INA219 doesn't response

[tcsaba101]

I am biulding a battery analyzer, charging, discharging unit.

This version is working more or less full time for 2 months.
Two weeks ago the INA219 current sensor started to stop, does'nt responsing the bus cycle, and this case the chipKIT I2C library remains in an infinite cycle (twi.c).
After a reset it started to work again.
The failure has been repeated in an irregular ferquency, sometimes once in two days, sometimes in ten minutes.

I have changed the iNA, but after a while the problem continued.

Yesterday a plenty of times repeated.

Today I decided installing the input filter to INA recommended by the data sheet. 
Disconnected the power supply, started to solder and there was some sparks on the iron, because I haven't disconnected the battery.
I did and finished the filter.

When I switched on the system, the chipKIT uC32 board has been warm up, the MX chip was 80-90 °C drawing  much current, the 3,3v goes to 2,7. The card was destroyed.
Cheking the input supply there was no visible failure what could be the cause. I have changed to a chipKIT Uno32 board that was ok, but the INA219 was destroyed, and the program stucks in an infite loop elsewhere in MPIDE.
So the system collapsed.
Right now I dont have more INA sensor to continue.

Then I was thinking about the sparks through the soldering iron.
The iron is grounded, I have had the sparks from the battery + terminal, soldering to the high side shunt what means my system is grounded.
I thought my system is not grunded, at least I intended not to be.

Actually I supposed that notebooks are not grounded, but they are.
There are 3 notebooks connected by USB to my system: a datalogger, one for serial display, and the debugger. All grounded to mains and connected into loops though my system.
Creating nice groud loops.

Do you think this grounding mistake is enough explanation to the INA219 stucks and the damages?
After 3 months of operation?

What do you recommend to solve this?
I could detach the mains ground of notebooks at the inlet, and ground them with a separate wire to my system in one point.
This wire would be along the USB cable to minimize the loop effect.

Or using USB isolators the only acceptable solution?
One notebook connecting directly two throgh isolator.

Thanks, Csaba

[mij59]

Hi,


Please post the schematic and a photo of the circuit you've build.

[tcsaba101]

Hi,


Please post the schematic and a photo of the circuit you've build.

I send you picture at the moment.

I have schematics in parts and hand drawn.
PIC32 controller board, 17V, 12A Programmable CC, CV generator, 200W Programmable load (programable version under construction).
The system wiring.

Now I use the West Mountain Radio CBAIV as programmable load and as a reference data logger.
http://www.westmountainradio.com/cba.php

Which part you are interested in?

[mij59]

Thank you for the photo.

Please post the schematics  from the home made boards , and  how the boards are connected.

[wraper]

I suspect you have a VERY nice ground loop between INA and MCU + high currents flowing around. Schematic which handles high currents connected by a long wires to MCU without any isolation + a nice ground loop = very bad.

[tcsaba101]

I suspect you have a VERY nice ground loop between INA and MCU + high currents flowing around. Schematic which handles high currents connected by a long wires to MCU without any isolation + a nice ground loop = very bad.

The design itselt is a Force Sense system, there is no high current on the MCU side.
I have designed it very carefully.

The MCU,  INA, DAC grounds are connected by a sense wire to the battery - terminal.
The plus sense is not working on the INA219, but I use to measure voltage also by the PIC32 ADC through the +sense.
(The plus sense will work on the INA226 what I will use in the future.)

But he notebooks grounding......., that was a blame.
Two notebooks connected to the MCU card, one to the CBA discharge unit.

I think if I isolate the notebooks it will work, and I can buid it back from the ruins....

First I will try unconnected ground as each of them would run from battery.

[tcsaba101]

Thank you for the photo.

Please post the schematics  from the home made boards , and  how the boards are connected.

I try to put together something what I have in hand and I think it undestandable.

[tcsaba101]

Thank you for the photo.

Please post the schematics  from the home made boards , and  how the boards are connected.

Today I can send you the general drawing.
If you unterested more, I will continue next week.

[mij59]

Thank you for the photo.

Please post the schematics  from the home made boards , and  how the boards are connected.

Today I can send you the general drawing.
If you unterested more, I will continue next week.

Thank you, 'll wait for the schematics.

[tcsaba101]

After I have killed 2 pcs of INA219 sensors and my uC32 board, due to a silly grounding fault, my system works again flawless.
No I2C stucks, no lost connection between REALICE and MPLABX.

I have learded again........ Just summarised the case:

I don't know why, but I thought notebooks are NOT grounded. But yes, the adapter's minus terminal connected to mains ground.
Three notebooks are connetcted to the system, datalogger, display, development system. All was grounded and connected through USB connections. The problems has started after I connected the development system. With 2 notebooks grounding loop it was working without major problems.
I have struggled 2 weeks, the system was breaked down in a stochastic behavior time to time, mainly when there was running 10 Amp though the INA sensor's shunt.

At the beginning it made only stucks on the i2c bus. I haven't get the message at the first INA death. Only when the uC32 card's PIC32chip has gone close to 100 °C and obviously breaked down.
Then I started to think and look for the cause......

Now I operate 1 notebook from ungrounded mains (like from battery), one trough USB isolator (the others don't work this way) and one is grounding the whole system.