Aggregating Current Measurements from Two Different Shunts

[waterman46]

I don't think you can wire shunts in series without additional electronics. The voltage from each end of each shunt is different and these differences are rather small voltages. Connecting any terminal of any shunt to anything other than a voltage sensing device of high impedance (normal voltmeter) will cause the shunt terminal voltage to be changed, which would put errors into the result you expect.

What you need is a separate op-amp sensing the voltage of each shunt. The op-amps have a very high input impedance and the output is proportional to the voltage difference at the terminals. Connect output of both amps together, connected to a resistor to ground. Voltage across that resistor should be proportional to the sum of the two shunt currents.

[Oceanride007]

^^^^ That would work, Op Amps could be made to do lots of things but operate in the Analogue world.
Also if your cabling would allow it, aggregate your Negatives to your Shunt to one battery then connect your 2 negatives together, then do the same to positive except this one would be connected diagonally. Of course cabling to the Inverter is the limitation.

[dfelsent]

Umm. The circuitry needed to do this properly is not hard, but subtlety is needed in the design.
The components required must have very low input offset voltage to avoid consistent errors in the reading. These errors will cause the battery monitor to act as if there is a charge or drain at all times.
It can be done for about (guessing a bit) $40 us of parts. A low offset dual opamp, some precision resistors, and circuitry to filter and protect the amplifier power supply. And circuitry to protect the inputs. And maybe another amplifier to properly drive the 50mv low impedance output. And protect the output as well.
Oh and add some money for the case.
I’d buy something if possible, the engineering to make this a solid reliable sub system is a pain.

[goboatingnow]

Originally Posted by Oceanride007

^^^^ That would work, Op Amps could be made to do lots of things but operate in the Analogue world.
Also if your cabling would allow it, aggregate your Negatives to your Shunt to one battery then connect your 2 negatives together, then do the same to positive except this one would be connected diagonally. Of course cabling to the Inverter is the limitation.

No need to join positives , you could use one negative shunt and between the shunt and the battery, connect a paralleling negative strap. One shunt will read the combined return current to both batteries.

That way any ordinary battery monitor will will work

[tanngrisnir]

Originally Posted by goboatingnow

No need to join positives , you could use one negative shunt and between the shunt and the battery, connect a paralleling negative strap. One shunt will read the combined return current to both batteries.

That way any ordinary battery monitor will will work

Ha, even better! A no extra component solution.

Been fun brushing up on my analog electonics after 35 years.
Easy to build a prototype with a lm752 or lm124 and a handful of resistors but as dfelsent points out making it work properly in the real world with noisy reactive loads is the what makes it challenging.

[Dockhead]

Originally Posted by dfelsent

. . . I’d buy something if possible, the engineering to make this a solid reliable sub system is a pain.

But what?


The Pico will work, but it's $$ and tons of functions I don't need.

[goboatingnow]

Originally Posted by dfelsent

Umm. The circuitry needed to do this properly is not hard, but subtlety is needed in the design.
The components required must have very low input offset voltage to avoid consistent errors in the reading. These errors will cause the battery monitor to act as if there is a charge or drain at all times.
It can be done for about (guessing a bit) $40 us of parts. A low offset dual opamp, some precision resistors, and circuitry to filter and protect the amplifier power supply. And circuitry to protect the inputs. And maybe another amplifier to properly drive the 50mv low impedance output. And protect the output as well.
Oh and add some money for the case.
I’d buy something if possible, the engineering to make this a solid reliable sub system is a pain.

It’s not difficult as that , firstly there is a whole series of fixed gain current sensing amps , which have very low offset currents high power supply rejection and protected inputs with very high common mode voltage ability ( 60 -80 v)

The summing takes place on the output and after a simple RC low pass filler. , I have several in various designs. Cost around $10 in total , a calibration procedure can remove the need for most precision components.

Quite easily done and not complex or expensive

[goboatingnow]

Originally Posted by Dockhead

But what?


The Pico will work, but it's $$ and tons of functions I don't need.

Ardunio and a current sense board or two.

[goboatingnow]

Originally Posted by tanngrisnir

Ha, even better! A no extra component solution.

Been fun brushing up on my analog electonics after 35 years.
Easy to build a prototype with a lm752 or lm124 and a handful of resistors but as dfelsent points out making it work properly in the real world with noisy reactive loads is the what makes it challenging.

That way you indicate whilst in theory is fine , is far too problematic in reality

What you would do is deploy dedicated fixed gain current sensing amps one for each shunt. And then sum the much higher sense voltages using a summing circuit like you show , this would provide drive current etc and a simple attenuator can restore the range to 50mV if required.

The current sense amps have robust input protection , good supply rejection and optional filtering etc. They require very few additional components

[dfelsent]

Originally Posted by Dockhead

But what?


The Pico will work, but it's $$ and tons of functions I don't need.

Yeah. So pico is actually not expensive if it solves your problem. Issue for me is this:
I can do a solid design in a few (3-5) hours. And do simulation testing in that time too.
(I am out of practice but this used to be what I did in my sleep)
Plus someone needs to build and package and test this. That’s another solid day of 8 hours.
At reasonable rates the pico is waaay less expensive. As in 2-5x. And my rates are unreasonable these days because there is no forward benefit if I do a loss leader.
If you are desperate I’ll draw something up but absolutely no promises and I’m seriously out of practice.
Really, the pico,if it does what it says, seems like the far best option.
If it were me, and I am quite capable of deigning and building and calibrating and testing and packaging the shunt combiner, I would would either buy the pico or use two bmv712 and do the math to add the two in my logbook.
A boat (for me) is enough of a project without adding custom electronics to the brew.
Sorry about rain storms and parades.

[hzcruiser]

Originally Posted by HeywoodJ

For dumb shunts, technically yes. The shunt output is just the voltage drop across a very carefully configured resistance. Most shunts we use are 0-50mV. So, if you wired them in series you could get out a 0-100mV signal. [...].

Wire two shunts in series?

[GeorgeW]

Originally Posted by hzcruiser

Wire two shunts in series?

The OP indicated "the OUTPUT of two shunts".

It's not an unreasonable question and under some configurations might actually work but would most likely cross-connect the load conductors with instrument gauge wires.

[GeorgeW]

If the use of two monitors and adding the amperage from each is an acceptable solution then some of the low-cost battery monitors that use Hall Effect transducers and are available on Amazon might work. They cost less than $50 each. Amperage and cable diameter would need to be checked.

[RaymondR]

You could use a couple of DVMs and do the addition in your head??

[GeorgeW]

This configuration would work, possibly with some constant error due to the series connection of the shunts. Connections to the shunts would be at the battery end and at the load end only, nothing between the shunts. Meter will see the combined voltage of the two shunts in series. A very low resistance connection between the shunts would be preferable to minimize error due to voltage from in the connections. Meter full scale will need to match sum of shunt maximum mv output or be custom calibrated (Sharpie?) or digital on appropriate scale.

. Panel1. . . . . . . . . . . . . . . . . Panel2
..... X ....................................... X
..... xxxxxxxxxxxxxxxxxxxxxxxxxxxxx
........................... X
........................... Xxxxxxxxxxxxxxxxxxxxx
....................... Shunt1 . . . . . . . . . . . . . X
........................... X . . . . . . . . . . . . . Meter
........................... X . . . . . . . . . . . . . .. X
....................... Shunt2 . . . . . . . . . . . . . X
........................... XxxxxxxxxxxxxxxxxxxxX
........................... X
..... xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
.... X ......................................... X
.... X .......................................... X
. Battery1....................................Batter y2


My carefully drawn ascii diagram is being mangled by having whitespace removed when posting. Ignore the periods - the X's show connections.