Previous installments of this series (1 and 2) have documented a frequency wobble I'd noticed with an SDR-1000 radio when trying to receive QRSS with it. This weekend got to working on it again and I believe I've traced down the source of the wobble!
The first thing I wanted to eliminate as a source of the wobble was the power supply in the radio. This radio has a 5 volt regulator chip that runs a bit hot (I measured 80 C at the heatsink mount), so I decided to make sure it stayed nice and cool. Did this by taking the regulator off the PIO board and mounting it on an old Pentium 4 heatsink and fan combo I had in the junk box.
As can be seen from the picture this has got to be one of the best heatsinked 5 volt regulators around. After getting the radio put back together I fired it up. With the heatsink/fan combo and the fan in the front of the case, the regulator chip stayed nice and cool. However after letting the radio warm up and stabilize, the wobble was still there. It did not seem as severe as it had been before the regulator change, but still on the order of 3-4 Hz.
On a whim, I decided to try running the radio with neither of the fans powered up. After watching the trace in Spectran drop a bit less than 200 hz over an hour I got a pretty straight looking trace -- no sign of the severe wobble I'd been seeing. The screen shot below is a Spectran trace (0.17 Hz resolution) of WWV on 15 Mhz after the radios been running for about an hour without the fans on.
After seeing this I plugged in the case fan and the wobble returns. Tried it with just the fan on the 5v regulator heatsink, wobble returns. So my conclusion is that the fan is the cause of my frequency wobble!
I initially thought the fans might be putting enough load on the 13.8 volt supplying the radio to make it a bit flaky, I changed supplies to a 30 amp linear supply I use with one of my other HF radios. Still wobbles if any of the fans are on.
So at this point I know what's causing the wobble, but not why. I've got three ideas for that at the moment. First is perhaps the fans are putting out enough electrical noise that they are interfering with the SDR-1000. Second possibility is that the vibration of the fans is disturbing the radio. The SDR-1000 is a direct conversion receiver and they have a reputation for being subject to microphonics. Final thought is that the fans keep the air moving around so much in the case that the circuits never really come to a stable operating temperature.
The SDR-1000 has a provision to drive it with an external reference oscillator. Thinking of trying either an external temperature compensated oscillator, or a GPS locked reference oscillator to see what that does.
I will say this bit of debugging has been interesting. The fan would have been my last guess for the cause of my wobble.
Update 3 October 2004: Did a bit more work on this and have damped out the wobble substantially. Here for more information.
Comments to: Bill Tracey (kd5tfd@ewjt.com)
Last Updated: 3 October 2004
Copyright © Bill Tracey 2004