• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 552 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

AMC3311: Offset problems due to interference and bad layout?

AMC3311: Offset problems due to interference and bad layout?

Erik Wallebom
Prodigy 60 points
Part Number: AMC3311
Other Parts Discussed in Thread: AMC3301EVM, AMC3301, AMC3330

Tool/software:

Sorry I took to long to investigate so the thread was locked.

To answer your questions:

Could you provide some more information on your system?  
What are the 3 AMC3311s being used to measure?
The AMC3311 is used to measure voltages 0 to 300VDC. Very slow voltage change (battery voltage) so the response does not need to be fast.
As seen on the screenshot of the schematic in original thread there is a voltage divider with 1Mohm resistors on each side of a 10kohm resistor. The AMC3311 measure over the 10kohm resistor.

Are you seeing the offset on all the devices?
Yes, but it varies between ICs. I have now desoldered the two left ones and only focused on the position to the right on the PCB. There is no improvement so my idea that they are interfering with each other was not true. I have tried two different ICs and one shows 1,5mV on output and the other shows 9mV on output (when shorted on input before the 1Mohm resistors).
If shorted after the 1Mohm resistors (closer to the AMC3311, same as shorting the 10kohm resistor) then there is very low output voltage (it is then within my offset requirement).

When you say 0V input, does that mean your input supply is off or is the input being forced to 0V?
I mean the ICs are powered, i.e. the PCB is on. But the input is shorted as described above. I have to force the input to 0V using a screw driver or some other way to short the traces. If I do not short the input (but still supply 0V over the resistor divider) then there is 1,5V or 9mV depending on IC on the output which means my product will not show 0V in the LC-display when the battery voltage is 0V.

As mentioned before, the ICs are very temperature sensitive. Putting my finger on the IC makes the output rise several mV.

I have ordered new AMC3311 chip. Also AMC3301 and the AMC3301EVM are on order. I will see if those are better.

  • 0
    TI__Expert 6510 points

    Hi Erik,

    When you see the 5-10mV offset with 0V input, where are you measuring?  Can you compare the measurements at the device input and across the 10k resistor to see if they are the same.  Also, do you know if the offset is DC or is there some small oscillation happening? 

    You also mentioned some of the debugging steps you tried in the previous thread, did you do those changes one at a time or were multiple changes done together(moving/removing 10uF cap, moving 100pF cap, shorting ferrites).  Have you tested the combination of shorting the ferrites (L2 & L3) and removing the 10uF (C7) cap?

    Thanks.

  • 0 in reply to Saleh Ahmad
    Prodigy 60 points

    Regarding your debug questions. I have tried removing different components without good result. Cannot say I have tried every possible combination though.

    Find attached screenshot of my measurements.

    Input of resistor divider = 0mV (shorted)
    10k resistor = 1.8mV
    Output of AMC3311 = 1.8mV

    Something is generating the 1.8mV. The mV varies if I put my finger on the AMC3311 (i.e. heating up).
    After measurements I changed to a new AMC3311 chip and now the measurements is 3.5mV so it's even worse.


    I have also received the AMC3301EVM board.

    I realize the input impedance of AMC3301 is for low impedance shunt measurement but I used a 1Mohm+1.5k+1Mohm resistor divider. One thing I notice is when connecting INN and HGND together (shorting) then the input is 42mV and output is 343mV (gain of about 8.2 which is correct). This is when the resistor divider is shorted (as depicted in the screenshot) So the AMC3301 has the same problem as the AMC3311. When separating INN and HGND the measurement goes to 0mV and works fine. Can this be related to the problem with the AMC3311?

    I wonder how AMC3330 behaves when shorting INN and HGND. I do not have that chip at hand right now.

  • 0 in reply to Erik Wallebom
    Prodigy 60 points

    I removed the AMC3301 from the AMC3301EVM board and replaced it with a AMC3311. It has the same behavior still. A bit lower offset of 0.6mV this time. This offset (0.6mV) is within spec though. But when putting my finger on the chip it goes up several 0.1mV in a few seconds. If I short the input it goes to 0.0mV so this shows this is "injected" voltage and not official offset voltage error.

    I think my solution to all this may be to change the resistor divider to 10x lower values. Instead of 1Mohm + 10kohm + 1Mohm I will use 100kohm + 1kohm + 100kohm. This may introduce some other problems though. I have already tried those values and it now works good enough for my product I hope.

    If you have some input regarding this issue, please let me know. Should it not be possible to use resistor divider of 1 to 2Mohm with the AMC3311?

  • 0 in reply to Erik Wallebom
    Prodigy 60 points

    Another solution would be to add a front end amplifier before the AMC3311. I will look into that.