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.

OPA604: Batch with low slew rate and long settling time

Part Number: OPA604

We have a batch (92AE7) of OPA604AU devices that fail to meet their specifications in terms of slew rate and settling time when used as a voltage follower on +/-15V rails. The parts were sourced from a franchised dealer.

The example below shows the response to a +/-6V square wave input to both the suspect batch (92AE7, red trace) and a sample of another batch (68CKK, green trace). The vertical scale is 2V/div and the timebase 1us/div. The effective slew rate of the two parts is 13V/us for the suspect batch and 22V/us for the other batch. The datasheet value is 20V/us.

Is this a known problem with this batch, and, if so, how did it come about and how did it get through production batch testing?

  • Hello Peter,

    Sorry to hear you are having slew problems with the particular OPA604AU product you received. Seems odd that would be the case because the OPA604AU would have received slew rate testing in the final test.

    It does appear that the slew is low for the one batch, but before we can say for sure it should be observed that the OPA604AU states the slew rate conditions as 20 VPP, RL = 1 kΩ. Your testing was done with an output of 6 Vpp and I don't know what your load was. It would be helpful if you could retest the OPA604 AU slew rate under the datasheet conditions and then report back with those results.

    Thanks, Thomas

    Precision Amplifiers Applications Engineering

  • H Thomas,

    Thanks for the reply.

    I've taken a board from our lab that has not had its OPA604s replaced yet. I did a quick test using +/-10V input and 1KOhm load. I show both edges of the squarewave in this view:

    The rising edge is approximately 10V/us, half the datasheet figure.

    BTW, the text and the scope view in my previous post showed a +/-6V (12Vpp) squarewave.

    Regards,

    Peter

  • Hi Peter,

    Okay, then the OPA604AU slew rate does appear to be well below the 15 V/us minimum for those devices. TI wouldn't have knowingly shipped out of specification product so somehow these questionable devices got through to you.

    Since this isn't an applications issue and more so a product quality issue, I am working to find the best way within the TI system to assist you. I just sent an email to an engineering manager more closely associated with the manufacturing and test of our Precision Amplifiers products. Please allow him a little time for a response so that the issue can be correctly studied and addressed on our end.

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • Hi Peter,

    We are having some discussion about the questionable devices on this end. Are you wanting to have the OPA604AU product from lot 92AE7 replaced with product from another lot?

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • Hi Peter,

    Please send me a e2e friend request so that we can take this off the general e2e forum where it is in place to address applications inquiries. Once I have your friend request I'll accept and send you my email address so we can handle the issue in a more company to company manner.

    Thanks, Thomas

    Precision Amplifiers Applications Engineering

  • Hi Peter,

    It appears your responses regarding the OPA604AU matter didn't make it through to the e2e. Please try entering a response again.

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • Hi Thomas,

    Sorry about the aborted posts on the forum - the new TI E2E site is not behaving well on the PC I'm using (Win7/Firefox).

    I'm sorry also that I can't find a way of requesting an e2e friend on the TI E2E site. The FAQ example talks about a "Users" button, but this does not seem to appear on the current site, at least not on my browser.

    We have ordered another 450 parts of OPA604AU from our supplier to replace the 92AE7 batch, and requested that they check the batch code before despatch. Something like 240 parts from the suspect batch have already gone into our products delivered to customers around the world, as it's a fault that has largely evaded our production tests. It's going to be a hugely expensive operation to get those products back to check and replace the amplifiers. We know that not all these products have the suspect OPA604 batch, but have no real way of knowing in advance which do have.

    It crossed my mind, and no doubt yours as well, that the 92AE7 batch parts that we have could be counterfeit, rather than the problem being a different amplifier somehow being marked as an OPA604AU during the TI production process. Either way, all the parts I have tested from batch 92AE7 do not function well when required to output signals more negative than -6V, with some variation in the actual waveform distortion from part to part. The suspect parts work well enough in the range +/-5V to pass our standard tests, but we do not test so thoroughly at +/-10V. I did a rough measurement of the input bias currents on a couple of the suspect parts. The results were around 60pA, so at least the suspect amplifiers are likely to have FET inputs.
    As another illustration, the suspect parts react more violently than a good part to multiplexer charge kickback, even when it comes from a 0.4pC charge-injection device (ADG1208). The scope shot below is of the output pin of a multiplexer cycling round several channels that are all being fed with a constant -6V, each channel with an OPA604 as a buffer. The multiplexer returns to a grounded source between samples in order to reduce cross-talk. The left-hand low-going excursion is when sampling an OPA604 that has been replaced with a good part, and the second is sampling a suspect part that has not been replaced. The product design I tested has a 100 Ohm resistor as a damper from each OPA604 output to the corresponding input pin of the multiplexer. The multiplexer output feeds a high-impedance buffer.
    When I've got a moment, I'll do a further test of powering a suspect part from +/-24V, to see if the lid stays on.
    I hope this information is of use to you.
    Regards,
    Peter
  • Hi Peter,

    The e2e system was completely updated across the past few months and it appears that not everything works as it has in the past; including how one goes about adding friends. That feature doesn't appear to be working as needed at this time.

    Our Quality Engineering department has been investigating the date/lot code you provided, 92AE7. I can provide you an important finding regarding this lot in question.

    Do you have an email address I can use to contact you directly? I know that users sometimes don't want to provide an email address here on the e2e because it then becomes public.

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • If I may chime in, judging by the completely different settling behaviour of the red trace part below, I would suspect this is not OPA604.

  • Hi Thomas,

    I've been getting copies of your posts as emails from "TI E2E support forums", but replies to that address get lost. You can contact me at "peter@ced.co.uk". I'll take the risk of that being public.

    In order to get complete replies in this forum, I've been having to use MS Word to construct the complete reply and then a copy all and paste as one entry into the forum. It's not perfect, as you can see from the variable line spacing in my previous post (which was constant spacing in Word), but at least it does not abort the post when I try to include a picture.

    Here is a photo of an extracted OPA604AU from batch 92AE7:

    Regards,

    Peter

  • Hi Peter,

    Thank you. I'll contact you directly via your email.

    Thomas

    Precision Amplifiers Applications Engineering