My Pulse Width Modulated AM transmitter project required dummy loads to match the load impedance of the RF power stage (50 ohms) and the Pulse Width Modulator (12 to 16 ohms). Surely any self-respecting amateur radio station would have a decent 50 ohm dummy load, I hear you ask incredulously. I do have one at VK3HN … or I did.
A month ago, while testing the transmitter’s power supply (capable of sourcing 200 watts) into my best 50 ohm RF dummy load, I burned out the ceramic resistor 💥😫. It is a commercial unit, and unfortunately the ceramic power resistors I can get on eBay do not physically fit the machined mounting. So I needed a new 50 ohm load. And a separate 12 to 16 ohm load to test the Pulse Width Modulator. To the drawing board.
Two types of non-inductive power resistors are mostly used in homebrew dummy loads — ceramic and old school carbon ones. Ceramic power resistors with values of 50 ohms and power ratings of 100, 200, 250 watts and higher are plentiful on eBay. These appear to be mostly pulls, and anecdotal reports are that some are damaged after having been pulled a bit too forcibly. I’ve bought several lots of these tab-mounted resistors without really understanding how they are made or how tough they are.
A few internet searches suggest they are ceramic cases on various substrates including Beryllium Oxide (known to be slightly nasty), AlN (aluminium nitrate?) and Alumina, and may work as high as 50GHz. I had trouble finding any details whatsoever for the specific ones I obtained. A search on YouTube for ‘dummy load’ brings up videos of these devices being used for RF loads. Mostly, people simply bolt them to a suitable heatsink and pump SSB into them, with (presumably) a lack of smoke and safe results.
The old school approach is to parallel a big pile of carbon resistors. These cannot be surface mounted against a heatsink but if they are mounted in free space a fan can effectively cool a block of them.
I discussed the choice of resistor type with fellow QRO AM transmitter builder, Wayne VK3ALK, who is experienced in such matters. For AM, he thinks the ceramic tab-mounted ones are prone to hot-spots and burning out, and prefers to use a bank of old school carbon resistors. The 100% AM duty cycle punishes components, power supplies and dummy loads. I decided to follow his lead.
Finding good quality carbon resistors at any power rating above 2 watts was not that easy. I tried all the big online international suppliers and all the locals. Most suppliers stop at 1 watt carbon. I found some odd-ball Russian NOS ones, they looked promising, but the date of manufacture unknown, quality unknown, condition unknown. The last thing you want your dummy load to do is to fail under load. As well as having to rebuild the load you will probably have to rebuild your transmitter.
A dummy load must be reliable. With that in mind, the most suitable ones I could find were new 2 watt carbons from RS; even those were a stock-out, and the last 20 were on back order for several weeks.
50 ohm load
The 50 ohm load is made sandwich-style between two identical and symmetrical single sided PCB pieces, drilled with a 12×8 grid of holes. This accommodates ninety six 4k7 2 watt carbon resistors giving a theoretical resistance of 48.9 ohms and an actual measured DC resistance of 48.5 ohms.
When driven with my FAT5 160 AM transmitter, which delivers almost exactly 50 watts of un-modulated carrier, it gets hot after about 2 minutes, so with twice the RF power, a fan will be needed.
I boxed it up in a nice ex-Rockby Retek light aluminum enclosure with 60mm fan (from a computer power supply) and DC socket on the rear panel, and a fan switch on the front panel, offset to leave space for a future power meter. An old plug pack supplies 12 volt DC power to the fan. An intake hole with a suitable wire mesh will be added to the front panel.
PWM dummy load
In the AM transmitter I’m building, the Pulse Width Modulator must sink modulated DC to a non-reactive load of about 12 ohms. At rest (un-modulated) a PWM switching a 120 volt DC HT set to a bit less than 50% duty cycle and operating at around 90% efficiency will deliver slightly less than half the HT, say 50 volts, into 12 ohms at the far end of the modulator’s LPF. That’s over 4 amps, or 200 watts. Modulation adds four times this power at peaks (Peak Envelope Power) in the combined sidebands.
If the 12 ohm dummy load is replaced with a class D or E PA operating at a conservative 80%, it will deliver up to 160 watts of carrier into a 50 ohm load. Designing an AM transmitter for these power levels ensures it will run cooler and less stressed at the legal limit, 120 watts of carrier.
I decided to use three of the ceramic 250 watt 50 ohm resistors from eBay for this load, giving 16 ohms, which would limit the power levels in test. The 4 ohm mismatch between the LPF and the load would likely impact the filter’s performance, but I don’t have the test equipment to assess this. A fourth ceramic resistor can easily be paralleled down the track. I reused a junk box heatsink with convenient M3 tapped holes.
With the dummy loads done, it’s time to get the modulator and transmitter PA up and running to generate some heat!