0W output power on all bands after removing a turn from L31

Post Reply
agrif
 

Posts: 7
Joined: 10 Jun 2022, 19:25

0W output power on all bands after removing a turn from L31

Post by agrif »

After testing the power output on all 5 bands, my 40m was a little weak (3.5W on 12V) and I could not separate the coil any further, so I removed it and took out a turn.

Now I'm reading 0W on all bands, with 0% efficiency on 12V and nan% on USB power. Receive still works fine, and even transmit appears to do something -- I can see the signal on my SDR. NanoVNA said my filters were still intact, so I transmitted a 20m WSPR beacon and a few stations picked me up, but not many. So it looks like the transmit machinery is all still intact, except with very very low power.

So: My BS170s died, right? Is there anything else I can do to diagnose this problem? Also, if removing one turn from L31 blew out my amplifier, how can I improve output on 40m without exploding FETs?
agrif
 

Posts: 7
Joined: 10 Jun 2022, 19:25

Re: 0W output power on all bands after removing a turn from L31

Post by agrif »

An update, in case anybody finds this in a search.

It was the BS170s. They were measuring about 300 ohms between gate and drain, both ways, and the diode between source and drain was completely blown. Replacing them with new ones from Mouser worked perfectly, and after resetting L31 to the original number of turns, I'm back on the air.

Desoldering the BS170s was a huge pain, though. The pad connected to ground refused to give up its solder. I ended up having to drill those out -- and almost completely wrecked my pads in the process. It's a miracle I didn't break anything. If anybody else does this, be careful -- there's a small trace going to one of the relays directly next to these holes. In hindsight, I wish I used a socket for the transistors, in case they blow again.

I doubt my RF board can survive another round of this, so I'll content myself with the performance I can get by fudging the coils around without rewinding.
s57l
 

Posts: 56
Joined: 30 Dec 2021, 21:30

Re: 0W output power on all bands after removing a turn from L31

Post by s57l »

Unsoldering fets is actually very easy. Cut them off and unsolder each leg separately. Than clean the holes with a vacuum pump. If you had problems with the ground pad, you need stronger soldering iron. Drilling is definitely NOT an option, as you must be very lucky to make it work and you destroy the pcb. Even more so with multilayer boards, as this one. If you just think a little, element legs are much harder material than the pcb and the solder, so you can not drill the leg out, but you will always drill a new hole next to it!
I find it very surprising how many people are afraid of destroying pcb with the heat and than drill it…
John
 

Posts: 49
Joined: 19 Feb 2022, 08:56

Re: 0W output power on all bands after removing a turn from L31

Post by John »

It's possible that when you removed 1 turn from the coil you did something that changed the circuit the FET's were driving through and a high VSWR blew up the FET's. Perhaps there was still insulation on the wire of the toroid, or a small solder bridge occurred when you removed the wire from the toroid and resoldered it. Or it wasn't well connected to a dummy load, or it was attached to an antenna that had a high VSWR on the frequency of the band you were testing the rig on. After making changes it's always a good idea to test the rig first on USB power to help keep the FET's from popping in the event of a problem. They simply will not survive a high VSWR. Class E is very efficient, which means you can get away with using low power RF transistors that are handling higher power, and no heatsink. But if something keeps that power from leaving the FET's like it should, then they will fry in seconds.

If the FET's pop, just snip the FET's off and leave the leads sticking up. Then solder the new FET's leads onto the old leads. It's not pretty, but it is the safest way to replace the FET's if you aren't experienced and well equipped for the job. I wouldn't use sockets because they just add capacitance and reduce performance to some degree at the higher frequencies.
agrif
 

Posts: 7
Joined: 10 Jun 2022, 19:25

Re: 0W output power on all bands after removing a turn from L31

Post by agrif »

Thank you both for your suggestions!

I got the pins out fine, and the other plated through holes were no trouble. But somehow I got a thin surface of solder right in the middle of the board, inside the grounded holes, and no amount of cajoling with iron, vacuum, or wick could get it out. I tried for so long that the solder mask was lifting off the traces around the holes! I don't recommend drilling, but I was out of options. I used a tiny sub-mm bit and a hand drill and patience. No pins in the way, and the holes were empty enough to act as drill guides.

I like the idea of stacking the new FETs on the old pins. How much capacitance does a socket add, compared to a longer stacked pin? I'm new to RF circuits, but it seems short enough and the board already uses headers for two RF signals (labeled "RF" and "Drive").

I was a bonehead and ran the radio off 12V to test. The FETs must have blown instantly -- it read 0W immediately. I think in my head I was expecting a few second warning... now I know better.
elTrauko
 

Posts: 30
Joined: 04 Jan 2022, 21:58

Re: 0W output power on all bands after removing a turn from L31

Post by elTrauko »

Always, always the tests or measures of ROE in 5v from USB
Post Reply

Users browsing this forum: No registered users and 3 guests