Strange power loss with Sabvoton 96120 and QS273 (SOLVED)

[Santacruz]

Hi all,
I am hoping I can get some help with the following problem.
I have a strange power loss problem with my set up. Apart from this, it is running perfectly and doing what it should.

The problem I am having:-
When fresh off the charger 90.2v (22s @ 4,1v per cell), I get maximum power. Drawing around 100amps I have about a 6v voltage sag.
Once the unloaded voltage gets to around 81v, then on riding, it is as if a limiter has been switched on. Max draw is about 30amps. Voltage sag is still around 6 – 8 v. Speed does not change, only acceleration.
It can be cured by charging back up to 90.2.
So it is a rinse and repeat cycle.
But I don’t understand what is causing it.
I would have thought I would still have good power draw down to my cut off settings.
My low voltage cut off setting in the Sabvoton is 66v, on the cycle analyst it is 68v and on my smart BMS is 60.2v. (2 separate stages to control power, before 3rd one by BMS cuts off completely)
Battery balance is near perfect, temperature are good. After running controller is cold, motor is cold, batteries are cold (Temp sense at 4 points).
Cell balance at charge / 0 – 0.002v
Cell balance at 81v / 0.013v

Set up:-
Battery, 22s5p 26650 5000mah cells. (Parallel connections 10mm x 0.2 nickel strip x 2)
(Series connections 5 x 2 of 10mm x 0.2 nickel and 4 x 10mm x 3mm nickel plated copper bars).
300amp smart BMS
400 amp shunt
Awg 6 silicone power wires.
Sabvoton 96120 controller.
Cycle analyst CA3 (set up to give a legal pedal only setting and 2 other settings for power) PAS works perfect on this set up.
Motor QS274 40H.

Runs smooth and quiet. Acceleration is very good. I get about 25km from fresh charge until the limiting appears. Even then, everything runs nice and smooth, the only thing noticeable is the lack of acceleration power.
Being in Netherlands, everything is on flat roads / cycle paths of tarmac.

 

[Jonno]

Sounds like your LoVgain is set too high on your CA - try lowering it.
Section 5.0 in the manual.

Jonno

 

[Santacruz]

Sounds like your LoVgain is set too high on your CA - try lowering it.
Section 5.0 in the manual.

Jonno

Thanks for the answer.
Not too sure if it will solve my problem.

On my CA, the low voltage feed back gain is set to 400.
However, the problem happens when I am at 81v and with sag, the lowest I have run is about 75v. That is nowhere near the low voltage cutoff I have the CA or controller set to.

 

[amberwolf]

You'd have to check the Sabvoton thread(s), but there is a setting in there that I don't know the name of that rolls back power when you drop below a certain voltage. Both the voltage and the amount of rollback are things you can change.

Sorry I don't know what it is called. It's a wierd name.

I found one thread for an MQCON/Risunmotor version that uses an app on a phone, and in that app is a setting called "current-limiting voltage", which is probably the one you need to change.
https://endless-sphere.com/forums/viewtopic.php?f=2&t=108997&hilit=sabvoton

 

[Santacruz]

Thank you amberwolf, I will look into it.
If I discover anything, I will post my findings here.
At the moment, as long as I keep the battery pack charged, there is no problem.
It is more of an inconvenience not being able to use a larger range of my battery pack, (that is at normal power levels).
With the power drop in the 80v area, it is still rideable, just not as (insane?) as with full power and at nearly 66yo, it's probably my only real fun outside of designing and building

 

[Santacruz]

Update, I have bypassed the CA3 and it has made no difference.
I see this limiting kicks in around 81v -82v.
At the moment the bike is sitting around 80v and I get about 25 - 30 amps on full throttle, yet it still gets up to speed.

For now, I plan to run the battery pack down low (66v ish) to see if anything else happens. (Plus I can get an idea of range with my riding style).
When I have done that I will check the Sabvoton software and see if anything shows up (Maybe an error code or something).
Also to note, I am using the Sabvoton PC version MQCON V2.1 software. (It is the only one that has the setting for throttle ebrake and motor temp), if I use the android app, it turns off the regen, so now I just use the PC.
I went through all the parameters last night and cannot see any settings that would cause the problem I have. Basically, everything is set to max of the controller (locked version from QSmotor) and my power control is set through the CA.
When I next put the program on I will post screen shots of the settings.

 

[amberwolf]

So there's no setting called "current limiting voltage" or anything like that?

If there is, it needs to be set down near your LVC. If it's set near your full voltage, then it'll cut power back pretty quickly as the voltage drops to that level.

 

[Santacruz]

So there's no setting called "current limiting voltage" or anything like that?

If there is, it needs to be set down near your LVC. If it's set near your full voltage, then it'll cut power back pretty quickly as the voltage drops to that level.

No,
On the MQCON software I have, are the following settings:-
lack volt which I set to 60v,
Over volt which is set to 110v (To cover brake regen on full charge of 90.2v).
DC limit current set to 120a (controller max),
Boost DC current set to 120a
Rated phase current set to 120a
Max phase current set to 300a
Protected phase current set 400a
Basically all the max settings of the controller, as I use the CA3 to control the power how I want.

Now when charged, everything behaves as it should. On hard throttle I have pulled about 110a according to the CA3. When I am in the 81v area, then the limit to something like 30a starts.
I did think it might be temperature roll back from the controller getting hot and while it was stone cold to touch, that doesn't mean it is not hot inside at the time. But leaving the bike over night and using it next morning it is still limiting. Charge it up and it's fine again.
I have it in the back of my mind that maybe there is a setting left over from the android app when I first set everything up. As I see on some posts that the app has the current limiting voltage and as I have said I stopped using the app, because it did not have the throttle brake regen setting.
So I plan to run the battery pack low to guesstimate my range capabilities, then charge up and first connect the android app to have a look at what settings it shows, to see if indeed there is a voltage current limit setting that is not on the PC software.
It may just be a case that I have to use 2 different software setups (android and PC) to get the settings correct.
I will post my results when that is done.

 

[Santacruz]

I did another 25km run, average where I could 60kph, drawing 23amps. So, still slow to accelerate, but gets up to speed okay.
Down to 79v, so did program check first with MQCON app and then with MQCON pc software for comparison.
The app indeed shows current limiting voltage, whereas the PC software shows DC current limit in amps.
Apart from the app not showing ebrake throttle connection, everything else has the same readings and sadly the current limiting voltage is 2v above the minimum voltage setting, so it's not that.
Here I will show the app and pc screen shots. Maybe I am missing something obvious.

 

[Santacruz]

Also. 2 more pictures. PC software reading with bike on full throttle and one of the BMS showing how well the cells are balanced before I start charging.

 

[Santacruz]

Update.
I took the plunge and did a factory restore. Which worked fine.
I have reprogrammed with some minor alterations.
One thing I have done, is disabled the motor temp, as I noticed in the live running, the temp goes all over the place (I think due to electrical noise inside the motor) and only settles to a fixed reading when the motor has stopped.
So I wonder if this could be something to do with limiting cutting in. Maybe at 80 - 81v the noise is worse.
After the test I hope to know more.

New settings have smoothed the power delivery, power is still good, just delivered smoother.
Unfortunately, on hitting 81v, the power dropped off. Most I could pull was 26amps, Stopping and giving full throttle from standstill - 26amps.
So, my idea of noise on the motor temp sensor is not the answer.
Now I am at a loss. I cannot even work out where the 26amps comes from. I have no settings for that.

Does anyone have an idea what is causing this?

 

[amberwolf]

The only thing I can think of is that the controller's actual settings, as it responds to them, are *not* the same as what the setup program "thinks" they are, either because of a bug in either the firmware or the setup program, or a mismatch between versions (whcih you probably have no way of determining).

So that say the setup shows you 62v for the CLV, but in reality inside the controller, the value placed there by the setup program is read by the firmware as say, 82v. To test this, you could try setting the CLV to 42v instead, and see what happens. (or even lower)

If it does make a difference, then it could be that only that setting is mismatched between the FW and setup, but others could be too, and the only way to determine whihc ones are is by experimentation (which has the risk of damaging something).

 

[Santacruz]

Thank you Amberwolf.
That is a good idea.
I still have the BMS to protect while I test the controller with low cutoff settings and if that proves correct, for future use, I can use the CA3 as a low voltage rollback before the BMS shuts power off.
I will give that a try and post my results.

 

[amberwolf]

Thank you Amberwolf.
That is a good idea.
I still have the BMS to protect while I test the controller with low cutoff settings and if that proves correct, for future use, I can use the CA3 as a low voltage rollback before the BMS shuts power off.
I will give that a try and post my results.

It shouldn't affect the LVC (meaning, the voltage at which the contorller stops trying at all). Just the CLV, meaning, the voltage at which it lowers current (to minimize voltage sag as the cells get close to empty)

And if it works like I speculated, then it *will* still roll back power under that condtion, at the correct voltage, just that you have to experimentally determine what actual value to put in there, rather than it being what it says it is.

 

[Santacruz]

I have changed the low voltage to 40v and the current limiting voltage to 42v.
After testing with the battery pack at 82v I have regained full power. No more limiting.
Some fairly heavy pulls (limited a bit by the weather) from 81v, I got 94.5amps (not full throttle).
Voltage sag to 66v.
I consider this solved and have marked this post as such.
For the battery, I will control the low voltage cutoff and the low voltage current roll back with the CA3 and this will be set just above the BMS cut off.
So I still have 3 safety settings.
I don't ride daily giving it everything it has got, these power runs were for testing. But this means I have more range before needing to charge, which suits me fine.
Thank you to Amberwolf for the suggestion. It proved to be the answer. (Also to someone else I know who had the same suggestion).

 

[amberwolf]

Glad it worked--it is certainly an oddball bug in their software. :/