Balmar SG-200 Battery Monitor

[john61ct]

Dug up from my past notes on "how does Gibbo's gadget actually work?"

AC impedance spectrography: put AC voltage across the bank, measure the phase angle and amplitude of the current, derive the impedance of the battery. Use a wide variety of frequencies and analyse the results against an internal database of known chemistries using sophisticated proprietary mathematical algorithms.

______
SG is making on the order of 20,000 calculations per second. In addition to smart analysis of voltage and frequency response, referencing an internal database of various LA chemistries.

______
electrochemical impedance spectroscopy, small "load tests", high frequency conductance tests involving micro-burst current pulses, measuring voltage changes at millisecond intervals, derive internal resistance. that's why you have to use #14 AWG wire

______
Above just from references to the Merlin SmartGauge.

Personally I don't feel compelled to try to understand the inner workings of it or the Balmar SG200,

happy to just verify accuracy "as a black box"compared to other known-good benchmarking tools and techniques.

Balmar wants to label that side of its workings

"Active Impedance Compensation"

for purpose of discussion,

and that's just fine with me, fact that its results output is now integrated with traditional coulomb-counting makes it unique

and a decent estimate of SoH is not just icing on the cake, but a huge bonus.

[sailinglegend]

Originally Posted by john61ct

....its results output is now integrated with traditional coulomb-counting makes it unique...

As I've said many times before "this is not unique" as Merlin have been doing this for years with their Data II cell technology on Abrams tanks and another military fighting vehicles.

Now it is available cheaply to the masses, but are they equipped to decipher the results and what will they do when they are told the SoH is 73%?

Probably absolutely nothing if the batteries still seem to be working OK!!!!

[john61ct]

Yes I meant unique in our consumer-level world, military gear is another whole level innit.

And of course what decisions result from the better information is up to each owner, someone heading out on long expedition in primitive areas does not have the same needs as and occasional weekender.

[nebster]

Originally Posted by witzgall

The SG200 does in fact measure impedance. I won't be able to provide any more insight into the methods or process other than this statement. I know this won't satisfy your curiosity, but at least it may help a bit?

Thank you for clarifying this.

I wonder if the active method this monitor uses (pulsed AC injection) can trigger some interaction effects with the high-impedance monitor attached to a downstream shunt. That seems improbable, but Balmar doesn't have any guarantee as to what components are in parallel with that downstream shunt.

I'll continue to posit that absolute impedance alone is not a useful metric for lithium, because the range of possible values is so huge to start with. So there must be a derivative (with respect to time, or current, perhaps) in there to make use of it. That would jive with Balmar's observation that cable length doesn't matter much.

[lateral]

A bit of a drift but if you have to cut +ve cable to shunt solder or crimp?

[AndyEss]

Originally Posted by nebster

I definitely think the first sentence (two sense wires) would be valuable. I don't think the second one would, without a shunt for every string in the bank, because the net current will deviate wildly from the internal currents in each string. (I know this for a fact, because I have watched the currents on each string in my bank before in comparison to the bus current, and the internal ones fluctuate substantially as the strings re-equilibrate themselves

Interesting measurement Nebster.

When I was writing my thought on using the differential voltage measurements, I had also considered having a shunt at each battery for simultaneous measurements of current at each battery. As this would increase product cost substantially, I thought that current flows within the bank would be relatively constant between each battery and the net flow, sonI dropped this branch of thought.
You have actual measurements of wild divergences while each battery in the bank re-equilibriates. Most excellent.

Did you measure these current transients over time? Were they exacerbated when circuits were turned on and off, and then quickly damped out to a more steady state?

If the perturbations were of short duration, maybe just the single shunt measuring net bank current might be sufficient. On the other hand, if the inter-battery currents are continuously variable due to minor differences in between the bank batteries, the single bank shunt wouldn't be useful to possibly monitor individual battery SOH.

Again, the differential battery voltage measurement can be easily and cheaply performed.

I'm just cogitating out loud here as I'm lying on my back post shoulder surgery yesterday. I tore a muscle loading water onto my boat from my dinghy while there was a bit of sea running. I can't wait to get off my a$$ and get back on my boat.

[nebster]

Originally Posted by AndyEss

You have actual measurements of wild divergences while each battery in the bank re-equilibriates. Most excellent.

It's definitely interesting, but I suppose some would worry that this excess flow contributes to battery wear and tear. I tend to think that as long as things stay thermally calm, it probably doesn't matter. But I don't know.

Did you measure these current transients over time? Were they exacerbated when circuits were turned on and off, and then quickly damped out to a more steady state?

I made a couple measurements for a period of several minutes, switching from string to string in a 6p pack. In each case, there was either a large net charge or discharge going on. So maybe 100A of discharge net, and I would see discharge on a string varying between 6 and 20A kind of randomly and somewhat quickly. Sometimes it would sit still for a few seconds and then move a good bit. Other times, one string would sit right at a steady 1/6 of the load for a while.

I didn't study it long enough to notice whether things stabilized or not. I think all of my measurements were within 10 minutes of initiating a large change in total demand.

If the perturbations were of short duration, maybe just the single shunt measuring net bank current might be sufficient. On the other hand, if the inter-battery currents are continuously variable due to minor differences in between the bank batteries, the single bank shunt wouldn't be useful to possibly monitor individual battery SOH.

Yeah, it might be sufficient. With identical cells, and cell-level voltage monitoring available, I'm not sure how valuable per-string SOH calculations are to me personally. On the other hand, there's no way I'm wiring in n expensive shunts just so that a new system can pulse in AC more easily for impedance measurements.

Again, the differential battery voltage measurement can be easily and cheaply performed.

Right!

I'm just cogitating out loud here as I'm lying on my back post shoulder surgery yesterday. I tore a muscle loading water onto my boat from my dinghy while there was a bit of sea running. I can't wait to get off my a$$ and get back on my boat.

Feel better soon.

[transmitterdan]

Originally Posted by nebster

I'm not sure how we can assume this device attempts to find the internal impedance of the battery. It's an interesting idea, but... (a) it's really hard to calculate electrochemical impedance outside of a lab, and (b) it's not even particularly indicative of the first-order thing most users care about. (See below.)

Now we know for sure it does use impedance.

Originally Posted by nebster

Sure, but... it seems like a rounding error. A good shunt can impose a negligible additional impedance.

Snip...

Finally, it appears that Balmar has already said something along the lines of "the length of the wire doesn't really matter that much." So that kinda undercuts the whole theory that wire impedance matters.

I don’t agree with that as already explained. But even so, there are other problems that can be caused by putting other shunts between the SmartGauge shunt and the battery. Many digital ammeters create noise from their digital circuitry which could interfere with another sensitive voltage or current measurement. So I think Balmar engineers are smart to head off problems like that by their requirement to connect their shunt directly to the battery minus terminal. Good practice would be to make that wire short and fat.

I have not seen them say any old length of wire is ok. My guess is that their engineer(s) would not like to see many extra milliohms of wire or other shunts. And they provided an excellent way of monitoring banks that are separated by simply using multiple shunts.

I’m not fighting with you. I’m just trying to explain why Balmar have valid engineering reasons for their recommendations. If you don’t want to follow them I don’t really care.

[nebster]

Originally Posted by transmitterdan

Now we know for sure it does use impedance.

We do, and half of us are skeptical that it helps a wit, but it at least sort of makes "sense" of some of the other weird things they've stipulated.

But even so, there are other problems that can be caused by putting other shunts between the SmartGauge shunt and the battery. Many digital ammeters create noise from their digital circuitry which could interfere with another sensitive voltage or current measurement.

Agreed; that's the conclusion I wrote above, too. It still doesn't explain how they expect an SG200 to be used with a lithium battery that has an integrated BMS, however.

I have a suspicion that they are being overly conservative, but we'll learn more if these devices are as magical as it sounds like they are, because they're going to be deployed widely if so. And not everyone will be able to follow the directions.

I have not seen them say any old length of wire is ok. My guess is that their engineer(s) would not like to see many extra milliohms of wire or other shunts.

I think the comment regarding the length was up-thread here, or maybe it was in the other monster lithium thread.

They can't manage impedance much by stipulating wire length anyway or shunts anyway. It's a second-order value on any reasonable installation with any larger lithium pack.

I’m not fighting with you. I’m just trying to explain why Balmar have valid engineering reasons for their recommendations. If you don’t want to follow them I don’t really care.

No worries; you're debating. It's one of the best ways to learn!

And, in that spirit: I very much agree that they have "engineering reasons." It's the "valid" that I'm much less certain of. People just say stuff all the time, often for another layer of CYA, without understanding what they're actually observing or without full understanding the consequences for their customers. Especially early on with new products.

[john61ct]

Eliminating as much as possible any variables that even "might" interfere with optimum performance, is IMO valid.

It really is in all our interest for SoC measurement to get more accurate, or at least as accurate as Merlin's benchmark plus the added value functionalities.

[sailah]

So I have had the SG-200 installed for a couple weeks. I'm on shore power behind my barn on the hard. Go in the water on Saturday.

All batteries are brand new, with date codes from the last 2 months. Group 31 FLA start, and 2 GC2s in series for house.

The SoC seems pretty accurate and matches the True Charge 2 when it tapers off. I have unplugged from shore power for hours at a time, using my fuel polisher and all interior lights as load while I work.

At one point SoC went down to 56% so I plugged shore power back in. Otherwise I usually keep shorepower on, other than when I need extension cord for heat gun or something.

I have configured both battery charger and SG200 to FLA and capacity of bank. I have a Blue Seas ACR between House and Start. All loads other than starter, go to house bank.

My SoH registers at 90% and has not changed after cycling a few times. I know the SG200 takes some time and cycles to become accurate but I feel I have done that?

Since the ACR combines batteries under charging loads, does the SG200 "see" this start bank? I have the orange voltage lead from shunt to start battery +. Since I have started the engine all of once (and it started way faster since upgrading cable to 2/0) I can't see start bank being depleted much at all.

Long-winded way of asking, is SoH calc to be expected at this point? I would think I would see higher SoH with brand new batteries. Be patient? Equalize bank?

Thx

[rgleason]

My guess is that they are not fully charged. What is the trailing amps at the end of the charge?


I also have a SG-200 with (2)T105 in series. Charge to 100% SOC with SOH 100% and remove power. In three days they show 88%SOC. Is this normal for FLA? I did not read the voltage this time.

[witzgall]

Often new flooded batteries can take many cycles before they test up to their rated capacity, assuming they are quality batteries. Of course I cannot say for sure if that is the case in your installation.

Chris

Originally Posted by sailah

So I have had the SG-200 installed for a couple weeks. I'm on shore power behind my barn on the hard. Go in the water on Saturday.

All batteries are brand new, with date codes from the last 2 months. Group 31 FLA start, and 2 GC2s in series for house.

The SoC seems pretty accurate and matches the True Charge 2 when it tapers off. I have unplugged from shore power for hours at a time, using my fuel polisher and all interior lights as load while I work.

At one point SoC went down to 56% so I plugged shore power back in. Otherwise I usually keep shorepower on, other than when I need extension cord for heat gun or something.

I have configured both battery charger and SG200 to FLA and capacity of bank. I have a Blue Seas ACR between House and Start. All loads other than starter, go to house bank.

My SoH registers at 90% and has not changed after cycling a few times. I know the SG200 takes some time and cycles to become accurate but I feel I have done that?

Since the ACR combines batteries under charging loads, does the SG200 "see" this start bank? I have the orange voltage lead from shunt to start battery +. Since I have started the engine all of once (and it started way faster since upgrading cable to 2/0) I can't see start bank being depleted much at all.

Long-winded way of asking, is SoH calc to be expected at this point? I would think I would see higher SoH with brand new batteries. Be patient? Equalize bank?

Thx

[sailah]

My SoC goes to 100% routinely, and SG200 reads no amps in or out after sitting on charger overnight. Charger LEDs indicate float charge.

I have tested the accuracy of the shunt by testing amp draw on my Fluke first and then seeing what the SG200 says. Pretty much spot on for the ~1.5A load I tested so I feel the shunt is accurate. So I also believe the SG200 when it says the bank is not taking any more current and SoC is 100%.

My question is more along the lines of why isn't my SoH reading higher than 100% on brand new GC2s. Mine are Duracell and date codes are 3/2019.

I've drawn the bank down to 56% once, and other times to mid 90s %. But I really haven't drawn it down/cycled like I would when I would cruise.

My suspicion is that I haven't cycled it enough for the SG200 to learn, just wanted to hear others experience.

[two-rocks]

Originally Posted by sailah

Long-winded way of asking, is SoH calc to be expected at this point? I would think I would see higher SoH with brand new batteries. Be patient? Equalize bank?

Thx

Similar situation, minus the shore power (50w solar to keep things topped up until the other 300w go back up).

My FLA batteries are 8+ years old and show 90% SOH. Cycled similar to you.

Going to wait until a few consecutive deep cycles (launch this Friday).

Hope it is not accurate - lithium batteries were the next project ;-)