INA139 current monitor draws too much current

[Odysseas]

Hi everyone,

I'm trying to build a solar charger, using a simple boost configuration. To sense the current coming from the solar panel and going into the battery, I use two INA139 current monitor chips. See schematic. The supply voltage V+ is derived from V+_BATTERY_H through a little bit of filtering (V+ is about .7V below V+_BATTERY_H). The output current from the INA139 is run through a 100k resistor each to generate an analogue voltage proportional to the current (1V = 1A). HS1 and HS2 are just heat sinks for the diode and MOSFET.

However, the current monitors start drawing way too much current the instant power is applied. I tried ramping up the battery voltage through a PSU (no solar panel connected) and at one volt, the current draw is already about 100mA. I checked for short circuits and everything, nothing. Initially, I had only the right hand side INA139 populated, then only the one on the left side of the schematic - both show the same behaviour. One has now disappeared in smoke. At one point, D4 wasn't populated yet and the left hand side INA139 didn't draw too much current. This would suggest that the problem only appears when pins 3 and 4 on the chip see any voltage above ground.

What could be going on here? Thanks for any help.

 

[Nigel Goodwin]

Where is your V+ coming from for the chips?, are you feeding them from too high a voltage?.

I've not used the 139, but i have used the 219, and that is essentially a TTL/CMOS supply type chip, with the usual supply limitations.

 

[rjenkinsgb]

There is nothing wrong with your schematic, it ties up perfectly with the INA139 datasheet.

As long as the supplies are no more than 40V, it _should_ work fine.

As it's not, I'd very carefully check the pin connections and orientation of the ICs, a track fault or pin swap is the most likely problem.
https://www.ti.com/lit/ds/symlink/ina139.pdf

Interesting device, not one I've seen before!

 

[Odysseas]

The V+ voltage is derived from the battery voltage, less one diode junction (0.7V). The battery voltage will not exceed 30 volts.

I'll triple check the pinout again.

 

[Nigel Goodwin]

The V+ voltage is derived from the battery voltage, less one diode junction (0.7V). The battery voltage will not exceed 30 volts.

I'll triple check the pinout again.

Yes, I see from the post above that the 139 works from higher voltages, where the 219 doesn't.

I agree that it's most likely connected wrongly.

 

[gophert]

Post a good photo of front and back of your PCB (or breadboard). Try to keep front and back photos the same size.

 

[schmitt trigger]

From the data sheet:

Quiescent current: @ VSENSE = 0, IO = 0. typ 60, max 125 µA
Yes....micro amps. Not even a single miliamp.

So, to determine whether the IC itself is defective, tie both inputs together (essentially shorting the current sensing resistor) and see whether the IC repeats the anomalous behavior.

 

[Odysseas]

I've dug up the project. I managed to turn one of the current monitors into smoke. The other one was already defective, so I replaced it and it works now, with useful accuracy. It may well be that the ones I put on the board before were damaged. It's a bit of a mistery but at least it works for now. Thanks everybody for their help!

 

[crutschow]

Did you get them from a reliable chip supplier?

 

[Odysseas]

I had the same thought, but yes, I think it's a reliable supplier (www.reichelt.de).