• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 424 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TLV3801: TLV3801 input noise

Thomas Lehner
Prodigy 20 points
Part Number: TLV3801


Hello,

I have got a question regarding the TLV3801.

We want to detect very small and very fast pulses. Height about 5 mV, length a few ns.  The TLV3801DSGT seems to be perfectly suited for this task.

We see two solutions:

1.) Use the TLV3801 directly on the small pulse.

We would need a reference voltage that is adjustable in very small steps but I see no problem there. The software could adjust the voltage to detect the pulse reliably and to have optimal jitter. Since the hysteresis is 2 mV, we should always be able to trigger a 5 mV the pulse somewhere, right?

2.) We use a high speed opamp to pre-amplify the pulse, say times 10.

This might be the more sensible solution, but we hate to do that because it takes away flexibility. In other conditions we want detect large pulses and solution 1 would allow us to do so. Also, since this unit is present 128 times, this adds cost and complexity.

My main question is about noise. According to my understanding, a comparator is very similar to an operational amplifier. Therefore I expect the input noise voltage to be similar. Is this a sensible assumption?

The input noise is not specified in the data sheet, but there is some (thermal) noise, right?

When the input noise of an opamp is significantly smaller, pre-amplification would decrease the overal noise. Noise translates to jitter and we want the jitter to be as low as possible.

Basically my question is this:

Will the jitter be larger without opamp? (How much?)

Thanks

Thomas

  • 0
    TI__Mastermind 26460 points

    Thomas

    Thanks for your post.  All excellent questions and definitely a challenging application.  With such a small input (assuming no amplification), I suggest you consider our TLV3811C which has zero hysteresis.  An application similar to yours is LIDAR where are customers attempt to detect signals as small as possible since the amplitude of refections can sometimes be quite small.  For a reference, possibly you will consider a DAC.  Possibly you can divide the DAC down to create more resolution and reduce its noise.  I will see if my design team can tell us the approximate input noise of our comparator.  Different than amplifiers, comparators aren't really intended to have the inputs with a near short across them (very low differential input voltage).  For example, if you were to short our inputs, assuming a well matched differentail input stage to our comparator, you would see something that looks like white noise.  Without any internal hysteresis, we will essentialy look like a linear amplifier.  Only when the differential input is large enough will the output saturate in a high or low state.

    In regards to your final question, I am not sure how to answer.  Minimum jitter would be obtained with the sharpest edges where the noise at the transition voltage is minimum.  If the amplifier helps create such a waveform from your input source, that should improve jitter at our comparator output.   Hopefully you can obtain our EVMs to run some experiments.

    I will try to get back to you with any noise performance details that my design team can provide.  Hopefully I can get some information by Monday.

    Chuck

  • 0 in reply to Chuck Sins
    TI__Mastermind 26460 points

    Thomas

    I heard back sooner than I expected.  Input noise is 250uVrms with 3GHz bandwidth.  My intuition is that adding an amplifier will minimize your jitter but I can connect you with our high speed amps experts if you wish to explore further.  Adding a little gain also allows you the opportunity to filter the signal a bit.  I assume at the least you can filter all the low bandwidth noise and maybe even bandpass filter the signal.  Just let me know and we can take this offline and I will loop in the appropriate folks.

    Thanks again for sharing your very interesting application.  Sounds like medical Slight smile

    have a nice weekend, Chuck