Cycle Analyst V3 preview and first beta release

[NeilP]

I've got another question for you. Why does the MPH indicator blink sometimes? Is it because I didn't have the Speed Throttle turned off? I've now turned it off following these instructions"

If Speed Throttle and maximum speed limit enforcement are not required, disable the speed control logic:set SLim->MaxSpeed to the maximum value, IntSGain = 1, PSGain = 0, DSGain = 0.

You should set IntSGain=100 as per the correction in the Errata document.

Regard the blinking: What NeilP said.
Raise the setting on SLim->MaxSpeed.

...but I am having a job getting my head around how it works too from just the manual.

Hmm - okay. I'll have at look at making it clearer in the manual. Meanwhile:

• Bottom line: SLIM->MaxSpeed sets a target limit for speed limiting. Setting this to a very high value will work to 'disable' speed limiting for most moderately powered bikes although the PID controller software (as controlled by the three Gain adjustments) is still active - it just never reaches the max speed so it seems '

• 
  
  I am working from manual only here at moment. Confined to barracks with arm in plaster, due operation. I don't have powered up V3 here on a bike to play settings with, only going from manual. But what I was looking for after reading the section above was two different settings.
  They being:
  a setting for MaxSpeed
  And
  a setting for SLim
  
  
  Maybe there is, and I just have not dug deep enough in both menu & levels.
  On an epic to try and create a nLite Slipstreamed Win XP + SP3 + all further updates at moment. A bloody 2 x 14 hr struggle so far!

 

[sulanino]

In particular - the symptom you have sounds like a bad ThrI->FaultVolt configuration which would be indicated by a flashing throttle bar graph on the Main screen.

First - this was a bad analysis since this symptom would only show up at WOT, not all the time... oops!

Thank You.

I will work on it tomorrow. Its 8 PM and i have other things that needs my attention.
Put my kid to bed, dishes, laundry...
Hmmmm... Why did i get married?

Good evening teklektik and others watching!

Finally! Flashed to the newest prelim.

It took some time. Had to order a cable, and buy a new house, so time suddenly went fast.

Now the bike can run again. So somehow the CA did something wrong and needed to be "reset".

Regards, Lasse.

 

[NeilP]

It took some time. Had to order a cable, and buy a new house, so time suddenly went fast.

buying a new house just to keep the new CA and e-bike dry..now that is taking things to the extreme

 

[rowbiker]

I'm posting this in the hopes of getting responses from ES users who have experienced similar issues as I have. Based on the reading I've done of many previous posts dealing with the new CA, my hardware is most likely fine, and the "fix" will probably be to reflash the firmware on the CA. At any rate, this a copy of the email I recently sent to ebikes.ca and Ed Lyen:
---------------
Justin, et al.

Thanks for all of the excellent hardware and service supplied to our project. We've been impressed with the quality of all items from you and Ed Lyen, and the wealth of information available via the folks on endlessphere.com I probably missed something in following directions that were out there, but I'm now stuck at this point and don't want to break anything while trying to "fix" it. Hence, my cry for help from you. I'm cc'ing Ed Lyen on this in the event that he might have an insight to what I did wrong.


Issue/problem:
CA v3 Prelm6 Cal->RShunt value always resets to 0.000mOhm after being hand edited and saved

Hardware:
CycleAnalyst Ver3 DP, firmware Prlm6 (purchased from ebikes.ca, Sep 2013)
LiMn 13s 48V battery pack (purchased from ebikes.ca, Sep 2013)
12 FET 3077 MOSFET Extreme Modder Controller LYEN Edition (purchased from Lyen, Sep 2013)
Crystalyte HT3525 rear hub motor (purch. from Lyen, Sep 2013) in 20" spoked wheel (laced by us)

Connections:
eBrakes (2) input to CA only, via factory connector (modified connectors on brake side wires)
throttle (1/2 twist, hall effect) input to CA only, via factory connector
CA to controller via standard JST-SM 6 conductor wire
no jumpers or crossover wires installed

History:
Bike was tested with controller, battery and motor (w/o CA) yielding expected results. Brake and throttle inputs went directly to controller, and worked as expected (smooth response from throttle, high torque, expected RPM's, quiet running, no heat). Battery was fully charged, and resting voltage of the pack was at 52.56VDC, as received from ebikes.ca When CA was first plugged into system, the CA booted to display "Low V", and the throttle, when connected via the CA, would not activate the motor. I (mistakenly?) assumed that the CA would display the "live" voltage from the battery, and would default to a reasonable low voltage cutoff value.

Using the two buttons on the CA, I began programming the CA. I entered the battery chemistry (LiMn) and cell count (13s), and the CA correctly calculated it to be a nominal 48V battery pack. However, the "Low V" still displayed. Thinking it might affect this, I tried to reprogram the Cal->RShunt value from 0.000 to 1.333mOhm, which is what Lyen recommends for this particular controller. Unfortunately, each time (tried this more than a dozen times) I changed the value, it would return to 0.000.

To get past the "Low V" display, which prevented the motor from being turned on, I changed the Cal->V Scale setting from the default of "00.00 V/V" to to "52.55 V/V" (actual battery voltage). This indeed eliminated the "Low V" display, but resulted in battery voltage reading on CA now being 89V, and even 91.6 after the battery's first full charge after doing an 8 mile test run. I left the Cal->Zero Amps setting unchanged at "2.50V 2.50V".

While the system "works" in that the throttle and brakes now make the bike go and stop, the voltage display is obviously incorrect, and the current/amps display shows all zero ("0.00") values, probably due to the incorrect RShunt setting. My speedometer settings also work perfectly, although they don't appear to be "remembered" by the CA at the end of a trip.

From my studies of the posts on endlessphere.com, which have been extremely helpful, it would appear that I should reflash the CA's firmware. To prepare for that, we got the CA programming cable from ebikes.ca, and I have downloaded the correct software, including the latest? (Prelim 6) hex file. I have *not* yet changed the firmware on the CA, as delivered from ebikes.ca in Sept 2013. I do not want to run the risk of "bricking" the CA, if that is possible -- as it is with some other devices whose firmware can be reflashed.

Please advise what I should do, given the above circumstances. Thanks in advance!

HJ Schmidt
Debora Slee
The Tringa Group
St Paul, MN USA
---------------

 

[NeilP]

I am no CA V3 expert not having mine connected to a bike yet..but re flashing..

I did mine the other day, just because I wanted to..have had it for over a year and bike still not built.

Unlike a controller that needs to be powered off and caps discharged to re upload data, the CA needs power...I had none...except a 9v battery, a 7.2 volt battery from a camera, a length of scrap wire and some wire strippers.

powered up the CA with just the batteries and twisted wire around the terminals of the batteries and flashed it up..no issues. So I reckon firmware flash is pretty stable.

just make sure you check the site again for latest flash software/drivers etc. I'd boot computer from cold boot just to be safe.

 

[rowbiker]

Thanks, Neil, that's reassuring info. Is there an "authortitative" location we can go to which is guaranteed to have the latest firmware code and update tool app available?

 

[NeilP]

links on page one of this massive thread

http://www.ebikes.ca/store/store_CAV3.php

 

[teklektik]

Looks like you need to reflash for sure - your EEPROM is corrupted.

What's wrong?
The newest p6 firmware won't edit a Calibration value (e.g. RShunt) if it can't understand it. Your stored value for RShunt is wonky so the firmware won't let you fix it.

How do we fix this?
The newest downloader allows the firmware to be flashed while retaining (undisturbed) the existing Calibration parameters - including RShunt. The new 3.0p6 firmware is just such firmware and lacks default Calibration values because it thinks they will already be there. If we just flash p6, the wonky values will remain and things still won't work properly. But - 3.0B19 has workable defaults so we flash it just to get the Calibration values in place, add the newest code over it (3.0p6), and we are ready to go....

Please download these three items:

• Grin Tech V3 Page: the newest downloader app
• Grin Tech V3 Page: the older 3.0B19 firmware
• Unofficial Guide page: the newest 3.0 PRELIM (3.0p6) firmware

The newest firmware is always available via a link on the same page as the Unofficial Guide. Subscribe to that page to be notified of updates.

1. Install the downloader app.
2. Flash 3.0B19
3. Flash 3.0p6 immediately on top of the B19 flash
4. Check the RShunt value - it should be defaulted to 1mOhm and you should be able to edit properly. If not, then something else is awry.

Normally, you just flash over what is there with the new firmware, but this case is special because the data is corrupt so we do the double-flash trick.

BTW - The 3.0B19 flash should put a workable default in place for your Cal->VScale value as well (about 31.05V/V) so your voltage readings should be okay.

 

[rowbiker]

teklektik -- have you ever thought of writing user manuals, product documentation, and the like...? Oh wait -- you already have!

I will follow your instructions to the letter, and report back the result. My gratitude knows no bounds. Also, thanks!

 

[Snickers]

Just a question to the CA expert...
I have a CA V3 B19 connected to a Thun torque sensor.

I try to have assistance for starting from 0 speed by pushing on the left pedale. I read the CA document but I doesn't find the parameter to have only torque assistance. In every case, It seems we need to reach minimal speed before the assistance will start. :?
The interrest of a torque sensor are proportionnal assistance 8) and the help for starting at 0 speed.
About the help for starting at 0 speed, is the function missed or I have to update my CA?

 

[teklektik]

I have a CA V3 B19 connected to a Thun torque sensor.

B19 is very old. You really need to update to 3.0p6.

I ran B19 for very long time using the throttle - I had no PAS at the time. It worked well and was very stable. But - all the PAS functionality including Thun support has been substantially changed and improved since B19.

However, even with the newest firmware, assist will not start immediately. You may be able to get it going in a revolution if you are in a low enough gear to start. It does not start just on torque (pedal pressure) alone - it must see the crank turning at 'some' configurable speed. So, you need to get the pulses occurring fast enough to 'wake up' the assist.

 

[Snickers]

Thanks, unfortunately this confirmed my meaning. I can't understand that Justin forget this function. Is there a incompatibility with other function ?
I ride a recumbent bike and sometime it's impossible for me to start due to the hill if I don't anticipate by changing to the right gear. With ebike we use generally the higher gear ratio and I use the motor at starting without changing the gear ratio or smal change...
On my bike, there is no twist throttle because I use essentially on the street and bike lane. In Germany, they are not so much tolerant with people don't respect the rules…

There no possibility to starting at 0 only by pushing the pedal currently with current V3 version. Majority of ebikes with torque sensor have this function. I'm little bit disappointed...
The alternative to the starting from 0 in legal mode could be a button with a limited speed at 6 km/h without pedaling.
This is authorized by the law and also many pedelec bikes have this function.

It's seems to be possible:
"Set SLim->MaxSpeed to limit the maximum assisted speed. The throttle is enabled without pedaling at speeds up to PAS->MxThrotSpd, while at higher speeds pedaling is required to enable the throttle; setPAS->MxThrotSpd to zero to always require pedaling or to a very high speed to never require pedaling. SLim->MaxSpeed in conjunction with PAS->MxThrotSpd can provide compliance with a variety of pedalec legal requirements."

This is not my prefered solution but I don't need to punch the button for every start.

 

[teklektik]

Thanks, unfortunately this confirmed my meaning. I can't understand that Justin forget this function. Is there a incompatibility with other function ?

It was not forgotten, it was eliminated on purpose so that pedaling is required to eliminate false starts from just putting a foot on the pedal. It also allows the WGain to be set for gentle engagement - if the PAS pedaling detection is not used, the assist tapers off according to WGain instead of being shut off fairly quickly when pedaling stops - there would be a sort of conflict between a 'slow' WGain setting for getaway and a 'fast' WGain setting for stopping. This is sort of an 'implementation reason' and something might be done to change this behavior, but....

You might send an email to Justin about your issue. I don't think there will be much more tinkering with 3.0 before the production release, but at least your voice would be heard for consideration.

 

[teklektik]

I see that a big issue is the 'button-push' - and I also see that you have a 'scope in your profile so I'm guessing you know your way around circuit fabrication....

This is a more difficult means to achieve what you want, but maybe it would be appealing as a project:

You could add a little external circuit powered from the throttle or PAS 5v to send in PAS pulses to the CA instead of sending the Thun PAS pulses directly. To avoid having a button, the unit would need to be 'stateful' - but the idea is simple. You send a train of PAS pulses so the CA thinks you are pedaling - as soon as you apply torque the CA immediately applies power. From that point, there are some options, here's one: When the circuit detects THUN PAS pulses it sends those instead of the artificial pulses so the CA is working normally. (Remember that the minimum cadence the CA uses for power calculations is 55rpm, so this switch-over should appear seamless). This will allow the CA to shut down assist fairly quickly as soon as the real PAS pulses stop. Your circuit might then wait until there was 3-5 seconds of NO-THUN-PAS pulses before it started once again generating it's own pulses - re-arming the torque-only getaway mode.

This could be done with a tiny Arduino or a some 555 timers as was done in this ebike simulator.

This is certainly not a 'small' workaround, but perhaps others are interested in the functionality - and if you are looking for a winter project, well....

 

[rowbiker]

teklektic/et al.
The "Double Flash" as defined above by teklektic was a perfect fix for my issues with the CA v3 settings. The only adjustment to your excellent instructions I would make is that your "Unofficial Guide" link is actually a better 'one stop shop' because the grin web page showing "CA V3.0 Prelim Hex File April 2013 <-- Use This One " link actually still points to the #5 prelim, NOT the #6. I suspect that Justin or his web staffers will eventually update this link, but in the meantime users should just be careful to actually look at the version number in the file name to verify they are using the 'latest' hex file.

 

[rowbiker]

while on the subject of links, the "ebike simulator" link above in teklektic's post gave me a 404/not found result.

 

[teklektik]

..., the "ebike simulator" link above in teklektic's post gave me a 404/not found result.

Oops - Fixed!

...the grin web page showing "CA V3.0 Prelim Hex File April 2013 <-- Use This One " link actually still points to the #5 prelim, NOT the #6.

Very good catch!
Tinkered the directions to download the 3.0pxx firmware via the Guide page link and pinged Justin to repair the Grin Tech V3 page.
Thanks!

Glad your issue was resolved

 

[wyvernwaddell]

Hi folks!!
I'm having a minor problem with the Prelim6 software..The flash was fine, settings are good (I wrote down all the old ones first this time) The bike rides fine..but the battery icon doesn't show discharge if I have amps showing on the main display. If I go into setup menu-preferences-main display and switch to watts, the battery icon shows drain. If I switch it to amps, it doesn't show drain. Is this a known thing? The software updater verified the flash as good-so I'm confuzzled...I even reflashed it-same results.
Help?

 

[Architectonic]

Thanks, unfortunately this confirmed my meaning. I can't understand that Justin forget this function. Is there a incompatibility with other function ?

It was not forgotten, it was eliminated on purpose so that pedaling is required to eliminate false starts from just putting a foot on the pedal.

How much pedalling is actually required though, some commercial ebikes use half a crank to get started and I think we should have this option available (if it isn't already an option).

I have been considering the torque sensor option too. But I want a fairly quick response...

 

[izeman]

Thanks, unfortunately this confirmed my meaning. I can't understand that Justin forget this function. Is there a incompatibility with other function ?

It was not forgotten, it was eliminated on purpose so that pedaling is required to eliminate false starts from just putting a foot on the pedal.

How much pedalling is actually required though, some commercial ebikes use half a crank to get started and I think we should have this option available (if it isn't already an option).

I have been considering the torque sensor option too. But I want a fairly quick response...

i don't use a torque sensor, and it wouldn't work well with a mid drive setup, but when i wanted to buy one for my other ride i thought that starting the motor JUST by pedal pressure is the absolut benefit of the sensor. with every torque sensor equipped commercial bike it's like this. put a foot on the pedal and off it goes! no need to turn the crank half a turn to start the motor. launching uphill is sometimes impossible. if the rider was strong to easily launch uphill he probably wouldn't need the motor.
if you don't want the bike to launch with your foot on the pedal - pull the brake and cut motor's power. it's as easy as this.

 

[Snickers]

@teklektik, Good idea, it’s possible to developed but I brought a CA V3 and a Thun torque sensor to remove a home made PAS electronic card I made… 8)

i don't use a torque sensor, and it wouldn't work well with a mid drive setup, but when i wanted to buy one for my other ride i thought that starting the motor JUST by pedal pressure is the absolut benefit of the sensor. with every torque sensor equipped commercial bike it's like this. put a foot on the pedal and off it goes! no need to turn the crank half a turn to start the motor. launching uphill is sometimes impossible. if the rider was strong to easily launch uphill he probably wouldn't need the motor.
if you don't want the bike to launch with your foot on the pedal - pull the brake and cut motor's power. it's as easy as this.

+1, you resume exactly the problem. I also use the brake cut off when I stop and keep one foot on the pedal…
I have a recumbent bike and in uphill, I cannot launch the bike few meters with my foot like you will perform with standard bike.
Picture from my bike to understand the situation I could be to start sometime in uphill. To prevent starting just by touching the pedal, It has to be a minimal value to be reach like minimal torque>15N.m? and a minimal time >100ms? before the power will come.
I will contact Justin...

 

[teklektik]

@teklektik, Good idea, it’s possible to developed but I brought a CA V3 and a Thun torque sensor to remove a home made PAS electronic card I made… 8)

Yep - well, that was not trivial workaround but I thought I'd throw it out there - some folks just love these little 'science project' undertakings.....

+1, you resume exactly the problem. I also use the brake cut off when I stop and keep one foot on the pedal…

I think you can see the safety issue that caused this mode to be omitted, but it's always good to talk with Justin about this stuff - sometimes alternate solutions can come to light.

However, if you are happy with your ebrake procedure above, you can still do the external circuit and just continuously run in fixed frequency PAS pulses to activate the CA 'pedal detector' - this is much simpler - it would only take a 555 and a two or three parts - smaller than a pack of gum. This would give you the starting behavior you wish, but would of course, disable the the rpm-scaling. However, with torque-scaling still active, this may prove entirely workable. I'm not pushing for this at all - just pointing out a much simpler workaround that might be effective and could be fabbed pretty quickly.

I actually use a related hack for TRQ-PAS support: I have a PAS wheel and use a pot to adjust the 'virtual torque' in place of the Thun sensor. Similarly, in your case, if you don't have AuxPot free for PAS Level (maybe using it for 3-speed or 'preset' switch), you could put a pot on the multivibrator to adjust the 'virtual cadence' to get variable assist. The CA scales assist in the min/max range 55rpm to 120rpm.

Anyhow - some hacking options for the tinkerer....

BTW - very cool bike!

 

[teklektik]

How much pedalling is actually required though, some commercial ebikes use half a crank to get started and I think we should have this option available (if it isn't already an option).

I have been considering the torque sensor option too. But I want a fairly quick response...

Justin has spoken of starting in the neighborhood of half a revolution, but my bike is adjusted somewhat differently so I can't verify this. Other members may be able to step up. Behavior is dependent on other bike/rider/technique attributes in addition to the settings, so firm start times can be a little tough to pin down.

The Thun is a bit pricey and if you have strict requirements, you might consider picking up one of the plug-and-play PAS wheels (PAS_12P) from Grin Tech as a first step. I mounts without even pulling a crank and plugs into the CA directly. Since the CA 'pedaling detector' is tied to PAS pulses not torque, this would allow you to verify the behavior is appropriate before the major Thun commitment. A PAS wheel can also be set tp to scale with rpm (AutoTorqPAS) which is a sort Thun-wannabe strategy, but it works fairly well.

 

[wyvernwaddell]

Hi folks!!
I'm having a minor problem with the Prelim6 software..The flash was fine, settings are good (I wrote down all the old ones first this time) The bike rides fine..but the battery icon doesn't show discharge if I have amps showing on the main display. If I go into setup menu-preferences-main display and switch to watts, the battery icon shows drain. If I switch it to amps, it doesn't show drain. Is this a known thing? The software updater verified the flash as good-so I'm confuzzled...I even reflashed it-same results.
Help?

Ok, since I've fiddled with this a bit more, I have more information. The battery is an Ebay deal 48 volt 20 amp hour battery, LiFe chemistry. I have it programmed correctly in the CA. I ride the bike around and the volt reading on the CA shows correct voltage drop but the "gas gauge" shows full all the time. I tried swapping the battery out with another of the same type, size and chemistry with a known voltage level (reads about half full on my other CAV3 equipped bike) and the gas gauge on this CA reads and stays full.
At this point it doesn't matter if I have the main display showing watts or amps-the gas gauge reads full all the time regardless.
Is there something that could be wrong with the hardware to cause this? It worked fine on the prelim5 software. This didn't happen until I flashed the prelim6 software.
I flashed it back to B22, set it up and checked it and the issue persists.
I tried connecting the other battery up to my CA and it read correctly there,
So I think something must have happened to the CA, however minor, since it's tested as showing a full "gas gauge" for two different batteries, one of which is known to be at half charge.
Anyone have any ideas?

 

[teklektik]

I ran some tests last night and this morning and could not duplicate the Watts/Amps difference - but it seems that symptom was really not in play - so, okay on that count.

The Gas Gauge is a software artifact so if your CA is showing the proper voltage, there is no hardware failure that could reasonably be causing this (I don't think...). Since it persists across a reflash, it sounds like an EEPROM or configuration issue.

I know the State of Charge gets written to EEPROM on power down, so I was wondering if you had some hinky value stored away that is giving it problems. But - the SOC value should be blitzed on a re-flash since it's not in the calibration parameter section of memory - so that sort of makes the EEPROM idea invalid.

But - the Gas Gauge uses a kind of interesting algorithm that combines historical Ah and open circuit voltage to track the SOC under all circumstances. It actually switches modes and looks at the open circuit voltage when the bike has no (little) current draw since that's the only time the voltage is close to 'open circuit'. It then looks up the OC voltage in a table (by chemistry) and determines the SOC. It generally takes a bit of time when it's making adjustments to the accumulated Ah reading (which is error-prone) to correct it to the more accurate SOC derived from the OC voltage.

• So - I''m wondering if you need to zero your amps so the CA can get a proper opportunity to do the OC voltage measure/correction to the SOC. If you have a non-zero 'idle current' it may not be doing the gauge update because it thinks the voltage reading will not really reflect OC.
  Just a guess - try the Zero Amps trick.

Failing that - maybe some more info:
Do the voltage and Ah read correctly?
What is the computed RBatt after a ride?
What are your battery config settings?