• State Resolved
  • Locked Locked
  • Replies 13 replies
  • Subscribers 48 subscribers
  • Views 1969 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.

LMH6505 and LMH6714 - internal delay time(latency)

Hirotaka Matsumoto
Guru 27525 points
Other Parts Discussed in Thread: LMH6505, LMH6714, LMH6401, LMH6517

Hi all,

Would mind if we ask LMH6505 and LMH6714?

<Q1>
How long is the internal delay from input to output?(ns? us? it means latency with internal processing) 
The datasheet P2 shows Time Domain Response 2.1ns(LMH6505) and 1.5ns(LMH6714) , is it the delay time of internal processing?

<Q2>
It is relation to Q1, if there is the internal delay from input to output, how much is the tolerance of the internal delay?

<Q3>
LM6505 is auto gain type, does this(auto gain) affect the internal delay?
And then, LM6717 is set gain 7.92(Vout/Vin=7.92)

Furthermore,

Application: Ultrasonic flowmeters
Constraction of Application: Receiver(sensor) --> LMH6505 --> LMH6505 --> LMH6714 --> ADC --> FPGA 

Kind regards,

Hirotaka Matsumoto

  • 0
    Guru 27525 points

    Hi team,

    We guess that your team is so busy, but could you let us know it?

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Mastermind 22825 points

    Hi,

    Sorry about the delay in response as I was out of office until today. I will try to answer as best as I can:

    <Q1>
    How long is the internal delay from input to output?(ns? us? it means latency with internal processing) 
    The datasheet P2 shows Time Domain Response 2.1ns (LMH6505) and 1.5ns (LMH6714) , is it the delay time of internal processing?

    Answer: No, the values you have shown are the rise and fall times to a step input. They are not the input to output delay time. If it is the delay you are concerned with, that's something normally not measured on Analog parts like this and we'd have to measure these on the bench. 

    <Q2>
    It is relation to Q1, if there is the internal delay from input to output, how much is the tolerance of the internal delay?

    Answer: There will be a finite delay and we have not characterized / measured that delay tolerance. I have a couple of questions:

    a) Can you provide the worst case value of delay that you can tolerate?

    b) Do you measure the delay using a small step input (such that Slew Rate limitation does not affect readings) and then look at the time delay from the input mid-point to the output mid-point?

    <Q3>
    LMH6505 is auto gain type, does this(auto gain) affect the internal delay?
    And then, LM6717 (6/8/15 Hooman: I think you meant to say LMH6714 here?) is set gain 7.92(Vout/Vin=7.92)

    Answer: I understand you like to know the variation of delay with gain setting. Once we decide that it is necessary to measure this delay (if the answer you provide to my questions in <Q2> shows that your delay time target is close to the capability of the device), we could also measure relate to gain setting (VG voltage), and also include the effects of temperature (if any).

    I see that your signal path consists of two LMH6505's cascaded together and then followed by the LMH6714. So, the delay would have to be measured from the input to the 1st LMH6505 and all the way to the output of the LMH6714. The delays will be cumulative.

    Please let me know the question I've posed above so that we can decide on the next best course of action.

    Regards,

    Hooman

  • 0 in reply to Hooman Hashemi
    Guru 27525 points

    Hooman san,

    Thank you for your reply!

    We will check following your questions.
    a) Can you provide the worst case value of delay that you can tolerate?

    b) Do you measure the delay using a small step input (such that Slew Rate limitation does not affect readings) and then look at the time delay from the input mid-point to the output mid-point?

    kind regards,

    Hiroaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    Guru 27525 points

    Hooman san,

    We are sorry for the late response.
    We will send followings;

    a) Can you provide the worst case value of delay that you can tolerate?
        We confirmed it from our customer. The value was the following;
        The worst case value of delay: Maxium tolerance is 56psec(peak, it means as the range of vibration)

    b) Do you measure the delay using a small step input (such that Slew Rate limitation does not affect readings)
        and then look at the time delay from the input mid-point to the output mid-point
        No, our customer stage is in the development stage.
        Actually, they don't measure it.

    Kind regards,

    We appreciate your help!

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Mastermind 22825 points
    Hello Matsumoto-san,
    The input / output delay number that you need (56 ps peak) is beyond the capabilities of the LMH6505. Also, the 150MHz bandwidth of the LMH6505 means that even with a very short input / output delay time, the mid-point voltage crossing of the output will be 2,333ps (=0.35 / 150e6) later than the similar input crossing point, which must be too much lag if you are in the sub-100ps delay range.

    Sorry I cannot be much help, but you may want to consider much faster DVGA's such as:
    LMH6401 (5GHz)
    LMH6517 (1.2GHz)

    Here are more Variable Gain Amplifiers to choose from:
    www.ti.com/.../variable-gain-amplifier-products.page


    Regards,
    Hooman
  • 0 in reply to Hooman Hashemi
    Guru 27525 points

    Hooman san,

    Thank you for your prompt reply and your kindness!

    We confirmed our customer again, and we got following;
    MAX tolerance of the variation width: 56ps
    It is possible to adjust the absolute MAX delay.

    So, they want to reduce the variation width within 56ps.

    In this case, will the answer be the same as you mentioned? 

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Mastermind 22825 points

    Hello Matsumoto-san,

    To be more clear, I've done this drawing that shows the relationship between the Input and output transition and the time delay "Delta_t":

    Can you please clarify what the spec is on this image? Is it the variation between "Delta_t" of individual devices or is it the value of the "Delta_t"?

    In my earlier post, I had assumed that "Delta_t" would be the value of the delay. If you are talking about the variation of delay of individual devices (production spread), the answer might still be that it is too tight of a spread given the fact that the device output rise time is so slow (in 1000's of ps) in-comparison and the random variation of device bandwidths alone would be too much to guarantee a spread less than 100ps.

    However, it would still be good if you clarify the definition of Delta_t that you are using?

    Regards,

    Hooman

  • 0 in reply to Hooman Hashemi
    Guru 27525 points

    Hooman san

    Thank you for your excellent reply!
    We apologize that our explanation makes you confuse.

    Our customer would like to know the variation(tolerance) of internal delay Delta-t(Δt)(as you mentioned) in case of from input to output.
    So, as you mentioned, if it is 1000's of ps, it is impossible to adjust within 56ps. 
    Is our recognition correct?

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Mastermind 22825 points
    Hello Matsumoto-san,
    Thanks. I now understand what you are looking for.

    For the reason I gave earlier, that the LMH6505's rise / fall times is in 1000's of ps, I don't think the maximum 56 ps input / output delay time variation is something that can be supported. I've forwarded your query to other people as well, and if they come back and tell me that this is something we can investigate and generate data, I'll post a reply here.

    Regards,
    Hooman
  • 0 in reply to Hooman Hashemi
    Guru 27525 points

    Hooman san,

    Thank you for your kind reply!
    OK, at the first, we will convince our customer of this circumstance.
    After that, if we have some update, we will inform you.

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hooman Hashemi
    Guru 27525 points

    Hooman san,

    If our customer realize variation 56ps, do you have any recommended devices?(except LMH6401, LMH6401 is a bit expensive,,,)
    If you have some advice, please let us know it.

    We appreciate for your help.

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Mastermind 22825 points

    Hello Matsumoto-san,

    I recommend you describe in more detail what you are trying to accomplish and either I or my colleagues might be in a better position to recommend something that might work.

    So far I know that you need gain adjustment and that maybe why you are looking for the LMH6505. I'm not sure why the requirement for input / output delay or any other details.

    Here is my email address if you like to send it to me directly:

    hooman.hashemi@ti.com

    Regards,

    Hooman

  • 0 in reply to Hooman Hashemi
    Guru 27525 points

    Hooman san

    OK, we effort to confirm our customer why the requirement for input / output delay or any other details.
    Certainly, we lack of the reason why our customer need internal delay as you mentioned.
    Furthermore, our customer choiced LMH6505 and LMH6714 by themself.

    After that, if we have some feedback, we will contact you directly.
    Thank you for your help!

    Kind regards,

    Hirotaka Matsumoto