Power Amplifier


I run my amplifier on 24 VDC rather than 28 to give it a little bit more margin on breakdown voltage. To adjust bias, I have a 0.1 ohm resistor in the power return lead, and I set the bias at about 10 mA for each transistor. Performance-wise, distortion is low up to a total current of nearly 4 A, so I generally tune up my transmitter by adjusting the Drive pot for about 3.5 A. Heat has not been a problem at all during normal CW operation. I do limit my testing duty cycle so as not to push my luck. My heatsink is very large but thermal conductivity within the devices is the weak point.

The only failure I’ve had is when driving the amplifier into a lowpass filter that is too low in frequency. In that case, the SWR is infinite and both IRF510s fail shorted. Running with a poorly-tuned transmatch has not been a problem, but I cautiously do my initial tuning with minimal drive


WA2EBY MOSFET Power Amplifier

After a market search, I decided on the WA2EBY power amplifier design, based on cheap IRF510 MOSFETs. This amplifier is featured in the ARRL Amplifier Classics book and some edition(s) of the Handbook. It’s also available in kit form from HF Projects as the HF Packer Amp; they also offer the amplifier board kit alone, which is much cheaper than the full kit. I built mine from the original article on double-sided copper clad, using the Sharpie method. A full ground plane is highly desirable for any RF design, and especially so for power amplifiers. 

My general board fabrication procedure is this: First, draw the wiring side layout on 0.1-inch grid paper. Second, drill all the through holes. I use a Moto-Tool in its drillpress stand and HSS drill bits from #75 on up. Third, mark the holes on the component (ground plane) side that do not connect to ground. These I disconnect from the ground plane using Stahler drill mill cutters, which are no longer available. You can also use a ball-end cutter on the Moto-Tool to remove the copper immediately around the hole. Fourth, I paint the ground plane with fast-drying auto primer, which is the fastest way to entirely cover the board with resists. Fifth, I draw the pattern with a fresh black Sharpie. Finally, drop it in etchant. I like ammonium chloride because it’s clear. To remove the resist, wash with lacquer thinner. Parts that have a connection to ground are soldered directly to the ground plane on the component side. This avoids an extra wire path through the board.

Below is the complete power amplifier (left) and 2-W driver amplifier (right). I used this peculiar orientation because it made access to the IRF512 tabs very easy when (not if) they need replacement.

This data was taken with the second set of IRF510s. Distortion with my first set was notably higher. Perhaps this pair was (accidentally) better matched. When building any push-pull amplifier, one should use matched pairs to minimize second-order distortion.

Here is the frequency response with a maximum output power of 10 W at 14 MHz:

Not a bad amplifier at all considering its low parts cost and small size.


Here are some distortion plots with power supply current around 3.5 A.  More power is available if driven harder, but with an increase in distortion.





Xmtr HomeTransmitter.html