Internal BMS and Victron SmartBMS - not matching

[archjeb]

Hi folks.

I have a trolling motor that is 36v. I have 3 SOC 100Ah batteries wired in series. I have a Victron SmartShunt on the Neg terminal near the battery.

When I look at the Bluetooth interface between the two management interfaces, there is a discrepancy in the actual power usage.

The batteries were capacity tested...so this 114Ah Capacity is legitimate. But...I'm using 100Ah for all my calculations, including what I set the size of battery for my Smart Shut.

The SOK shows 46% SoC. So I'm calculating that I used 61.56Ah of the batteries capacity. Assuming 46% left of 114Ah as this BMS is using.



But when I look at the Victron's information, it is showing that I consumed 78.6Ah. The State of Charge %...that is based on a 10% of extra headroom I gave myself. Basically, I set the Smartshut config so that my 0% discharge floor is really set at 15%. I wanted extra so I didn't run the battery down to 0 when on the water. So that percentage on the State of Charge on the Victron...you could add 15% for a full discharge to compare to the SOK's BMS.



At any rate, the issue I have here is with the CONSUMED Ah numbers.
Victron is saying 78.6Ah. While from the SOK's BMS perspective, it looks to be around 61Ah. Granted, the SOK BMS we have a little interpolation here...because I'm calculating this based on the 114Ah capacity of the battery (remember I load tested this to confirm that is legitimate).


When I used a load tester (one of those cheap off Amazon 150w testers) over the entire battery bank...there was a slight discrepancy between the two. But as the current draw increases it seems like the numbers diverge even more so. I get more Ah drift. Also, the current load (real time) between the two also gets bigger (greater delta) as the load increases.

Any thoughts on why such a discrepancy? Who do I really trust? The Victron or the SOK BMS here?


-J

 

[Ampster]

Any thoughts on why such a discrepancy? Who do I really trust? The Victron or the SOK BMS here?

I had difficulty reconciling the two values with the different sources and the distinct settings. My suggestion is to try to calibrate them by charging the pack to 100% and comparing the Ahrs consumed, then charge and compare value for Ahr charged. The value you give to pack size should not affect Ahr readings. I would disregard the percentages until you are sure the inputs are matched. In addition they may be using different methods to calculate percentage. The current measurements should be close unless the current measuring devices are in different spots on the circuit and picking up additional loads. I don't know any reason that the charging values would be different.
Incidentally, from what little I know of the Victon device is that it is just a shunt with Coulomb counting capability but not a BMS.

 

[archjeb]

I'll recharge and set the Victron discharge floor to 0% - to make things a little more comparative. But I notice that even when I put a load on the battery pack....the instantaneous Current draw numbers of the SOK BMS and the Victron are off. The higher the current draw...and the greater the delta. Seem weird to me. They should be the same...they are in series.

Maybe, when I'm back out on the water...I can run the boat motor at full throttle (which is about 50Amps of current draw) and use a clamp on Ammeter to see what the actual current draw is and compare that with the Victron and with the SOK BMS to see which one is off?

 

[Ampster]

But I notice that even when I put a load on the battery pack....the instantaneous Current draw numbers of the SOK BMS and the Victron are off. The higher the current draw...and the greater the delta. Seem weird to me. They should be the same...they are in series.

It could be the SOK batteries in series. Any input from SOK about how they work in series. I would start with the assumption that the Victron Shunt is probably more accurate.

Maybe, when I'm back out on the water...I can run the boat motor at full throttle (which is about 50Amps of current draw) and use a clamp on Ammeter to see what the actual current draw is and compare that with the Victron and with the SOK BMS to see which one is off?

Sounds like a plan. In that use case, the important issue is range or knowing how much of available energy you have consumed so you can plan enough to get back to the dock.

 

[archjeb]

Any input from SOK about how they work in series

I haven't asked them directly...other than, they support up to 4 in series...

 

[archjeb]

I think I'm going to have to call SoK on this. I did another test yesterday. I used the trolling motor enough to get a reasonable test here for discharge.

If I put the clamp on Ammeter, both the internal BMS and the Smartshunt are pretty close. They have a delta - maybe a half an amp when looking at the continuous power draw at steady-state. What I mean by steady-state, is having the motor set to a constant speed so its drawing mostly a steady current draw.

The State of Charge for the SOK BMS just doesn't add up. Even if the 'Calculated Capacity' is higher than 100Ah (and a bench capacity test confirmed I was getting the stated capacity)...the numbers still don't add up. The SOK BMS is basically telling me I used only around 20Ah's while the SmartShunt is showing significantly more consumed.

Really weird. They show the same current draw in 'realtime'; but the SoC calculations are way off.

 

[RCinFLA]

First picture says Smartshunt has reference battery capacity set to 78.6 AH / 0.82 = 95.85 AH
Second picture says SmartShunt has reference battery capacity set to 42.9 AH / 0.44 = 97.5 AH.

These two numbers match reasonably as well as can be expected.

On SOK BMS readout, if there has been zero current load or charge on battery for at least 3 to 5 minutes, and cells are balanced with cell voltage readout reasonably accurate, having a rested open circuit cell voltage of 3.294 vdc is not 83% SoC. That OCV for cell is about 40% state of charge.

Another issue with SOK readout is if total battery voltage is 13.19v then average of four series cells is 13.19v / 4 = 3.2975v which is greater than the listed max and min cell voltages. Still in the 45% state of charge range, definitely not 83%.

Don't believe the SOK BMS readout.



All Columb counters need to reset their count periodically by a full charge to clear out accumulated error build up. The criteria for 'full' is different for different battery monitors and BMS's. It should jump to 100% SoC when it triggers a Columb counter 'full' reset.

The SoC also needs the correct total battery capacity AH amount.

 

[archjeb]

On SOK BMS readout, if there has been zero current load or charge on battery for at least 3 to 5 minutes, and cells are balanced with cell voltage readout reasonably accurate, having a rested open circuit cell voltage of 3.294 vdc is not 83% SoC.

So, this is the OCV after 24hours with no load (other than the micro load of the Victron SmartBMS inline). Everything else is completely disconnected for 24hours.
So that is indeed the resting voltage.


This begs the question: Why is the SOK BMS so far off?

When I did a resistive load test (capacity test), they seemed to be pretty accurate and in agreement. But now with 'real world use' they diverged significantly.

Maybe because the trolling motor is using PWM and the current is actually more variable in nature?

-J

 

[RCinFLA]

So, this is the OCV after 24hours with no load (other than the micro load of the Victron SmartBMS inline). Everything else is completely disconnected for 24hours.
So that is indeed the resting voltage.


This begs the question: Why is the SOK BMS so far off?

When I did a resistive load test (capacity test), they seemed to be pretty accurate and in agreement. But now with 'real world use' they diverged significantly.

Maybe because the trolling motor is using PWM and the current is actually more variable in nature?

-J

You need several inputs for SoC Columb counter to be accurate.

1) Charging and discharging current measurements must be accurate. Most BMS's perform a number of series samples running average so PWM motor should not affect average current reading.

2) There must be a reference point for actual SoC of battery. This is usually a full charge condition representing 100% SoC.

3) The BMS or battery monitor must have a value for battery total AH capacity.

This is either manually setup by user or an 'automatic' determination of battery capacity is done. The 'automatic' usually means an initial rough estimate based on zero cell current voltage reading. It uses this estimate until a total discharge depletion of battery occurs, followed by a full charge, to get an actual AH measurement.

In this case, I would guess the automatic battery AH capacity number is off. You should make sure cells are balanced, then totally deplete battery, followed by a full recharge to 100%. This will install a fresh battery AH capacity number.

 

[archjeb]

Ya, SOK suggested, when I first bought the battery, to do several full discharge/charge cycles so that the BMS could calibrate the SoC correctly.
On one of the other batteries in series, I cycled it a complete 5 cycles. Its showing the same thing...

 

[RCinFLA]

You should not need to cycle more than once.

The balance condition should be evaluated at full absorb charge voltage when all cells get above 3.50v and charge current tapers off to low level.
You may need to hold absorb voltage for a couple of hours to balance the cells.

If one cell rises to 3.65v and trips BMS cell overvoltage, it means cells are misbalanced. BMS will bleed down the overvoltage cell in a few minutes and start charging again.

Since it is hard to predict method BMS resets Columb counter, just make sure it shows 100% when at full absorb when charge current tapers off.

 

[archjeb]

What is really weird, is that when I do a full discharge with a steady state resistive load, the Victron & the internal BMS is pretty close.
i.e. when using one of those adjustable load tester off Amazon.

But in the real world application with the trolling motor...that is when the two diverge.

Strange...

If the 'Columb Counter' is off on the internal BMS; I'd imagine it would have been off on both the 150w resistive load and the trolling motor. But its not; its just with the trolling motor.