Clifton Laboratories 7236 Clifton Road  Clifton VA 20124 tel: (703) 830 0368 fax: (703) 830 0711

E-mail: Jack.Smith@cliftonlaboratories.com

I've received a copy of a Celwave catalog page from Tom, N0SS, with the two dummy loads I've identified as 75 watt and 200 watt devices, based on the Res-Net microwave catalog. The two model numbers are BL060 and CL150, with power ratings of 60 and 150 watts respectively. Celwave's ratings make more sense to me than the values I obtained from the Res-Net pages. (The actual power ratings depend on the heat sink, and I believe Res-Net makes the resistor element, to which Celwave adds the heat sink. I imagine Res-Net's ratings are for an infinite heat sink, which these two loads most certainly do not have.)

23July 2006

I've added data for two loads, my venerable Heathkit HN-31 Cantenna and a 500 watt "Bird-style" termination. The 500 watt load was part of a non-working military surplus termination wattmeter I acquired many years ago. There's no manufacturer's ID plate on the load, so all I can say is that it is a "Bird-style" load. The defect was in the wattmeter, so I continue to use the load.

I recently purchased several 50 ohm terminations or "dummy loads" from Ridge Equipment and today compared their performance to my "reference standard" and my 100 watt Bird Electronic Tenuline model 8323 attenuator. 

The measurements are made with an HP8752B network analyzer over the frequency range 300 KHz - 3 GHz. My calibration standard for a "perfect" 50 ohm load is a MiniCircuits Laboratories KARN-50-18 termination. The KARN-50-18 termination has reasonable performance over the frequency range up through 3 GHz.

Return loss measurements with a network analyzer are made with respect to a standard; the calibration process assumes the standard is the truth and calibrates out its parasitic inductance and capacitance to the extent they are defined in the network analyzer's calibration standard setup. In other words, the return loss measurements show how well the dummy loads match my KARN-50-18 standard. (I also have a N "short" and an "open" I use in the calibration process.)

Ideally, I would use an Agilent calibration standard set, comprised of an "open" a "short" and a "load" each with known parasitic parameters. I don't have a calibration standard set, so take these measurements with a bit of caution, particularly above 500 MHz or so. 

The loads I tested are shown in the photograph. They are: 

  - Bird Electronic 100 watt Tenuline model 8323 attenuator

 - Res-Net Microwave 100 watt dummy load model RFT100NFE. This is the rectangular, black load at the left center of the photograph. (The 12" ruler will give you an idea of the size of the various loads.)

 - Unidentified 75 watt load (but appears to be a Res-Net Microwave RCX150NM conduction load mounted on a small heat sink.) This load is said by Ridge Equipment to be rated at 75 watts. I would not trust it at that power level unless it were mounted on a large metal plate to increase its heat sink area. I believe these loads are normally mounted on 19" rack panels providing additional heat sink area. The basic load part is rated by Res-Net at 150 watts.

 - Unidentified 200 watt load (appears to also be made with Res-Net parts) with a signal pickoff port.

 - 15 watt Celwave AL015B 15 watt load, integrated into a male Type N connector.

I measured the "return loss" of the loads. Return loss is the ratio in decibels of the power reflected from the load to the power incident to the load. It is related to the more commonly used amateur radio term "SWR" by the equations shown to the right. 

I measured a return loss of 40 dB or better up through about 1.5 GHz.There's a tiny blip up at 300 KHz, which may be more an artifact of measurement than anything else. With an SWR of 1.02 or better up through 1 GHz, there's a lot to like about this load. 

Based on its physical size and weight  I believe 100 watts is an honest rating for this load. At 100 watts, it will run quite a bit warmer than the much larger Bird 100 watt load. However, the ceramic resistor element used in these loads is rated for continuous operation with a heat sink temperature of 100° C. That's the boiling point of water at sea level.

These loads are rated at an SWR of 1.10 : 1 up through 2 GHz, and my measurements show that is a conservative rating.

These are exceptional bargains, at $9.00 each, so I bought two.

One load shows better than 40 dB return loss up to 1 GHz and the second is not quite as good, but still excellent, particularly considering the bargain price.

These loads have a heat sink that's on the small side for the rated 75 watts, and I expect they would handle the rated power, but with a heat sink temperature that's higher than comfortable for normal use.

The specification on these loads is shown below (use the 150 watt specification). My measurements show these specifications are met.

This is a cute little load, with a built-in Type N male connector, rated at 15 watts. I can't find formal specs on this load, but its measurements show it to be an excellent performer up into the GHz range..

This is another load without specifications, other than what is on Ridge's web page. It's rated at 200 watts and appears to be a Res-Net product, based on its general design and the internal parts.

The load has a signal sample port providing an attenuated signal at a BNC connector.

The signal sampler is just a short length of wire, spaced a half inch or so above the 50 ohm termination resistor. The stray capacitance between the wire and the input line and resistor provide loose coupling.

Channel 1 shows the return loss, which again is excellent up into the 2 GHz area.

Channel 2 shows the coupling loss, i.e., the loss between the sample port and the input port. The response axis is in decibels and the frequency axis is logarithmic so a simple capacitive coupler will have a frequency response that is a straight line, with a slope of 20 dB/decade. The response tracks theory at least up through 2 GHz, where there is a kink in the response, probably due to stray inductance starting to become important. 

At the low frequency end, where the coupling factor is 90 to 110 dB, the network analyzer exhibits noise in the signal.

If one were to use this load strictly for HF use, say below 30 MHz, the coupling arrangement could be altered to increase the sampled signal level, perhaps replacing the stray capacitance with a physical capacitor of 1 or 2 pF. Or, a high value resistor could be used.

Of course, a frequency flat sampler can be built either with resistive elements or with a toroid current sampler.

The 8323 is an older design high power attenuator manufactured by Bird Electronics, the same company that makes the legendary Model 43 wattmeter. (The company was founded by James Bird, hence the name.)

The introductory photograph shows the 8323 is physically the largest of the loads; indeed it is at least three times the physical size of the 200 watt load.

In part this is due to a very conservative specification by BIrd and in part due to it normally being run with a temperature far below the 100°C rating of the Res-Net loads. Bird loads and power attenuators of this era are filled with transformer oil to improve heat transfer from the resistive elements to the metal housing.

The return loss is more than acceptable into the GHz range. The 8323 is not rated for performance above 500 MHz or so, and it's certainly excellent in this range

The attenuation is nominally 30 dB, and it is within 0.25 dB of that value from 300 KHz through 1 GHz.