delay measurements

[mourilo]

I'm trying to make a device that could measure the time delay between two signals. I have a reference signal of 10MHz and I have another delayed 10MHz signal. I need to make a precise measurement of this delay between two square wave signal which can as small as 0.17ps.

Most of the components that I've seen so far are too slow for this purpose (their rise time is in nano sec...wont have enough time to respond). Do you have any idea of which component can be accurate enough to measure a delay of 0.17ps.

Thank you for help in advance

 

[nbucska]

Is the signal digital or analog ? if dig. what logic family?
If analog, waveform, amplitude ?

What kind of resolution and accuracy do you need ?
One device or mass produced ?


<nbucska@pcperipherals.com>

 

[mourilo]

In my circuit I have two switching MOSFETS (vn0300l, they are not connected together), each MOSFET is eather on (0v across S and D of the mosfet) or off (5v). They are switching at the same frequency (exactly 10MHz) but are out of phase.
Having this two waveforms (more or less square waves coming from two different mosfets) I need to measure the delay between two waveforms. I need to detect the delay as small as 0.17ps, so the precision is a big factor.
One device or mass produced is not a factor
thank you

 

[47gandalf]

You are undertaking a big job to measure these within 170 attoseconds. Look into jitter measurements for high speed optical communications systems.

You're looking to measure to 1.7 ppm, or 170 parts per 100,000,000.

Good luck. It won't be cheap. And ask yourself how you will calibrate and verify it.

 

[Heydave]

The timing resolution you are trying to measure is about the time it takes a signal to travel .002 inches. How do you expect to probe such a signal and keep such perfect mechanical accuracy even if you figure out how to measure it. One possible scheme is to use an optical interference process and mechanically adjust the difference to balance the phase difference. My guess is that various sources of phase jitter in your signals will overwhelm your desired phase resolution. Nice challenge.
Good luck, you'll need it.

 

[melone]

Why does it have to be so exact? Can you just drive the outputs from a single clock source, and just use the current clocks (actually the gate voltages) to gate the output?