[time-nuts] Re: Measuring oscillators against GPS - Trimble NetRS

[Tom Wallace]

I've been using the NetRS for timing for a couple of years now. They are 
inexpensive and work well when properly configured, but they do have 
their quirks. Some notes I've found useful:

1. The NetRS needs a dongle to provide primary power and the Ethernet 
interface. Approximately 100% of the ex-USCG receivers on the well-known 
auction site come without the dongle, so plan to locate one elsewhere or 
build one (pinouts are in the manual).

2. UNAVCO's resource page on the NetRS 
(https://kb.unavco.org/kb/article/trimble-netrs-resource-page-471.html) 
is a trove of useful information (manuals, FAQ, etc.) If that page 
didn't exist, it would be a major undertaking to get a NetRS working well.

3. The latest firmware (1.3.2) is available from UNAVCO 
(https://kb.unavco.org/kb/article.php?id=770). It adds the ability to 
receive the L2C signal in addition to the L2P(Y), and corrects some 
security vulnerabilities. It can be applied no matter what the unit's 
warranty status (all the surplus ones are out of warranty).

4. For timing use, clock steering on the NetRS must be turned off. 
Otherwise, the clock values are smeared by the steering and don't 
reflect the state of the 10 MHz input.

5. Converting the binary output of the NetRS to RINEX is a multistep 
process: use runpkr 
(https://kb.unavco.org/kb/article/trimble-runpkr00-latest-versions-744.html) 
to convert to .tgd format, then teqc 
(https://www.unavco.org/software/data-processing/teqc/teqc.html) to 
convert to RINEX.

6. Computing clock stability from PPP data can be done over the net 
using NRCan (https://webapp.geod.nrcan.gc.ca/geod/tools-outils/ppp.php) 
or locally using gLAB (https://gage.upc.edu/gLAB/). Comparison of these 
results gives good agreement (~1 ns) in the magnitude of the clock 
variations.

7. The absolute offset of a clock from the reference source is not 
easily determined using the NetRS; there is no consistency in the clock 
offsets computed by 2 NetRS systems using the same antenna and fed by 
the same 10 MHz reference. Interestingly, the differences are not even 
the same modulo 100 ns (the period of the 10 MHz signal).

8. The NetRS handled the last WNRO with no problem. However, I have been 
unable to find information on whether the system uses a pivot point to 
handle WNRO. If it does, and the pivot is the date of the last software 
release (October 2012), then the systems are good for another decade. 
However, I have no information on this, and would be interested to hear 
from anyone who does.

9. The sawtooth in the NetRS PPS output is very large (+/- 30 ns), and 
the output voltage is over 5V (it may have been designed to deliver 5V 
into 50 Ohms). If you want to discipline the oscillator you're 
measuring, I'd suggest checking the input level it can tolerate, or 
using a modern receiver like the uBlox F9 (which has around 1/8 the 
sawtooth variation).