In March 2024 I purchased a second hand Yaesu FT991 (known as a “shack in a box radio” because it works all bands, HF 100w to VHF 50w / UHF 50w + C4FM) from a second hand dealer who described it “as in NEW condition”!
It arrived and performed very well on VHF and UHF. The only antenna available at the time was the flowerpot. It was in daily use in the shack and the C4FM worked well, until just before the SERG convention. I did notice the chassis got very warm with long VHF overs, even at 5 and 25W! No sign of the fan running at all either.
Lets check this out.
Top cover off, another cover off and all is revealed. Huh, why is the fan plug dangling in free air? Removed by someone because it was noisy? Or some other reason? The reason (I surmise) becomes evident later on in the story. Plug the fan back in, it works just perfect, only runs when the chassis is warm, quiet at low speeds and then increases speeds with overall chassis temperature. Nothing wrong here. Chassis is a massive block of cast and milled metal, with careful design of the internal cooling fins, fan placement, and of course electronic control as required. Note: The fan plug was removed and the cable wound around other cables to keep it out of the way. There was no way it could have fallen out from the original socket. Fee Fi Fo Fum, I smell a human at work! Two nice shiny dual HF Mosfet amplifiers look like they may have been replaced at some time too.
I then decided to do some bench testing on HF into a dummy load, prior to installing a HF antenna.
Everything worked well, BUT within seconds there was a VERY STRONG burning smell, on the 7Mhz band only (Why is there always a BUT?) The internal tuner appeared to tune all the bands ok, including the smelly one, and it was, I guessed, using only about 10 watts. Further testing TX CW or FM from 5W to 50W was perfect on 7Mhz, then increasing the power further produced a tiny squealing noise from inside the radio ( a cry for help?), power went lower, and then danced around rather than up, DC current did some strange things, not consistent with the power output too. Hmm, not good. Back in the cupboard, feeling a bit disappointed at the time.
Recently with some time on my hands, it seemed opportune to delve further.
First thing, examine the block diagram in conjunction with the detailed circuit diagram. The RF output Mosfets are common to all the HF + 50Mhz bands. So what is specifically just for the 7mhz band? The tuner block is switched in or out by a pair of relays, and it had no affect on the problem = not that.
Divided into 7 bands, each has a bandpass filter taking up quite a bit of PCB real-estate. Solid coils, 3 per stage, in and out relays, and capacitors. Perhaps it was a stuck relay? No sign of anything evident physically on the top pcb area. Thats a pity.
Two options left.
1) Use the TX until the smoke is visible or a small bush fire erupts, showing the real fault? Don’t laugh. This technique had to be used more than once locating faults in “the old days” in TV sets. Arcing triplers, tube flashover, sometimes the only way. Fault finding requires using ALL the senses, hearing, sight, touch (ouch that thing is bloody hot!) and even listening to the customer, in extreme cases.
2) Or preemptively remove the board for examination underneath, which may or may not be a waste of time?
Well, in case option one may kill a pair of expensive output fets, there was nothing for it but to remove the whole HF PA / bandpass filter assembly. Yes, I was dreading that at first. However with 99% of the connections being plugs and sockets on all control and RF cables, and only one PCB to PL259 socket needing desoldering, it didn’t take too long. Just a bunch of screws to store.
Like the magician, the “trick” was revealed as soon as the pcb was out and turned over.
There was even a solid carbon track wending its way directly to the fan. That had to be scraped off the pcb. So this was something that had been happening for a long time, just getting worse and worse.
It was revealed that the two SMD capacitors, in parallel with the third inductor, had become a charcoal BBQ brick, cracked in half and still refused to die, sat there arcing / burning merrily away into the pcb, which was fast on the way to become a blob of charcoal too.
Now WE know the problem. I guess someone may have replaced the output Mosfets, which didn’t fix the problem. Then the fan was unplugged (to stop the smell reaching the outside world too quickly). Then the radio was moved on.
Examining all of the nearby bandpass components, it seemed feasible to pinch the components from the next range, then replace them later on, as a way to effect a repair to the 7Mhz band, for now. That was quite easy, and checking my stock, it seemed I had very similar sized and rated components anyway. Putting the original Yaesu components in for 7Mhz, and the slightly unknown ones in the 3.6Mhz block (80M seems to be a noisy band at this location 24hrs of the day anyway) would be worth a shot. That was completed in quick time, the PCB was re-assembled back in the diecast box, cleaned and new heatsink compound applied, screws torqued down nice and tigh. Then it was then testing time.
What a joy. Now we have ALL bands CW/FM performing from 5 to 100watts, no interior noises at all, and DC current following along as it is supposed to do. Better still, NO burning smell, or in fact any smell at all. So now the extreme test. 100W on 1.8Mhz then 14Mhz then 50mhz, for 40 minutes each time and repeat. (Rubber band on the microphone PTT trick and disabled TOT). No sweat. Fan operation and the cooling design is absolutely superb, just raised the chassis by a few degrees C. I could hardly feel any increase on the case and rear accessible diecast chassis.
Remember I said look for HOT things. That was the 150W dummy load! It probably rose about 50C. Ouch. That’s why it has a RED warning, “this item will have elevated temperatures in use“. NO sugar Sherlock.
After a few more hours of TX (at full power) and RX sensitivity testing, I am prepared to say, job done.
Note: While this may read a straight forward repair, the reality was anything but. More than two days (and nights) studying the block diagram and cct. Guesswork and luck over the best path to take, after considering the known symptoms. Then the luck at the level of damage with the chance of repair, followed by the actual repair.