the shunt if its possible to do so? Shunt specs says 50A continuous and 100A max.

B) The POWER LIMITS on the CA only go to 99A and 9900W. Are these max cutoffs or will higher power limits go over these limits but not show up on the display?
Will a 10000 watt + system be displayed properly? Or will it only display up to the 4 digits of 9999?

Thx
Last edited by ebike11 on May 06 2019 10:41am, edited 1 time in total.

User avatar
amberwolf   100 GW

100 GW
Posts: 27915
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by amberwolf » May 06 2019 12:01am

Lowtech_velorizontal wrote:
May 05 2019 4:58am
Just tried to modify the W/rpm, with several values until the max of 40 W/rpm but still the same problem.
Hmm. NOt sure on that--you might directly contact ebikes.ca, if you haven't already.

I suppose it's possible that eeprom settings/etc are corrupted, and causing the problem. I replied in a very recent (days ago) thread about a cycle analyst issue (voltage reading?) that links to the fix for that, if it is your problem.

drbenjamin   1 mW

1 mW
Posts: 17
Joined: Jun 17 2014 2:22pm
Location: Redmond, WA USA

Re: Cycle Analyst V3 preview and first beta release

Post by drbenjamin » May 22 2019 1:08am

Pls disregard this post, I have figured out how to use the potentiometer in CAV3
Last edited by drbenjamin on Jun 09 2019 9:04pm, edited 1 time in total.

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Sempu goes crazy

Post by izeman » May 24 2019 10:12am

I did install a sempu torque bottom bracket, and it was working more or less fine.

Two things sometimes happen:

a) the bike doesn't react on pedal pressure at all
b) the bike over reacts and shoots off with fully open throttle

So i made this video showing the setup screen of the sempu sensor. Sometimes it shows 0Nm if no pedal pressure is present, which is fine. Then you put your foot on the pedal and it shows a WAY too high amount of torque (leading to way too much throttle). But worst of all the torque value sometimes doesn't go down to zero after you release pedal pressure. Or it shows negative torque, which is impossible (at least if you're not pulling the pedal up (what i don't do).
Please see the video https://youtu.be/HN0bD33ZWPk
Any advise welcome.

kevinscargobike   1 W

1 W
Posts: 55
Joined: Nov 01 2017 6:47pm

Re: Sempu goes crazy

Post by kevinscargobike » May 24 2019 5:04pm

izeman wrote:
May 24 2019 10:12am
Sometimes it shows 0Nm if no pedal pressure is present, which is fine. Then you put your foot on the pedal and it shows a WAY too high amount of torque (leading to way too much throttle). But worst of all the torque value sometimes doesn't go down to zero after you release pedal pressure. Or it shows negative torque, which is impossible (at least if you're not pulling the pedal up (what i don't do).
The good folks at Grin have replaced two of them for me that had behavior roughly like this.

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by izeman » May 25 2019 1:34am

So you say it's a hardware issue?

kevinscargobike   1 W

1 W
Posts: 55
Joined: Nov 01 2017 6:47pm

Re: Cycle Analyst V3 preview and first beta release

Post by kevinscargobike » May 25 2019 2:04am

izeman wrote:
May 25 2019 1:34am
So you say it's a hardware issue?
My memory of it the first time was that they had me double check that the dot was in the right place relative to the crank arm, and then just mailed a new one to me - which I guess must mean they thought it was a hardware problem?

The second one was more strange, it was on a bike that usually used a 36v battery and worked fine, but would go negative and read high only when I was using a 52v battery instead. That one was mailed back to them, but I didn’t hear anything about the cause.

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by izeman » May 25 2019 2:36am

The sensor is from sempu directly as it's a T4 with special spindle lenght which ebikes.ca doesn't sell.
I was hoping for some electrical problem like adding a cap or resistor may help probably to signal noise issues or something.

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by izeman » Jun 03 2019 9:55am

I did some more investigation and the sensor gives some strange readings.
First of all it now shows 1.88V resting voltage and 20Nm w/o touching the pedals (where as it was 1.5V before). When i set sensor to "custom" i can reset the ZERO torque voltage, and then it shows 0Nm again.
And when i ride it works quite ok, but suddenly shows 600-1000hW and the bike boosts away as it tries to multiply that. Of course my input is minimal (<100W) at that time.
To me it looks like two possible reasons: a) broken sensor b) noise on the wire or some grounding issues?!?
Any more ideas?

drbenjamin   1 mW

1 mW
Posts: 17
Joined: Jun 17 2014 2:22pm
Location: Redmond, WA USA

Re: Cycle Analyst V3 preview and first beta release

Post by drbenjamin » Jun 09 2019 9:03pm

Hi All

I'm setting up my new ebike w/ a CA. I got everything working great using the Aux input (a potentiometer) to control the power limit during PAS. I am now trying to set up a preset to use the pot as a speed limiter (have the CA keep constant spped by varying power). From reading the docs on Grin's website it looks like in version 3.0 we could set the AuxA->Function on a per preset basis, but not in version 3.1. Does anyone know why this change was made and is there any way to revert? I expect I'll have to do some PID tuning to get the speed limiting working well and it would be nice to be able to switch a preset back to using the pot for power.

Thanks!

User avatar
danielrlee   1 kW

1 kW
Posts: 490
Joined: Mar 29 2013 8:30am
Location: Wiltshire, UK
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by danielrlee » Jun 10 2019 2:18am

drbenjamin wrote:Hi All

I'm setting up my new ebike w/ a CA. I got everything working great using the Aux input (a potentiometer) to control the power limit during PAS. I am now trying to set up a preset to use the pot as a speed limiter (have the CA keep constant spped by varying power). From reading the docs on Grin's website it looks like in version 3.0 we could set the AuxA->Function on a per preset basis, but not in version 3.1. Does anyone know why this change was made and is there any way to revert? I expect I'll have to do some PID tuning to get the speed limiting working well and it would be nice to be able to switch a preset back to using the pot for power.

Thanks!
The change was made in order to reduce precious EPROM usage. Unfortunately, there is no way to revert to the previous behavior without a firmware downgrade.
https://TorqueTech.co.uk : Electric bike spares, parts & accessories : 10% discount code ’ENDLESS10

drbenjamin   1 mW

1 mW
Posts: 17
Joined: Jun 17 2014 2:22pm
Location: Redmond, WA USA

Re: Cycle Analyst V3 preview and first beta release

Post by drbenjamin » Jun 10 2019 8:46pm

danielrlee wrote:
Jun 10 2019 2:18am

The change was made in order to reduce precious EPROM usage. Unfortunately, there is no way to revert to the previous behavior without a firmware downgrade.
Rats. OK, thanks for the reply!

JanComputerman   1 W

1 W
Posts: 63
Joined: Mar 08 2018 3:08am
Location: Rhode Island, USA

Re: Cycle Analyst V3 preview and first beta release

Post by JanComputerman » Jun 17 2019 5:19am

Wow plenty of reading here! Just flashed last week 3.14 from 3.12 on my CA3 and got the max speed 654 a couple of times and not once...not a big problem. I run a slim throttle (fantastic!) and current control (torque drive) instead of speed control which allows nice controllable takeoff acceleration, throttle proportional Regen with a left brake lever ebrake switch on the front brake. All good, however, I see the kick of motor power is still present when I forget to release the throttle before I release the ebrake switch. I have throttle ramps setup but they apparently are not used with the ebrake switch which would fix the kick. It has the unwanted effect of high axle torque on the rear dropouts as well as a kick in the butt from the seat. I see it was a PAS problem for some time too but I don't use PAS (5v sensor). Oh and 2133 miles on the clock in a year with no riding for many wet months this year.
How about it?

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by izeman » Jun 17 2019 5:24am

Just wondering: Is there still some development done? I don't see any progress or anything since teklektik left the forum.
Is it just me that has this feeling?

User avatar
tomjasz   10 GW

10 GW
Posts: 4146
Joined: Mar 29 2014 1:45pm
Location: Out riding, MN USA
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by tomjasz » Jun 17 2019 10:40am

izeman wrote:
Jun 17 2019 5:24am
since teklektik left the forum.
What happened to teklektik?
Thanks Justin_le we're here thanks to you. All the best to the mods for their tireless work keeping it on an even keel.

Marc S.   100 W

100 W
Posts: 236
Joined: Jul 04 2012 12:23pm
Location: Hamburg, Germany
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by Marc S. » Jun 18 2019 2:53am

izeman wrote:
Jun 17 2019 5:24am
Just wondering: Is there still some development done? I don't see any progress or anything since teklektik left the forum.
Is it just me that has this feeling?
Nope, you are right. Since Tek left, this forum thread is way less useful.

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by izeman » Jun 18 2019 3:02am

tomjasz wrote:
Jun 17 2019 10:40am
izeman wrote:
Jun 17 2019 5:24am
since teklektik left the forum.
What happened to teklektik?
Can't say for sure, but i heard there were some "different view points" between him and other user(s). So he left. But this is only hear say. I tried to contact him twice through different channels, but sadly he didn't react. :(

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by izeman » Jun 18 2019 3:03am

Marc S. wrote:
Jun 18 2019 2:53am
Nope, you are right. Since Tek left, this forum thread is way less useful.
Yes. He was such a supportive person with such a bunch of knowledge. It's a shame he left.
Maybe we should start a petition to bring him back? :wink:

User avatar
rowbiker   100 W

100 W
Posts: 125
Joined: Jul 16 2013 1:36pm
Location: Saint Paul, Minnesota USA
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by rowbiker » Jun 18 2019 9:46am

+++1

User avatar
justin_le   10 MW

10 MW
Posts: 2233
Joined: May 28 2007 3:27am
Location: Vancouver
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by justin_le » Jul 04 2019 4:52am

izeman wrote:
Jun 03 2019 9:55am
I did some more investigation and the sensor gives some strange readings.
First of all it now shows 1.88V resting voltage and 20Nm w/o touching the pedals (where as it was 1.5V before). When i set sensor to "custom" i can reset the ZERO torque voltage, and then it shows 0Nm again.
And when i ride it works quite ok, but suddenly shows 600-1000hW and the bike boosts away as it tries to multiply that. Of course my input is minimal (<100W) at that time.
To me it looks like two possible reasons: a) broken sensor b) noise on the wire or some grounding issues?!?
This is a direct quote from Tang of Sempu
We had a batch last year. One of the electronic components was of poor quality, which led to the problems that had occurred. Now the problems have been solved.
So almost all of these mystery behaviors and inconsistent performances of the CA3's torque PAS mode with the Sempu BBs seems to be from a bad batch of sensors. What we would see on probing devices on our end is that the torque signal would often end up randomly going to either 0V (resulting in negative human watts), or jumping right up to like 4.5V (resulting in very high human watts and the power boost), and then jump back to normal operation. Though sometimes they would only go back to normal after unplugging and replugging the Sempu from the CA in order to reset it.

We've been trying to get Sempu to send us a large number of replacement sensors for which the "problems have been solved" which we'll then be able to send out under warranty to anyone who is still experiencing erratic Sempu torque signal issues.

And what I would advise in the meantime for people currently having this problem is to change the CA3's PAS control mode from Torque Assist to Basic PAS for the time being, and then at least the behavior will be nice and predictable :-)
Previously competed in the Suntrip race on a back to back tandem solar powered row/cycle trike. 550 watt solar roof, dual Grin All Axle hub motors, dual Phaserunner controllers, 12 LiGo batteries, and a whole wack of gear.

Now back in Vancouver learning to be a dad with my Big Dummy Frame (yes This One, thanks ES!) with GMAC 10T rear hub motor, Phaserunner controller, and 52V 19Ah EM3EV pack
My website: http://www.ebikes.ca
Please contact via email, info@ebikes.ca, rather than PMs, which are disabled

User avatar
justin_le   10 MW

10 MW
Posts: 2233
Joined: May 28 2007 3:27am
Location: Vancouver
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by justin_le » Jul 04 2019 5:00am

JanComputerman wrote:
Jun 17 2019 5:19am
Wow plenty of reading here! Just flashed last week 3.14 from 3.12 on my CA3 and got the max speed 654 a couple of times and not once...not a big problem.
Yes, unfortunately this is a bug that seems to have slipped in undetected with the most recent release. It doesn't affect any of the average speed or distance computations and we'll be having a 3.15 build that will fix this. In the meantime if someone needs to routinely check their max speed reading, I'd recommend either installing the 3.13 firmware, or press and hold the right button to do a reset when you are on the screen with the max speed display. This will _only_ reset the max speed value to 0 without resetting your trip Ah and Distance, so you can use it to clear the 654 reading without affecting anything else.
I run a slim throttle (fantastic!) and current control (torque drive) instead of speed control which allows nice controllable takeoff acceleration, throttle proportional Regen with a left brake lever ebrake switch on the front brake. All good, however, I see the kick of motor power is still present when I forget to release the throttle before I release the ebrake switch. I have throttle ramps setup but they apparently are not used with the ebrake switch which would fix the kick.
Can you tell me do you still get this kick if you change over to pass-thru throttle mode, or is it only experienced with the amps throttle? And if you do experience it only with the amps throttle, can you change to a watts throttle (basically the same thing) and see if it is also present then? We implemented more special windup control behavior in the power limit feedback loop (for instance with PAS) and may have overlooked doing the same for the amps feedback loop.
Previously competed in the Suntrip race on a back to back tandem solar powered row/cycle trike. 550 watt solar roof, dual Grin All Axle hub motors, dual Phaserunner controllers, 12 LiGo batteries, and a whole wack of gear.

Now back in Vancouver learning to be a dad with my Big Dummy Frame (yes This One, thanks ES!) with GMAC 10T rear hub motor, Phaserunner controller, and 52V 19Ah EM3EV pack
My website: http://www.ebikes.ca
Please contact via email, info@ebikes.ca, rather than PMs, which are disabled

User avatar
justin_le   10 MW

10 MW
Posts: 2233
Joined: May 28 2007 3:27am
Location: Vancouver
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by justin_le » Jul 04 2019 5:10am

drbenjamin wrote:
Jun 09 2019 9:03pm
From reading the docs on Grin's website it looks like in version 3.0 we could set the AuxA->Function on a per preset basis, but not in version 3.1. Does anyone know why this change was made and is there any way to revert?
There were a number of reasons for this but one of them came down to limited eeprom memory space for additional parameters as we added all the new features in CA3.1 (like digi-aux). There was also the interplay with now having both digital and analog aux controls being able to change per preset, when there is logic we have to exclude both aux devices from simultaneously controlling the same preset.
I expect I'll have to do some PID tuning to get the speed limiting working well and it would be nice to be able to switch a preset back to using the pot for power.
What we'd recommend here is just having your auxA control the speed limit, and then use the digi aux to control your power limit, or visa versa. Is there a reason that this approach wouldn't work fine for what you're trying to do?
Previously competed in the Suntrip race on a back to back tandem solar powered row/cycle trike. 550 watt solar roof, dual Grin All Axle hub motors, dual Phaserunner controllers, 12 LiGo batteries, and a whole wack of gear.

Now back in Vancouver learning to be a dad with my Big Dummy Frame (yes This One, thanks ES!) with GMAC 10T rear hub motor, Phaserunner controller, and 52V 19Ah EM3EV pack
My website: http://www.ebikes.ca
Please contact via email, info@ebikes.ca, rather than PMs, which are disabled

User avatar
izeman   10 GW

10 GW
Posts: 4929
Joined: Jun 21 2011 8:25am
Location: vienna, austria
Contact:

Re: Cycle Analyst V3 preview and first beta release

Post by izeman » Jul 04 2019 5:11am

justin_le wrote:
Jul 04 2019 4:52am
So almost all of these mystery behaviors and inconsistent performances of the CA3's torque PAS mode with the Sempu BBs seems to be from a bad batch of sensors. What we would see on probing devices on our end is that the torque signal would often end up randomly going to either 0V (resulting in negative human watts), or jumping right up to like 4.5V (resulting in very high human watts and the power boost), and then jump back to normal operation. Though sometimes they would only go back to normal after unplugging and replugging the Sempu from the CA in order to reset it.
Exactly the behaviour I noticed. Thanks for confirmation. Sempu in the meanwhile sent a replacement for FREE, and it seems that this is working fine. It does look a bit different as well (the wire going into the BB is a bit different, missing the braided sleeve and it glued in more nicely) - maybe this is how they can be identified - in case the new sensor is from a different batch, which doesn't have those issues.
Manufacturing date (if that is) is only 8 days apart 20190115 vs 20190123.

Capture.JPG

Post Reply