SoC% false accuracy? State of Charge

Announcement - Help us fight the BOTS! Please report all spam including stuff in your inbox!

Wolf

Active member
Joined
Sep 25, 2018
Messages
1,581
OffGridInTheCity said:
Seems high but maybe its because its reflective of the overall system instead of just the 18650s in isolation.
Try this

Go into Batriums Telemetry tab and select daily history extract that to file.

image_tbzqbc.jpg


Import into excel or some sheet.
Then add the cumulative in and out do some quick math and this shouldgive you a more accurateindication.

image_wgeqvl.jpg


Wolf
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
OffGridInTheCity said:
[...] I simply recorded the SoC at the same (resting) voltage at 6:00am each morning. This is a 1560ah battery bank @ 14s lithium-ion.

Started at 8/28@6:00am 50.1v and set SoC =29.3% explicitly
On 9/1@6:00am 50.1v the SoC had risen to 33.4%.

SoCwent up by 4.1% (64ah higher out of 1560ah battery bank). During this period,125kwh where drawn from the battery bank. [...]

So you're getting about 1% drift each day from Batrium's coulomb counting fuel gauge (due to charge inefficiency, self-discharge and current integration errors). That amount of drift is not unusual for a simple coulomb counting algorithm. To counter it you can recalibrate daily by forcing the software to reset the SOC to 100% when you charge up to the max voltage in the limited range used (= "virtual battery"). Then the SOC of your virtual battery will likely never be off by much more than 1% (actually slightly more since you want to set the capacity in Batrium to the lower virtual battery capacity (Ah in used range), not the full physical battery capacity).

As the battery ages the physical capacity in your virtual range may decrease beyond what you set, in which case you'd need to update the Batrium settng in order for the virtual SOC to remain accurate (if the software doesn't do that automatically).
 

OffGridInTheCity

Active member
Joined
Dec 15, 2018
Messages
1,746
@Wolf - taking your suggestion....
My Batrium dump seems to stop at July 22, 2020 (247 records) - don't know why. Is that a max buffer in Batrium?. But regardless, what's interesting....

My overall (241 records) CumulAhDischg (199,994.976) is *higher* than my CumulAhCharge (195,957.98)! I summed a smaller set - July 1 -> 22 - because the data looks clean and its still higher on Discharge (13,286) than Charge (12,889) - 397ah difference.

I can't even think of atheory as to why my discharge ah would be > charge ah... :huh:

@Daromer
I don't think I'm mixing ah and wh. I'm noting that a kwh @ 52v = 19.2ah. Therefore, if I have a 0.5ah difference in SoC for each kwh taken out of the battery - I can do a generalized 0.5ah/19.2ah to get 2.6% over the time period.

@guass163 - I never get to 100% SoC, so I don't have a way to 'reset' the SoC under Batrium and Batrium doesn't have any other means. This makes SoC drift each day - fairly substantially... to the point that its not useful for me 'operationally' day to day.
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
OffGridInTheCity said:
@guass163 - I never get to 100% SoC, so I don't have a way to 'reset' the SoC under Batrium and Batrium doesn't have any other means. This makes SoC drift each day - fairly substantially... to the point that its not useful for me 'operationally' day to day.

You do, because here "100% SOC" refers to the SOC of your virtual battery defined by the limited range you use in your physical battery. E.g. if you charge only to 4.0V/cell and wish to use at most 123Ah from a 200Ah physical battery then your virtual battery consists of the 123Ah range below 4.0V in your physical battery. I suspect you can configure Batrium to report SOC of this virtual battery (vs. the entire physical battery), by setting the capacity to 123Ah and having it recalibrate to 100% at 4.0V at end of charge.
 

OffGridInTheCity

Active member
Joined
Dec 15, 2018
Messages
1,746
gauss163 said:
OffGridInTheCity said:
@guass163 - I never get to 100% SoC, so I don't have a way to 'reset' the SoC under Batrium and Batrium doesn't have any other means. This makes SoC drift each day - fairly substantially... to the point that its not useful for me 'operationally' day to day.

You do, because here "100% SOC" refers to the SOC of your virtual battery defined by the limited range you use in your physical battery. E.g. if you charge only to 4.0V and wish to use at most 100Ah from your physical battery then your virtual battery consists of the 100Ah range below 4.0V in your physical battery. I suspect you can configure Batrium to report SOC of this virtual battery (vs. the physical battery), by setting the capacity to 100Ah and having it reset to 100% at 4.0V.
This won't work for my off-grid situation. While I do have an operational max 56.6v(4.04v/cell) -on a regularbasis the top voltage never get's there andvaries significantly depending on PV input and consumption. For example, we had a cooler day last week and the battery reached 54.1v (3.86v/cell) but yesterday was quite hot and the battery topped out at 52.0v (3.71v). In the spring/fall, there are several days where it will go all the way to 4.04v/cell because consumption is at its lowest.

To recap -I have a fixed bottom at 49.5v (3.54v/cell) where the inverter is turned off each cycle but the top varies all over the place between 52v and 56.6v dependingon PV in and consumption out for that cycle. If Batrium had a 'reset at bottom' then I'd be in good shape.
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
OffGridInTheCity said:
[...] To recap -I have a fixed bottom at 49.5v (3.54v/cell) where the inverter is turned off each cycle but the top varies all over the place between 52v and 56.6v dependingon PV in and consumption out for that cycle. If Batrium had a 'reset at bottom' then I'd be in good shape.

Yes, it appears you need bottom vs. top recalibration. Are you sure Batrium does not offer such? Usually both are available in most fuel gauges - even the simplest ones. I'd suggest contacting their tech support to inquire about that.
 

Redpacket

Member
Joined
Feb 28, 2018
Messages
1,309
Another factor that may be involved is described in this thread:
https://secondlifestorage.com/showthread.php?tid=8482
Basically some inverters generate high levels of electrical noise which introduces significant errors into shunt measurements - which affect SoC readings....

re top & bottom SoC reset limits, the Batrium gear does have both but in OffGrids' case, not sure that would help reset the drift as he's running in the centre (ish) of SoC range?
Screen shots posted here:
https://secondlifestorage.com/showthread.php?tid=7153&pid=66982#pid66982
 

OffGridInTheCity

Active member
Joined
Dec 15, 2018
Messages
1,746
Redpacket said:
re top & bottom SoC reset limits, the Batrium gear does have both but in OffGrids' case, not sure that would help reset the drift as he's running in the centre (ish) of SoC range?
Can you tell me exactly what "Re-Calibrate at Low SoC" actually does? I can't figure it out. Let's say that at my low/cutt-off of49.5v I have an SoC of 25%. So I know the 1st cycle it will go down to at least27% at 49.5v cutt-off.....
If "Re-Calibrate at Low SoC" is30%.... (and 27% is < 30%)then what happens? what setting does it use for recalibration.... 0%?:huh:

YIKES - I just set"Re-Calibrate at Low SoC" to 30% and then forced the Metrics to 25% and it went to 0% immediately. Unfortunately - this caused the shunt to stop reporting Shunt V, A and W!!! they went to 0v, 0a, and 0w.
Doesn't look like a good solution as I do care about Shunt V, A, and W a lot more than SoC:(
 

daromer

Moderator
Joined
Oct 8, 2016
Messages
5,659
If you dont got to 0% or 100% the current built in function Will not work. The function in the bms relies on that it cannot go further so it resets. Its not about resetting att a det voltage. Its a plain counter Up and down that relies on reaching One of the endpoints often.
 

gauss163

Member
Joined
Jun 28, 2020
Messages
278
OffGridInTheCity said:
Can you tell me exactly what "Re-Calibrate at Low SoC" actually does? I can't figure it out. [,,,]

In fuel gauge terminology an SOC recalibration usually means that when you are at some known SOC you reset that gauge to this known value. In the simplest casethe SOC is "known" by mapping a (near)resting voltage to SOC (based on known V to SOC tables). This serves to eliminate any errors that have accumulated over time.

Often the "known SOC values" used are those near empty or full, but they can be any value in the used range (= "virtual battery") that is attained frequently enough at (near) rest so that the SOC can be accurately calculated from the voltage (if they are not attained frequently enough then the errors can accumulate to large values).

So in your case when you reach the empty voltage of your virtual battery (49.5v = 3.54v/cell) you want to reset the fuel gauge to 0% in order to recalibrate the SOC of your virtual battery.

As for what is actually implemented inBatrium's simple coulomb countingfuel gauge I can't say, since it is very poorly documented, and I don't have one to reverse engineer it, and not enough has been said above to infer such. If I owned one I would contacttech support to attempt to obtain better documentation. You certainly deserve such given the high price point.

To learn much more about fuel gauge algorithms (including much more accurate algorithms) see the Barsukov book linked here.
 

Redpacket

Member
Joined
Feb 28, 2018
Messages
1,309
OffGridInTheCity said:
YIKES - I just set"Re-Calibrate at Low SoC" to 30% and then forced the Metrics to 25% and it went to 0% immediately. Unfortunately - this caused the shunt to stop reporting Shunt V, A and W!!! they went to 0v, 0a, and 0w.
Doesn't look like a good solution as I do care about Shunt V, A, and W a lot more than SoC:(
That's more like a bug & not supposed to happen!

Do you have gear (eg an oscilloscope) to look for HF noise at the shunt particularly when the inverter is running?

I just did the daily history export from Batrium (for ~2.5 years data) & got
Cumulative Charge Ahrs = 141044
Cumulative Discharge Ahrs = 139688
So about 99.03% efficiency reported (edit, corrected %)
 

OffGridInTheCity

Active member
Joined
Dec 15, 2018
Messages
1,746
Thanks to you all - I think I've got the outlines of Batrium SoC operations. Operating in the middle v range / off-grid style with varying max v(s) per cycle + an unusable Batrium feature / suspect shunt situationhas rendered SoC 'not good' for me at this point.

Fortunately I have perfectly good metrics with out Batrium SoC and my voltage curves (due to such a large battery / relative steady consumption) are quite steady - so operationally I'm good. But I understand better now why many focus on SoC and I don't :)


P.S.
>Do you have gear (eg an oscilloscope) to look for HF noise at the shunt particularly when the inverter is running?
Thanks for the suggestion but no scope.
 

400bird

Member
Joined
Apr 27, 2020
Messages
97
To pick back up an old thread, sorry.
I am seeing the same issue with my barium shunt/SOC.
Offgridinthecity did you ever figure this out? Or just continue to ignore SOC?
Is your shunt connected in reverse? I wonder if somehow their coulomb counting doesn't work quite right when you change the shunt orientation. Seems odd, but it feels more likely a software issue than something external.

Seems like there is some suspicion of the inverter being at fault. I have a Schneider XW6848 pro.

I do have an oscilloscope, I'd be surprised to see anything in voltage if I graph the DC pack voltage, there's a giant battery right there working to stabilize the voltage. Maybe I need to graph DC voltage, current, and grab both sides of the shunt?
 

OffGridInTheCity

Active member
Joined
Dec 15, 2018
Messages
1,746
To pick back up an old thread, sorry.
I am seeing the same issue with my barium shunt/SOC.
Offgridinthecity did you ever figure this out? Or just continue to ignore SOC?
Is your shunt connected in reverse? I wonder if somehow their coulomb counting doesn't work quite right when you change the shunt orientation. Seems odd, but it feels more likely a software issue than something external.

Seems like there is some suspicion of the inverter being at fault. I have a Schneider XW6848 pro.

I do have an oscilloscope, I'd be surprised to see anything in voltage if I graph the DC pack voltage, there's a giant battery right there working to stabilize the voltage. Maybe I need to graph DC voltage, current, and grab both sides of the shunt?
It just doesn't work for me - the values are all over the place. I ignore it.
 
Top