Racal syncal30 PA fault

[neon indicator]

I have one of these possibly made by Grinel

It would occasionally blow the finals if a good aerial was used on 80m rather than a 6' to 8' whip.

I stripped it and tried various revisions of the schematics I have. Basically unless the aerial is very poor, the PA is increasingly unstable below 10MHz, such that if set up properly, into a dummy load on 3.6MHz, the PA self oscillates at about 40W as soon as TX selected.

I even replace the PA transformer with exact original and also a modern design with same turns ratio using Binocular core. (I have the original type in now).

The various revisions around PA and feedback and even series resistors put in and taken out on various schematics suggests an ongoing design issue.

Has anyone experience of this lower frequency self oscillation problem or any ideas I can try?

I replaced all the decoupling caps on PA board and added extra ceramic ones.

I rebuilt the matching/tuning coil and also now have a cleaner replacement. The more I cleaned up the Aerial Matching/Tuning Coil the more unstable it became at lower frequencies!

Suggestions appreciated. I'd like to use this rig for 5Mhz channels.

I'm using 18 x NiMH 4A "C" cells as power instead of original 19 x NiCd "D" cells.

[G8HQP Dave]

I can't comment on this specific unit, but I understand LF oscillation is a common problem with transistor PAs. The problem is that the device gain increases with lower frequency. Only a smidgen of feedback is enough to cause oscillation.

Cleaning the matcher reduced losses so reduced the load, hence more oscillation. There are really only two options: reduce LF gain, or reduce feedback. The trick is to do this without messing up the required operation. Add a lossy choke somewhere?

My apologies if you already knew all this.

[neon indicator]

It's always good to get confirmation of your own thinking

I built a PA using a pair of IRF510. It's not got much power on 12V, but about 20W or more at 24V, which happily this radio uses, but it needs more drive. IRF510's are also 1/10th of price of the MRF parts in the official PA. The radio is supposed to output about 22W.

[GMB]

I haven't had any such problems at all with my TRA931X and I have often used it with precisely tuned dipoles when I'm operating "in the field".

[neon indicator]

Mine's South African and the Schematics and Parts in PA don't quite match the actual Racal TRA931 schematics.

I have 3 versions of the SA schematics The PA PCB /parts doesn't 100% match any of them.

[G8HQP Dave]

I used the IRF510 for a QRP class E PA to check the maths in my PhD. A nice cheap switching FET but it had quite high capacitances. Even at 2MHz it needed a bit of driving. Afterwards I wondered if it might work better as a cascode - lose a bit of efficiency but no longer have a Miller-multiplied capacitance to feed.

[neon indicator]

I'm driving at about 12 Ohms source I think via a toroid step down and splitter from 50 Ohms. Two in push pull with bifilar contra-wired toroid for +24V supply on drain, then a binocular core on output as common mode current choke as balun with 2 x series caps for DC isolation. Rather than using common centre tap transformer as feed and push pull combiner at same time. Gate bias a pair of 3.3V zeners in series, on on each FET as slight temp comp. then 2 x presets off single 6.6V supply to adjust class AB bias of FETs individually.

There is a 50MHz PA design using EIGHT IRF510 (4 in parallel each half of push-pull). I can't remember where I saw it.

At higher supply volts the fets work better as Drain/gate capacitance decreases. I got 100W from pair of IRF740 with 36V supply and about 6 Ohms source drive on gate @ 14MHz. mounted on old PII/PIII heatsink all wired valve base style.

I think you can in theory get 30MHz from the IRF740 (probably want 40V to 90V PSU) and possibly 70MHz from an optimised IRF510 design.

Of course no problem using the cheap 2000pF gate capacitance 15A 900V parts for 137kHz or 500KHz. I think actually possible up to 4MHz. Maybe an interesting 160m/80m lightweight (no mains transformer) 1kW linear direct from rectified 230VAC (300VDC) with suitable Mains isolation on the input and output RF transformers for push-pull

[G8HQP Dave]

I only used 12V supply, which was too low really, but it did the job I wanted. I was investigating the carrier frequency response, and drain modulation AM frequency response. Class E is counter-intuitive, because you don't tune for maximum power but maximum efficiency. Or at least, you should - my guess is that most people running Class E just peak the output because they are not too bothered about efficiency.

[yesnaby]

Hello,

I'm glad you guys know what you are talking about

Michael