Post Go back to editing

Strange AD595 behavior

Hello,

I'm using a K-type thermocouple and an AD595 to measure temperatures in the range +20 C to -100 C.  The system that I'm measuring has rapid temperature changes, covering the entire temperature range in as little as 2 seconds.

I'm powering the AD595 with +/- 12V (from two lead acid batteries), my thermocouple has an ungrounded junction and I've made the ground connection as described in the AD595 spec sheet.  The output of the AD595 goes directly into a NI USB data acquisition card.  The batteries are also powering 2 pressure transducers.

The problem that I have is that once the temperature drops below 0 C, I get very large oscillations on the output.  I've atached a screenshot of some of the results.  The frequency of oscillations is larger than my sampling frequency (10 kHz).  I have no idea what's causing it or how to fix it.  If it weren't for the fact that this only occurs at negative temperatures I would just assume I need some low pass filtering and be on my way, but I'm confused by this situation.

Any help that anyone can provide would be very much appreciated!

on Jan 5, 2012 3:45 AM
0

Jonahz,

  It would help if you also posted your schematic with component values.  I suspect oscillation, so currentleak may be on the

right track.  Put the 50 ohm in series with the output of the 595.  Also read page 5 of the d.s. on adding caps depending on

what your gain is.  Finally, the AD595 is over ten years old.... You may want to look at the AD8495.

Harry

  • Harry & currentleak,

    Thank you both for replying.

    I tried adding the 50 Ohm resistor and it's definitely improved the response.  I've attached two plots of some results, along with a schematic.

    Harry - if I'm using the default gain of 247.3, I shouldn't be adding capacitors for additional frequency compensation, right?

    Thanks again,

    -Jonah

  • Hello jonahz,

    Is the reference junction (cold junction) temperature pretty stable or is it also moving around some?

    One idea is to place a 1MOhm resistor between the thermocouple and where you ground it (so ground that leg of the thermocouple through a 1MOhm resistor), see if this makes a difference.

    Also, as Harry suggested, take a look at the AD8495...

    We do also recommend at filter at the inputs of the amplifier, this might help you.

    https://ez.analog.com/cfs-file/__key/communityserver-discussions-components-files/382/e998adadd58b9c7876c151f7b8e6e2cf.html

  • Hi Reemo,

    I should have mentioned that the reference junction temperature is farily constant (17 C +/- 1 C).

    I originally had the 1 MOhm resistor -> ground as you described, but I found that I had much lower noise when I grounded it directly.

    So I took a look at the AD8495 and for my application it doesn't seem like it would be a substantial improvement because of my small measurement range, the +/- 12V supplies I already have in use for this system, and everything is done on strip boards so I'd need an adapter for the MSOP anyway.

    However, I did like the details in the spec sheet about filtering at the inputs.  I implemented that (setting corner frequencies to ~200 Hz) and now my response looks beautiful.  I've attached a plot.

    Thank you all for your help!

    -Jonah

  • Excellent, happy to hear that you found a solution.

    Thanks for asking the question.

    Reemo

  • This question has been closed by the EZ team and is assumed answered.