• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 49 subscribers
  • Views 592 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.

LMH6881: Group delay

Sven Probst
Expert 1630 points
Part Number: LMH6881

Dear experts,

how much is the group delay from INxD to OUTx? Could you maybe provide a step response plot containing input and output signal?

Thank you in advance!

  • 0
    Guru 48395 points
    Hello Sven,

    The TINA model seems to be giving 70psec, but normally there is more than that in package prop delay - so this would need a measurement. No phase plots in the datasheet that I can find. Either that or matched cable length step response prop delay would be needed.
  • 0 in reply to Michael Steffes
    Expert 1630 points
    Hello Michael,
    thank you for the prompt answer. I'm not sure if the spice model is usable / exact w.r.t. latency / group delay. A measurement would be more trustworthy. I don't need a very precise number. I simply need to know if it is <1ns, 2ns or maybe 10ns.
    Maybe someone could measure?
  • 0 in reply to Sven Probst
    Guru 48395 points
    No the model will not give you the right answer. I was looking at the EVM board, unless you cal to the device pins it will produced a measured delay probably dominated by the board prop delay.
  • 0 in reply to Michael Steffes
    Expert 1630 points

    Hi Michael,

    do I understand you correctly that you have an evaluation board available? As I already mentioned I don't need precise numbers: +-1ns would be fine. This corresponds to about 20cm cable length, thus the board delay might be small enought w.r.t. my required accuracy. If you could estimate the trace length on the board and would know its material (FR4?) I could estimate the board delay.

    I was very happy If you could provide such a measurement.

  • 0 in reply to Sven Probst
    Guru 48395 points
    here you go, I don't have a network analyzer, but if you do - you could get this yourself,

    www.ti.com/.../lmh6881eval
  • 0 in reply to Michael Steffes
    Expert 1630 points
    Hi Michael,
    thank you for the link.
    I'm currently comparing a few different options and LMH6881 is only one candidate, thus 200€ and three days delay till delivery is not what I was looking for.
    Is there anybody (maybe from TI itself) who could provide the step response or the group delay time??
  • 0 in reply to Sven Probst
    TI__Genius 15975 points
    Hi Sven,

    Sorry for the delay in reply.
    We are taking a look to provide the step response or group delay data for the LMH6881 and should be back with a reply soon.

    Best Regards,
    Rohit
  • 0 in reply to Rohit Bhat
    TI__Genius 13935 points

    Hello Sven,

    I measured group delay for three gain settings (Max, 0dB and Min).  At maximum gain, the group delay at 1.5GHz was approximately 310ps with respect to the low frequency delay.  The group delay dropped to 0 at 2.3 GHz and then started to diverge reaching a maximum of -900ps at 3GHz. 

    We always measure group delay with respect to the low frequency value, network analyzers do not go to DC, and group delay has no definition at DC.  For this reason, the EVM does not contribute to the measurement because it forms part of the low frequency base delay. In this case the base delay was 1.39ns.

    The group delay does not change appreciably with gain. 

    These measurements are on the EVM using input and output baluns.  The baluns may contribute some group delay at higher frequencies. 

    Regards,

    Loren

  • 0 in reply to Loren Siebert 1
    Guru 48395 points
    I think what he was after was the group delay through a physical device on a board. Normally, I used to do group delay in frequency span of interest where for this part maybe 10MHz to 1GHz linear sweep - then it is the measured phase. For the device itself, that 310ps at 1.5GHz backing out the board delay sounds reasonable for the device given sim showed 70psec and we always get a package prop delay.