We all know about the Roaming Gnome. He's the little guy who is never still, but rather travels the world sending back post cards from the numerous exotic places he visits, when in fact he should be sitting still. Rock still, in the garden. The gnome's travels may be humorous but when your radio frequency travels, all humor goes out the window. The dial is supposed to stay put, just as the little concrete guy is suppose to stay put.

A couple of years ago I obtained a very nice and very clean Yaesu FT-902DM. Now this is one fantastic radio. Not only does it, unlike so many other transceivers, have the power supply built right inside the case, it covers all of the WARC bands - all bands from 160 through 10 meters - and operates all modes including AM and FM. Great for 10 meter operation. I won't outline all of the features. Simply take it that it's a very nice radio. It's from an earlier era so it does contain valves (ok, tubes for us on this side of the pond) in the transmitter (driver and PA) so more "tuning" is required than for current solid state rigs. Yes, the receiver is all solid state. It's been a fun radio to operate and makes a great rag chew rig for such bands as 17 meters.

Alas, all is not perfect with this radio. It suffers from a bad case of frequency drift. It requires a very long warm up time to stabilize enough to hold comfortable QSOs. Naturally following a station that drifts away from you while you're trying to carry on a chat is annoying.

The drift is downward in frequency and amounts to several KHz over a period of time, and is consistent on all bands. I determined to resolve the problem by finding the source of the malfunction and correcting it. I said that because some of the advice I received when I inquired on a couple of Yaesu email lists was to take the easy way out and simply get and use the external VFO which mates with the rig. Now having that VFO is certainly ideal and someday I'd like to get one but not for the reason given. The external VFO would allow split frequency operation further adding to the rig's utility.

I monitored the output of the internal VFO and noted that it increases in frequency. By the way, I must add that the drift is heat related. With the filaments turned on and especially while transmitting, the heat buildup is greater and thus the drift is accelerated. That told me on thing. The VFO is using temperature compensating capacitors and the compensation is too much. Surely, it did not leave the factory in that condition, so one must presume that the values selected for the circuit are correctly arrived at. Probably, component values have changed with age. Something has gone bad.

The original complement of capacitors within the tuning portion of the VFO, besides the variable ones, consisted of (in parallel) a 33 pf NPO, a 15 pf NPO, a 3 pf N750, and a 12 pf N750. A total of 15 pf N750 temperature compensation exist.

I ran a couple of tests. I removed both of the temperature compensating caps. but to insure that the total capacitance remained the same, I temporarily tacked in a 15 pf NPO. As expected, the VFO drifted down in frequency. Remember, the VFO had drifted higher in frequency originally. I then added only the 3 pf N750. The VFO still drifted up, but of course not as much as with no temperature compensation. I needed more.

I replaced the 3 pf with the 12 pf N750 and now the drift is downward again. Not nearly so much as at first. Now the drift is much slower (less distance over time) and ends up no more than about 1.5 kHz over a 12 hour or more time period. Not bad, but still not good enough. Far from it.

By rights, something less than 12 pf and more than 3 pf of temperature compensation is needed. Either a capacitance value of perhaps even one of like capacitance but a lower compensation. A N220 perhaps.

I don't know what the final solution will be yet. My next steps will involve some experimentation with different values of temperature compensation caps. At the same time, I will be looking further into the existing components seeking a clue about the real cause of the problem. The caps. I've checked so far have measured very close to their stamped values. That does not mean (if it's possible for this to happen) that one of the temperature compensation capacitors has changed its characteristics. I'll update this article as work progresses.

August 13, 2004. I'm a week or two behind in adding this update. The good news is that the radio is back to normal operation. I removed the recently installed 12 pf N750 and put the 3 pf N750 back in. As before, the drift reversed direction. Drifting up in frequency. I installed - mainly because it was the closest I had for something less than the previously installed 12 pf, an 8.2 pf N220. I'm not sure I could have come much closer had I a large assortment of sizes to work from. Drift now for all practical purposes is gone. After about an hour of warm up time, it's up in frequency only 300 Hz. The manual lists the stability as less than 300 Hz after 10 minutes of warm up and less than 100 Hz after 30 minutes warmup.

I set the dial to read 14.275.0 (no particular reason for picking that frequency) and after 24 hours, the frequency was rock solid at 14.274.8. The warm up time included having the filament switch on to help insure that the heat build up within the radio would be similar at least to normal use. The 200 Hz change, I can live with. It's certainly a vast improvement over the original drift which would have been perhaps a couple of kHz or more.

Now with the radio back in good working condition there is one more thing I am going to have to attend to. The two PA tubes appear to be soft. In the tune mode, the power after tuneup reaches perhaps 90 watts but falls off within a few seconds to perhaps 60 to 70 watts. So an investment in a new matched pair of 6146 tubes is in order.

Questions? Comments? Contact me.