Batteries SOC - State Of Charge

[Philtao]

Originally Posted by goboatingnow

I would agree where you accuracy measure charge current into the battery , and have the ability to reset the amp counter , the resulting meter is more accurate then a voltage meter because voltage and SOC have a very complex relationship varying with discharge Rate , SOC , temp , age and battery type

In my case I wrote the software for the meter and the charge efficiency is modified as is the peukert exponent based on averaging the recharge amp count to estimate decreasing charge efficiency

No data yet as it’s just fitted so I’ll report back in the autumn when I have some data.

I agree that is a breakthrough SOC estimation for LA technology. I have developed something similar for Lithium, but I think it is much easier. Some of the key points to make SOC estimation reasonably accurate (and also to know when it might not be accurate):

• High accuracy current measure. shunt with an integrated auto-ranging instrumentation amplifier (like a multi meter) - the mini current I measure is 5 mA, resolution 1 mA at low current, error < 1%

• The ability to set criteria for full and empty states so that the system has fixed reference points

• energy in and out is counted as mA/s with cumulative counters

• Every time the battery is full according to the defined criteria and as long as it has been discharged to at least 70% SOC, the battery efficiency is updated (based on running average) so that accuracy increases over time (as long as there is no drastic change in usage pattern?)

• When the criteria for a full state are met, charging must stop (no float), so my system outputs a signal when the full charge criteria are met (used to command relay, or better direct control of chargers)

• SOC is displayed in a different color depending on the age of the last 100% SOC reset (same if SOC has been reset manually)

If due to weather conditions the battery does not regularly fully charge, I force (generator) a full charge to 100% about every month, and when the SOC resets to 100%, the estimated SOC is out by no more than 5%. That is good enough for me, but as a matter of habit I always correlate with the voltage - just to be sure!

The system also measures the battery effective capacity in accordance with the defined criteria for full and empty states (I am conservative in my criteria, so it is less than the battery announced capacity). Need to run a complete charge and discharge cycle when the battery is installed, then once every 6 - 12 months. The SOC is based on that measured effective capacity. The capacity fading is also integrated as one input in the estimation of the SOH (state of health).
Accurate current measure is the key!

[Sternwake]

When charging, knowing how many amps the battery is accepting indicate relative state of charge.

If Say one has a 40 amp charger, seeking 14.7v, and a healthy 100 amp hour battery.

When turned on, how long does it take 40 amps to achieve 14.7v at battery terminals.

My Northstar 103AH battery can take 35 minutes of 40 amps before reaching 14.7v

If 40 amps nearly instantly brings the battery to 14.7v, then about 24 amps are required to hold 14.7, It is generally 28 to 20 amp hours from full.

If it only requires 10 amps to hold the battery at 14.7v, it is in the 10 to 15Ah from full range.

When amps taper to 0.5 or less at 14.7v I consider it full,
but amps will taper to 0.0x if held at 14.7v with enough time
Reducing voltage to 13.6v once amps taper to 0.5@ 14.7v, it will still take 0.3 amps at 13.6v, and 5 to 10 hours later this tapers to 0.0x amps.

My previous northstar, at end of life, ~ 6 years and 1200 deep cycles, would never taper to 0.5 amps,@ 14.7v It would taper to 0.8 and then start rising, and 5 months later would taper to 1.2 and then start rising.

HOw many amps the battery accepts at a certain electrical pressure is very revealing, but more so with more experience watching what that specific requires.

There are inexpensive amp hour counters now, and some with hall effect sensors, rather than a shunt. The HE sensor can be slid over a + cable and might be way simpler than a big fat expensive shunt on the (-).

As long as one is aware of their limitations, and the counter is starting countdown from actual full charge, they can be more than good enough, and a million times better than trying to estimate state of charge by loaded voltage alone.

The cheap HE ammeters I have experience with, read very closely with other shunted wattmeters, and 2 different clamp on meters in the 0.4 to 100 amp range.

I've not yet played extensively with the amp hour counting hall effect 200 amp ammeter I purchased for under 30$

[witzgall]

Dockhead,

Chris from Balmar here. The product you refer to is the SG200.

We have tested our product extensively against a very popular amp-hour counting product currently in the market. Out of the box, both do a very good job starting with a newish battery and 100% charge, tracking SoC. This was not surprising to us, as this is just simple amp-hour counting.

Where things get interesting is when the battery bank is aged or goes through aging, and when PSOC cycles occur.

The SG200 starts with battery profile info we created in our labs - this is our "Battery Chemistry." The SmartShunt augments that data with voltage and calculated impedance information saved at points of rest. This is the "Learning" we talk about in the marketing stuff.

So when a battery has been "Learned" and PSOC cycles happen, we can correct for the errors inherent in coulomb counting, so the SoC does not experience synchronization drift.

The biggest challenge for us has been when large Lead-Acid banks are being measured and they already have significant loss of SoH when the system is installed. Getting a good SoH can take a long time. It eventually gets there, but we are working on making it faster/better.

We have made some significant changes in the chemistries and operating firmware since it was introduced over two years ago. Early on there were some use cases that we had not replicated or anticipated in the lab, and we had to work those out.

And here is the general disclaimer. We have found that many of the customers that experience issues have some sort of installation issue that gets in the way. You would be shocked at some of the installation photos sent to us. Most are well done, but then some, not so much. One that we see all of the time - the negative cable from the starter battery is connected directly to one of the house bank battery studs. If the starter battery is lead and the house Lithium, this essentially creates a hybrid battery chemistry. In speaking with our reseller and techs, it is clear that across all SoC meters, incorrect installation is a big problem across all MFGs.

Chris

Originally Posted by Dockhead

In my experience, amp-counting meters are basically useless, and besides being wildly inaccurate, they tend to err on the reckless side, so much less useful than simple voltage readings. I threw mine away and installed a Smart Gauge, which is supposed to do some kind of sophisticated analysis of voltage, but I see no evidence of that -- it always agrees with my own analysis of system voltage.


To the OP:


I think your state of charge when a charge source is connected is basically unknowable. Nor do you really care to know this accurately?


What you really want to know is when you need to charge. This is easy to know -- provided there is no charge source attached, no matter what the load, if the system voltage is above 12.10, you cannot have less than 50% SOC. If the bank is under much load, you will have more than 50%. But if you make a habit of charging any time you see system voltage under that (or better, under 12.24 for 60%), you can't go wrong. Only caveat here is you must be sure there is no surface charge -- so you cannot have been charging recently.



If you've got a charge source attached, then system voltage doesn't tell you the SOC of the batteries, but you STILL know whether you're OK or not. You can tell by system voltage whether the charging source is actually putting charge into the batteries besides just supporting loads, and you can tell whether the loads are too big for the source and your batteries are being drawn down.


So in my opinion system voltage will give you the information you need to make all critical decisions in operating a lead-acid battery bank.


For more convenience, so you don't have to use a table, you can use a Smart Gauge, but in five or six years of experience with mine, I never found it telling me anything I couldn't know with system voltage and a table.



If for some reason you really need to know how far you are getting, charging on solar, and if your solar charger doesn't give you this information, then you could use an amp-counting meter specifically for this purpose, but NOT relying on it to tell you when you need to charge.



Balmar has a new version of the Smart Gauge which I believe combines amp-counting with the original Smart Gauge principle. I haven't used it so can't comment, but maybe that's the best solution of all.

[Boatwright]

Originally Posted by goboatingnow

equalizing is great for accessible cells in a wet acid battery , its not a good idea in anything sealed

like most question and answers , it depends is often the situation

DIY is fine if you understand what you are doing. its when you exceed your knowledge that things go wrong , sadly male humans in particular tend to be over confident in their own abilities

IMO: Following the KISS principal, we have 4 Trojan 105 wet cell batteries. These are heavy duty, deep cycle units. We are careful to not discharge them below 75% and are able to do this by understanding our energy budget, using low current devices such as LED lighting, a heavily insulated top loading refer, simple nav gear, etc.. We use our inverter infrequently -- vacuum cleaner, small power tools, etc. and then for only a few minutes of on time.

Our simple system does require understanding what one is doing, checking water levels regularly, and keeping an eye on the digital volt meter through the day -- all of this is part of our ship keeping routine, and no more trouble than checking the oil in the engine.

I agree with the last post 100%. And will say one more time that installing a battery monitor won't substitute for fundamental understanding.

BTW: Our bank is now in its 6th season and we are still at 90% + of the original total capacity, as measured with a load tester at commissioning this spring. I fully expect to get 10 years out of these batteries.

[Dockhead]

Originally Posted by witzgall

Dockhead,



Chris from Balmar here. The product you refer to is the SG200.



We have tested our product extensively against a very popular amp-hour counting product currently in the market. Out of the box, both do a very good job starting with a newish battery and 100% charge, tracking SoC. This was not surprising to us, as this is just simple amp-hour counting.



Where things get interesting is when the battery bank is aged or goes through aging, and when PSOC cycles occur.



The SG200 starts with battery profile info we created in our labs - this is our "Battery Chemistry." The SmartShunt augments that data with voltage and calculated impedance information saved at points of rest. This is the "Learning" we talk about in the marketing stuff.



So when a battery has been "Learned" and PSOC cycles happen, we can correct for the errors inherent in coulomb counting, so the SoC does not experience synchronization drift.



The biggest challenge for us has been when large Lead-Acid banks are being measured and they already have significant loss of SoH when the system is installed. Getting a good SoH can take a long time. It eventually gets there, but we are working on making it faster/better.



We have made some significant changes in the chemistries and operating firmware since it was introduced over two years ago. Early on there were some use cases that we had not replicated or anticipated in the lab, and we had to work those out.



And here is the general disclaimer. We have found that many of the customers that experience issues have some sort of installation issue that gets in the way. You would be shocked at some of the installation photos sent to us. Most are well done, but then some, not so much. One that we see all of the time - the negative cable from the starter battery is connected directly to one of the house bank battery studs. If the starter battery is lead and the house Lithium, this essentially creates a hybrid battery chemistry. In speaking with our reseller and techs, it is clear that across all SoC meters, incorrect installation is a big problem across all MFGs.



Chris

Thanks for weighing in.

What you write makes sense, and adds to my conviction about the fundamental uselessness of amp-counting meters.

Your new smartgauge sounds great. I might try it myself except the way my system is configured, there is no way to get all the current through a single shunt.

And anyway I don't really need it.

[Dockhead]

Originally Posted by Boatwright

. . .



I agree with the last post 100%. And will say one more time that installing a battery monitor won't substitute for fundamental understanding.. . .

:

Amen to that!

In fact I would go further - an amp-counting meter plus lack of fundamental understand is a fast way to destroy a battery bank.

[goboatingnow]

Originally Posted by Dockhead

Amen to that!

In fact I would go further - an amp-counting meter plus lack of fundamental understand is a fast way to destroy a battery bank.

I think that comment was referring to mine

Amp counting is fine if you understand the issues

Voltage readings are fine if you understand the issues

If every case batteries can be ruined by a lack of understanding

[Dockhead]

Originally Posted by goboatingnow

I think that comment was referring to mine

Amp counting is fine if you understand the issues

Voltage readings are fine if you understand the issues

If every case batteries can be ruined by a lack of understanding

That was exactly NOT referring to your system.

Your system sounds more like the new SmartGauge. Totally different from simple amp-counters like my utterly useless Victron one, which is what I was referring to.

I would buy one of yours if you ever sell them! And if your system could aggregate the data from two shunts.

[sailinglegend]

Originally Posted by goboatingnow

Amp counting is fine if you understand the issues

Voltage readings are fine if you understand the issues

If every case batteries can be ruined by a lack of understanding

That's why the original Merlin Smartgauge was invented 14 years ago.

It only measures the voltage - over 1000 times per second - and even when charging can give a reasonably accurate result for SoC - more accurate than any shunt-based system once the batteries start to age.

So smac999 that answers the OP's question.

...how do you determine SoC while charging?

Balmar also licensed the Smartgauge but are trying to be clever and integrate their new Balmar version with a shunt. Witzgall in post 33 has admitted:

We have found that many of the customers that experience issues have some sort of installation issue that gets in the way.

I told him two years ago that this would be their problem, which is why Merlin refused to introduce a shunt with their original Smartgauge. MaineSail who has been testing the SG200 for Balmar has for a long time quoted over 90% of shunt-based BMs are badly installed - more recently his feedback has been more shockingly surprising - to use his own words:

....we have yet to come up with more than a handful of properly installed and properly calibrated Ah counters..... and why programming is critical to the performance of your Ah counter

The SG200 manual is 36 pages long!

So what hope is there for owners of the shunt-based SG200?

[goboatingnow]

Originally Posted by Dockhead

That was exactly NOT referring to your system.

Your system sounds more like the new SmartGauge. Totally different from simple amp-counters like my utterly useless Victron one, which is what I was referring to.

I would buy one of yours if you ever sell them! And if your system could aggregate the data from two shunts.

The system supports multiple charge and discharge shunts (connected via a one wire LIN bus )

Shunt accuracy isn’t really too important if the fully charged trigger is regularly activated.

High accuracy on high currents is tricky to achieve , I aim for 0.5 percent , manual high precision calibration could improve that to 0.1% , ranging would be needed to improve over that

It remains to be seen if the approach actually results in better SoC predictions, it will take time to determine that.

[Dockhead]

Originally Posted by goboatingnow

The system supports multiple charge and discharge shunts (connected via a one wire LIN bus )

And will aggregate the data?


Are you selling them?

[goboatingnow]

Originally Posted by Dockhead

And will aggregate the data?


Are you selling them?

The current software reports both data types , ie aggregate charge and aggregate discharge ,as well as reporting selected shunt currents , the meter is actually the master controller for all the shunt “ slaves “

Selling things is complex , if it works I’ll give you one for the price of the components.

Right now however I don’t know if the “ ageing” parameters really work and if they do, how much they actually improve the meter

The shunt boards are programmable so they work with any existing shunt and can be added in parallel to existing gauges etc. a single wire connects back to the master ( meter ) unit.

[Dockhead]

Originally Posted by goboatingnow

. . .Selling things is complex , if it works I’ll give you one for the price of the components.
. . .

Wow, that's a nice offer! Sign me up. I'll write an article about it for you if you like.

[goboatingnow]

Originally Posted by Dockhead

Wow, that's a nice offer! Sign me up. I'll write an article about it for you if you like.

No probs. But it will be a bit of a wait and of course it could be a whole bag of nothing

I’ll detail the system a bit more on here soon

[Yerffoeg1]

As mobile phones have run battery management for years to unsophisticated end user(no meters and dummy loads) Why not boats.
Measure input and output and tell the end user what is left in the battery. AI will learn the battery and user. (Twenty first century is here)