• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 46 subscribers
  • Views 238 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AMC7834: Current Sensing is drifting far away with higher temperatures (>roomtemp)

Marks Greulich
Intellectual 530 points
Part Number: AMC7834

Hi there,

we were lucky when we measured an external current via the Current Sense Pins and a shunt (200mR). Under roomtemperature it seems to work. 

But then we tried to measure that same current under hot temperatures up to Tambient=125°C and had a massive shift in the reading value vs. the reallity. The current reading from the AMC told us a value which was almost as double as high as the reality was! We have seen a wrong temperature reading from around 50°C on...

So again - the AMC is telling us a too high temperature so he thinks the differental voltage across the shunt (200mR) is higher then it is.

Shunt is specified up to 175°C. 

Thermalpad of AMC is on Vss not on GND. The AMC is working under 125°C (SPI working, bipolar DAC driver are working).

So not sure if the AMC has a thermal problem? But what is against this thesis is (SPI, DAC working, shift already from lower temperatures...).

Do u have any idea what it could be?

Thanks,

Markus

  • 0
    Intellectual 530 points

    Made some investigations but for me its not a round story so far:

    • Was afraid that the internal reference Voltage (2,5V) was drifting away. But the bipolar DACs are stable all over the temperature range.
    • I measured the U-Ref over temperature and I have a delta of around 1mV from AMC-internal-Temperature (32°C to 122°C).
    • I realized that the AMC-Temperature values over 125°C are getting nonelinear

    My questions:

    1. What is the problem here if it is not the internal reference voltage? The internal OP-Amps drifting away?
    2. Why do I realize that behavour so early at such low AMC-Temperatures (e.g. I see a delta-I between real current and AMC-Current of 12mA @AMC-Temperature =85°C). So far away from the specification maximum. At 122°C AMC-Temperature I got a delta of 26mA in the CS-Value!!!
    3. Does a drift of the capacitor-values at the reference pins or other caps at the AMC causing such problems?

  • 0 in reply to Marks Greulich
    TI__Guru 52683 points

    Hi Markus,

    How are you powering the device? Are you using a stable supply? Is there significant switching noise or ripple on the supply pins? If you are using a switch-mode power supply, do you have an LDO as well?

    Is the DC value of your supplies shifting across temperature?

    What is your common-mode voltage for the CS inputs?

    Thanks,

    Paul

  • 0 in reply to Paul_Frost
    Intellectual 530 points

    Hi Paul,

    we realized that we only measured the reference voltage pin and not the compenastion pin where these >4.7uF caps is on. There we saw that it shrinkend a lot down to 1.xxx V!!!

    So that cap (4.7uF Tantalum) was not the right one and changing it to an 22uF Elko or ceramic the problem was totally gone. It seems to be stable under temperature with a hot gun.

    So the power supply could be a bit noisy but should be stable. There is a DCDC one at the +5V & -5V, and an LDO at the 3v3 side. LDOs/DCDC is not within the critical temperature.

    I guess the "compensation" cap >4.7uF is kind of lowpass filter for the reference voltage or how does it work exactly? Perhaps the filter curve of our 4.7uF tantalum shifted in that way that there are coming several spectral destortions. The cap change had an massive impact.

    Cheers,

    Markus

  • 0 in reply to Marks Greulich
    TI__Guru 52683 points

    Hi Markus,

    Good news I suppose.  

    I am not sure about the behavior of the cap, but if the value decreases with temperature, then it is possible the reference sample current was collapsing the reference voltage, though I doubt it.  I may be mistaken, but tantalums usually behave as a short when the begin to fail.  Maybe the cap is not specified for this temperature?

    Thanks,
    Paul

  • 0 in reply to Paul_Frost
    Intellectual 530 points

    Hi Paul,

    your right - it looks really like the tantalum behaves like a short and the 2,5V dropped down massively. Not run out of spec from my side but perhaps it was damaged during assembly process.

    BR,
    Markus