开云体育

Did anyone test at what power output those “reverse biased” diodes begin to conduct? At that point I think the RF amp output is practically shorted to the ground unless operating on the lowest band of the filter bank and may blow the finals.

If real PIN diodes are used this would be almost nonissue but with a rectifier pn junction the reverse bias voltage is often the limiting factor.

Rich Evan thanks for the pointers. I’ll have to research the past archive from my desk later

I was trying to see the safe limit if the output power the other day by increasing the powe and supply voltage. Around 12V supply I got 10W output. A bit above the finals blew and I was doing the postmortem analysis. The final transistors should be able to push more power easily and I was thinking of a possibility of what went wrong. I suspect the diodes of the “off” LPF branches conducted and the RF current went straight to the ground. 10w on 50 ohm is about 25v rms so that may be the problem.

So using LDMOS and 1:1 output transformer there’s great amplifier reserve capacity but the danger line is also right there.

It’s a bit of heavy duty soldering work to replace the finals but I’d rather know the limits before I take it to the field. I’ll need a supply voltage limiter once I find out a good supply voltage with enough safety margin. The burnt chip smelled bad and the drain contact that I soldered looked like spot welded - I couldn’t unsolder it.

Thanks JZ. I just don’t want to find out the limits at 3000 feet elevation and hours away from my bench.

I think with a pair of high performance RF transistors and 1:1 output transformer, there may not be enough safety margin at any supply voltage if the diodes are reverse biased at -2Vcc. If the rev bias can’t be increased, I might have to drop the final stage supply voltage with series diodes to create the margin.

If anyone else finds their favorite final transistor, be careful high performance choices will run into this issue quickly. If the high band transmit power performance is worse than lower bands I’d carefully look into heavy load situation by unwanted conduction. LPF input above the cutoff frequency will look shorted to ground from the source.

At that point I think the RF amp output is practically shorted to the ground unless operating on the lowest band of the filter bank and may blow the finals.

If real PIN diodes are used this would be almost nonissue but with a rectifier pn junction the reverse bias voltage is often the limiting factor.

This is a question of the power amp output voltage swing vs the diode reverse voltage. Not the circuit topology or the diode choose. Depending on the vdssat and Ron of the final transistor the RF voltage can swing bigger than with some other transistors.

So the reverse bias is actually generated from the RF output transformer, not a fixed voltage, so it is not like having a fixed reverse bias. Analysis continues....

Evan, thanks for the pointer. That document is an excellent description of the design.

There are two statements that I am not fully convinced, though.

1. the voltage doubler produces a voltage greater than the Vp-p of the RF power
2. the bias voltage is always ahead of the RF power envelope while the power is taperd up or down.

Yeah I understand what you are saying but the scope of this thread is the switching behavior during TX so I omitted the details for the RX state.

Allison, I am not blaming anyone! Please stop distorting my discussion points.

I don't have any drive issue. That issue is for some switching MOSFET with a bit higher Vth like FDT86256, which I am not using in my QMX.