80m Transmit Power

[NZ0I]

The second build of a P1 prototype board has reached testing of the 80m transmit circuit. It worked fine, but I've been a little disappointed in the maximum transmit power. It has tended to max out at just barely 1.5 Watts. It seemed that the beefy IRF610 power FET should be able to deliver much more than that.

I'd tried to minimize the loss through the low-pass filter. And indeed that is part of the down-side to this simple transmitter design. A very aggressive LPF is needed at the output in order to remove the strong harmonics generated by simply switching the FET on and off. The aggressive filter drops the harmonics by over 50 dB but also reduces the fundamental by about 7 dB. It seems that the only way to compensate for the LPF loss is to increase the drive voltage applied to the FET's drain. But there are limits to how high the switching supply can boost the LiIon battery - and 12V seems to be close to the maximum voltage that is prudent.

But it turns out that there is a straightforward way to boost the power delivered by the FET without increasing the drain voltage. A simple RF step-up transformer brings the RF signal voltage up. An updated schematic shows the transformer as L104: https://drive.google.com/open?id=1yk6tQTWNo6adcF94_UHlC6f1EtOyIl8k. With the substitution of the transformer for the 5.8 uH inductor used before, the output power from the transmitter approximately doubled.

The problem then became that the power supply was unable to supply enough current to the transmitter, and would shut the transmitter down (as designed for safety) before reaching 2W RF output. By temporarily disabling the current limiting an RF output of 2W was achieved, and that limit was due to the 11.6V upper limit of drain voltage available from the power supply. With a 13.8V supply I estimate that the transmitter would put out 2.5W or more, but I haven't tested that. A few tweaks to the design and I think it could put out 5 watts... but not with the current power supply.

Some of the transmitter and power supply components get warm to the touch after 30+ seconds of key down. The output power also begins to droop by several dB. I haven't determined the cause of the droop. It isn't the power supply since it measures a solid regulated voltage that doesn't droop measurably. Cooling various components with compressed air seemed to have no effect - I tried it on all the warm components I could identify. The power droop isn't much of a concern since there is no requirement for the transmitter to operate continuous key down. But still, I'd like to understand the cause.

A little more software work and then on to testing the updated 2m transmitter design.

[NZ0I]

Upon further investigation, it was found that the upper limit of output power was about 2W when using a 12.8V power source (an SLA battery). An obvious droop in power was still evident: the output would fall to a little over 1W after a minute or so of continuous transmit. The power supply was ruled out as a source of the power droop.

Looking into the power droop I discovered that the two 80m LPF inductors were getting very warm. This had been observed when testing earlier boards, so I had substituted larger Coilcraft inductors capable of handling more power. These larger inductors are also getting very warm. When I cooled them off using some compressed air the output power of the transmitter rose, going up close to the original 2W of output power. That confirms that the LPF filter inductors are responsible for the power droop.

Each stage of the LPF introduces about 3dB of signal loss. So when we are measuring 2W of output power, the power is 4W at the input of the final filter stage, and 8W at the output of the IRF610 power FET. That means that the first LPF stage is dissipating about 4W of power, and the second LPF stage is dissipating about 2W of power. Virtually all of that power is being dissipated in the inductors since they are far lossier than the capacitors in the LPF - and that is born out by the heating of the inductors. The first LPF inductor was noticeably warmer than the second one, again supporting the analysis.

Apparently, as the inductors' temperatures rise their loss increases, thereby causing the observed power droop. From what I read, that is true for all ferrite materials. Trying to "shove" more power through them would only result in thermal runaway, destroying the inductors. The only solution is to decrease the inductors' loss, their temperature, or both. Redesigning the LPF might accomplish that, but a very aggressive filter is required due to the ragged signal that the switched FET produces. So a lower-loss LPF design will most likely mean a redesign of the entire final output stage. So I see two better alternatives:

1. Decreasing core losses by lowering the flux density, which means using physically larger inductors.

2. The other option is to simply accept that the maximum output power from this transmitter is about 2W, and it might drop by several dB when it is operated at high duty cycles.

I think I'll go with option 2, but I will also change the inductors' footprint to provide the builder with the option to install either the surface mount components or wind his/her own larger toroids. Other suggestions?

73,

Charles

NZ0I

[NZ0I]

Problem solved. The solution: don't use ferrite. And use larger cores. 

I wound two 2.7 uH coils using two T60-2 cores. It required 20 turns of #24, which fit easily onto a T60 toroid. After replacing the small ferrite coils with the two hefty powdered iron cores the output power jumped quite remarkably. I am now seeing 5W of output power using the 12.6V battery power source. 9.8V is being supplied to the IRF610 power FET to get 5W out. The droop has totally disappeared. Spectral output still looks good - better than 45dB of harmonic suppression - but a bit higher than using the 5% tolerance ferrite toroids. The price of using the hand-wound toroids is roughly the same as using the larger Coilcraft ferrite inductors. The powdered iron inductors still get slightly warm to the touch, but much cooler than the ferrite inductors did. The dummy load gets much warmer :-)

The downside to using the powdered iron toroids: 1) the builder must wind them by hand; 2) the inductance tolerance is relatively high: maybe +/- 15% if wound carefully, and +/- 10% if then tweaked using an inductance meter; 3) the inductors are larger so take up more space on the board and more volume in the box; 

73,

Charles

NZ0I