[time-nuts] Measuring coax temperature coefficient with a TICC

Tom Van Baak tvb at LeapSecond.com
Thu Apr 20 16:00:28 EDT 2017


> Please forgive my ignorance but what is a TICC?
> Regards
> John P

Short answer:

https://www.febo.com/pages/TICC/

https://www.tapr.org/kits_ticc.html

Long answer:

When working with precise time one of the very first instruments you'll want is a frequency counter. But when your clocks get accurate enough that your counter always reads 9.999 or 9.999999 or 10.000000000 a different measurement approach is called for. This is where a phase meter or Time Interval Counter (TIC) is useful: instead of making a frequency measurement for a fixed duration (gate time), you can instead continuously monitor the slow and steady drift in phase (time) between two sources; over seconds, over hundreds of seconds, even over days. Programs like Stable32 or TimeLab can make phase, frequency, or stability plots from the data set.

That's why a TIC can be used to measure coax temperature coefficient. A frequency counter can't do that.

Most of us get our counters from eBay where the favorite TIC's are SR 620, hp 5335, hp 5370, hp 53132. But it has always been a challenge to make a modern, cheap, homebrew TIC that matches commercial instruments in performance. Over the years a number of time nuts have tried to make sub-nanosecond counters. The most recent time interval counter by John Ackermann, the "TICC", is a winner. See the links above for details.

By using a recent off-the-shelf TDC (time to digital) chip from ti.com John avoided the complexity and analog calibration of earlier amateur designs. In order to allow arbitrary range of interval, he actually uses two TDC chips; one to precisely time the start pulse, and another to time the stop pulse.

Moreover, John's "TICC" exposes the underlying start and stop channel timestamps so that the board not only acts like a simple start=A stop=B TIC, but can also be used as a dual-channel time-stamping counter (TSC). Unlike commercial counters, both the A and B inputs can collect data simultaneously and independently if you want. This avoids certain deadtime, collision and sampling issues that many commercial TIC's have. So it's a very nice design, and low cost, and open source.

/tvb




More information about the time-nuts mailing list