First IF Stage testing and debugging

By | 2nd July 2017

Following the assembly of the first IF stage PCB I began testing and as expected it did not work.

The first set of bugs turned out to be the FET’s were in the wrong way, I placed them according to the legend in my PCB software but the device legend was incorrect – thankfully no damage was caused and once I re-placed them in the correct position I started to see life, however the gain was not very good as I should have seen around 50-60dB from the first amplifier to the last stage but at best was only around half that. A slight improvment has been noted since changing the L-Match inductor but nothing to get excited about so the debugging will continue.

One of my readers, Andy G6LBQ noted that the L-Match inductor was incorrect as I had, for some unknown reason, chosen 3.3uH instead of the correct value 6.0uH, so I wound 4 turns on an FT37-43 toroid which gave me 5.89uH according to my LC Meter – close enough for me.

So here is a quick video of the IF stage in action, I injected 450KHz plus 10.250MHz from the AD9959 DDS unit in to the mixer stage and displayed the output on the Spectrum Analyser, when I looked at the output on the scope I noted around 960mV Peak to Peak which does not seem right – barely even 0 dBm.

By my calculations I am measuring 42mV peak to peak at the Crystal Filter output in to the first amplifier stage and seeing 960mV peak to peak at the output so that would suggest a voltage gain of 27dB which does not come close to the spec in the original design I used as the cornerstone.

Having spent many hours debugging this board over the weekend I am no further forward, 48mV from the Crystal Filter and only 960mV at the final output – not acceptable.

Next step will be to change the FET and bipolar devices as I suspect they may not be genuine parts as I bought them from Ebay, I may be proved wrong howver this game is process of elimination so I will order some new devices from Radio Spares or Mouser.

Andy G6LBQ mentioned the AD603 device which looks to be a good option however at £12 each these seem rather rich, that being said I will buy one and do a lash up and if it performs as well or beyond expectations I may well change the design to accommodate. It is such a shame the MC1350P is obsolete as this would have been just the ticket, if I can find a stock of these at sensible prices then I maybe persuaded.

Moral of the story – breadboard the design before committing to a PCB!

Leave a Reply