The update notifications of all my antenna pages on this website are now listed here.

[dummy loads - general]      [dummy load components]    [construction]

[measurements & performance]    [power measurement with RF-probe]

A dummy load is just one of those standard tools that you need in the shack. It's a load that has a purely resistive impedance of 50 ohms at all frequencies of interest. You don't need to pollute the airwaves while just tuning or testing your transmitter!

Dummy-loads are easy and relatively inexpensive to build yourself, using one or more suitable resistors. How inexpensive, depends on the dissipation you need (QRP (5-10 W) or (much) more) and how long you want to be able to apply power without frying the dummy load and setting your shack on fire.

So, the basic issues are:

	how to get a wide-band purely resistive 50 ohm load with SWR very close to 1:1.
	how to dissipate the heat that is generated in the resistor(s).

The first issue is easy to address: use non-inductive resistors, and keep wiring short to minimize capacitance. Options:

single 50 ohm resistor with sufficient dissipation rating

multiple resistors with lower dissipation rating, in parallel or series-parallel configuration. E.g., 20 parallel metal film resistors of 1 k ohm and 2 W rating to get 50 ohm and 40 W. I used four 50 ohm hybrid resistors with 100 W rating (acquired via eBay; total $11 + $15 S&H, mid-2009). With these, the only configurations that work are "2 parallel of 2-series"  and "2-series of 2-parallel". I have chosen the latter.

The second issue is not quite as easy to solve. Without additional measures, the resistors may overheat and be destroyed in relatively short time, even at nominal power. My hybrid resistors can only be used anywhere near their power rating, if 1) they are mounted on an adequately dimensioned heatsink, and 2) there is good thermal contact between the resistors and the heatsink. So: the resistors must be mounted on the heatsink with thermal paste (a.k.a. thermal grease or thermal compound).

Dissipation can be further improved if the resistors and heatsink are actively cooled with a fan, or they are immersed in oil. Pure mineral oil (a.k.a. paraffin oil) is a good choice, and is available in drug stores and pharmacies. It is safe for human consumption (though it does have laxative effects...). Baby oil usually is just pure mineral oil with a fragrance (check the label). In the past, traditional transformer oil was used. That type of oil often contains highly toxic PCBs! The resistors and heatsink (if you use one) can be mounted in a tin can (from loose-leaf tea, or an empty paint can - some point stores sell clean, empty cans).

With proper dissipation, using four resistors with 100 W rating should allow me to use my dummy load for 400 W. I have mounted the resistors on a heatsink, immersed in mineral oil in a paint can of 1 quart (almost 1 liter).

Also, I decided to install both a BNC and an SO-239 coax jack in parallel. With few exceptions, I use BNC connectors on my coaxes. But I want to retain the ability to use it with a PL-259 connector without needing adapter plugs. 

 Electrical and construction diagram of my 400 W dummy-load

Unlike "regular" resistors, these hybrids are mounted on a small metal tab. The tab is mounted on a heat sink, and is also one of the two leads. Note: in the configuration above, the common point of the four resistors is connected to the heatsink! However, the ground of the connectors (and the can itself) is connected to the leads of only one resistor pair. We do not want to short-circuit the resistors, so: the heatsink shall not touch the can!

References:

	"Oil filled, power measuring, dummy load" by Dave Rudd (AI4JI)
	"Baumappe: 100 W-Last-widerstand mit 40 dB Auskoppeldämpfung", Funkamateur, 080306, 2 pp.
	"Building a Dummy Load and Measuring Power Accurately", Ken Kemski (K4EAA)
	"N5ESE's QRP Dummy Load", Monty Northrup (N5ESE)
	"A dummy load and power meter for HF", Jim Tregellas (VK5JST), Amateur Radio, April 2004, pp. 9-12

COMPONENTS

Components of my dummy load:

	4 hybrid 50 ohm resistors with 100 watt rating
	1 quart (≈ 0.94 liter) paint can: $2.50 at ACE Hardware (mid-2009 pricing) the inside of the can and lid are coated; the coating must be removed from the rims of the can and the lid..
	1 quart mineral oil (a.k.a. paraffin oil): $7.50 at Rite Aid pharmacy (mid-2009 pricing) yes, rather expensive; you may want to check dime stores or "99 ¢" stores. note that you can also use "baby oil", as this is just mineral oil with a fragrance (check the label!)
	heat sink (as big as will fit through the opening of the can)
	small tube of thermal grease (a.k.a. heatsink grease, heatsink compound), $3 at Radio Shack
	ten M3 screws + flat washers + lock washers for the resistors and wire clip
	BNC coax chassis jack
	SO-239 coax chassis jack
	four M3 screws + flat washers + lock washers for the SO-239 connector
	1 ring-tongue terminal (to connect a wire to the "chassis" ground at the SO-239 connector
	silicone caulk (to prevent oil seepage via the connectors in the lid of the can)

The components of my dummy-load and RF-probe
(the four hybrid resistors are in the foreground)

CONSTRUCTION

Construction of the dummy load is relatively simple. As always, it helps to have decent tools!

The basic steps are:

	mounting the connectors on the lid of the can
	shaping the heatsink and adding threaded holes for the mounting screws of the resistors.
	adding insulated brackets to keep the heatsink off the bottom of the can (to improve oil circulation), and prevent the heatsink from banging around inside the can
	mounting the resistors onto the heatsink and wiring to the connectors
	installing the heatsink into the can
	adding the oil and closing the lid

Installing the connectors is straightforward: mark the center of each connector on the lid, e.g., with a permanent marker (felt pen). I placed mine off-center, to have shorter wires to the resistors. Then mark the holes with a center punch, and drill the holes:

	Avoid destroying the lid when marking the holes with the punch and when drilling the holes!
	Use proper backing of the lid (e.g., with flat piece of wood in a bench vise).
	Use a drill-press.
	Start with pilot holes, then work your way up to the desired diameter.
	Make round holes, by drilling through a small piece of folded cloth or rag.
	I used a conical grinding bit to get to the final diameter of the hole for the SO-239, and to ensure it is round.
	install the SO-239 and BNC connectors; use the ring-tongue terminal on one screw of the SO-239 connector.
	Scrape the coating of the rim of the lid and of the opening of the can.

The lid of the can, ready to install the connectors

My heatsink was quite a bit of work. It was rectangular and too large to fit through the top of the can. So I had to give it a round shape, by shortening the fins of the heatsink. I used standard L-brackets from the DIY store to prop up the heatsink. They are thick enough to tightly fit between two fins of the heatsink. I drilled a 3.2 mm hole through them and through the outer pairs of fins, and then tapped the holes with an M4 thread. The brackets are glued onto a thin strip of circuit board material without copper cladding. The ends of the strip are shaped to have the same radius as the inside of the can (measured at the outside of the bottom of the can). I used Gorilla Glue^®.

Heatsink and brackets (before shaping)

After shaping the heatsink, I marked the eight mounting holes for the resistors (the holes are spaced 27 mm, marked the holes with a center punch, drilled the holes, and tapped them with an M3 thread. Likewise I made two threaded holes for a strain-relief clip.

To avoid the resistor leads and associated wiring from shorting to the heatsink, I installed a small strip of bare circuit board material between the resistors. I glued it to the heatsink, and it is also kept in place by the mounting screws of the clip.

The leads of the hybrid resistors are very fragile! Be careful when handling them. Apply heatsink grease to the entire surface of the back of each resistor, and install the resistors with M3 screws, flat washers and lock-washers. Excess grease will be squeezed out from underneath the resistors. This is OK. If it is really a lot, just scrape it up.

I stripped 5 cm (2") of insulation of a short piece of coax cable, tied the braid into a pigtail and removed most of the insulation from the center conductor. The braid is soldered to the leads of one resistor pair, the center conductor to the other pair. The coax is fixed in place with the clip.

At this point, we can do a quick check with an Ohm-meter to make sure the connections are OK.

Heatsink and brackets (after shaping), with resistors and connectors installed 

Cut the coax to the right length, remove insulation of the unconnected end (at least the distance between the connectors), tie the braid into a pigtail, cut the pigtail short and solder it to the ring-tongue terminal on the SO-239 connector. Cut the center conductor to size, solder it to the center pin of the SO-239 connector. Use the cut off piece of (still insulated) center conductor to interconnect the center pins of the SO-239 and BNC connectors. Lavishly apply silicone caulk to the connectors, such that they (and the mounting screws) are completely covered. This will avoid the oil from seeping out.

With the brackets mounted, the heatsink will not fit through the opening of the can. Insert the brackets (glued to the shaped circuit board material) into the can, such that they stick out above the can. Hold them in place, slide the heatsink over the brackets and install the mounting screws. Now lower the heatsink into the can. Fill the can with oil for about 7/8. Do not top off the can with oil! Leave about 1 cm (≈⅜") at the top. The heatsink should be fully submerged.

Now we can close the lid. Careful not to bend the lid when pressing it onto the can. The whole thing weighs in at 1.4 kg (3 lbs).

Done. Fire away! If you have access to an antenna or network analyzer or an SWR meter, use it to check the dummy load before applying power.  See below.

The lid is on the can - ready for action!

MEASUREMENTS AND PERFORMANCE

The frequency sweep diagrams below show that my dummy load really has excellent characteristics up to at least 30 MHz (see frequency plots below):

	impedance is very close to purely resistive (Xs is very small)
	within 1½ - 3% of 50 ohms
	SWR is less than 1.03.

Couldn't ask for anything more!

I have run the dummy load for several minutes with my transceiver at 100 W (its max power), and there was absolutely no noticeable rise in temperature of the can.

Frequency sweep of my dummy load (1.8-30 MHz)

Full frequency sweep of my dummy load  (1.0-180 MHz)
 

POWER MEASUREMENT

We can measure the power that is dissipated by the dummy load, by measuring the voltage across it. Clearly, this will be an RF-voltage. The relationship between the power and the voltage is given by:

So we need to determine V_RMS across the load. This can be done by measuring V_peak or V_peak-to-peak across the load when we excite the load with a pure sine-wave (e.g., by keying a CW transmitter, or sending a single tone with an SSB transmitter):

Of course, with an oscilloscope, we can measure V_peak-to-peak directly (via an appropriate voltage divider or 10x probe).

I have an excellent Fluke 8050 Digital Volt Meter (DVM) that can measure "true RMS" AC-voltage. Like most DVMs, it only does so up to a certain frequency, well below RF. For my Fluke the specified limit is 50 kHz for the lower voltage range settings, and only1 kHz for the 200-750 volt range.

So, if we want to use a DVM, we need an RF-probe that "demodulates" the RF voltage to a  V_peak value. In its simplest form, such a probe consists of a half-wave diode rectifier and capacitor filter/smoothing. The output voltage is measured with a DC-volt meter with a high input impedance, e.g., a  DVM.

Here is a standard circuit:

A simple RF-probe

Obviously this circuit will be fooled by a DC-offset on the RF signal. We can fix this by swapping the diode and the capacitor. We can also make life a little easier by including a voltage divider with a scaling factor that is equal to the reciprocal of √2. Then the output voltage will be the RMS value that we are after. We can make a voltage divider where one resistor is the input impedance of the DVM. My FLUKE DVM has 10 M Ω input resistance. The second resistor should be 4M14 Ω, so 3M9 + 220k = 4M12 would be a good choice. This approach is shown below.

 RF-probe with DC-block and peak-to-RMS scaling

The voltage drop across the diode does affect the measured voltage:

Without correction for the small voltage drop across the diode:

So, we prefer a diode that has:

	low "forward voltage" (Vf, voltage drop), as the "knee" voltage of the diode determines the lowest RF-power that can be measured (relatively) accurately.
	high "max (repetitive) reverse voltage" (Vrrm) or "peak inverse voltage" (PIV), as this determines the max power level that can be measured with the RF-probe, without destroying the diode.

The options are basically Germanium, Schottky, and silicon diodes. Germanium diodes tend to have the lowest forward-voltage, silicon diodes the highest. So: small signal silicon diodes such as the ubiquitous 1N4148 should not be used.

Note that all these diodes are available from several manufacturers and the Vf and Vrrm may vary slighly between them. For my 100 W transceiver, 90 V is sufficient. I opted for the OA91 diode. The 1N34A is OK for QRP.

References:

	"N5ESE's Classic RF Probe", by Monty Northrup (N5ESE)
	"Diode Selection for RF Probes", by Jack Smith (K8ZOA), 11 July 2008
	"Oil filled, power measuring, dummy load" by Dave Rudd (AI4JI)
	"Building a Dummy Load and Measuring Power Accurately", Ken Kemski (K4EAA)
	"A dummy load and power meter for HF", by Jim Tregellas (VK5JST), Amateur Radio, April 2004, pp. 9-12
	"Absolute RF power measurement using simple techniques", by Adam Farson, VA7OJ/AB4OJ
	"A Simple SSB PEP Measuring Procedure", by Adam Farson, VA7OJ/AB4OJ

 
My RF-probe

©2009 F. Dörenberg N4SPP

   top of page