• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 49 subscribers
  • Views 537 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.

THS4521 Settling time

tony holt
Prodigy 180 points
Other Parts Discussed in Thread: THS4521, THS4541

The High-Speed amp product selection guide states a settling time of 10ns to 0.1% but the data sheet spec is 10ns to 1%. Which is correct?

How could I estimate the settling time to .0015% (16-bit)? Is there a simple way to estimate how much this is impacted by increasing the gain to say 5x?

  • 0
    TI__Genius 13935 points

    Hello Tony,

    Looking at Figure 3 (Large signal pulse response) confirms that the datasheet number of 1% is most likely correct.  Looking at that figure the scale is too coarse to be sure what the 0.1% settling would be. 

    Unfortunately there is no way to estimate settling time to 16bit accuracy.  Second and third order effects come into play with these levels of accuracy. 

    I would expect increasing the gain by 5x to increase the settling time to around 40ns, mostly due to lower bandwidth at the higher gain.

    Regards,

    Loren

  • 0 in reply to Loren Siebert 1
    Prodigy 180 points

    Thanks for the prompt response.

    Loren Siebert 1 said:

    Unfortunately there is no way to estimate settling time to 16bit accuracy.  Second and third order effects come into play with these levels of accuracy.

    Ok, I rather thought that would be the answer. So how reliable are the Spice models in predicting settling times? Are they likely to vary widely for different amplifiers and are there any clues as to which models might give reasonable settling time results?

    Is there any particular logic wrt settling time specifications? When specified at all, datasheets seem to specify 1%, .1 %, .01% or sometimes more than one; I'd guess its down to the intended application for the device but it doesn't always seem obvious to me as to why the particular values are chosen.

    It would be helpful if it could be standardized a bit better - two values would be very helpful to give some idea as to how to extrapolate to different accuracies before undertaking a bench evaluation.

    Tony

  • 0 in reply to tony holt
    TI__Genius 13935 points

    Hi Tony,

    I would trust TINA models to a first order of accuracy ~10% timing accuracy on a 1% settling time measurement.  I would not use TINA for 0.1% or 0.01% timing.  There are too many variables that you can't model (like PCB trace impedance and ADC input reactance). 

    The faster the amplifier is the harder it is to measure accurate settling times, so the trend has been toward less accuracy rather than more.  Also, differential amplifiers are very difficult to measure to a high degree of accuracy since it's the combination of two outputs, so now the accuracy requirment of the measuring instrument must be 2x that of a single ended amplifier since the errors could add.  Another reason we do not specify highly accurate settling times is that with high speed amplifiers the PC board design and test equipment interface can dominate the response.  Many ADC inputs are highly reactive (ususlly capacitive) this will greatly change the settling time. 

    With single ended op amps there are tricks that can be used, but these tricks don't work for differential amplifiers. 

    Probably one of the best rules of thumb is if in doubt use a faster amplifier.  For example, our THS4541 will settle much faster than the THS4521 and the THS4541 has been optimized for temperature drift as well.  (Note that the THS4521 is specified for a 1k load while the THS4541 is specified at a 200 Ohm load.)   

     

    Regards,

    Loren

     

  • 0 in reply to Loren Siebert 1
    Prodigy 180 points

    Thank you, that's very helpful. I was wary of looking at even faster amplifiers as the input impedance usually tends to get very low, potentially requiring an extra buffer to drive it, but I note that the THS4541 actually has a slightly higher input resistance of 110k ohms compared to 100k for the THS4521. I'll give both a try.

    Tony

    PS. The resistance specification for the THS4521 at 3.3V is missing in the input impedance spec - the datasheet only says .7pF