The Topper Amplifier Revisited

The amplifier built manhattan island style on a copper PCB I thought my 200mW Madagascar Mighty Mite (MMM) would benefit from some sort of afterburner, so I dusted off a project shelved in 2011: the Texas Topper amplifier. I had built based on a design by Chuck Carpenter and kitted by Rex Harper. I ran into a couple problems back then, including some difficulty getting the bias right on the mosfet at the heart of the amplifier. In another brilliant move, I managed to burn out said mosfet by grounding it while trying to get it and its heat sink to fit into a metal box.

So, having established that I had once upon a time purchased this in kit form from QRPme, I didn’t feel too bad about going back to that website and looking at the schematic for the most recent design iteration, which includes a means of dialing the bias up and down on the gate of the mosfet. I still had the kit board around, but thought that rather than try to rework them, I’d have more flexibility to experiment if I laid it out on a copper board, as I had done with the MMM. The only slightly tricky bit was gluing the 12V relay to the board and getting all of the connections to it correct looking at it upside down.

Another shot of the amplifier from a different angle

I built the RF detection and relay circuit first, and verified that was working by driving it with my FT817 set to lowest power (0.5W) with a dummy load connected to the output of the relay. Next, I added the biasing circuit and tweaked the pot for 2.7V at the mosfet gate, and then added the mosfet. Finally, I tacked on the LPF.

Photo of the MMM, amplifier and dummy load connected by BNC cables.
The MMM hooked up to the amplifier and a dummy load.

I initially tried the amplifier with my MMM attached and was underwhelmed by the output: for 210mW in, the scope showed 5.27Vrms out, the equivalent of 555mW, a gain of 4.2dB. At least the output looked smoothly sinusoidal. I noted that even at that low power level, the mosfet and its heat sink got hot quickly. With 15 seconds of key down, I had gone from an ambient 30 degrees C (it is winter in Madagascar, but the garage still gets toasty in the mid-afternoon) to 38C. Towards 30 seconds, the rate of temperature increase began to take off and hit 60C. In retrospect, I probably should have stuck some sort of buffer stage between the MMM and the amplifier. I did the rest of my testing using attenuated output from an FT817 to drive the amplifier.

Two BNC connectors joined by a pi-style resistive pad.
Cheap 3dB attenuator. The metal mounting bracket is a computer expansion slot cover.

With just a bit more drive, the amplifier performed better. I set my FT817 for 0.5W output and ran it through a 3dB attenuator, i.e., yielding 250mW at the same frequency as the MMM (3.579 Mhz). Under the same conditions, I measured 10Vrms from the amplifier, so 2W (9dB gain). At that power level, the mosfet hit 50C with 15 seconds of keydown, and 60C at 20 seconds. I think I need a bigger heatsink.

Out of curiousity, I took out the attenuator to see what the amplifier would do with 0.5W input, and measured 23.8Vrms = 11.3W output (13.5dB gain).

Pencil drawing of component layout.
Layout of components on a PC board.

(Addendum, August 2017)

I was not comfortable with how hot the FET was getting with its screw-on heatsink, so I moved some components around and added a beefier block of aluminum. It still gets hot, but takes longer and dissipates more rapidly. If I ever stick this in a box, I’ll want that heatsink exposed to air flow.

Leave a Reply

Your email address will not be published. Required fields are marked *