Isolated voltage readings

[wizpic]

I have a project where I need to measure 4 x 24v battery banks which are isolated from each other. I really only want to use 1 Arduino to record the data onto single SD card. I've used an ADS1115 (18 bit converter) to read the voltages at the moment I'm using a dual power supply reading the 4 channels with common ground but I can't have a common GND has this would cause issues. I've looked at galvanic isolation but they seem to use transformer and other components and the more I looked the more confusing it's causing and not sure if it would have an effect on the readings or cause un-stable readings. The code is all working well and had it on soak test for nearly a week to make sure there are no issues or bugs but it's time to go for isolation method.
Which would be the best approach any idears or input would be grateful.

Thanks
Wizpic

 

[Warpspeed]

Its a bit Mickey Mouse, but how about the "flying capacitor" method ?

The basic idea is to connect a fairly large capacitor across the voltage you wish to measure. Then disconnect it from the battery, and connect it across your voltmeter.

You can use either mechanical relays, or high voltage mosfets to do the connection/disconection.
Anything else starts to become either complicated or less accurate.

 

[wizpic]

After carrying out some more searching I think the best way would be a dc/dc converter isolated(va0505a) si8600(I2C isolator) and a single ADC converter for each channel which will give me complete isolation not the way I wanted to go but better and more reliable I think

 

[Warpspeed]

Why do you need a dc/dc converter ?
Each ADC could be powered off the battery it is monitoring.

I have a very similar problem, but with thirty Lithium cells to monitor.

Thirty different ADCs would not all track each other to the accuracy I want, so that would still be a problem.
The flying capacitor method and one ADC should give me much more consistent results with lower overall cost and complexity.

 

[wizpic]

It's just has easy to power it from the master unit as either way I still got to drop it to 5v to power the ADC converter from each battery and the best way to do that is a dc/dc converter

 

[FvM]

It's impossible to decide about best or even suitable isolation methods without a common mode voltage specification. If common mode voltage is relative low, a non-isolated differential voltage divider and amplifier is often the best (accurate and simple) solution.

 

[chuckey]

how about running a multivibrator on each battery then the AC O/P will be Vsupply-.2V P-P, feed to peak detector circuit via a DC blocking cap.
Frank

 

[BradtheRad]

Consider using an optocoupler across each battery.

By installing zeners inline, you obtain expanded scale readings.

- - - Updated - - -

On second thought, it will be tedious and tiresome to calibrate all 4 opto readings.

 

[wizpic]

It's impossible to decide about best or even suitable isolation methods without a common mode voltage specification. If common mode voltage is relative low, a non-isolated differential voltage divider and amplifier is often the best (accurate and simple) solution.

That sounds like a good option I've had a quick browse around looks simple, I will do some more home work on that solution

 

[D.A.(Tony)Stewart]

Long ago I designed a SCADA system for payloads on long umbilicals with 10 remote battery chargers for each experiment battery. In order to get the isolation from noise I successfully used mercury wetted relays to differentially MUX the remote voltage to an ADC with about 1 sweep of 96 inputs per second.

Today remote ADC's are very low cost with serial data.

otherwise with parallel wiring, the flying capacitor S&H method works well in either electromechanical Reed Relays rated for 1e6 operations, or High Voltage Analog Switches up to 100V or so, with DPDT Muxing and resistive grounding between sources near common voltages may be practical.

details on layout are required for EMI and CMRR issues.