Cycle Analyst V3 preview and first beta release

[larghio]

When ever i change my pas assist level with auxd switches the motor drops out for what feels like 1-2 second.
This was not the case with auxA on b3
kind of annoying when I need to increase assist for a hill and loose alot of momentum.

Is this normal ?

 

[teklektik]

When ever i change my pas assist level with auxd switches the motor drops out for what feels like 1-2 second.

Ya, it's a bug...

Lightrush also brought this to our attention and the problem has been addressed in b5. Lightrush has been super about testing preliminary versions of b5 and reports the throttle glitch is now gone on his bike and the AuxD stepping works smoothly as with AuxA hookups in previous 3.1 versions.

The plan is to do some additional in-house testing on this and get b5 out very shortly - it has a couple of other other minor DigiAux upgrades as well.

As always guys, thanks for the test reports!

 

[teklektik]

v3.1 Beta 5 Released (More DigiAux...)

3.1b5 is here to get the DigiAux feature ready for prime time with a bug fix or two and some new tweaks.
A Special Thanks go out to lightrush's bug reporting and testing assistance to get the annoying PAS power surge and power dropout bugs spotted and squashed!

One interesting alteration was suggested by Paul at EM3EV: the addition of a DigiAux setting to determine the initial AudD level on power-up: either Level #1 or the last level on power-down. This is the same idea as that used for Presets. The configuration now looks like this:


Another visible change is improved responsiveness to rapid AuxD button pushes. This should allow you to more quickly click to the target level. There was a little tweaking to the AuxA pop-up as well so it's displayed with a smaller adjustment and allows finer setting tweaking before disappearing.

Although most of the focus is on finalizing the DigiAux feature, mrbill's suggestion to display negative values of instantaneous Wh/mi (due to regen) in the Custom Field also got the nod, and the fix is in for that as well. Thanks mrbill!

• This release is compatible with the previous 3.1b3/b4 releases so you can flash the NoEeprom version and preserve your present settings.

Here's a summary version of the included Release Notes:

Note: V3.1b5 Setup files are **incompatible** with firmware or Setup files of any release prior to 3.1b3 including all earlier 3.1 versions.

IMPORTANT: Please install this package according to README_install.txt. Failure to do so may result in improper operation of the Setup Utility.

IMPORTANT: After installing this zip package on a PC, the Setup utility will no longer be able to properly utilize Setup files prepared for versions prior to 3.1b3. Settings in files prior to 3.1b3 will not display or edit properly and will download to the CA incorrectly resulting in improper CA operation. Similarly, setup files prepared with this release cannot be used with earlier CA 3.1 firmware versions. 3.0 Setup files are unaffected.

>> Any incompatible existing CA setup must be recorded and manually reconfigured over a fresh 'CA3-1b5_firmware.hex' installation.

>> All existing CA Setups prior to 3.1b3 must either be:

1. exported manually from the CA console (at least the statistics)
2. OR read by the Setup Utility and preserved in screen snapshots PRIOR to installing this package. Saving to a 3.1 Setup file and reading the file after installing the zip package will not work.

>> Existing 3.1b4 setups are compatible with this release, allowing re-flash of 3.1b5 firmware using the 'xxx_NoEeprom.hex' version to preserve an existing configuration. See the 3.1b4 Release notes below regarding 3.1b3 compatibility if you are jumping from 3.1b3 to 3.1b5.

IMPORTANT: Because the firmware flash will load default settings, throttle voltage configuration may be incorrect. Treat this as an initial CA installation and place the bike on a stand or otherwise elevate the drive wheel to prevent a runaway when the flash completes and before a proper configuration can be restored.

(1) 1730 - The UP BTN threshold voltage is incorrectly set to 0.35V
(2) 2394 - Watchdog timer fault may occur on power-up
(3) 1730 - DigiAux 'Current level' Setup parameter is not useful
(4) 1730 - (new) Add Setup parm for initial AuxD level on power up
(5) 1730 - DigiAux button response is sluggish
(6) 2390 - Instantaneous Wh/mi Custom View displays '---' for regen
(7) 2414 - Digital Aux has power burst when start pedaling w/TorqPAS Assist=0
(8) 2440 - PAS has brief power dropouts as AuxD buttons are pushed

So, as always: "Have at it!" and thanks for all the terrific beta tester support!

Download!
<<< CA3-1b5.zip withdrawn - see 3.1b6 >>>

(Previously implemented 3.1b4 beta features (part of this build) described here and in the Release Notes.)
(Newer 3.1b6 beta release available here!)

 

[teklektik]

There is a new posted version of the DigiAux Notes available >>here<<.

This is just a trivial change to update the Setup Utility image to show the new parameter as shown above.
No other content is altered.

 

[Marc S.]

Did a clean FW update to v3.1b5_firmware.hex with fresh setup file and I think I found a bug:

When turning the Grin poti, the AnalogAUX screen is not anymore displayed with b5.
Only if you are in the AnalogAUX setup screen and leave Setup, the AnalogAUX screen will appear for a second.

After that, I reverted back to v3.1b4, loaded the setup file and when turning the poti, the AnalogAUX screen is back again.

Doing a second attempt with v3.1b5_firmware_NoEprom.hex right now...

Nope, AnalogAUX screen is gone again with b5.
I'm going back to b4.

 

[teklektik]

Did a clean FW update to v3.1b5_firmware.hex with fresh setup file and I think I found a bug:

When turning the Grin poti, the AnalogAUX screen is not anymore displayed with b5.

Yep. You did... Rats.

It's actually a little more far-reaching than you report: AuxA-only mode (where AuxD is configured OFF) is non-functional. The pop-up isn't really missing - it's just that AuxA is effectively OFF. AuxA works fine if AuxD is not OFF which, not surprisingly, was the focus of the testing.

It's fixed here and I will get it posted up shortly.
Thanks for the bug report - appreciated!

 

[teklektik]

v3.1 Beta 6 Released (Repaired b5)

Thanks to MarcS's bug report this morning on the showstopper AuxA issue, we have a new 3.1b6 release.

I'm going to forego posting the usual Release Notes summary here and just note that this release specifically targets the b5 AuxA issue so this is really just a 'fixed b5' that should work as originally advertised. There are no other changes.

Sorry about the b5 snafu guys, we're sort of stuck between trying to keep a steady flow of releases coming this riding season and the sizeable overhead of full regression testing. This was a case of "...damned if you don't".

So: "Have at it! (again)" and thanks for the bug reports.

Download!
View attachment CA3-1b6.zip
(Previously implemented 3.1b5 beta features (part of this build) described here and in the Release Notes.)
(Newer 3.1b7 beta release available here!)

 

[izeman]

http://www.aliexpress.com/item/bike-Scooter-Wuxing-biswitch-original-headlights-button-hedlight-switch/32599530173.html

received my order today, and wanted to note, that these are NOT momentary switches.
those are really made to turn on/off a light or something like that. so press it, and it's locked in.
just a warning.

 

[Merlin]

perfekt for my elmoto thx J

 

[larghio]

http://www.aliexpress.com/item/bike-Scooter-Wuxing-biswitch-original-headlights-button-hedlight-switch/32599530173.html

received my order today, and wanted to note, that these are NOT momentary switches.
those are really made to turn on/off a light or something like that. so press it, and it's locked in.
just a warning.

my green dk-11 from ebay are momentary. not latching

 

[Ecyclist]

Hi guys,
I have a GNG 1500W mid drive kit on my bike and I hooked up CA3 to enhance throttle operation. Can you post settings for CA3 running a GNG motor or a small block motor so I have some kind of starting point? Should I upgrade to the newest firmware or do it later? I still have a steep learning curve ahead of me and I'm hoping for a small shortcut.
Thank you.

 

[johnnyz]

Hey there

Having what I thought was a throttle problem with replacing the stock controller with my 18 fet crystalyte controller on a stock ebike. Between 30 kph and 47 kph very twitchy throttle..when I clamp down on the throttle and handle bar with my hand so it doesnt move I can take most of the twitchiness out of it. I ordered a CAv3. It appears to be a direct plug in to my controller (the CA v2 was and worked well). I want to utilize the throttle ramping features of the CA so that I can take away some of the torque (At initial starting speed and for cruising).

Question; is this unit plug and play with my Crystalyte controller? and can you suggest any presets with respect to limiting the ramp on the throttle?

Thanks

John

 

[Alan B]

I haven't kept up with the updates, so perhaps this has already been fixed, but I've seen no discussion on it so I'll ask. On my Cycle Analysts, all versions, the peak readings are so frequently bogus that I have learned never to trust them. The peak reading logic needs to have better filtering against bad ADC readings. Peak speed, current, power, minimum voltage all seem to exhibit this to one degree or another. Has this been reported, addressed, improved or even discussed? The values are often nonphysical, such as an impossible peak speed.

Thanks!

 

[MrDude_1]

I haven't kept up with the updates, so perhaps this has already been fixed, but I've seen no discussion on it so I'll ask. On my Cycle Analysts, all versions, the peak readings are so frequently bogus that I have learned never to trust them. The peak reading logic needs to have better filtering against bad ADC readings. Peak speed, current, power, minimum voltage all seem to exhibit this to one degree or another. Has this been reported, addressed, improved or even discussed? The values are often nonphysical, such as an impossible peak speed.

Thanks!

the impossible peak speed is likely a poor speedometer pickup... a double pickup on a magnetic one for example, or noise on a controller hall output... etc. Its only as good as the data you feed it.

 

[Alan B]

I haven't kept up with the updates, so perhaps this has already been fixed, but I've seen no discussion on it so I'll ask. On my Cycle Analysts, all versions, the peak readings are so frequently bogus that I have learned never to trust them. The peak reading logic needs to have better filtering against bad ADC readings. Peak speed, current, power, minimum voltage all seem to exhibit this to one degree or another. Has this been reported, addressed, improved or even discussed? The values are often nonphysical, such as an impossible peak speed.

Thanks!

the impossible peak speed is likely a poor speedometer pickup... a double pickup on a magnetic one for example, or noise on a controller hall output... etc. Its only as good as the data you feed it.

No speedo pickup involved. This is a DD hubmotor bike.

However you are right, speedo pickups are noisy and need good software filtering.

 

[MrDude_1]

I haven't kept up with the updates, so perhaps this has already been fixed, but I've seen no discussion on it so I'll ask. On my Cycle Analysts, all versions, the peak readings are so frequently bogus that I have learned never to trust them. The peak reading logic needs to have better filtering against bad ADC readings. Peak speed, current, power, minimum voltage all seem to exhibit this to one degree or another. Has this been reported, addressed, improved or even discussed? The values are often nonphysical, such as an impossible peak speed.

Thanks!

the impossible peak speed is likely a poor speedometer pickup... a double pickup on a magnetic one for example, or noise on a controller hall output... etc. Its only as good as the data you feed it.

No speedo pickup involved. This is a DD hubmotor bike.

However you are right, speedo pickups are noisy and need good software filtering.

you cant filter everything in software. If you're getting noise on your hall sensor signal, that needs to be fixed. unfortunately I dont know how to check that other than to use a scope.

 

[Alan B]

I've seen bad readings with different bikes, motors, controllers and different Cycle Analysts, not all of them even mine.

Perhaps we should all start looking at the peak/min values and report on anomalies. I can't be the only one experiencing this, in fact the most recent reports I have are from others for extremely high battery current peaks that I just don't know if we can trust or not. Other examples are min voltage occasionally reports something ridiculously low, impossible, my 32AH LiPo pack would melt the wiring before the voltage would get that low. That is totally internal to the CA, so I can't "fix the signal". Peak speeds are often impossible. Peak currents that seem ridiculous, though it is harder to be certain they are wrong, but with the massive capacitor bank in the controller these currents are probably physically impossible as well. Occasionally I see extremely high regen currents. I know how hard it can be to cleanse readings and get good values. These signals are inherently noisy, as is any ebike. ADCs are noisy. The data must be processed carefully, you cannot trust every conversion. The software can't fix everything, true, but it should do the best it can, is it doing that now?

This noise may be causing other problems in the data processing, such as driving the feedback loops or occasionally mistriggering features. Hard to be certain. Cleaning it up may have other benefits to CA operation.

These are my observations. It doesn't happen all the time. But we should review these readings and see if we can give useful feedback on it. I just checked the Borg and the values today look reasonable. Peak current is 85 amps, a little high but hard to know for certain, though this is a high end sinewave controller and they tend to not have the big current peaks beyond settings, which is 80A in this case. Speed 38, not very likely but possible on a big downhill. So I don't have specific extreme examples to report today.

Ride Safe.

 

[IdleUp]

I'm sure this questions has been answered but this 100+ page is difficult to search.

I have a modern version Bafang 750 and a few guys told me only Version 1 would work with my motor is this correct?

Thanks in advance!

Mike

 

[amberwolf]

I have a modern version Bafang 750 and a few guys told me only Version 1 would work with my motor is this correct?

Version 1 of what?

And why would only that version work?

 

[IdleUp]

I have a modern version Bafang 750 and a few guys told me only Version 1 would work with my motor is this correct?

Version 1 of what?

And why would only that version work?

Cycle Analyst of course - I can't answer the question - That's why I'm asking . . . .

 

[teklektik]

I have a modern version Bafang 750 and a few guys told me only Version 1 would work with my motor is this correct?

I think you mean CA version 2.4 not version 1.

Short form:
No, the Bafang runs with a V2 fairly easily, but a V3 can be operated in V2 mode and so 'works' as well.

Long form:
The meaning of 'work' can be slightly elusive...

V2 - V3 differences
Although it has a basic Current Throttle mode which requires some custom wiring, the V2 is fundamentally a monitoring and limiting device. It gives battery stats and can limit current and speed as well as providing configurable LVC and a true current-based 3-position switch (unlike the usual control PWM '3-speed' switch). On the other hand, the V3 is fundamentally a control device that provides all the basic V2 functions plus throttle/controller voltage matching (no dead zones), throttle ramping, temperature power rollback, autocruise, pedal assist, Current/Power Throttle, and integrated ebrake support. It also throws in support for lifetime statics for two battery packs. The V3 is a smart device that works best when driving a dumb controller so all the features are integrated within the CA itself.


Conflicts with the Bafang
Since the V2 is doing basic monitoring and limiting, it lives happily with the Bafang which is large unaware of the CA.

The issue with a V3 and the Bafang is that the Bafang has a smart controller that provides pedal assist, gear shift detection, etc. that is integrated with throttle operation. This subsumes some of the V3 functionality and getting them to play nice together in that arena is not a plug and play affair. There are threads on this here on ES (Kepler) and a custom plug and play solution is available commercially from at least one source.

Dumb Down the V3
That said - the V3 can always be operated in V2 mode which will give you basic V2 operation plus temperature rollback and multiple battery support. If you run the V3 in 'normal' (not legacy) mode you can pick up the throttle/controller voltage matching and throttle ramping - here you are steering away from the pedal assist and fancier throttle modes that can conflict with the Bafang (i.e. you run the throttle in PassThru mode). In either case, the effectiveness of the temp rollback feature depends on where the thermistor is attached - ideally it's epoxied to the windings, but attached to the motor case can also be workable.

In the case of either a V2 or V3, the Bafang does not have a CA_DP connector with CA power, throttle, and shunt connections, so some bit of tinkering is required - this is not a plug-and-play installation. The easiest way to go after this is explained in "Appendix D. Adding a CA-DP Connector to a Generic Controller" of the V3 Guide. That solution works for either CA model. I recommend looking at the Bafang conversion/integration threads on ES as well.

 

[teklektik]

Question; is this unit plug and play with my Crystalyte controller?

I've answered this in response to your PM, but for the rest of the readers: it's unclear. There are more than one version of the CA interface even though they have the same connector. The Guide describes a simple process to determine which type you have as the first step of the installation procedure. Once the type is determined, there are instructions for connecting the V3 to your controller accordingly.

and can you suggest any presets with respect to limiting the ramp on the throttle?

This is sort of like asking "What's the best flavor sundae?". The CA comes with default settings that are workable for bikes of 'average loaded weight' with low to moderate power' (so - "hot fudge"). Beyond that, the whole point of the CA is to allow you to tune performance to your vehicle, terrain, and personal taste.

Follow the basic setup and try reducing the ThrO->UpRate. Once basic setup is complete, if this doesn't do what you need, consider getting a 3-postion switch so you can have different current/power ranges for different applications (e.g. bike path, road, boost). You can also try Current Throttle as the surest way to smooth throttle operation, but this requires some adjustment of the PLim->AGain. Current Throttle can result in a less responsive throttle - this is fine for most applications but technical riders may find it objectionable.

 

[teklektik]

On my Cycle Analysts, all versions, the peak readings are so frequently bogus that I have learned never to trust them. The peak reading logic needs to have better filtering against bad ADC readings. Peak speed, current, power, minimum voltage all seem to exhibit this to one degree or another. Has this been reported, addressed, improved or even discussed? The values are often nonphysical, such as an impossible peak speed.

Perhaps we should all start looking at the peak/min values and report on anomalies. I can't be the only one experiencing this, in fact the most recent reports I have are from others for extremely high battery current peaks that I just don't know if we can trust or not.

The software can't fix everything, true, but it should do the best it can, is it doing that now?

All CAs share a great deal of common sampling code although the V3 has better handling of speed signals. ADC measurements undergo averaging and there are no cases where a result is derived from a single ADC sample. So, there is some basis for observing similarities in behavior across the family, however it is unlikely to be attributable to the simple cause you have suggested.

Although there have been reports of aberrant speed readings and attendant power loss because of speed limiting violations, I don't believe we have any posted reports of these other voltage and current anomalies. That said, we have had examples of odd behavior on 'big current' bikes due to electrical noise from long signal wire runs and/or poor wire routing or where the CA was housed inside the metal body of the bike with controller and phase wiring. In those cases, adding a cap to the signal wires remedied the issue.

I would not be surprised to see an occasional high max current reading on big DD installations that have Xie Chang type controllers (or similar) that make use of 'block time'. During this period, the programmable controller current limiting is 'blocked' and the controller is essentially running in unlimited mode. The CA current limiting will override this, but it might be possible for an initial very high current spike to sneak through before the CA can squelch it. This is just a WAG that goes primarily to the point that at least some popular controllers are normally configured to allow extremely high current in a fairly 'hidden' manner.

In any case, as always, the feedback is appreciated and I will open a ticket on the matter. If you have a particular recurring issue, we might look into instrumenting some code for you to flash to grab some data from these transient episodes.

 

[Alan B]

Thanks. That is good to hear. I will look for specific examples which might be useful.

While it is true that during block time the limiting is suspended, this typically occurs at zero speed and very low back EMF, and so while the output phase current can go really high, the actual battery current drawn to produce this is generally not all that high. Peak battery current is generally in the midrange and higher. But it is a good point to keep in mind.

 

[amberwolf]

I have a modern version Bafang 750 and a few guys told me only Version 1 would work with my motor is this correct?

Version 1 of what?

And why would only that version work?

Cycle Analyst of course - I can't answer the question - That's why I'm asking . . . .

A CA v1 would be rather hard to find nowadays, so strange that they would require it.


Any version of a CA would work as a monitoring device, and perhaps with slight modification of "CA connector" wiring, and/or using external shunt, any version would work as a limiting device as well, with any controller, AFAIK.


As to why, that is something you have to ask the people that told you that (I'd assumed they had also told you why). Only they can answer it. I asked you so you can ask them if they hadn't already told you.