[Lowfer] WE2XGR/2 frequency & sampling rate cal signal test

[Garry Hess]

Jay, I tried to make the measurement you were after but ran out of time.

I've been using an SDR-IQ exclusively for some time now and have it 
clocked by an external 50 ppb OCXO. Not quite as accurate as GPS, but 
darn close. To calibrate the setup a GPS-referenced 10 MHz signal is 
used to determine the clock frequency to within 1 Hz (that's the 
precision for the clock frequency using SpectraVue 3.00 software). Then 
with a GPS-referenced 100 kHz signal the sound card sampling rate is 
adjusted so Wolf indicates zero error. This was last done Jan 18, 2012. 
Tuning in your test signal on 506.5 kHz AM a picket fence of 120 Hz 
harmonics was noted but nonetheless Wolf picked up the 800 Hz tone and 
gave the following:

Loop antenna to SDR-IQ 506.5 kHz AM detector

2012-03-04 01:02:45 >WOLF10  -r 11100.274 -f 800 -m 60 -w 0.0000
01:03:45 f: 0.004 a:-0.5 dp: 22.5
01:04:45 f: 0.004 a: 0.8 dp: 20.1
01:05:45 f: 0.004 a: 2.1 dp: 18.6
01:06:45 f: 0.004 a:-2.9 dp: 21.0
...

Next, the sound card sampling rate to be used with line-in rather than 
USB input was checked. This was done by applying a GPS-referenced 1000 
Hz tone to line-in and measuring the frequency with Wolf:

1000 Hz GPS-referenced tone to line-in

2012-03-04 01:14:25 >WOLF10  -r 10800.668 -f 1000 -m 60 -w 0.0000
01:15:25 f: 0.000 a: 2.0 dp: 64.7
01:16:25 f: 0.000 a: 1.9 dp: 64.7
01:17:25 f: 0.000 a: 1.7 dp: 64.7
01:18:25 f: 0.000 a: 1.6 dp: 64.7
...

Now the analog receiver was turned on, an Icom 706MkIIG, and that was a 
problem. That receiver never stabilizes well because an internal fan 
kicks on and off, even when in receive mode. The LO frequency spikes 
from that fan can be avoided by blowing air over the radio with an 
external fan that is continuously on (confirmed by nice straight lines 
using ARGO on QRSS60 sources). However, it takes quite awhile for that 
to stabilize and there wasn't enough time. On top of that, this radio is 
quite deaf below 1 MHz so AM detection didn't produce sufficient tone at 
800 Hz. Instead, the radio dial was set to 506.3 kHz USB to give an 
expected tone frequency of 1000 Hz, subject to LO drift. Wolf results were:

Loop antenna to IC706MkIIG 506.3 kHz USB detector so 800 Hz tone
on 506.5 kHz carrier appears as a nominal 1000 Hz tone

2012-03-04 01:48:16 >WOLF10  -r 10800.668 -f 1000 -m 60 -w 0.0000
01:49:16 f: 0.245 a:-2.0 dp: 37.6
01:50:16 f: 0.247 a: 2.4 dp: 38.9
01:51:16 f: 0.245 a: 1.4 dp: 39.7 external fan turned on
01:52:16 f: 0.222 a: 0.4 dp: 39.4
01:53:16 f: 0.210 a: 1.6 dp: 42.0
01:54:16 f: 0.203 a:-1.0 dp: 39.8
01:55:16 f: 0.194 a:-1.1 dp: 35.7
01:56:16 f: 0.181 a:-2.7 dp: 31.0
01:57:16 f: 0.183 a: 2.6 dp: 35.4
01:58:16 f: 0.181 a: 2.3 dp: 38.3 ~ stabilized by now??? Not really!

A final measurement using the IC706MkIIG tuned to 99 kHz with a 100 kHz 
GPS-referenced signal gave:

GPS-derived 100 kHz to IC706MkIIG 99 kHz USB detector to produce
nominal tone of 1000 Hz

2012-03-04 02:03:13 >WOLF10  -r 10800.668 -f 1000 -m 60 -w 0.0000
02:04:13 f: 0.002 a: 3.1 dp: 76.0
02:05:13 f: 0.001 a:-2.3 dp: 76.0
02:06:13 f: 0.001 a:-1.9 dp: 76.0
02:07:13 f: 0.000 a:-1.4 dp: 76.0 still slowly drifting down in frequency

One might expect the tuning error to be proportional to carrier 
frequency but the results at 100 kHz are pretty good while the results 
at 506.5 kHz indicate substantial tuning error from the dial display. 
But due to insufficient warmup I don't trust the latter measurements.

The 706 radio uses a 30 MHz LO and the SDR-IQ external clock of 60 MHz 
is derived from a 10 MHz OCXO multiplied by three, then doubled. Thus, a 
good 30 MHz signal is already on hand to injection lock the 706 LO. One 
of these days I may get around to doing that so drift isn't an issue any 
more. Even if the radio isn't used at LF it's a good performer at HF and 
there the drift issue is magnified.

-- 
73, Garry, K3SIW, EN52ta, Elgin, IL