• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 33 subscribers
  • Views 794 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.

ISO724x latch up(?) of noise test

Shinichi Inoue
Guru 19651 points
Other Parts Discussed in Thread: ISO7241M, ISO7240M, ISO7241C, ISO7641FC

Dear Specialists

My customer is evaluating the ISO724ｘ.
The device was latched up, when the noise was impressed.
What reason is considered?
Could you please see the following.

Conditions
ISO7241 was installed with FR4 4layers and 6 layers, Vcc1=3.3V,Vcc2=5.2V
When radiation noise 0.5kV was impressed from 1cm of distance, the output of secondery side  was halted.
At the same time, Icc2 was increased from 16mA(normaloperation) to 680mA.
This phenomenon was happened at ISO7240 and ISO7241.

Question
Why this phenomenon is happened?
The customer is considering this may be latch up.
What measures are considered?

Best regards,

  • 0
    TI__Genius 9985 points

    Is the customer seeing this problem on M-grade devices (ISO7240M & ISO7241M)? 

    If yes, then this could be due to lack of input noise filters on the M-grade devices. In that case, we'd recommend the customer to use C-grade devices since they have integrated noise filters on the inputs.

    M-grade devices are very fast (150 Mbps) and therefore we couldn't put noise filters because that'll slow down the speed.

  • 0 in reply to Saleem Marwat
    TI__Genius 9985 points

    Alternatively, the latchup that the customer is talking about is likely not a logic latchup, but PNP-NPN substrate latch-up, causing a supply-ground current increase.

     We could recommend:

    1. Better decoupling capacitors on the supply close to the isolator – do they have 100nF close to our isolator?

    2. 10 to 20 Ohm in series with VCC2 to limit latchup current – though this will degrade VCC2 and hence VOH levels on the output of the isolator by 160mV to 320mV (since their normal icc2 is 16mA). Even higher series resistance could be tried if a further drop in VOH is acceptable. Configuration is attached.  

  • 0 in reply to Saleem Marwat
    Guru 19651 points

    Dear Saleem

    Thank you for your reply.and great help

    I confirmed the device version.

    Those are ISO7240CDWR and ISO7241CDWR.

    I'll feedback the recommendation to the customer about decoupling capacitor and additional current limit resistor.

    At the same test condition, Maxim（MAX1493x） was not happened the latch up.

    Best regards,

    Shinichi

  • 0 in reply to Shinichi Inoue
    Prodigy 120 points

    Hello,

    I have the same problem with a iso7241m (same power supply, same result). There is a decoupling capacitor of 100nF, close to the device (2mm, maybe too much ?, i will try to add more capacitors directly soldered on the leads). Do your customer sold his problem ?. Could you give me the exact reference of the maxim device that you tested ?

    Best regards

    Ludo

  • 0 in reply to ludovic garcin
    TI__Genius 9985 points

    The other option is to try the C-grade devices (ISO7241C or ISO7641FC) since they have input noise filters integrated.

  • 0 in reply to Saleem Marwat
    Guru 19651 points

    Dear saleem

    Thank you for your reply.

    I am waiting for new device that input filter integrated.

    I appericiate your kindness help.

    Beat regards,

    Shinichi

     

  • 0 in reply to Shinichi Inoue
    Prodigy 120 points

    Dear Saleem,

    Thanks for your help. Today, it's not possible for us to use a C grade, because we drive an adc, and with C grade, the propagation delay (2*42ns (side 1 to side 2 then come back to side 1), worst case) will be too long.

    The device iso7241m seems to be very fragile regarding supply bypass... (we use it in a new device. we started production 2 months ago. We have a problem with the iso7241 on 2 boards (about 50 boards produced))

    We are going to test a Si8641 from silicon labs.

    Best regards