Amphours remaining jumps around a lot?

[malkierian]

So I recently replaced my AGM battery bank with a larger lead acid bank. I had a Wbjr attached to one of my classics for about 8 months before that. Amphours remaining was always steady during that time. However, now that the lead acid bank is in place, suddenly I will get random jumps in Ah remaining to the tune of 10% of my entered capacity, such as is shown in this graph:
https://i.imgur.com/wxCUl8H.png
That percentage is calculated off of reported Ah remaining from that classic divided by the capacity I told it. Literally nothing else has changed in that time range. Amp draw has been nearly constant, voltage is still the same. I know the remaining capacity fluctuates somewhat based on the BTM readout, but BTM readings are steady in that period, too. I just can't for the life of me figure out what is going on. As you can see, during charging, it often jumps up like that, too. I have my efficiency set at like 89%, but I can't see how that would apply to this problem. The bank's capacity is 1955 Ah at 24v.

[Vic]

Remaining Capacity  will change,   as the battery is being  discharged  --  this causes a reduction in the Remaining Ah.   This value will,   and should,  increase,   as the battery is charged.

The Remaining Capacity should be calculated,   based on the Net Ah removed,   or replaced from the battery,   compensated by the battery temperature value,   as read by the Classic BTS,  plus,  the battery Charge Efficiency compensation.   This value should be reset when the Classic transitions (on its own),  from Absorb,  to Float.   At this point,   the Remaining Ah,  should be the Capacity that was used in the Setup,  but compensated by the battery temperature value,  as read from the BTS.

The graph that you included,   shows this quite well.  for a battery that is being cycled.

The real question really seems to be,   why was this Ah value not changing  very much with the previous AGM battery.   If this was the case,   then,  either that battery was not being cycled very heavily,   or,  perhaps,  there was some installation or setup issue with that installation.

You mentioned that the AGM battery was replaced by a "lead acid bank",   which is assumed to be a Flooded LA battery.   An Ah Charge Efficiency value of 89% may be a bit high,   even for a young FLA battery.   This value could have been provided by the FLA battery manufacturer (and they should know this value for their batteries).   We usually start with this value at about 80% for our FLA (Lead Antimony) batteries  ...

IMO,   FWIW,   Vic

[malkierian]

I appreciate the response, but nothing you mentioned seems to explain those two straightline jumps from below 85 to above 95. They were from one second to the next. Nothing charges like that, and no amount of temperature shift can account for what would be an instantaneous 200 Ah increase in remaining capacity while being discharged, as illustrated by the two points circled here:

https://i.imgur.com/r6NjDOm.png

Also, as far as the efficiency, yes, these are FLAs, and yes 89 would seem a bit high, but I already did the experiments with specific gravity to ensure that what it said was 100% before was actually 100%, and anything less always lagged behind. These particular FLAs have an added technology that the manufacturer said increased the efficiency and decreased internal resistance (for measuring end of charge conditions) compared to standard FLAs, and they were right.

[Vic]

Apologize for having reiterated what you already knew.

From the perspective here,  the horizontal time scale was difficult to interpret,  down to the second.  Sorry.

Have little to offer.    Good Luck,  Vic

[malkierian]

Sorry, the way I worded that response was pretty hostile. I edited it, hopefully reducing or eliminating that.

[Vic]

malkierian,

Notta problem.    All the best,   Vic

[ClassicCrazy]

What are you using to get the data from the Classic to make your graph ?
Have you tried graphing the SOC value from the Classic ?

Larry

[malkierian]

What are you using to get the data from the Classic to make your graph ?
Have you tried graphing the SOC value from the Classic ?

Larry

Apologies for the very late reply, I didn't think I was going to be getting any help. I have that python script that's floating around polling and reading from the charge controllers to get their information and publishing it to a Graphana/InfluxDB setup. That script is translating the raw data into readable values, and I'm simply grabbing the Classic's "ahRem" value that's provided by whatever infernal internal formula the Classic processes the Whizbang's data with. For the record, I'm still getting that jump in the middle of the night. Between midnight and 4 AM, when nothing is coming in and output is pretty flat, the ahRem value (again, that the Classic itself is supplying) suddenly gains 10-25 % of its maximum capacity as still available (i.e., max of 2000 amphours, one such jump went from ~1600 Ah remaining to ~1970, and then when the solar panels became energized again, they charge from that level, somehow thinking that it only had like 5% to refill). I can't find any external explanation for this jump. No change in voltage, no change in battery temperature, no sudden power output, nothing, and I can't find this formula explained anywhere, so I don't know where else I could look to find why it's doing this.

[ClassicCrazy]

Bob is probably the only one who might be able to answer this . So maybe he will see your post.

Larry

[boB]

malkierian

Do you have Auto-Restart (reset) enabled to happen at midnight ?    I wouldn't think that would be it but if it WAS, I would think the remaining Amp-Hours would go DOWN and not UP.

Wondering if this is happening at any certain time of the night or is it any randomish time between around midnight and 4 AM ?

Can you show a graph of A-Hours and WB current ?

Also, which register number or address are you reading ?  Whether it is a register number or modbus address could be important and would be "off" by one of course.

[malkierian]

Auto-restart? Meaning the whole classic restarts? Or just the AH reset? If the former, no, I don't have them restart automatically. If the latter, how do I check that?

It seems to be pretty random. Over the last 7 days, I've had it happen 3 times around midnight, twice around 2 and once at 4:30. And I looked, the week before that it didn't happen once.

Voltage, Ah remaining, and WB current: https://i.imgur.com/kq4vTKn.png

I'm not sure how to answer the last question. I haven't manually coded anything, I'm using the script mentioned here: http://midniteftp.com/forum/index.php?topic=3875.msg45306#msg45306. Although, I did go and dive into the code and find it processed the signal like this: https://pastebin.com/WueTAjbc
I'm using the "RemainingAmpHours" field at the moment.

[malkierian]

I should also mention, I have a separate battery monitor that I check quite often, and up until this point where the SOC and Ah remaining jump up, the WB SOC matches the percentage stated by this other battery monitor.

[boB]

I should also mention, I have a separate battery monitor that I check quite often, and up until this point where the SOC and Ah remaining jump up, the WB SOC matches the percentage stated by this other battery monitor.

What I think of when you mention this is that "maybe" the data is not getting to the server and then it suddenly gets reported and all of a sudden it does..

But I have not seen a graph of battery SOC so I am not sure.

[malkierian]

I should also mention, I have a separate battery monitor that I check quite often, and up until this point where the SOC and Ah remaining jump up, the WB SOC matches the percentage stated by this other battery monitor.

What I think of when you mention this is that "maybe" the data is not getting to the server and then it suddenly gets reported and all of a sudden it does..

But I have not seen a graph of battery SOC so I am not sure.

The middle graph in the image I posted IS the SOC. Literally one report to the next it goes up. Straight line.
EDIT: I was using Ah remaining and dividing because it gave a much more even, accurate representation of SOC than the classic's own SOC which rounds to the nearest whole percentage, but the classic's SOC still matches my math.

[boB]

The middle graph in the image I posted IS the SOC. Literally one report to the next it goes up. Straight line.
EDIT: I was using Ah remaining and dividing because it gave a much more even, accurate representation of SOC than the classic's own SOC which rounds to the nearest whole percentage, but the classic's SOC still matches my math.

Are you referring to the middle of the 3 graphs that is labeled "power"  ?   From 0 to 100 ?

Maybe you can post again if I'm not getting it.

If that graph I am looking at is AH-Remaining then I don't know why it would do that.

Does your other SOC meter show that ramp downward before the straight up ledge around 4 AM or so ?