SOC meter that uses amperage acceptance instead of voltage?

[Maast]

Voltage is a terrible way to keep track of SOC unless you're already deep into the 'knees' and if thats what you're trying to avoid its already too late.

I suppose if you had a really LARGE bank the inputs and outputs would still keep the C rate low enough its would be sorta accurate without the readout voltage being jerked around by charging/discharging. Something like 1000ah for 100a charge/discharges would be my guess.

I know that even with my current bank of FAIMM FLX 500s AGMs I never know when they're actually full unless I let them sit for a few hours but they never actually sit without something going on for more than a few mins. With a LFP bank I imagine it'll be even worse with their flat charge/discharge curves.

IMO AH counting is better, but not by much and would need continual recalibration on both the high and low end.

I''d love to see a SOC meter that uses amperage acceptance to actually know when full is actually 'full' but I havent seen one and I dont think they exist.

Anybody know of one?

 

[sajjen]

Current acceptance of the cells will not decrease until you are "in the knee".

 

[Maast]

Current acceptance of the cells will not decrease until you are "in the knee".

Absolutely true, however it does definitively tell you if you're actually full or have undershot it. Or if you're full and you're still throwing voltage at it and damaging the batteries.

 

[sajjen]

If the cells are not full, you can push a lot of current and the voltage stays below the knee. If they are getting full, the voltage starts to rise. After a while, the current starts dropping. The rising voltage is a very good indicator that you are getting close to full. As long as your charging source is capable of delivering a significant charge rate, you will not overcharge as long as you stop once you start climbing the knee. If your charge source is very weak, compared to the battery capacity, you will have a hard time determining when to stop charging.

 

[Maast]

If you stay out of the knees thats true, but not the point. LFPs need the occasional completely full charge and voltage is a terrible way to judge state of charge, especially if installed in a system and there are loads. To get a actual full charge you have to watch the current taper to 0 and then STOP. Applying a "float voltage" after a short few mins absorb phase damages batteries.
Read the below:

Marine lithium batteries in operation | Nordkyn Design

This article details the relations between voltage, state of charge, current and temperature for lithium iron phosphate (LFP or LiFePO4) battery banks.

nordkyndesign.com

It's not the only source that says the same thing (in fact I've found many papers and studies) but it's the clearest.

 

[sajjen]

I'm not sure if you are misreading me or if I'm unclear. Have you any practical experience with LFP? Your concerns seem based on theory.

Going into the knee is a reliable way of seeing when your cells are starting to get full. Couloumb counting SOC meters reset at knee voltage because it is reliable.

Keeping your cells out of the knees at all times is a certain way of making sure you have little knowledge of the SoC. It also keeps you from noticing imbalances, and has many other disadvantages.

Going in to the knees gives a lot of info and if you terminate your charge/discharge properly it poses little risk.

 

[Maast]

I understand you, but you're evidently not understanding me. Yes okay, daily cycling to the knees is fine/desirable. Not the point.

LFPs require the occasional FULL charge to 3.55v/3.6V. Using voltage to monitor charge to the edge is unreliable. This I have seen myself, personally, on my test bench with the big blue cells I've been testing the snot out of and establishing cycle life at various C rates. The voltage bounces too much and if you attach a O scope to it you can see ripple and transient highs and lows far faster than a multimeter can read it.

AH counting is unreliable because it's not fine grained enough. Again I have seen this personally on the test bench with the shunt-based AH meter displaying a Full cell even though amps are still going in and the batt was showing 3.6v. It's close but not perfect, better than using voltage anyway.

If the cell is accepting amps it's not full.

The only reliable and precise way to know you're at a full charge but not overcharge is watch the current taper to zero.

Floating at full voltage for more than a few mins is bad . Damages batteries. Read article.

 

[sajjen]

No, you are not understanding what I'm thinking. I am saying you should charge into the knee, not stop before it. That way your SoC algorithm can calibrate.

Counting AH is unreliable IF it is not fine grained enough. Bad coulomb counting is bad. Good counting is good. What make of meter are you using? I've used cheap ones that makes no useful numbers.

Getting any meaningful SoC numbers from just the cell voltages in a system that is actually used is impossible. Counting is the only way. But it does need calibration by entering the top and bottom knees regularly.

What are the ripple levels you are seeing with your scope? What kind of charger are you using?

Floating at full voltage for more than a few mins is bad . Damages batteries. Read article.

I fully agree, you should never float LFPs at all. But letting the charge go into the knee does not equal floating. Neither does letting the current taper of to something like 0.05C and then cutting it off.

 

[Steve_S]

Hmmm interesting timing for this topic. I just finished installing the Chargery BMS on my LFP pack and it can & does track SOC & DOD and displays such on the display, albeit the crappy shunt they sent is being replaced with a deltec unit (solid brass & far more reliable). Luckily it means I did not have to buy another monitoring device for that use. Technically, if it was my only battery system, I could also program it into my Midnite Classic controller's wizbang but as there s FLA in my mix, that's a problem.

The FLA is where I am tripping a bit because I was looking at this DROK Digital DC Multimeter 0-90V 100A Voltmeter Ammeter 12v 24v 30v with a Hall Sensor that I could use to monitor the FLA bank on it's own (FLA is backup power bank once the LFP is in place) allowing me to free up the Wizbang and program it for the LFP pack(s). I would prefer to avoid using more shunts as every connection is a weak point and potential for issues. I know that with FLA the only real way to tell the condition & state is to take electrolyte readings for every cell which is time consuming and messy so it's not something to do daily or for a quick estimate of battery state. I simple Volt Meter is fairly useless for them.

Spoiler: WizBangJr trick to limit current

** This was found on a YouTube video related to programming a Classic for LFP / LiIon **

Setting the WBjr current limit on the Classic
The purpose of this document is to explain how to set up a current limit on the Classic based off
the WBjr current. This will only control the specific Classic that has the WBjr connected to it so
if the desire is to limit more than one Classic each Classic shall have its own WBjr and need
these adjustments made.
The Classic MUST have firmware version 2096 or newer.
For the purpose of this example we are going to say we want to limit the charge current into
the battery bank to 50 amps max.
1- On the Classic that has the WBjr connected to it we need to press the “Main Menu” button several times until Charge is highlighted.
2- Scroll to the right to highlight “Tweaks” and press Enter.
3- Press the right upper button under the word “More”
4- Hold the left arrow and press the upper right button 2 times (You should see a screen like this)
5- The number under “address” should be highlighted, Scroll up or down to get to 4405.
6- Now use the right arrow to scroll over and highlight the value under “Value”
7- Change this to the current limit you want times 10. As mentioned above we want 50 amps so 50 x 10 = 500.
8- After setting the current limit to the value you desire hold the upper left button and press enter. (You should see “DATA SENT AND SAVED”)
9- Press Status to return to the main screen

Now whenever the current into the battery exceeds 50 amps this particular Classic will back
its current down to maintain 50 amps into the battery. It will even put out more then 50 amps to compensate for the load.
Link to the Forum where I asked and found out about this solution https://www.solarpaneltalk.com/foru...-unlimited-solar-panels-overcharge-guaranteed

 

[fat_old_sun]

Voltage is a terrible way to keep track of SOC unless you're already deep into the 'knees' and if thats what you're trying to avoid its already too late.

I suppose if you had a really LARGE bank the inputs and outputs would still keep the C rate low enough its would be sorta accurate without the readout voltage being jerked around by charging/discharging. Something like 1000ah for 100a charge/discharges would be my guess.

I know that even with my current bank of FAIMM FLX 500s AGMs I never know when they're actually full unless I let them sit for a few hours but they never actually sit without something going on for more than a few mins. With a LFP bank I imagine it'll be even worse with their flat charge/discharge curves.

IMO AH counting is better, but not by much and would need continual recalibration on both the high and low end.

I''d love to see a SOC meter that uses amperage acceptance to actually know when full is actually 'full' but I havent seen one and I dont think they exist.

Anybody know of one?

Bogart Engineering's TM-2030 Battery Monitor & its companion SC-2030 Solar Charger use Ah counting to determine a fully charged condition. In addition, they can be programmed to deliver X% more charge than what was taken out during the previous discharge. For example, the batteries were taken down by 55 Ah overnight and the overcharge % is 8%, then the batteries will be charged until they receive 55 Ah plus 8% of 55 Ah, before the batteries considered fully charged. This additional controlled overcharge acts as a mini equalization step and fully converts any trace of sulfation that may have occurred overnight.

 

[jasonhc73]

As of yesterday, my batteries have all equalized to 2 decimals voltage. I charged them bulk at 56.5V (28.25V), after the bulk charge, they "absorb" to 53.0 or 53.1. (I set the float at 51.2, which I plan to never use)

They were as much as .65 off for the pair, now one pair is 0.03 is the worst. Pair 2 are equal, the other two are 0.01 different.

26.55 26.52 | 26.53 26.53 | 26.53 26.54 | 26.54 26.53

Today after 4 hours of use they show 52.6 V. Yesterday after 9 hours, 51.9V.

The % on the charger doesn't have any idea how to show a SOC, SOC really only works on AGM/Lead.

Having an AH gauge does make sense, if I can define what the AH is. The meter would then measure how much power was used.

 

[Dzl]

Hmmm interesting timing for this topic. I just finished installing the Chargery BMS on my LFP pack and it can & does track SOC & DOD and displays such on the display, albeit the crappy shunt they sent is being replaced with a deltec unit (solid brass & far more reliable). Luckily it means I did not have to buy another monitoring device for that use.

What is the issue with the Shunt Chargery provides? Are readings accuate after replacement or do you have to do some sort of calibration?

 

[Steve_S]

What is the issue with the Shunt Chargery provides? Are readings accuate after replacement or do you have to do some sort of calibration?

One because I prefer to know the quality of the parts what will be handling the loads and overall safety as well.

Chargery's is a painted metal (not brass) 300A/75mv unit.

The Deltec & Bogarts are brass with base to mount them. 500A/50mv

 

[Dzl]

One because I prefer to know the quality of the parts what will be handling the loads and overall safety as well.

Chargery's is a painted metal (not brass) 300A/75mv unit.

The Deltec & Bogarts are brass with base to mount them. 500A/50mv

Yeah, the pictures sorta say it all. The Deltec looks waay more robust. I might try to save a bit by trying to buy the Chargery without a shunt and buy the deltec separately.

 

[jasonhc73]

Ok now...

Why don't you just use a hall sensor?

 

[Dzl]

Ok now...

Why don't you just use a hall sensor?

What are the pros/cons of each method of measurement?

 

[jasonhc73]

What are the pros/cons of each method of measurement?

The shunt is another piece to increase ohms. On very small amps, you just pass the current through the ammeter directly. Low cost.

A Hall sensor is outside the cable and only magnetic. It can be moved/removed without turning anything off or disconnecting anything. Can be much more expensive, anywhere from 3 to 10 times more than a shunt or direct.

 

[grizzzman]

Bogart Engineering's TM-2030 Battery Monitor & its companion SC-2030 Solar Charger use Ah counting to determine a fully charged condition. In addition, they can be programmed to deliver X% more charge than what was taken out during the previous discharge. For example, the batteries were taken down by 55 Ah overnight and the overcharge % is 8%, then the batteries will be charged until they receive 55 Ah plus 8% of 55 Ah, before the batteries considered fully charged. This additional controlled overcharge acts as a mini equalization step and fully converts any trace of sulfation that may have occurred overnight.

Yep, it is a good system, but the real beauty of the system is its adjustability. An example is its ability to end charge at an amp set point. And adsorb time to say 0 and float to 0 if needed. Works well for lifepo4 charging needs.

 

[SCClockDr]

The shunt is another piece to increase ohms. On very small amps, you just pass the current through the ammeter directly. Low cost.

A Hall sensor is outside the cable and only magnetic. It can be moved/removed without turning anything off or disconnecting anything. Can be much more expensive, anywhere from 3 to 10 times more than a shunt or direct.

One thing to keep in mind when using Hall Effect sensors in a DC environment is that the sensor is susceptible to the earth's magnetic field. Which can induce significant errors when the coil is moved. RV's are not good environments for their use.

 

[John Frum]

One because I prefer to know the quality of the parts what will be handling the loads and overall safety as well.

:|