Isolated battery voltage measurement with adc

[digitalcat]

Hello everybody,

I wanted to show you an isolated battery voltage measurement system that I am designing for a client.

the circuit is based on tlp7820 isoamp and ltc2344 adc.
tlp7820 has an optimal dynamic range of 300mV, so I use a voltage divider to convert 30Vdc (max) to 300mV ...
Tlp7820 also has a fixed gain of 8.2 and the 30Vdc input output is about 2.5vdc ...
I use a ltc2344 adc which has a differential input so it can be used connected directly to the output of the TLP7820,
ltc2344 can be configured with internal or external reference voltage, for an input range of 0-2.5vdc,
so every analog range seems to match.
I think it should work as I checked from the IC reference designs, I tried to put everything together as I wrote previously, i've never used this ic's and this kind of analog measurement, so I hope you can help me and confirm that I am not doing something. wrong..
Below are schematics of analog isolated frontends with inputs from batteries, and adc input from analog fronteds..
There are a total of 12 inputs, 3 x adc Lt2344, with 4 inputs each one..
Thanks

 

[FvM]

According to your schematic, you want to use internal reference. Measurement range can't be 2.5 or 2.4 V in this case.

Where do you get 12x +5V_ISO_AIN?

 

[digitalcat]

According to your schematic, you want to use internal reference. Measurement range can't be 2.5 or 2.4 V in this case.

Where do you get 12x +5V_ISO_AIN?

Hi FvM, thank you for the reply,
you're right,
i have to change the voltage divider ratio in order to fit the 0-2.048vdc of analog input range allowed by that configuratio (001).
Anyway i'm thinking i'll use external 2,5vdc reference, in order to match the same range of the isoamp.
Regarding the 5V_iso, the net ha global scope, i'm using the same reference for each input, i have not enough space to place 12x isolated dc/dc converter..

 

[FvM]

Regarding the 5V_iso, the net ha global scope, i'm using the same reference for each input, i have not enough space to place 12x isolated dc/dc converter..

In other words, you are shorting ISO_AINx_n terminals through GND_ISO_AIN. Doesn't sound like a good idea as long as ISO_AINx_n aren't connected to a common star point. Otherwise you'll bring up balancing currents through GND_ISO_AIN, at least measurement errors caused by voltage drop and possible cable overload.

I would think about differential voltage dividers for the inputs.

 

[digitalcat]

In other words, you are shorting ISO_AINx_n terminals through GND_ISO_AIN. Doesn't sound like a good idea as long as ISO_AINx_n aren't connected to a common star point. Otherwise you'll bring up balancing currents through GND_ISO_AIN, at least measurement errors caused by voltage drop and possible cable overload.

I would think about differential voltage dividers for the inputs.

touchè... .. this is exactly where I assumed the circuit might be weak..
The customer initially asked for 12 x differential adc input.. afaik it wants to measure voltage level of batteries that probably have same gnd.. anyway i would to prevent the problem you correctly described above, then i placed a footprint for a resistor between iso_ain and GND, to replace eventually with a some kilos resistor...
Do you think it could be enough ?

--- Updated Jul 29, 2022 ---

another issue.. in according with the datasheet of tlp7820:

This means that the following configuration should be better...

(sorry for the low resolution of my draft)

at least 60k of isolation between each iso_ain- and not-to-high Vin+/Vin- respect to gnd...
Any suggestion ?

 

[FvM]

"batteries that probably have same gnd" is just a guess. You can't avoid a clear specification of battery connection and possible potential difference. Worst case, batteries might be series connected, the standard situation for a battery management system in a high voltage battery module.

To achieve maximum mode range, you'll use separate p and n voltage dividers with low side tied to center of the differential amplifier common mode range. Unfortunately TLP7280 datasheet doesn't clearly specify the common mode range, probably -0.3...Vdd1-2V. Similar devices from other manufacturers like TI AMC1100 have up to Vdd1 common mode range.

 

[digitalcat]

"batteries that probably have same gnd" is just a guess. You can't avoid a clear specification of battery connection and possible potential difference. Worst case, batteries might be series connected, the standard situation for a battery management system in a high voltage battery module.

They are not in series,
the system has different batteries in different points.. the worst case is they are not connected to the same gnd or the grounding is weak

 

[FvM]

I would at least consider bat+ ground short as possible scenario.

 

[digitalcat]

I would at least consider bat+ ground short as possible scenario.

yes it could be possible,..
what do you think about my sketch ?
Vin- shorted with gnd and 30k input resistance for each line.

 

[FvM]

what do you think about my sketch ?
Vin- shorted with gnd and 30k input resistance for each line.

The asymmetrical topology translates common voltage into differential input. Use symmetrical divider with grounded midpoint.