question

bons avatar image
bons asked

Smart Shunt underestimates remaining charge

I have a Smart Shunt and a paralleled pair of 300 ah LiFePo batteries with Bluetooth BMS built in. I often run a load of about 100 amps with about 50 amps simultaneously charging from solar. After 6-8 hours of operation my Smart Shunt estimates that I have about 10% charge remaining but my Bluetooth BMS indicates that I have about 30% remaining. At the start of each day I'm certain that I have 99-100% charge and both the Smart Shunt and BMS agree. I have my Peukert factor set to 1.05 and the amp hour settings are correct @ 600 ah. I'm trying to figure out why there is such a large disagreement. A 5% error is reasonable, I guess, but 20% seems too large.

SmartShunt
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

2 Answers
Alexandra avatar image
Alexandra answered ·

@BonS

Check your discharge floor setting. It will count own to that, not to empty, unless you set it.

2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

bons avatar image
bons answered ·

My discharge floor is set to 10%. When I reach the floor my battery says I still have about 30% remaining. It seems to me that the Smart Shunt is in error. I haven't set the Zero current calibration yet, so I'll do that and see, but it's hard to imagine that a little offset will create such a large discreppancy at high current draws of over 1kw.

7 comments
2 |3000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Alexandra avatar image Alexandra ♦ commented ·

@BonS

With a disagreement of two items, it is hard to know who is correct. Does the bms have a shunt in it? How does it calculate soc?

A zero current calibration may help.

It most likely you will need to change the peukerts a bit. Not all lithiums are 1.05

0 Likes 0 ·
bons avatar image bons Alexandra ♦ commented ·

I presume that the each battery BMS does, but the manufacturer, LifeBlue, doesn't say how they arrive at SOC. I've always trusted the battery for their SOC as I didn't have an external shunt. I will do the zero current calibration. After the next heavy cycle, I will also measure the zero-load battery voltage to get another estimate of their SOC. So, my Fluke, will be the 3rd party arbitrator and should point the finger at the source of the error.

0 Likes 0 ·
Alexandra avatar image Alexandra ♦ bons commented ·

@BonS

Voltage and SOC are very much not the same thing, so a Fluke will not help there.

https://www.youtube.com/watch?v=H2c_suhosOI Andy from The Off Grid Garage does a neat set of videos on battery bank stuff.

At a rough guestimation though. 100A drain with 50amps of charging over 6 hours is 300ah. (50x6=300) So you battery bank would have been depleted closer to the 10% than the 30% in any case, even if you have a good C rated battery. I am more inclined to believe the Smart shunt here, just based on your example figures and good old maths.

0 Likes 0 ·
bons avatar image bons Alexandra ♦ commented ·
I don't think that the Smart Shunt just looks at voltage for its SOC estimate. Its setup info doesn't even include the type of battery that it's monitoring. Rather, it integrates current flow over time to create its estimate for the consumed amp-hours. All setup really knows is the amp-hours of my batteries and when they're estimated to be 100% charged. From there it integrates current flow in & out of the battery to estimate SOC. That's why the my meter voltage is the truth teller. If the Smart Shunt is in error, then I can fudge the amp hour rating of my batteries or play with the Peukert factor but that's a non-linear adjustment and not very likely to give satisfactory results under all load conditions such as adjusting the amp-hour rating. BUT - I really shouldn't have to do this if the Smart Shunt resistance is within specification. There's an old saying to never put two thermostats in the same room as they'll never agree with one another.
0 Likes 0 ·
kevgermany avatar image kevgermany ♦♦ bons commented ·
You're right about the smart shunt calculation. What Alexandra is saying is that the terminal voltage isn't an accurate indicator of state of charge. If you do use this as an estimate, allow the batteries to rest for half an hour before measuring. I assume you have an accurate discharge curve for the batteries at your load. Which C rating are you using in your calcs?
0 Likes 0 ·
bons avatar image bons kevgermany ♦♦ commented ·
Thanks for your response. I am sampling my batteries at rest, without load or charging, and in a steady state. I do have good discharge voltage curves for my LiFePo batteries. I'm not using C ratings or calculations as I am doing my comparisons with the batteries fully at rest. My Peukert factor is so small that it can't account for such a large discrepancy based upon my discharge rate of around 1.8 C. With all that said, I won't be in a situation where I can replicate my service conditions until early April and even then my testing is a bit weather dependent. I'll be cleaning and checking tightness of all my connections between now and then.
0 Likes 0 ·
bons avatar image bons kevgermany ♦♦ commented ·
With regard to my load, I estimate that I was drawing about 50-55 amps from the batteries while the rest of the load is sourced by the PV's. Of course the PV current is variable but their average was around this amount on that day (I have 1440 watts of PV panels.) So 50 amps for 8 hours is 66% of my 600 ah available so my remaining ah should be around 30%.
0 Likes 0 ·