diy solar

diy solar

Accuracy of Overkill BMS SOC

rootusrootus

New Member
Joined
Jan 17, 2021
Messages
55
Is there a consensus on the accuracy of the Overkill BMS state of charge? I am trying to decide if there is any value at all to getting a Victron SmartShunt so I can see with bluetooth the state of the battery. The BMS has bluetooth, also, but is the SOC measurement accurate enough? I gather from some Internet searching that even the Overkill guy (Steve? He's on here, right?) isn't 100% sure how the BMS is tracking SOC and whether or not it's coulomb counting.
 
Mine came in a 100 AH pre-built battery I bought ..... so, im not 100% sure if it's the same as what Overkill sells, but does use the same App.

I had to calibrate the current on mine to get it to be accurate. Zero, charge, and load.

The other thing to be aware of is that until the battery goes thru a full charge cycle it displays SOC based on a voltage guestimate .... then it tracks based on current in and current out. It works very well for me since I calibrated the current.
Also be aware that writing any parameter change to the BMS resets it back to a voltage guestimate til you go thru full charge cycle again.
 
Is there a consensus on the accuracy of the Overkill BMS state of charge? I am trying to decide if there is any value at all to getting a Victron SmartShunt so I can see with bluetooth the state of the battery. The BMS has bluetooth, also, but is the SOC measurement accurate enough? I gather from some Internet searching that even the Overkill guy (Steve? He's on here, right?) isn't 100% sure how the BMS is tracking SOC and whether or not it's coulomb counting.
The BMS does coulomb counting. Assuming both are set up correctly, I believe the Victron is more accurate, but I don't think by very much. It does take a couple cycles for the BMS to calibrate and settle in. At first it was way off, but is now within 1% of the Victron.

I think the real advantage of the Victron is with the other features. It will communicate with Victron MPPT controllers, drive relays, and with the BMV-712 instead of the smart shunt, it has a display so you don't need to rely on a smartphone app.
 
I chuckle because I'm relying on the Victron smartphone app so I don't need a display. lol
Yeah, I actually went cheap at first, and picked up the AiLi battery monitor. Then I realized that I'd rather have it be bluetooth, not a dinky display somewhere. Figured I could do the BMV-712, but for my purposes I really don't care much about the display, so might well get the SmartShunt. My only remaining concern, based on reviews, is if the BMV-712 maintains historical information where the SmartShunt does not.
 
Yeah, I actually went cheap at first, and picked up the AiLi battery monitor. Then I realized that I'd rather have it be bluetooth, not a dinky display somewhere. Figured I could do the BMV-712, but for my purposes I really don't care much about the display, so might well get the SmartShunt. My only remaining concern, based on reviews, is if the BMV-712 maintains historical information where the SmartShunt does not.
Just looking at the manual, it seems the smart shunt has the same history data available.
 
I've been using the Overkill BMS (JBD) with a victron BMV-712 and find that the Overkill BMS SOC reading feels very unreliable to me. It seems to do a combination of per cell voltage in brackets combined rough coulomb counting. For example, the BMS has configuration for arbitrary 20% 40% 60% 80% voltage thresholds, which is not really a high resolution way to determine SOC. hope this helps.

BMV-712 is great shunt and has high resolution of current and seems to count discharged amp hours from a full charge really well.

edit: haven't done the current calibration zeroing yet. that certainly should help the BMS SOC estimation!
 
combined rough coulomb counting.
There is no mechanism in the OverKill BMS to count coulombs. It is as you describe in the next sentence, up to the user to determine what SoC to display for the different whole battery voltage. It is only as "arbitrary" as the data and effort the user puts into determining the SoC voltages.
 
I am running an Iceco 45 qt compressor/cooler off of my 280 amp hour battery I built. I have been monitoring the amp usage for the last 2.5 days. The BMS (overkill) last night said I had 229 amp hours left. The Iceco has been using approx .8 amps an hour until the last 12 hour period, this morning I went out and read the numbers and they had dropped to 183! The BMS claims I used 46 amp hrs over that period. Obviously, I did not use any more amps than previously but the Overkill somehow came up with the wild claim I used 46 amps in 12 hours. The same items have been running the whole time, the cooler and two LED lights on low setting. Does anyone have an idea why it would show these numbers?
 
There is no mechanism in the OverKill BMS to count coulombs. It is as you describe in the next sentence, up to the user to determine what SoC to display for the different whole battery voltage. It is only as "arbitrary" as the data and effort the user puts into determining the SoC voltages.
I have proved for myself that the Overkill does count coulombs.
 
I am running an Iceco 45 qt compressor/cooler off of my 280 amp hour battery I built. I have been monitoring the amp usage for the last 2.5 days. The BMS (overkill) last night said I had 229 amp hours left. The Iceco has been using approx .8 amps an hour until the last 12 hour period, this morning I went out and read the numbers and they had dropped to 183! The BMS claims I used 46 amp hrs over that period. Obviously, I did not use any more amps than previously but the Overkill somehow came up with the wild claim I used 46 amps in 12 hours. The same items have been running the whole time, the cooler and two LED lights on low setting. Does anyone have an idea why it would show these numbers?
The coulomb counting in the Overkill works ok, but isn't great. My guess is that over a few days it wasn't accurate, and it reset to what it thought was more accurate. Or, maybe you used more power than you think.

I also have a Victron 712, and after a week, the BMS is 10% different than the Victron. When I recharge, the battery reaches full charge and charging stops just as the Victron meter reaches 100%. The BMS is about 90% at that time, but will suddenly jump to 100% when it realizes that it is full.
 
here is no mechanism in the OverKill BMS to count coulombs
Are you sure on this, with my system it tracks the Victron BMV within 5% most of the time regarding capacity. At this very moment in my RV the displays are showing, Victron BMV, 96.5%, BMS, 94%.
There are odd occasions where the BMS seems to lose the plot and the capacity reading is incorrect. This is resolved by charging or resetting capacity via the app. If it's not counting coulombs its best guess is quite acceptable.

Mike
 
I have proved for myself that the Overkill does count coulombs.
While it does not exactly count coulombs, it does try to keep track. (see explanation from Steve from Overkill)

I found this tidbit on the inter web (cleaned up some formatting):
>>
Steve from Overkill Solar3 months ago

Long story short, we don't know. It's a mystery buried in Chinese firmware.
https://www.reddit.com/user/erskie/
3 months ago

I am not sure how it works, but I believe it is based on an ongoing tally of amps, resetting whenever the battery is recharged to the cutoff point.
In terms of accuracy .... generally it is ok, but the longer the time period between full charges the less accurate it will be. I have 2 24V batteries set up in parallel with identical cells, and identical BMSs. In addition to the data from the BMS there is also a Victron Shunt measuring the current to and from the batteries. Generally the two BMSs agree with each other (within 5%) while the Victron is further off. However, after a three week period in which no full charge occurred, not only were the BMSs far away from the Victron shunt, but they significantly disagreed with each other. Although the voltages of the two batteries were within a few hundredths.
Since then we had a sunny day, everything charged all the way up, and all the devices agree with each other again.....
Hope that helps?
 
@MisterSandals thank you for more light shone on this topic.

In my experience, JBD BMS does not register small loads of <0.5A reliably, eg i see motor spinning and only connected to battery but xiaoxiang ios app updating in real time shows 0W

BMV-712 registers very fine resolution even below 0.5A which led me to trust it more.

This reminds me. The JBD BMS and the BMV-712 indicate slightly different voltages. I ought to ensure to set the “full battery voltage” parameter such that they represent the same real voltage. This might mean setting one to 14.1V on one and 14.13 on the other or something. This would ensure that both begin counting down amp hours from the same moment/point. Probably best to use a trusted Digital Multimeter to establish the “true voltage” for such an exercise.
 
@MisterSandals thank you for more light shone on this topic.
This is all new to me, i only knew it did not actually count coulombs (requires Hall sensor or shunt). I "thought" it merely mapped user input voltages to show SoC. I wonder how the user values are used in conjunction with how the BMS calculates usage based on amps and time. If its a mystery to Steve then thats good enough for me to stop as I'm sure he put a lot of thought into this already.

I have 2 of these and have not RTFM yet. My greenhouse is pretty autonomous so
the BMS, once setup is just a battery protector. From time to time i glance at a LED
voltage display (cheap aliexpress unit) to make sure its between 13.0v (getting low) and 13.4v (pretty full)...solar charger set to stop at 13.8v and it settles tp 13.4v quickly.
 
This is all new to me, i only knew it did not actually count coulombs (requires Hall sensor or shunt). I "thought" it merely mapped user input voltages to show SoC. I wonder how the user values are used in conjunction with how the BMS calculates usage based on amps and time. If its a mystery to Steve then thats good enough for me to stop as I'm sure he put a lot of thought into this already.

I have 2 of these and have not RTFM yet. My greenhouse is pretty autonomous so
the BMS, once setup is just a battery protector. From time to time i glance at a LED
voltage display (cheap aliexpress unit) to make sure its between 13.0v (getting low) and 13.4v (pretty full)...solar charger set to stop at 13.8v and it settles tp 13.4v quickly.
There is a "shunt" integral to the BMS. That is how it is able to know the current going in or out. In my testing (partly by entering nonsense data into the voltage map) the BMS most certainly *does* count coulombs. It does some other stuff, too, which I don't fully understand. I believe the voltage numbers are used as a check to prevent the counter from getting too far out of sync, but I'm not sure. I also think the BMS can update those numbers on it's own as it learns the actual capacity of the battery if you do a few 100% discharge cycles.
 
@MisterSandals thank you for more light shone on this topic.

In my experience, JBD BMS does not register small loads of <0.5A reliably, eg i see motor spinning and only connected to battery but xiaoxiang ios app updating in real time shows 0W

BMV-712 registers very fine resolution even below 0.5A which led me to trust it more.

This reminds me. The JBD BMS and the BMV-712 indicate slightly different voltages. I ought to ensure to set the “full battery voltage” parameter such that they represent the same real voltage. This might mean setting one to 14.1V on one and 14.13 on the other or something. This would ensure that both begin counting down amp hours from the same moment/point. Probably best to use a trusted Digital Multimeter to establish the “true voltage” for such an exercise.
My 712 reads exactly the same voltage and current as the BMS. Perhaps that is just my good luck, but the BMS voltage and current meter can be calibrated. And you can "zero" the shunt on the 712. I have not needed to do any of that though, if you mess it up, it could be problematic.
 
I can only speak to the BMS I have which came in the small 100 AH prebuilt battery I ordered .... I'm pretty sure it's the JBD (Overkill) BMS since it uses the same app.
When the BMS is initially powered up .... or any time a parameter setting is changed... the BMS calculates a guesstimated SOC based on the pack voltage. This guesstimate is derived from the cell 20-80% Cap. Voltage parameters ... and these parameters can be adjusted.
Once one cell reaches the cell full voltage, the BMS resets to 100% SOC and starts tracking via charge and discharge current ... which is coulomb counting.

I don't know if there is an internal shunt .... or if it is hall effect.

I had to calibrate the zero, charge, and discharge current thru the config menu in order to get the tracking to be accurate, and as someone else stated, it's not good at detecting small current draws and accuracy can suffer because of that if it sits with small loads for a period of time .... but ..... once fully charged again, it resets ... and now that mine is calibrated, it tracks pretty good. I just use that pack to power my boat electronics and the SOC tracking is working very will for that application.

Other shunt based BMS do the same thing .... some just do it better than others. I know my Chargery BMS, which is shunt based works pretty much the same way and has the same issue with not measuring small loads accurately. They have improved that, but it still isn't as good as it should be ... maybe in the new BMS if they ever get around to releasing it.

The Victron seems to be the gold standard for SOC tracking .... but my round Chinese knock off that looks a lot like the Victron works very good also and measures very small loads.
 
Last edited:
The coulomb counting in the Overkill works ok, but isn't great. My guess is that over a few days it wasn't accurate, and it reset to what it thought was more accurate. Or, maybe you used more power than you think.
I have cycled my cells several times over several days and the ah's hardly changes. It's always within a couple of ah's but this is with low C rates. The built in coulomb counter might not be the most accurate but it is consistent in my case.

Based on my testing it might be off by a few ah's and this was compared to a DROK I have which uses a hall effect sensor. I don't have a Victron or anything that would be more accurate and I don't need it. The BMS is good enough for me and comparing the voltage to the ah's used I believe it is fairly accurate.

As was pointed out, changing any of the parameters will throw off the SOC and coulomb counting and the battery must be fully cycled in order to get accurate readings.
 
Back
Top