Running a house bank flat...

[ayates]

For anybody who ever wondered what it looks like to run a house bank flat The intermittent load is a fridge. Learning point is that I need to check/adjust the setting for the Victron BMV-702 low voltage alarm.


Allan.

[cr180]

I’m old school, but have always tried to keep house bank above 12vdc. Actually 12.3 if I can.

[smac999]

50% Soc at 9v means the batteries are toast and need to be replaced. They now have way less ah in them then new. Or the ah in the meter is set wrong...

You should still be at ~12v at 50%.

If you have an ah graph as well you’ll know how many ah you actually have now.

[rslifkin]

Originally Posted by smac999

50% Soc at 9v means the batteries are toast and need to be replaced. They now have way less ah in them then new. Or the ah in the meter is set wrong...

You should still be at ~12v at 50%.

If you have an ah graph as well you’ll know how many ah you actually have now.

Seconded. The batteries going flat at such a high SoC means something is definitely configured wrong or not performing up to spec. Or there's a hidden draw that's bypassing the measuring shunt.

[ayates]

Originally Posted by smac999

50% Soc at 9v means the batteries are toast and need to be replaced. They now have way less ah in them then new. Or the ah in the meter is set wrong... You should still be at ~12v at 50%. If you have an ah graph as well you’ll know how many ah you actually have now.

Originally Posted by rslifkin

Seconded. The batteries going flat at such a high SoC means something is definitely configured wrong or not performing up to spec. Or there's a hidden draw that's bypassing the measuring shunt.

Very good questions... I have always wanted to dig into the SOC accuracy, but have never done so...

This is a flooded lead acid battery bank, five years old, that is rated (new) at 740 AH.

The current draw for the entire duration averaged approximately 10A. This is well below the 20 hour rate of the battery bank.

There are definitely no loads before the shunt. The BMV was definitely set with the correct battery bank size.

We theoretically started at 100% SOC.

____Date/Time___ Volt SOC _AH Duration
---------------- ---- --- --- --------
2021-10-14 16:00 13.1 100 __0 _0 hours
2021-10-16 06:00 12.0 _73 377 38 hours
2021-10-17 21:00 _8.9 _50 705 53 hours

This would appear to imply that something is way off.


Allan.

[Dockhead]

Originally Posted by smac999

50% Soc at 9v means the batteries are toast and need to be replaced. They now have way less ah in them then new. Or the ah in the meter is set wrong...

You should still be at ~12v at 50%.

If you have an ah graph as well you’ll know how many ah you actually have now.

You can safely ignore SOC calculated by amp-counting meters.

[Dockhead]

Originally Posted by ayates

Very good questions... I have always wanted to dig into the SOC accuracy, but have never done so...

This is a flooded lead acid battery bank, five years old, that is rated (new) at 740 AH.

The current draw for the entire duration averaged approximately 10A. This is well below the 20 hour rate of the battery bank.

There are definitely no loads before the shunt. The BMV was definitely set with the correct battery bank size.

We theoretically started at 100% SOC.

____Date/Time___ Volt SOC _AH Duration
---------------- ---- --- --- --------
2021-10-14 16:00 13.1 100 __0 _0 hours
2021-10-16 06:00 12.0 _73 377 38 hours
2021-10-17 21:00 _8.9 _50 705 53 hours

This would appear to imply that something is way off.


Allan.

The batteries actually gave you 705 AH of real actual power so they are still in decent shape, or were at least before this episode.


The meter is simply not calculating SOC, but that is no surprise for simple amp-counting meters. You would be better off logging AH out vs system voltage and then make a table. 12v at light loads is roughly 50%, which is all most people need to know -- just charge before you see 12v.



Or if you want to know more precisely what kind of SOC you have, then you will want one of the SmartGauges, which analyzes voltage. Amp counting is a fools errand.

[smac999]

Originally Posted by ayates

Very good questions... I have always wanted to dig into the SOC accuracy, but have never done so...

This is a flooded lead acid battery bank, five years old, that is rated (new) at 740 AH.

The current draw for the entire duration averaged approximately 10A. This is well below the 20 hour rate of the battery bank.

There are definitely no loads before the shunt. The BMV was definitely set with
We theoretically started at 100% SOC

____Date/Time___ Volt SOC _AH Duration
---------------- ---- --- --- --------
2021-10-14 16:00 13.1 100 __0 _0 hours
2021-10-16 06:00 12.0 _73 377 38 hours
2021-10-17 21:00 _8.9 _50 705 53 hours

This would appear to imply that something is way off.


Allan.

Did you enter 1480ah instead of 740 into the meter?

The voltage and ah is pretty closly matched so the batteries are only down a little capicity.

But That graph says you are at 50% at -700ah So you have set the ah wrong.

At 50% you should be at ~370 ah Which is the lowest you should be drawing the bank.

At -700 you’d be near 0%

You got less then 700ah from the bank. (A test would shut off a bit earlier at 10.5v not down to 9). So consider setting the gauge to 680ah now instead.

[smac999]

Originally Posted by Dockhead

You can safely ignore SOC calculated by amp-counting meters.

They work great as long as the batteries are fully charged every so often so the meter can reset it self.

If constantly partial charged. Each little charge and drain will drift it wrong a little bit.

[rslifkin]

Originally Posted by Dockhead

The batteries actually gave you 705 AH of real actual power so they are still in decent shape, or were at least before this episode.


The meter is simply not calculating SOC, but that is no surprise for simple amp-counting meters. You would be better off logging AH out vs system voltage and then make a table. 12v at light loads is roughly 50%, which is all most people need to know -- just charge before you see 12v.



Or if you want to know more precisely what kind of SOC you have, then you will want one of the SmartGauges, which analyzes voltage. Amp counting is a fools errand.

Amp counting meters will never be super accurate, but if they've got appropriate correction factors (for things like Peukert effect, temperature vs capacity, etc.) and get a reset at full charge often enough, they're at least close enough to give a good idea of how much charge is left. If one is so far off as to be useless, then it's either a terrible meter, or you never reach full charge, or something else that's throwing it off.

[Dockhead]

Originally Posted by smac999

They work great as long as the batteries are fully charged every so often so the meter can reset it self.

If constantly partial charged. Each little charge and drain will drift it wrong a little bit.

They work great as long as:


1. The batteries are fully charged frequently
2. The real capacity of the batteries is equal to the nominal capacity (it's not)
3. Temperature is correctly compensated (it's not)

4. Peukert is properly compensated (it isn't)


In other words, they don't work great.

[Dockhead]

Originally Posted by rslifkin

Amp counting meters will never be super accurate, but if they've got appropriate correction factors (for things like Peukert effect, temperature vs capacity, etc.) and get a reset at full charge often enough, they're at least close enough to give a good idea of how much charge is left. If one is so far off as to be useless, then it's either a terrible meter, or you never reach full charge, or something else that's throwing it off.

In real life (as opposed to ideal lab conditions), an amp-counting meter will almost never be as close as a simple voltage reading.



I think that's good grounds to consider them useless.


A SmartGauge is a much better bet.

[Cheechako]

With no monitor, just a good voltmeter I would run to 11V or even less now and then by the AM. Lack of charging time, back to the boat late, fridge on etc.
I didn't notice any battery failure from it though.

I did have one of the earlier amp counting devices. It seems I spent more time fiddling with it than getting anything useful or reliable from it.

[rslifkin]

Originally Posted by Dockhead

In real life (as opposed to ideal lab conditions), an amp-counting meter will almost never be as close as a simple voltage reading.



I think that's good grounds to consider them useless.


A SmartGauge is a much better bet.

Voltage readings are tough in many systems. If you can't stop the load or rely on a steady draw, it's hard to get a good voltage reading.

When it comes to compensation, the Victron gear does have configurable settings for Peukert and capacity loss at lower temperatures. It's still not perfect, but once that and the charge efficiency is adjusted, you'll at least be close provided you hit full charge often enough. In my case, it's very rare to have more than 1 partial cycle before I get the batteries topped off, so the meter is at least close enough.