Problems with frequency stability
Good afternoon, all
I am using my Kiwi for some QRSS beacon monitoring. Unfortunately, I am seeing frequency variances of several hertz that happen pretty regularly over the course of a minute. Obviously, these could be due to temperature changes where I have the unit located...but I have the GPS antenna hooked up and GPs locked in and it still wobbles in frequency. Yes, i know the drift is only 10 Hz, but that's a lot for QRSS.
How can I get additional stability?
Mark Lunday, WD4ELG
Comments
Hi Mark,
If the GPS is locked, the error should be well below 1 Hz on 14 MHz.
I remember that you wrote on the wsprdaemon mailing list, and they found that your frequency was off. Have you found out what the reason for that was?
Is the GPS locked all the time, with about 30 fixes per Minute?
Hi Stefan. The frequency errors appeared to be fixed once I upgraded to the latest version of wsprdaemon.
The GPS lock comes and goes....sometimes I am getting 30 fixes per minute, other times like now I am not receiving any GPS satellites at all.
I have an extra antenna, I am wondering if that might help to replace the original antenna.
Mark
I have a Leo Bodnar GPS standard also in the shack...is there a way to feed that output into the KiwiSDR in place of its onboard GPS receiver?
The GPS needs to be locked all the time, otherwise it will drift very quickly.
You can feed an external 66.666 MHz clock to the kiwi, on the kiwi 1 you have to solder a cable or connector to the J5 pad on the PCB:
https://www.klofas.com/blog/2022/adding-bodnar-gpsdo-to-kiwisdr/
I have a Leo Bodnar GPS standard also in the shack...is there a way to feed that output into the KiwiSDR in place of its onboard GPS receiver?
Kiwi-2? Yes. There's a connector marked "EXT CLK".
Kiwi-1? More difficult. Requires you soldering a JST connector (very tiny) onto existing pads on the Kiwi-1 PCB. Or perhaps solder the end of a small coax (e.g. RG174) to the pads (coax must be properly strain relieved so as to not pull the pads off the board).
But if you have a good number of GPS sats being received you should not be seeing 1 Hz deviations. I would like to see a stability plot of WWV/H instead. Are you using the supplied "hockey puck" GPS antenna? Does it have a clear view of most of the sky?
Thanks for the replies.
I was using the patch GPS antenna that came with it, which looks exactly like the one:
Active GPS Patch Antenna with SMA Connector : Leo Bodnar Electronics
It's definitely not the hockey puck like this
Garmin GPS 18x GPS Puck Receiver with USB Connection 010-00321-31 | eBay
(Does the antenna type really make a difference? They are both non-directional, so it should not matter I would think)/
The antenna is sitting on an inverted cake pan right outside my 2nd story shack window, with a view of the southern sky from 090 to 270. A few oak trees at 090 to 110 AZ, up to 60 degrees elevation. Other than that, all other obstacles are at or below 20 degrees elevation (other houses, trees across the street).
There is a check-box labelled
Acquire if Kiwi busy? [n]
No is the default, but I must have checked YES in the past because it was checked. So I set it back to default (NO).
So whatever that did, combined with the new patch antenna, seems to be helping....the frequency seems to be much more stable; the QRSS signals I am monitoring are GPS-locked, so I expect them to remain constant/steady. I will do a test of WWV to confirm.
If I understand correctly, you only see the southern half of the sky. That could sometimes be tight, depending on how good the coverage actually is in that part. Maybe you can upload a screenshot of the GPS coverage (Az/El tab in the GPS menu)
The length of the coax to the antenna can play a role and of course if it is damaged.
The mentioned checkbox should not play a role in the problem, this is only relevant if there are enough satellites locked anyway.
Thanks Stefan. Attached is the GPS history for 5 days. The coax to the antenna (and the antenna) are brand new.
Also attached is the WWV for the past 12 hours, for 10 MHz (please ignore the absolute values in the graph, what's important are the variations, which are showing +/- 4 Hz). While there is always a doppler effect depending upon the WWV signal, I believe these are due to the loss of GPS lock.
You only have 3 sats in the az/el display which is the bare minimum to maintain a lock (assuming they're strong enough to be tracking & decoding properly).
What happens if you turn off GPS, do a manual calibration, and just let it free-run?
I will disable GPS and try manual calibration to see how that does. I will report back with results.
does your GPS puck have an omni view of the sky and if not, which direction does it favor.