[time-nuts] Question about effect of sample interval on ADEV
Tom Van Baak
tvb at LeapSecond.com
Tue Jul 9 16:37:10 EDT 2013
If you're familiar with the windows or unix command line then the tools and methods I use may be useful. Otherwise perhaps it's best to use standard GUI packages, like HyperTerminal, CoolTerm, Plotter, Stable32, TimeLab, etc.
Note that TimeLab includes data capture capabilities for all the common instruments that we use so you don't need the CoolTerm step at all. That is, Timelab will directly capture data from a hp 53132 either over serial or GPIB/Prologix. So I highly recommend this approach for newbies and oldtimers alike.
The 0.1 second jitter seems high. Yes, perhaps that is a problem with CoolTerm itself. Again, if you use TimeLab to acquire your data you can avoid using CoolTerm completely.
If you are a programmer or windows command line person, the tool I use to collect, timestamp, and log all my raw data is comlog. It's under my www.leapsecond.com/tools/ directory.
----- Original Message -----
From: "Hans Holzach" <h.holzach at vtxmail.ch>
To: <time-nuts at febo.com>
Sent: Monday, July 08, 2013 10:03 PM
Subject: Re: [time-nuts] Question about effect of sample interval on ADEV
>> FYI: I tend to record all serial (RS232/GPIB/USB/LAN) data from counters, analyzers, receivers, environmental sensors with a precision MJD prefix. This allows both tight correlation among different instruments in the lab and also allows ADEV-like tools to estimate, and then gradually refine to high levels of precision, the actual data rate, during data collection. Yes, it adds a few extra bytes, but it can be valuable information sometimes and storage is cheaper than it was a decade ago.
> as a newbie i'd be happy about a few hints how you do this. i use a prologix controller to read data lines from a 53132a counter. the data is then recorded by a terminal application (CoolTerm). the terminal application can add a timestamp to each line. however, even when in time arming or external arming (1pps from gpsdo) mode, the time stamp intervals vary significantly by up to maybe 0.1s, probably because the stamp is given by the terminal application and not by the counter. furthermore, ulrich's plotter program can't read CoolTerm's timestamp format, so i have to manipulate it in excel first before ulrich's plotter can process it...
> thank you,
More information about the time-nuts