Cycle Analyst V3 preview and first beta release

[teklektik]

Ah - okay.
This can be a little vexing to set up with a high power to weight ratio or a really heavy bike - low-powered bikes are a lot easier to tune. There really are no general settings that would be helpful.

Are you setting these because you need the speed limit to work for legal reasons? If you don't actually need the speed limit, disable the speed PID controller as described on the bottom of p30.

Otherwise, the tuning instructions in "Appendix F. Speed Gain Adjustment Procedure" is the closest thing available to a fixed cookbook tuning procedure. Your bike has lots of torque and you may find yourself running up and down the street many times to get this working satisfactorily. Some speed overshoot is not unusual, but it should be adjusted to be a single minor one.

Here's the values on my bike ONLY to show you relative value sizes since I have similar motors but a much heavier bike - these are only illustrative and it's very unlikely they would work on your bike: (P,I,D) = (0.41,9,2). As you can see the values are very small - tuning means making even smaller adjustments. These values arose from the tuning procedure mentioned above.

• I like to use a sort of 'binary search' technique for these sorts of things: Make an adjustment that is about 1/2 the available range. If it isn't enough adjust (+) 1/2 that step, if too much, adjust (-) 1/2 that much. I proceed this way going up or down 1/2 the previous step each time. This means that in 7 tries I have spanned 2^7 or 128 different values. Often something 'close enough' comes up much sooner. Anyhow - compared to just guessing bigger or smaller each time, this goes relatively quickly and directly homes in on the desired target value.

 

[Architectonic]

It seems you have two issues that you can address to held remedy the problem:
[*]Moisture: Moisture can certainly cause signals to scoot around the PCB and make the firmware see inputs that were never intended. This puts everything in question so getting the unit dried out is a first step.

Thanks.

The problem was likely due to moisture as I haven't seen the issue again. I am getting a replacement top cover (due to aforementioned button cover missing), so I'm going to try and make sure it is a little better sealed up than before...

 

[Santacruz]

Hello,
I have updated to CA3-!b3 and find it very good.
However, I am struggling to understand some of the technology used.
Oscillating and hunting under the gain adjustment section.
While I understand what I am reading, I do not know what I should be feeling from the motor when riding.
Can anyone explain these terms to me, for some reason, in this area I am a bit thick.
Also, if I use one of my presets which is fairly high power, I get a type of grinding feeling and sound from the motor when I throttle from standstill. I now pedal to get moving and then the throttle behave as it should. Is there settings I can make for this.
My bike is a self built MTB 26" With following specs.
Hallomotor 1500w rear BLDC with 10 speed cassette. 11t to 24t
Crank is a 60t.
Controller is Golden motor BAC-0501 connected by Cycle analyst v3 with the CA shunt. I have run and extra wire to the controller brake to use the regen function that works fine.
Power is 6 x 6s Zippy lipos. 3 parallel 2 series, so I have 48v at 15 ah.
On the hottest day about 28 centigrade and pushing the bike quite hard (55kph) I saw temp of about 60 cent' and the batteries never got warm. So I think my set up is reasonable.
So any help and advice on what the gain settings do and what I should hear and fell from the motor would be very helpful.
I am presently living and working in Holland, so the bike does all its work on cycle lanes and totally flat.
John

 

[teklektik]

The problem was likely due to moisture as I haven't seen the issue again. I am getting a replacement top cover (due to aforementioned button cover missing), so I'm going to try and make sure it is a little better sealed up than before...

Excellent. Thanks for getting back about the resolution to the issue. Appreciated.

 

[teklektik]

Hello,
I have updated to CA3-!b3 and ...

Hey John-
Glad the b3 release is working out for you.

As for the technical stuff with gains - well - the [strike]Un[/strike]official Guide is tough sledding for many - don't let it get you down - this thread is here to take up the slack where the Guide falls down... It's hard to write to please everybody, but frankly reading your post , you seem to have this pretty much under control - it looks like your CA hookup and the rest of the build are spot on.

So - there seem to be two issues:

1. "Oscillating and hunting under the gain adjustment section" and
2. "a type of grinding feeling and sound from the motor when I throttle from standstill"

#2 is an immediate concern and we need to get the motor running smoothly before trying to tune out the surging or oscillation.

This actually doesn't sound like a CA issue and I would recommend hooking the throttle directly to the controller and verifying that the motor runs 100% without the CA. If this is not practical, or you cannot make up the extension cables, etc to get it hooked up, you can try setting ThrI->CntrlMode = Pass-thru and following the directions on the bottom of p30 of the Guide to disable speed limiting. This should get the CA PID controllers pretty much out of the picture so you can repeat your experiment and see if the grinding noise persists. Check the Limit Flags on the Diagnostic screen and see if any pop up when the problem occurs. If you get flags then the CA is involved and we can pursue that; if not, then this is likely a PhaseRunner config issue and you will likely need to go to the PhaseRunner thread or contact Grin for controller tuning assistance.

Once the CA vs PhaseRunner matter is clarified we can go after the remaining CA issues.

 

[Santacruz]

Thank you for the information.
I am a little confused about the "phaserunner" issue.
I am running a Golden motor BAC 0501 controller and I would guess this has nothing to do with phaserunner. Or does it use the same software?

 

[teklektik]

I am a little confused about the "phaserunner" issue.

Ouch. So, apparently, was I...
It looks like speed-reading led me to a BAC-500 Phaserunner in stead of the BAC501 GM controller.

Okay - with a simpler less configurable controller, it's also less likely to be that. So looking back to the symptoms:

Also, if I use one of my presets which is fairly high power, I get a type of grinding feeling and sound from the motor when I throttle from standstill.

From a diagnostic perspective, I would still feel better if I knew this was caused by the CA - that is, if the problem goes away with the throttle going directly to the controller. If this test is not possible, then looking at the CA: it may be that hard acceleration is causing the issue and disabling the Speed PID controller as mentioned above will do the trick - give the test above a try. If the problem goes away, then it's almost certainly the speed controller. Try restoring the throttle to your preferred mode (perhaps Current Throttle) and try again. If there is still no grinding or hesitations then it's the speed controller, if not, then check the Limit Flags to see what is in play. If you can post back with the results we can look at what you found.

BTW - do you need to have speed limiting (legal reasons, etc) or is leaving it disabled okay? If you need it, the gains will need be tuned, otherwise you can just leave that part disabled.

Sorry again for the controller confusion...

 

[Santacruz]

The legal requirement here is max 250w and max 25kph. This I have dialed in for preset 1 and works fine.
Preset 2 I have on throttle with current control and pas disabled with a power limit of 40 amps.
Preset 3 I have open on all settings with throttle in current control.
The controller is set at 50 amps continuous with 75 peak, acceleration at 75% (Don't know what this is for, but 50% or less gives me a huge throttle delay with or without ramp up settings at either end of the scale on the CA3) and brake regen set at 60%. I have one extra wire from the brakes to the controller brake input for this and it works fine.
I will this weekend try direct throttle to the controller for a test.
It is possible that my weight is just a bit too much for the motor from standstill.
If I pedal and then throttle there is no problem.
My pas starts at about quarter crank turn and after one turn if I throttle it is fine.
I know I am probably pushing a bit too much in amps into the motor. 1500w @ 48v should probably be more like 31 to 35amps, but if you don't push it where's the fun.
I will report back after the weekend with anything I find.

John

 

[teklektik]

A new Preliminary Digital Aux doc has been posted for download here - now v1.0d.
This is a very small update to the adapter illustration on p5 to correct the UP/DN switch sense.
No other changes.

 

[Marc S.]

I've just made the update to v3.1b3 and come back a short 4km test run.
Everything works fine, but it seems the CA V3 doesn't calculates the resistance of the battery pack anymore and showes .000 ohm.

A known issue of v3.1b3?

 

[izeman]

Had a similar issue as well. The calculated value was frozen. Didn't change. Problem didn't appear with 3.0 release.

 

[teklektik]

Hmmm.
Okay, thanks guys. I put in a ticket to look into this.

This has been working fine here with hundreds of road miles, but it looks like there is an issue under the right circumstances.

• If possible, it would be good if you could post up here or PM me with your settings files so we can try to stage the failure mode with your configuration to get a leg up on duplicating the problem. No worries if this isn't doable...

Just to verify: these Setup configurations are brand new from-scratch 3.1 configs and not based on a 'corrected' existing 3.0 Setup file or an existing CA setup overwritten with a 3.1 NoEeprom firmware flash? I doubt this is the case, but have to ask since those scenarios could easily cause this wonky behavior...

 

[Marc S.]

Argh! Thats the error. I f***ed it up.
Thanks for the reminder!

I created a new 3.1 config file, but out of habit I choose the noEeprome.hex file...
Luckely made a screenshot.

How do I correct this?
Plug the CA V3 into the computer, read the config file, create a new 3.1 config file from scratch, type-in the values, ect, repeat the firmware update with the CA3-1b3_firmware.hex file, upload the new config file into the CA and I'm good, right?

 

[teklektik]

Argh! Thats the error.

No problem, this change to the EEPROM map is just a pain but we really needed to do it...
Sorry for the confusion.

How do I correct this?
Plug the CA V3 into the computer, read the config file, create a new 3.1 config file from scratch, type-in the values, ect, repeat the firmware update with the CA3-1b3_firmware.hex file, upload the new config file into the CA and I'm good, right?

Actually, the stuff in the CA is probably corrupted at this point and it's hard to tell what it will look like if you read it with the Setup utility. One problem is that the CA will tell the utility it's a 3.1 file even though it may be shaped like a 3.0 file, so the utility will misinterpret it as the wrong type anyway.

I would recommend pulling your lifetime statistics (mileage, battery cycles, total Ah) manually from the CA console. Then open an old 3.0 setup file in the utility. Start a second utility, create a new 3.1 setup, and manually transfer in the 'manually saved' lifetime stats and old 3.0 config. Save it, then shoot it into the CA after flashing the CA with the Plain Jane 3.1 'firmware.hex' file.

FWIW:

• As we move forward with 3.1, there will be additional EEPROM hiccups like this, but we'll try to minimize them (3.1b4 works with 3.1b3 setup files - Yahoo!). The problem is that when one of these tiresome episodes happens and you install a NEW 3.1 beta over and OLD 3.1 beta, it's no longer possible to examine or use setup files made with the OLD 3.1 because the utility can't differentiate old and new 3.1 versions. However, 3.0 setups still can be examined because they have a uniquely different designation. So - that was probably more than you wanted to know, but the point is we're aware of the difficulties but just now think it's better to work on CA stuff than put resources into tinkering up a smarter Setup utility. Hoping you agree...

BTW - Please report back if this resolves the issue (you, too, izeman ) so I can close out the ticket if this was a red herring. Thanks!

 

[Marc S.]

No need to apologise!
This is a beta FW version of an enthusiast device from an enthusiastic little company (with way better customer support than many big companys provide, I might add).

I've carefully read the v3.1b3 update instructions a couple of days ago when I updated the CA3 from my other trike as a test (its in pieces at the moment).
Well, I failed the test at the second approach, didn't I?

I agree completely about putting your resurces into the CA3 and deal with the CA setup utility later. Buy the guy a beer once in a while, will you?
(wasn't there a Phaserunner OS-X utility to finish as well? )

Will make a new FW update tomorrow, do a test ride and report back.

 

[Marc S.]

Did the new 3.1b3 FW update and loaded a brand new config file into the CA3.
Sure, it started to rain while the FW update was in progress. No test ride.

All seems fine and looks exactly like it should. Battery resistance field shows 0.199 Ohm

Marc

 

[teklektik]

Hey Marc-
Thanks for getting back on this.
If it goes all cattywampus on you after you get a chance for a real road test, just post up.
Glad it got sorted...

 

[Marc S.]

I will!

 

[justin_le]

Just a little pre-release teaser for the v3.1b4 firmware...
The firmware is actually ready but in-house road testing is backed up due to the phase of the moon and other seemingly cosmic events. Anyhow, it's the time of year for many to be working on builds or enhancements, so it seemed good to get this information out there ASAP with the firmware to follow shortly.

So:
Version 3.1b4 beta firmware introduces a new Digital Aux feature that allows the Cycle Analyst to actively control two different Aux Input features at once using two separate external controls. This applies across all presets and either input channel can be configured to control any of the existing 3.0 Aux Input modes (limiting or presets). This means that you can have one control for Amps limiting and another control for PAS Assist - both can be adjusted as needed.

Guys, you need to acknowledge how awesome this feature is that Teklektik has managed to pull off in the CA3 firmware! Basically, in the original CA3 layout I had figured that an analog control would be the most versatile way to allow people an infinite adjustment on any limit settings, but in practice the hardware for this is expensive (good waterproof pots being ~$10-20), and most people seem to prefer having several discrete levels instead of continuous adjustment anyway.

With the digital Aux input the same 3-pin plug on the CA is used, but it's hooked up in such a way that you can run the CA3 with only a set of digital buttons attached, or only an analog potentiometer attached, or both an potentiometer AND a pair of up/down buttons to control two different variables. Teklektik wrote a very nice PDF summary of how to this works and is wired here:
https://endless-sphere.com/forums/viewtopic.php?p=571345#p571345

But I'll place the relevant images on this post too for better visibility. A single button digital input (aux level increases to max and then wraps around to zero) only needs a button and a resistor. But if you include a diode in the wiring then you can have separate up/down buttons which would be the normal and preferred approach.


If you have an existing potentiometer or CA 3-position switch (which is internally wired like a 3 way voltage divider), then the following circuit placed inline to the existing analog aux line and adds the digital buttons while still preserving the analog behavior too. This requires two diodes in the circuitry.


What these circuits do is shift the analog range from being 0-5V into a ~1-5V range, and then an 'UP' button press becomes 0V, while a 'Down' button press is about 0.5V.


Here's what it can look like if you solder up a quick prototype of the above circuit using discrete tactile push buttons on vectorboard.

A nicer looking solution would be to use an (on)-off-(on) style rocker switch like this,
http://sell.lulusoso.com/selling-leads/384074/Rocker-switch-Electric-welder-switch-ON-OFF-ON-momentary-rocker-switch.html
But in any case it should be easy for people to prototype with spare items in their parts bin.

It's kinda ghetto but this suits everything on my own ebike at the moment and does the job zip tied on top of the now demoted 3-position switch.


I have a torque sensor on my ebike and have set the digital aux to control my pedal assist level, from 0 to 4 times in steps of 0.8, and whenever the button is pressed the screen momentarily updates to show this value


And the potentiometer is used to set the maximum amps draw from the battery pack, and any time this is rotated the new value is also shown on the display


It should only be a couple more days before we publish the 3.1 beta firmware which has this digital aux functionality, and I'm excited to get this into people's hands since admittedly it's been pretty long overdue.

 

[Arlo1]

I love it. Awesome work Justin!

 

[amberwolf]

A nicer looking solution would be to use an (on)-off-(on) style rocker switch like this,
<snip>
It's kinda ghetto but this suits everything on my own ebike at the moment and does the job zip tied on top of the now demoted 3-position switch.

You could replace the switch inside throttles (or other handlebar modules) that have a 3-speed switch with a momentary-rocker on-centeroff-on switch, too, if you want a momentary rather than rocker click-on-off-on that they come with.

There's also pre-existing modules for switches, like the one that came with my Fusin kit here:
https://endless-sphere.com/forums/viewtopic.php?f=4&t=39877&hilit=fusin+1000w
but are problaby available separately somewhere, that are momentary switches. (that one has three, but one could be left unused, or even custom-wired to parallel one of the CA's front panel buttons to scroll thru displays, or whatever. ).

 

[teklektik]

There were a number of circuit options, but early on we purposefully went for a solution that restricted the pushbutton hookup to be equivalent to a center-off rocker/toggle so that existing and readily available controller 3-postion switches could be converted by replacing the switch. Rockers and toggles can be pretty robust and easy to work even with winter gloves - something some of us need to consider (grumble grumble...).

 

[amberwolf]

AFAICR, that 3-button Fusin module was wired common on one side like your lefthand schematic, so it should work. (someday i'll get around to fixing (or sending in for repair) my CAv3 and be able to test that).

Right now, weather here means we need gloves, too--oven mitts to keep from being charred by the heat! (seriously, if it's metal in the sun, you could get burned to one degree or another, what with 120F+ in the shade today).

 

[Santacruz]

I haven't as yet got around to trying the throttle direct to my controller to check the grinding stalling noise I get from standstill.
My rear super lightweight rack fell to pieces, so my weekend was taken up fitting a better rack for the controller and batteries.
However, coming home from work last Monday there was a torrential downpour of rain.
The bike worked perfectly.
But the next morning on the way to work, I noticed it worked even better. Better acceleration, a bit smoother, faster etc.
On getting home at the end of the day I see that the CA3 had reprogrammed itself and the presets I was using had all changed in name and some values.
I have saved these to file and reloaded my previous setup and will now study the (I don't know where it came from) program file and see what I can integrate with my own settings.
But, could this change in setup have been caused by the rain?

 

[teklektik]

But the next morning on the way to work, I noticed it worked even better. Better acceleration, a bit smoother, faster etc.
On getting home at the end of the day I see that the CA3 had reprogrammed itself and the presets I was using had all changed in name and some values.
...
But, could this change in setup have been caused by the rain?

Seems unlikely, but I would like a quart or two of that rainwater to pour on my CA....

Seriously though, water does funny things to electronics, but re-writing EEPROM with new settings is more like of a detailed sequence of operations than a fluky bit of mis-operation. It's hard to discount the coincidence, but the 'how' part is the sticking point (but that could just be a lack of imagination on my part).

We'll discuss this here - meanwhile - for now let it go as serendipitous precipitation, enjoy the better performance, and post up if it reoccurs.