Balmar SG-200 Battery Monitor

[lateral]

All potted.
I guess for ip rating.

[AndyEss]

IP rating and protecting their proprietary design.
R&D does need to be paid for.
Good luck - my electrical system upgrade is several steps below this level, so I'm not learning anything more here that is of use to me.

[nebster]

Originally Posted by AndyEss

Is there any circuitry/microprocessors/ remote sensing leads on it?

From the manual, it looks like there are three positive-side, fused sense leads coming off of it. One for the primary battery, and two aux leads, presumably for just casual monitoring of two other batteries.

Like most shunt designs that simultaneously measure V, the device is using its current-carrying negative lead as the other half of the circuit with the main battery. (The diagram shows additional sense leads going to the two aux batteries if those do not share a common ground plane with the main battery.)

I don't think any of us want to reverse engineer and market our own devices, but it would be nice to know just why it needs to be the last one closest to the negative post.

Exactly. We need to understand the circuit design so that we can reason about how to use the module.

After a careful reading of the calibration routine, and especially the "if you want to quickly get an SOH estimate, do these special steps" section, it looks more and more to me like it is simply measuring voltage and current and being a lot smarter about inferring state. Which is a fine approach, but it's these extra requirements that are still leaving me unsure how to think about this device.

Put another way, the way I still suspect it works wouldn't require the things they're saying it requires!

[AKA-None]

If they were to say you can’t have a smart shunt in he mix then I could understand but a dumb shunt should look just like a piece of wire.

[transmitterdan]

It isn’t hard to understand why Balmar engineers don’t want extra shunts between the SmartShunt and the battery.

There are two circuit parameters changed by putting other shunts or long wires in the negative battery connection to the SmartShunt. One is extra resistance. The other is extra inductance. Either of these could reasonably affect an instrument that is trying to find the internal impedance of the battery. Any extra resistance in the connection between shunt + and battery - would be interpreted as internal battery resistance.

The only way to avoid this problem would be to measure battery voltage with differential wires. That would make the voltage and current measurement points independent. But it also would make installation more difficult. The SmartShunt doesn’t have such a wire. So it gets battery - voltage from the shunt + connection. Adding extra resistance such as another shunt or too small of a wire to battery - will be measured as part of the battery internal resistance. That is part of the formula for determining SoH for most battery types. For Lithium it is about the parameter to determine SoH. So for Lithium chemistries having extra resistance makes it doubly harder to determine degradation of SoH over time.

The simplest solution for two separated banks would be to put a SmartShunt at each bank minus terminal. I think you can have up to 3 for a single display.

[AndyEss]

From a little googlie research, LFP batteries in sizes typical for most house banks have internal resistances of 0.001 to 0.002 ohms. 000 gage cable has a resistance of about 0.0001 ohms. A 50mV, 100A shunt has a resistance of 0.0005 ohms.
10 ft of 000 cable has the same resistance of a single battery. Ten ft of cable before the smart shunt is more than a bit long, but even if it is 3 or 4 ft long, the cable's resistance is 30-40% of a single battery. Multiple batteries, wired in parallel would comprise a circuit of significantly less resistance than a single battery. Most house banks are multiple battery.
Introducing a dumb shunt of 0.0005 ohms is 25-50% of a single battery resistance, but about equal to a 4 battery (in parallel) house bank.
Easy solution- put the smart shunt as close as possible to the negative battery post.
Possibly a better solution - add one differential independent voltage input (separate from current) for the entire bank. Even better, add differential voltage measurements for every battery in the bank. You could monitor how each battery was degrading with time with this method.
Most A/D chips are multichannel, adding the small diameter wiring to each battery wouldn't necessarily be much of a problem to achieve.
Many A/D systems allow the user to choose single ended (like the Balmar SG200) seems to be, or differential inputs. These are almost always multichannel, typically 8 SE or 4 Diff.
My conclusion is Balmar's best practices recommendation is right on the money.

[witzgall]

Originally Posted by transmitterdan

It isn’t hard to understand why Balmar engineers don’t want extra shunts between the SmartShunt and the battery.

There are two circuit parameters changed by putting other shunts or long wires in the negative battery connection to the SmartShunt. One is extra resistance. The other is extra inductance. Either of these could reasonably affect an instrument that is trying to find the internal impedance of the battery. Any extra resistance in the connection between shunt + and battery - would be interpreted as internal battery resistance.

The only way to avoid this problem would be to measure battery voltage with differential wires. That would make the voltage and current measurement points independent. But it also would make installation more difficult. The SmartShunt doesn’t have such a wire. So it gets battery - voltage from the shunt + connection. Adding extra resistance such as another shunt or too small of a wire to battery - will be measured as part of the battery internal resistance. That is part of the formula for determining SoH for most battery types. For Lithium it is about the parameter to determine SoH. So for Lithium chemistries having extra resistance makes it doubly harder to determine degradation of SoH over time.

The simplest solution for two separated banks would be to put a SmartShunt at each bank minus terminal. I think you can have up to 3 for a single display.

The smartlink network supports up to 32 devices.

Chris

[sailah]

Got my SG-200 installed over last couple days in addition to a lot of other cabling/wiring.

First, the hole size is not 2 1/8" as other standard gauges. I had an original 2 1/8" battery voltage gauge at my nav and it wouldn't fit. I ended up chucking a wooden plug in my lathe, wrapped in 120G sandpaper and drove a lag into plug. Cut the head off the lag and then used my drill to "sand" the hole about 1mm wider. Other than this, install was about as expected. Instructions are clear and I followed them.

I ran a fused wire from my start battery (with common ground) to the shunt so I can measure voltage on my start bank. All charging sources come into house bank only with ACR so it seems to be pretty accurate when I put charger on as well as loads I have measured with my Fluke before.

Since I just installed it, and still on the hard, I don't expect SOH to work until it has cycled a few times.

So far it works well. I also had the display "dot" but I can say than when it's ON, you can't even notice.

I also replaced the SS fasteners with silicon bronze. Still need to tidy up a few wires, mount the fuse holder permanently etc.

Overall I like it. Does anyone know if the display has a power down feature? I know it gets dim if you don't play with it but wondering if it ever truly goes off and what the mA drain is of the display?

[Q Xopa]

Originally Posted by sailah

Got my SG-200 installed over last couple days in addition to a lot of other cabling/wiring.

First, the hole size is not 2 1/8" as other standard gauges. I had an original 2 1/8" battery voltage gauge at my nav and it wouldn't fit. I ended up chucking a wooden plug in my lathe, wrapped in 120G sandpaper and drove a lag into plug. Cut the head off the lag and then used my drill to "sand" the hole about 1mm wider. Other than this, install was about as expected. Instructions are clear and I followed them.

I ran a fused wire from my start battery (with common ground) to the shunt so I can measure voltage on my start bank. All charging sources come into house bank only with ACR so it seems to be pretty accurate when I put charger on as well as loads I have measured with my Fluke before.

Since I just installed it, and still on the hard, I don't expect SOH to work until it has cycled a few times.

So far it works well. I also had the display "dot" but I can say than when it's ON, you can't even notice.

I also replaced the SS fasteners with silicon bronze. Still need to tidy up a few wires, mount the fuse holder permanently etc.

Overall I like it. Does anyone know if the display has a power down feature? I know it gets dim if you don't play with it but wondering if it ever truly goes off and what the mA drain is of the display?

Nice work! Not many boats with wiring that neat.

[lateral]

Originally Posted by sailah

Got my SG-200 installed over last couple days in addition to a lot of other cabling/wiring.

First, the hole size is not 2 1/8" as other standard gauges. I had an original 2 1/8" battery voltage gauge at my nav and it wouldn't fit. I ended up chucking a wooden plug in my lathe, wrapped in 120G sandpaper and drove a lag into plug. Cut the head off the lag and then used my drill to "sand" the hole about 1mm wider. Other than this, install was about as expected. Instructions are clear and I followed them.

I ran a fused wire from my start battery (with common ground) to the shunt so I can measure voltage on my start bank. All charging sources come into house bank only with ACR so it seems to be pretty accurate when I put charger on as well as loads I have measured with my Fluke before.

Since I just installed it, and still on the hard, I don't expect SOH to work until it has cycled a few times.

So far it works well. I also had the display "dot" but I can say than when it's ON, you can't even notice.

I also replaced the SS fasteners with silicon bronze. Still need to tidy up a few wires, mount the fuse holder permanently etc.

Overall I like it. Does anyone know if the display has a power down feature? I know it gets dim if you don't play with it but wondering if it ever truly goes off and what the mA drain is of the display?

Power save in the menu.

Did you bother to do an equalize before hooking up?
ie fully charged.
I think its in the manual but I have done one recently.
Wish my wiring was that tidy.
Separate battery banks port /starboard messes things up.

[sailah]

Originally Posted by Q Xopa

Nice work! Not many boats with wiring that neat.

Thanks, it's my mission to get better with wiring. Nothing screams "this boat was well looked after" than a tidy wiring job. Unfortunately there are areas with typical wiring but anything I get my hands on is done to best of my ability. I do enjoy it.

Originally Posted by lateral

Power save in the menu.

Did you bother to do an equalize before hooking up?
ie fully charged.
I think its in the manual but I have done one recently.
Wish my wiring was that tidy.
Separate battery banks port /starboard messes things up.

I did not equalize, should I have? Maybe I didn't read that part in the manual. I have all new batteries, FLA, group 31 start, and (2) GC2 in series. I had charger on for maybe 30 minutes total so far just because I wanted to be on the boat when it was charging since I did so much new wiring.

I'm glad my small house bank is together. I moved my starter battery to the starboard side to be closer to switch and starter. Helped weight distribution and saved a ton in cables, esp 2/0

[ayates]

Originally Posted by sailah

First, the hole size is not 2 1/8" as other standard gauges. I had an original 2 1/8" battery voltage gauge at my nav and it wouldn't fit.

The great thing about standards is there are so many to choose from

I have seen the following hole diameters referenced for nominal 2" gauges:

2 1/16" (2.0625")
53mm (2.0866")
2 1/8" (2.1250")

I had read about somebody else running into hole size problems with the SmartGauge, so I had my new panel cut (Front Panel Express) with a 53mm hole. The SmartGauge fit perfectly. So I suspect it was designed to the metric dimension.


Allan.

[nebster]

Originally Posted by transmitterdan

It isn’t hard to understand why Balmar engineers don’t want extra shunts between the SmartShunt and the battery.

There are two circuit parameters changed by putting other shunts or long wires in the negative battery connection to the SmartShunt. One is extra resistance. The other is extra inductance. Either of these could reasonably affect an instrument that is trying to find the internal impedance of the battery. Any extra resistance in the connection between shunt + and battery - would be interpreted as internal battery resistance.

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.)

The only way to avoid this problem would be to measure battery voltage with differential wires. That would make the voltage and current measurement points independent. But it also would make installation more difficult.

I'd argue it would not make it more difficult, except in cases where the alternative is not having a customer. Just loop back the negative differential onto the power terminal by default. Let the customer run it back to the real battery terminal if necessary.

The SmartShunt doesn’t have such a wire. So it gets battery - voltage from the shunt + connection. Adding extra resistance such as another shunt or too small of a wire to battery - will be measured as part of the battery internal resistance.

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

That is part of the formula for determining SoH for most battery types. For Lithium it is about the parameter to determine SoH. So for Lithium chemistries having extra resistance makes it doubly harder to determine degradation of SoH over time.

Well, hang on a minute. Let's just agree, up front, that no one has defined what SoH means. About the only thing everyone agrees upon is that it is a unitless percentage of some baseline "specified capacity." Whether the underlying units are ohms, watt-hours, watt-hours-per-time lost, or something else, is open to everybody to decide.

So, yeah, if your goal is to calculate [impedance today] / [impedance at the beginning of the pack's life], that "impedance SoH" makes sense.

But I will suggest that what most of us actually care about is "capacity SoH", where the fraction we want to know is [energy capacity today] / [energy capacity at the start].

And, I'll go one step further and suggest that knowing that fraction is actually only the second most important health-related value most people care about when running an LFP battery pack. What we really want to know is what is the real capacity of my pack today? In other words, we want the numerator.

Knowing the numerator allows our state estimator to be more accurate, in a way that traditional current shunts can't be. And, since batteries deteriorate and lose capacity, having a constantly-updated capacity value avoids the user having to test and re-program like a traditional one.

Knowing the quotient (the "capacity SoH") is sorta useful, as a long-term piece of trend data. Yes, I'd like to know my capacity has fallen to 95% of what the manufacturer specified this year, and 92% next year, and so on. But it isn't actually that actionable in the day to day. Having a better SOC gauge that stays accurate day to day is the killer function.

That's a long-winded way of suggesting that I don't think this device is measuring impedance. I think it is re-estimating capacity and using that as the basis for its SoH quotient.

And, if my (absolute) guess is true, then I don't think impedance factors into the calculation at all. I never need to know my battery impedance to determine its capacity, nor to tell you its SOC if I have {instantaneous voltage, a stable current over the last few minutes}.

The simplest solution for two separated banks would be to put a SmartShunt at each bank minus terminal. I think you can have up to 3 for a single display.

Sure, but what if your bank is comprised of six managed strings? Or you have eight drop-in Battle Born or Trillium 12V packs, in 2p4s? Or some other combination? That's starting to sound really expensive!

The cell impedances, fusing, contactors, FET gates, and/or any other componentry in circuit will all vary considerably in these "batteries", and that's before we even talk about the wiring between the negative "terminal" and any measurement device -- which, as you point out, might be different length or gauge from boat to boat.

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.

?

[nebster]

Originally Posted by AndyEss

From a little googlie research, LFP batteries in sizes typical for most house banks have internal resistances of 0.001 to 0.002 ohms.

Those are typical impedances for a single cell. A 52V-nominal, 1p16s string will have a baseline impedance of somewhere around 10 to 40mOhm. (My own strings average about 29mOhm.)

000 gage cable has a resistance of about 0.0001 ohms. A 50mV, 100A shunt has a resistance of 0.0005 ohms.

3/0 is 0.06mOhm/ft, so ten feet is 0.6mOhm. Pretty much a rounding error on a 20mOhm battery. A shunt would be totally in the noise, which is why I'm surprised to see them call this requirement out.

Multiple batteries, wired in parallel would comprise a circuit of significantly less resistance than a single battery. Most house banks are multiple battery.

Don't forget fusing, interconnect, bus bars, contactors, and so on.

My own 6p16s pack is 7mOhm total impedance to the load busses.

There is no framework for inputting the pack configuration or the baseline impedance into the SG200. This, plus my own observation of how easy it is for my simple brain to estimate SOC from a voltage output and a look at the recent current demand, is why I wrote above that I think it's unlikely that this device tries to calculate battery impedance.

But, I've been wrong before.

Possibly a better solution - add one differential independent voltage input (separate from current) for the entire bank. Even better, add differential voltage measurements for every battery in the bank. You could monitor how each battery was degrading with time with this method.

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.)

But, anyway, a static impedance in circuit does not affect the characteristic shape of the charge curve, which is likely what the device is using to estimate SOC and to extrapolate SOH.

My conclusion is Balmar's best practices recommendation is right on the money.

You might be right, but I'm going to wager that you're not.

[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?

Originally Posted by nebster

But, I've been wrong before.