Its now February 2007, traditionally the time of year for those of us who construct our own equipment, to do the constructing ... Hmm?
Lets just say that the Ionica enclosure for my mast-mounted 9cm transverter failed the water ingress test . . . twice! But lets be fair about it, on each occasion, the rain had been accompanied by 50-60 mph winds . . . and I was partly to blame for the first incident by failing to seal off a couple of unused holes in the PA heat-sink. Prior to the second incident, Nick, GM4OGI and myself had been having great fun on 3.4GHz during a heavy sleet shower . . . the stronger the wind the greater the Doppler shift . . . the less intelligible the signal. The weather really was atrocious. But something was definitely wrong. After a while, maybe 30 minutes, my Local Oscillator began to make dramatic shifts during transmit, then reset itself during the Receive period to the extent that nick and I ended up chasing each other all over the band. I was monitoring my LO
on my IC706, so I knew the fault rested with me. After several days of torrential sideways rain, it was clear that the transverter was no longer generating RF, so during a temporary lull in the weather, I ventured up onto the roof and drilled a hole in the box. I estimate about 100ml of water poured out. I retested the box after allowing it time to dry but it still wouldn’t work. So It was back up onto the roof . . .
Condensation on the PA block and evidence of a fire!
The obvious question was ‘what was the most likely cause of the failure?’ The maximum current that the NMH range of DC-DC converters can deliver is 60mA per rail. Ideally I should have used an NMS1212 (84mA per rail) but I had an NMH1215 in stock, so I stepped the produced -15V down to -12V via a regulator. The NMH1215 showed no signs of distress during testing. However it is very possible that with the feed-through terminals sitting in water latterly, that extra current was drawn from the -12V line, over and above that which the PA requires. This extra demand may well have over-stressed the device resulting in a spectacular catastrophic failure. Fortunately there is a built-in fail-safe on the PSU board which automatically protected the PA. It is very clear that the structural integrity of the Ionica enclosure was suspect. It certainly had difficulty in fending of the Scottish weather, so I reluctantly opted to rebuild the entire system as a shack-based transverter. The coax run to the mast on the gable end is not too long so losses would be kept to a minimum. In stripping down the transverter I split the PA block so as to check for any signs of water ingress. Although there was no sign of direct water damage, there was a worrying amount of crystal growth on the inside of the milled aluminium lid. This had not been present before mounting the unit outside and was clearly the result of condensation. Looking back, and taking into account the size of the heat-sink and the amount of aluminium in the PA block itself, there would be a huge amount of thermal inertia involved. With temperatures at night dropping to close to zero Celcius, the PA block would require a lot of energy to warm it up. During the transmit cycle heat is produced, and since the PA block would be very cold, condensation would almost certainly be produced. This confirmed my decision to go for a shack-based approach.