Check the mounting screws of your transistors!

[Geert Jan PE1HZG]

Hi folks,

If you don't want to read long stories: may I suggest you open your HR500 and check, for each of the VRF2933 transistors, if both transistor screws are still fully tight. If you want to know why, read on...

I'm having a ball with my HR500, and it's a bit of a shame this group is so quiet. I've heard people complaining that Google Groups is hard to access - I had to go through a few hoops myself to be able to post this frankly - but I hope that isn't preventing you from enjoying the HR500.

Last week however, I had a rather unfortunate incident when one of the finals popped it's head. I was running a carrier of about 400W on 40m in a dummy-load for some tests. The levels were all safe, but the amplifier was running warm (60 degrees) and had put the cooling fans on higher speed.

Then I heard a "pop", the amplifier restarted completely and I quickly switched off!

Upon opening, I found that one of the VRF2933 transistors had popped it's lid off. Not nice, because these transistors are not cheap (but, in contrast to few months ago, at least they are available at digikey, mouser and the like).

After a night of mourning and asking Jim for advice (he was on his way to Dayton), I spent some time checking things out. With the failed transistor removed, and the amplifier running on a separate, current-limited power supply, I found that the other transistor would pull the bias current I expected so I hoped it wasn't damaged.

When removing the failed transistor however, I found that one of the two mounting screws was really loose. That could be an explanation: if the contact with the copper head spreader was impaired, the transistor would have run very hot.

I am certain tightened the screws really well (all four of them) but on the failed transistor, one was loose and the other three were not fully tight either. I am wondering if repeated temperature cycling had worked them loose a bit?

The HR500 manual just instructs to mount the transistors with  #4-40-x1/4" screws. Looking at the instructions of the Elecraft KPA100 kit I built some time ago, the transistors there were mounted using screws and split lock washers. Elecraft also put in a warning to re-tighten the screws after 5-10 hours of initial operation.
The HR500 instructions just say to use the screws, that's it.

With the way cooling for these amplifiers is done - with a sandwich of the transistor, the copper head spreader and the cooling fin, perhaps repeated heat cycles managed to make the screw work itself loose? There are different materials with different temperature coefficients in any case.

I tried to find pictures of other amplifiers and their mounting techniques. Most of them use a flat washer (to spread the force), a split washer and a screw or even a bolt and sometimes a torque driver (which I don't have).
This seems to be a bit of secret art because not many people describe how to build their amplifier, they just want to sell their box. Jim should be commended for guiding us to tell how to build ours.

I've changed my amplifier to use a flat washer, a split washer, and the earlier #4-40x1/4" screw on the places where the transistors are mounted. (would like to have used a bolt but getting imperial mounting hardware is difficult here).
Didn't use a torque tool because I don't have one, but I did use a PZ0 screwdriver that fit very well on the screw's head.

After replacing the transistor (remember to use a THIN layer of compound and to verify that the compound reached the full footprint of the transistor), and setting the BIAS again, all was well again.

Running power, indeed I found that the other, surviving final seems to be unscathed. After replacing the broken transistor, the PA is fully up and running again

I am explicitly not saying there is an issue with the HR500 and did not discuss this approach with Jim yet (Dayton outage) but if anything, I would suggest that people open up their HR500 and at least check the tightening of the transistors to avoid a costly experience like mine. Would like to read some more results on this one. And perhaps a few rings would save an expensive transistor swap.

73,

Geert Jan PE1HZG

[AC1KM]

Fascinating!

Mine had exactly the same failure last Sunday. (But mine also took out R10 and R12). I don't have any #4-40 lock-washers, but I applied some loctite blue after reading your post. My screws were still tight though when I removed the MOSFETs.

Quite the bang when one of those MOSFETs blow their tops.

Patrick/AC1KM

[Maurizio Brameri]

Good to know Geert.

I will check screws and I will try to use some threadlocker.

Meanwhile I will try to use a thermal past for microprocessors, I think better than the white silicon grease that is provided in the kit.

73 de Maurizio I2NOY

[Maurizio Brameri]

I opened this morning the amplifier to check the LPF and I find one screw that fixes the mosfet not completely loose, but not tight as I have tighten it about one week ago.

I will try with a drop of threadlocker.

73,

Maurizio I2NOY

[AC1KM]

Jim,

 Would you recommend refreshing the heat-sink compound at regular intervals? (biennial?)

Patrick AC1KM

[jcve...@gmail.com]

My HR500 has been in service since 2018 and I've never changed the heatsink compound. I don't use it all that much and hour a week or so.

I also haven't checked the MOSFET screws but it's been packed up since Dayton.

Jim.

[Mike Saeger]

My transistors blew and when I went to remove them, I discovered both screws were loose. I was pretty sure I torqued them down well.