N210 Unsuccessful Ref Lock

GNU Radio
1

Hi all,

My research group is continuing our setup of two servers with multiple
N210s for a GNSS application. Each server will have 4 or 5 N210s
recording
RF data at frequencies of interest. All are to be locked to the 10MHz
reference and 1 PPS output from a Septentrio PolaRxS. I have
successfully
tested our PPS arrangement using test_pps_input. However,
test_dboard_coercion sometimes reports an unsuccessful lock to our
external
oscillator. The number of unsuccessful locks varies with the inputs to
test_dboard_coercion:

Set1: “–rx --ref external” reports an unsuccessful ref lock at 1500MHz
Set2: “–rx --ref external --no_rx_gain” reports an unsuccessful ref
lock
at 1500, 1550, 1700, 1750, 1800, 1850, 1900, 1950, and 2000MHz.

These unsuccessful locks are intermittent. The results for Set2 don’t
worry me so much: we’re always running at high gain when recording.
What
is the cause and is there a good solution to this problem? Our plan is
to
have all N210s writing constantly to a ring buffer, but only writing to
file when triggered by an external source. The reference lock is
critical
since we’re studying GNSS signals. Any help is appreciated. I know
this
list usually hits it out of the park, so I’m looking forward to your
responses!

Thank you,
Ryan

Setup: Dell PowerEdge R510 running UHD 3.4.2 on Ubuntu 12.04 server.

2

help is appreciated. I know this list usually hits it out of the
[email protected]
Discuss-gnuradio Info Page
After splitting your 10Mhz reference, what is the power going to each
N210?

3

The clock signal has not been split yet: I’m only testing one N210 to
verify a reference lock. We will be using a SRS FS735 distribution
amplifier for clock splitting. It outputs 13+/-1 dBm to a 50 Ohm load
from
each output.

The current dboard is an RFX1800, but we will be using a combination of
1200s and 1800s depending on the signal we’re capturing. We changed the
sleep time in test_dboard_coercion to 5 seconds instead of 1. I’m going
perform some more thorough testing, but this looks like a good fix!

Thank you,
Ryan

4

On 06/11/2012 05:18 PM, Ryan Wolfarth wrote:

The clock signal has not been split yet: I’m only testing one N210 to
verify a reference lock. We will be using a SRS FS735 distribution
amplifier for clock splitting. It outputs 13+/-1 dBm to a 50 Ohm load from
each output.

The current dboard is an RFX1800, but we will be using a combination of
1200s and 1800s depending on the signal we’re capturing. We changed the
sleep time in test_dboard_coercion to 5 seconds instead of 1. I’m going
perform some more thorough testing, but this looks like a good fix!

Yea, sorry. It looks like there were a few things a little screwy with
the lock detect checking. I will push something tomorrow to make it a
little more consistent.

Thanks!
-Josh