Saturday 3 July 2010

Frequency Stability & Palm Trees in Stockport

Having reported making a start on the PCB-based version of the WSPR generator last week, I'm sorry to have to report today that the completed system was a great disappointment.

You can see the board here...

It worked - but not very well.

I found the AF oscillator to have very poor frequency stability (much worse than the initial version on the Breadboard).

In trying to track down the origins of the problems, I also discovered that the entire oscillator is (in my opinion) a bit "dodgy". The first cut of the circuit on the PCB didn't run at all - I found that this is due to a sensitivity to power supply voltage (which needs to exceed 10V for some embodiments of the circuit to oscillate reliably - my Breadboarded version was the exception rather than the rule in being able to run at 8V). The performance is a strong function of supply voltage - below a critical point (~8V) the oscillator doesn't run at all then, on increasing the supply voltage, it runs but with an irregular output (a sort of chaotic fluctuation modulating the desired periodic output). Pushing up the supply voltage still further (11 - 12V to be sure) takes the system into stable periodic behaviour.

Even once I'd realized that the supply needs to be high enough to get reliable oscillation, I still couldn't get any long-term frequency stability on the PCB-based version - so I gave up!

I turned my attention back to the breadboard version and connected up the minimum sufficient number of elements from my increasing collection of "beacon building blocks" (strictly, that's not true - I didn't need the bandpass filter, but it was already connected between mixer and amplifier, so I just left it there)...

There's (left to right) the WSPR Generator, the SDR-Kits USB Synth, the Mixer, the BandPass Filter, the 'Radio Times' Amplifier and the LowPass Filter. I turned on and was immediately rewarded by a spot from Reimund, DL6YCU...

The plan now (that I've learned the nuts and bolts from Gene Marcus, w3pm's design) is to program my VCXO-based system to generate the WSPR signal - I KNOW that has very much better frequency stability from the results I've achieved with QRSS and S/MT Hell. In fairness, Gene does offer a similar solution on his excellent WSPR Projects Page.

Today was the Stockport Radio Society's First Stockport Rally (the continuation of the annual event that used to be called the Reddish Rally). Usual mix of junk, amidst which I was delighted to find a Palm Radio Mini Paddle...

... which I picked up in "as new" condition for thirty quid. Not a bad deal, as Martin Lynch are today advertising them at £75.96.

Here's the new paddle atop my nearly-new FT 817...

Regular readers of these ramblings (haven't you got anything better to do) will remember I've been drooling over these keys for ages - ever since seeing Juan, ea5elp's videos of working /p on the beach. Great result - I look forward to the second Stockport rally !

...-.- de m0xpd

Update :-

I left the system on overnight and got the following outburst of acne (what else is the collective noun for a bunch of spots)...

Nothing too impressive (certainly not as good as previous H/B WSPR attempts) but proof of some limited level of success. Best dx was 2043 km to la9jo in arctic Norway.

1 comment:

  1. Nice find. I've been after one of those for some time and been put off by the price. So I'm a bit envious. But to be honest CW is out of favour with me at the moment so you'll probably make better use of it.

    Bit unusual, isn't it, having a rally on a Saturday?