Lead paint... Very thick Aluminum foil....
Budget "smart" 12V batteries all have the same cycling issue!
- Thread starter Will Prowse
- Start date
wpns
Solar Joules are catch and release
Turn OFF the output on the one that's taking all the load, and the other one should pick it up. When you turn it back ON, expect potentially large charge-one-battery-from-the-other currents till they equalize. Yes, this can trip overcurrent on one battery, starting the whole oscillation thing over again.
I just wanted to chime in on this current sharing issue as I've made some observations that I haven't seen mentioned elsewhere. (But please correct me if I've missed something or if you disagree with any of my conclusions.)
This is all based on my 12V system which consists of a mix of different "smart" batteries from WattCycle, GoKWh, PowerUrus (RoyPow), and LiTime, connected in parallel.
I've had the issue to a greater or lesser degree on all of the above batteries. So I can't speak to what might be better about other brands which I don't own. For example, none of these use a JK BMS but they ARE all fairly cheap yet decently made Chinese batteries.
Firstly, about that new "parallel mode" "fix" from WattCycle: As far as I can tell, all enabling it does is to change the "Cell OVP Release(V)" setting from 3.35V to 3.5V. (See screenshots.) That's it. To test this, I used the BMS Meta app to get to the battery settings to make note of them, then used the new version WattCycle app to turn on parallel mode. Then after going back into BMS Meta to check once more I saw the number had changed.
I'm sure the setting itself has been explained in the forum already, and by WattCycle themselves, but basically what it does is it turns off charge protection after it's been activated by cells exceeding the safe voltage limit (ie. 3.65V). They call this the "hysteresis" margin, designed to keep the charge protection up until the BMS decides it's safe to start charging once again, normally 0.3V below the upper limit. But somehow in all their wisdom, the BMS manufacturer (whoever they are?) also made it so that charge protection also hampers if not totally limits the ability to discharge at the same time. I'm not sure why as there is already a separate discharge protection function that works when the battery's SOC gets too low.
So by increasing the overvoltage release voltage to 3.5V, they're reducing the amount of time it takes to re-enable charging. But it's only a workaround, not a solution, because they haven't addressed the root problem which is that charge protection also affects discharging.
And it also reduces the effectiveness of the charge protection because 3.5V is still quite close to the 3.65V limit, especially in terms of the time it takes to come down from the limit after charging is stopped.
Related to this, I've found some cells take a very long time to come down to their resting voltage (around 3.33V) while others do it very fast. My two PowerUrus (RoyPow) batteries will go from 3.65V to 3.33V in about an hour, while all my others need at least a full day. So assuming the RoyPow BMS have a similar hysteresis of 0.3V (sorry I haven't tested that yet but they definitely have similar paralleling issues), they will do better in parallel than the other makes which are depending on the actively discharging batteries to drop so low that they start sucking large currents from the inactive ones.
Now, since we're stuck with workarounds (unless of course you send your batteries back where they came from, if you have that option): If you have direct access to BMS settings, you can change the overvoltage release directly. Then you don't need the "parallel mode" setting which is only available with WattCycle via their own app. I've done this with my GoKWh battery which uses the exact BMS as WattCycle but uses the BMS Meta app (or their own mostly same GoKWh app).
Or, as others have pointed out, you can simply avoid overcharge protection completely by using a lower controller/charger voltage (eg. lower from 14.6 to 14.2V). Voila, problemsolved worked-around.
I have to make one last point which I feel quite strongly about: I see that WattCycle has taken the brunt of negative PR for this issue, even though the problem is common to a huge number of vendors. I believe this is, unfairly, due to their having greater transparency in their BT apps and in their overall responsiveness after being called out. I can just imagine the people at LiTime patting themselves on the backs for issuing BS responses to the same issue for the past year+ and hiding all the cell voltages and settings in their app -- while their competition takes the hit and has all the accountability. Not to mention, the BMS maker gets to remain anonymous. Not sure what's behind that.
This is all based on my 12V system which consists of a mix of different "smart" batteries from WattCycle, GoKWh, PowerUrus (RoyPow), and LiTime, connected in parallel.
I've had the issue to a greater or lesser degree on all of the above batteries. So I can't speak to what might be better about other brands which I don't own. For example, none of these use a JK BMS but they ARE all fairly cheap yet decently made Chinese batteries.
Firstly, about that new "parallel mode" "fix" from WattCycle: As far as I can tell, all enabling it does is to change the "Cell OVP Release(V)" setting from 3.35V to 3.5V. (See screenshots.) That's it. To test this, I used the BMS Meta app to get to the battery settings to make note of them, then used the new version WattCycle app to turn on parallel mode. Then after going back into BMS Meta to check once more I saw the number had changed.
I'm sure the setting itself has been explained in the forum already, and by WattCycle themselves, but basically what it does is it turns off charge protection after it's been activated by cells exceeding the safe voltage limit (ie. 3.65V). They call this the "hysteresis" margin, designed to keep the charge protection up until the BMS decides it's safe to start charging once again, normally 0.3V below the upper limit. But somehow in all their wisdom, the BMS manufacturer (whoever they are?) also made it so that charge protection also hampers if not totally limits the ability to discharge at the same time. I'm not sure why as there is already a separate discharge protection function that works when the battery's SOC gets too low.
So by increasing the overvoltage release voltage to 3.5V, they're reducing the amount of time it takes to re-enable charging. But it's only a workaround, not a solution, because they haven't addressed the root problem which is that charge protection also affects discharging.
And it also reduces the effectiveness of the charge protection because 3.5V is still quite close to the 3.65V limit, especially in terms of the time it takes to come down from the limit after charging is stopped.
Related to this, I've found some cells take a very long time to come down to their resting voltage (around 3.33V) while others do it very fast. My two PowerUrus (RoyPow) batteries will go from 3.65V to 3.33V in about an hour, while all my others need at least a full day. So assuming the RoyPow BMS have a similar hysteresis of 0.3V (sorry I haven't tested that yet but they definitely have similar paralleling issues), they will do better in parallel than the other makes which are depending on the actively discharging batteries to drop so low that they start sucking large currents from the inactive ones.
Now, since we're stuck with workarounds (unless of course you send your batteries back where they came from, if you have that option): If you have direct access to BMS settings, you can change the overvoltage release directly. Then you don't need the "parallel mode" setting which is only available with WattCycle via their own app. I've done this with my GoKWh battery which uses the exact BMS as WattCycle but uses the BMS Meta app (or their own mostly same GoKWh app).
Or, as others have pointed out, you can simply avoid overcharge protection completely by using a lower controller/charger voltage (eg. lower from 14.6 to 14.2V). Voila, problem
I have to make one last point which I feel quite strongly about: I see that WattCycle has taken the brunt of negative PR for this issue, even though the problem is common to a huge number of vendors. I believe this is, unfairly, due to their having greater transparency in their BT apps and in their overall responsiveness after being called out. I can just imagine the people at LiTime patting themselves on the backs for issuing BS responses to the same issue for the past year+ and hiding all the cell voltages and settings in their app -- while their competition takes the hit and has all the accountability. Not to mention, the BMS maker gets to remain anonymous. Not sure what's behind that.
Attachments
wpns
Solar Joules are catch and release
Stop using the BMS for charge control by turning off the charge path when you hit COVP and you have actually solved the problem without any workarounds. Do you use the emergency brake on your car to come to a stop?
Good point, although an emergency brake doesn't auto-activate itself.
wpns
Solar Joules are catch and release
OK, do you just let off on the gas and let your anti collision system stop your car at stop lights in traffic? I mean it should work all the time every time, yes?
outsider
Solar Enthusiast
Having them come out and say, basically, "Y'all are stupid, this is a not a bug, it's a feature" is kind of off-putting.
wpns
Solar Joules are catch and release
Or they are the most egregious in their using crazy numbers for their default BMS settings, combined with turning off the discharge path on COVP. "Everyone else does it" has never been a reasonable excuse.
cs1234
Even a doofus can achieve high post counts.
- Joined
- May 9, 2022
- Messages
- 5,187
This one made me laugh.. just picturing it.
fuzzmunky415
Solar Wizard
Here and on their blog, they dug their own grave. Stiffed Will, called him 'unprofessional', then silence. Screw those guys.
As the proud new owner of a 100AH mini wattcyle, I noticed a different manifestation of this problem that will lead to the observed (strange?) behavior with ‘current sharing’. First, kudos to watt cycle for making it possible to open the battery with 6 screws! Thank you! I like the construction and capacity. The BMS seems to be a TDT-6056 100A type BMS with bluetooth.
To see a problem, simply connect a 1A load to the battery after charging to HVD – high voltage disconnect. Then, the battery terminal will be 0.7V or so lower than the internal battery voltage (as reported by Bluetooth app)!!
Attach charger set to 14.6V and charge until HVD (as seen on app). Then connect a 1A load to the battery.
On my battery the app will show 13.8V after a couple of minutes and the battery posts will be at 13.1V @ 1A. As voltage drops, the difference is maintained:
The battery terminal voltage is about 0.7V lower than the internal voltage! Why? (typically 0.7V -> silicon diode drop????).
A battery that does this will definitely not discharge when connected in parallel to a more normal battery until the more normal battery is significantly discharged - including an identical wattcycle.
From https://www.powersystemsdesign.com/...-management-charge-discharge-system/145/19032
Summary of problem as explained in the article:
Clearly, in the conflict state (0.7V drop), the charge mostfet seems to be off and we want it on.
I believe this is a common potential issue with almost all BMS of this type and is normally resolved in software by somehow turning the ‘charge’ mostfet on as soon as there is a discharge. I found something similar when I was changing internal settings on my Daly BMS and ended up with a bad choice of settings (not a Daly problem, my problem no knowing what I was doing….)
If a second battery, where the charge mosfet is on is connected in parallel, the output will be higher than 13.1V at low currents, and almost no current will flow from the wattcycle (or Litime or whatever is using this smart BMS).
Please note: it is the charge mosfet that should be on for discharge, in addition to the discharge mosfet.
If the discharge current is cranked up to 3A the voltage will suddenly jump by 0.7V, indicating that the charge mosfet has turned on.
Of if the cell voltage drops to around 3.33 V or so. Waiting for the drop to 3.33V can take a loooonnnnng time if there is a full parallel battery.
For my use case in a travel trailer, I absolutely need more normal current sharing (multiple 12V batteries of unknown origin in parallel - no problem until now...).
I did attempt to switch to parallel mode but nothing seemed to change. Not sure about meaning of 'comparability mode' or 'exact match mode'.
To see a problem, simply connect a 1A load to the battery after charging to HVD – high voltage disconnect. Then, the battery terminal will be 0.7V or so lower than the internal battery voltage (as reported by Bluetooth app)!!
Attach charger set to 14.6V and charge until HVD (as seen on app). Then connect a 1A load to the battery.
On my battery the app will show 13.8V after a couple of minutes and the battery posts will be at 13.1V @ 1A. As voltage drops, the difference is maintained:
The battery terminal voltage is about 0.7V lower than the internal voltage! Why? (typically 0.7V -> silicon diode drop????).
A battery that does this will definitely not discharge when connected in parallel to a more normal battery until the more normal battery is significantly discharged - including an identical wattcycle.
From https://www.powersystemsdesign.com/...-management-charge-discharge-system/145/19032
Summary of problem as explained in the article:
Clearly, in the conflict state (0.7V drop), the charge mostfet seems to be off and we want it on.
I believe this is a common potential issue with almost all BMS of this type and is normally resolved in software by somehow turning the ‘charge’ mostfet on as soon as there is a discharge. I found something similar when I was changing internal settings on my Daly BMS and ended up with a bad choice of settings (not a Daly problem, my problem no knowing what I was doing….)
If a second battery, where the charge mosfet is on is connected in parallel, the output will be higher than 13.1V at low currents, and almost no current will flow from the wattcycle (or Litime or whatever is using this smart BMS).
Please note: it is the charge mosfet that should be on for discharge, in addition to the discharge mosfet.
If the discharge current is cranked up to 3A the voltage will suddenly jump by 0.7V, indicating that the charge mosfet has turned on.
Of if the cell voltage drops to around 3.33 V or so. Waiting for the drop to 3.33V can take a loooonnnnng time if there is a full parallel battery.
For my use case in a travel trailer, I absolutely need more normal current sharing (multiple 12V batteries of unknown origin in parallel - no problem until now...).
I did attempt to switch to parallel mode but nothing seemed to change. Not sure about meaning of 'comparability mode' or 'exact match mode'.
outboard1196
I go victron victron victron, I'm a victron guy.
I have experienced this myself. Once.
outsider
Solar Enthusiast
A 0.7 V drop in a 12V battery represents a significant efficiency and power loss.
wpns
Solar Joules are catch and release
I think they fiddled with the BMS 'COVP-recovery' settings instead. Still kinda broken that their discharge path is off when they hit COVP, but that appears to be a hardware limitation, so there's no software fix.
12VoltInstalls
life passes by too quickly to not live in freedom
…and cheapo LiFePo4 12V batteries (not smart) just work fine and don’t care.
While Voltage ain’t a good state of charge indicator it’s still a useful parameter. It’s been adequate for me ~3 years now.
Sweet Tatorman
Subsistence farmer and Sweetpotato Evangelist
That 0.7V drop goes away as soon as the battery BMS senses that it is supplying power and turns the charging FETS back on. Therefore, efficiency loss is not an issue.
outboard1196
I go victron victron victron, I'm a victron guy.
Yeah I think this normal and all BMS have this quark. It's nothing to worry about.
The problem is that it does not go away unless (as far as I have been able to determine) one of 2 conditions are met.
These 2 conditions will turn on the charging indicator (light blue to dark blue).
So in my book it is just a bug they could fix (assuming off course that they can update all setting or maybe the firmware via a update rollout -it is after all a 'smart' battery)
On my wattCycle 100AH it goes away if the 'charge' button turns blue when
1. you draw a discharge current > 2.5A
OR
2. the voltage of one or all cells (not sure about exact value) drops below 3.35xx( something) Volts....
If you draw a lower current it takes a while. How long? I figure you need to discharge about 1AH(?) til you hit the number and suddenly your voltage will jump up to normal. Just make it jump up earlier -> much lower current or higher voltage! To fix the bug.
For now you can just apply 3 A discharge and the voltage "charging"
will turn on and voltage jumps up then and there.That 0.7V drop will cause nothing but grief if you have a parallel connection. Why? Because you have a race: the first battery that jumps in voltage (charging indicator enabled) will be the battery that will discharge 20+%!! If one of the batteries never hits the HVD that one wins the discharge race since it will be 0.7V higher (i.e. normal voltage)....
wpns
Solar Joules are catch and release
Don’t forget that when the second one cuts in it’ll dump a lot of current into the first one which is now at a much lower state of charge, and it might trip the overcurrent protection on one or both of them.
Sweet Tatorman
Subsistence farmer and Sweetpotato Evangelist
Yes, there is definitely a deadband around zero current where the BMS thinks it is in standby even though there is a small charge or discharge current flowing. Forum member RCinFLA has on several occasions given a good technical explanation why this is so. I will not repeat that here but if you are interested, I encourage you to search on posts by RCinFLA for discussion of the topic.
2.5A does seem like a fairly high threshold for sensing discharge but I have observed that up to an amp or so is fairly common.
Sorry for being dense, but I assume by technical explanation you mean the circuit for the output mosfets which I mentioned in my previous post. The 2.5A is high, but the 'killer' is the voltage for the HVD charge reset (the second condition). I think with more normal BMS this value is higher. I just measured the voltage to be 3.3350V (similar value that seemed to be discussed before in posts here):
Please note I am hijacking V1 to show the mosfet switch moment. Bit weird, but I needed that to hack the datalogger I had.
Why not raise the voltage say, 3.5V (like typically found with BMS) for reset and bump down the current a bit?
I can only imagine the confusion this will cause when a user asks: My app is reporting 13.6V but my battery only puts out 12.9V.....
Here is the terminal voltage graph at 1A discharge from above.
Never mind the confusion on the imbalance on parallel connected batteries.
Attachments
Sweet Tatorman
Subsistence farmer and Sweetpotato Evangelist
>but I assume by technical explanation you mean the circuit for the output mosfets which I mentioned in my previous post<
No.
Here is one of several RCinFLA posts that is relevant:
https://diysolarforum.com/threads/w...ltage-vs-soc-chart.105739/page-2#post-1454195
Yes, agree that both the 2.5A is annoyingly high and the 3.335V unnecessarily release is low. Of factory settings I've seen, 3.4V and up is typical. I have a 150 Ahr Eco-Worthy 12V I am currently testing where the release is set at 3.55V.
Last edited:
wpns
Solar Joules are catch and release
You are confusing:
1) BMS all have a deadband around zero, because they can't have infinite resolution to read tiny currents when they are reading hundreds of amps.
2) The BudgetSmart batteries disconnect the discharge path on COVP, which is the wrong thing to do, but is probably a hardware limitation of the CheapChineseBMS(TM) that they are all using, and don't reconnect it until the other parallel battery has discharged significantly. There's a partial workaround by tweaking the COVPrelease parameter (or whatever it's called), but that'll probably get you in trouble too. Really shouldn't be using COVP to signal 'charge complete', but here we are.
WorldwideDave
Solar Wizard
I’ll help you to summarize: the BMS is pure dog crap, and never run in parallel. Ya welcome.
I can’t believe it’s been almost 2 months since I started posting my issues and people are still talking about this. Just get your money back already.
I can’t believe it’s been almost 2 months since I started posting my issues and people are still talking about this. Just get your money back already.
Similar threads
