Group Actions
Engagement
  • Author Author: bernhardmayer
  • Date Created: Date Created
  • Views 3359 views
  • Likes 5 likes
  • Comments 6 comments
Related
Recommended

6.25 Watt Hamradio Dummy Load

bernhardmayer

In the last couple of weeks I had a little side project: a hamradio dummy load especially for the 2 m and 70 cm hamradio bands but it should also work on higher frequencies. A dummy load (https://en.wikipedia.org/wiki/Dummy_load ) is a handy device when it comes to testing receivers. It is connected to the HF output of the transceiver and burns all the HF energy in a resistor so that no HF signal is transmitted. The dummy load is actually a high power 50 ohm resistor. You only have to take care that parasitic inductance and capacity is low so that it acts as a resistor even at high frequencies. That's why SMD parts are more suitable than THT resistors.

 

image

 

i made my design with KiCad because it allows fast and easy working even when parts are placed in odd angles by using polar coordinates. It also has a nice 3D viewer to check the layout and mechanics of the board. The picture above is a screenshot of this 3D viewer.

 

The dummy load consists of 51 100 ohm 0805 resistors with a power capacity of 1/8 watts each. That sums up to 6.25 watts for the whole dummy load.

 

It also has an attenuator output (CN2) with a signal reduction of 20 dB. If this output is not used it has to be terminated with a 50 ohm resistor.

 

Input and output are connected via a SMA connector so that it easily can connect to the ArduTrx or any other hand held radio.

 

The real life hardware can be seen on the next picture. It looks very similar to the 3D model.

 

image

 

The hardware is released under CERN OHL v1.2 and all the data and documentation can be found on my project website ArduTrx - Dummy load There exists also a BOM with Farnell/Newark order numbers.

 

PCBs and parts for this project can also be ordered at Aisler: https://aisler.net/p/QEWPLPOW

 

If anyone rebuilds this project I wolud love to hear your experiences. If anyone has a network analyzer I would love to get a measurement and see how this dummy load performes in the GHz range. Just answer to this post or send me a message.

 

UPDATE july 24, 2019

 

I reworked the circuit according to the comments below the original blog post. The dummy load or attenuator is built now according to the following block diagram:

image

It now matches 50 ohm on the input and on the output. The downside of this is that the attenuator output is now -22.36 dB instead of -20 dB. The layout has only slightly changed.

image

 

There is now also a placement option possible where the device works only as a dummy load and has no attenuator output.

 

The updated data is on my website: ArduTrx - Attenuator or Dummy load

 

Nevertheless it still needs to be tested how the device behaves at higher frequencies.

  • in reply to bernhardmayer

    Hi Bernhard,

     

    As another option (maybe no need, but just a suggestion) personally I'd still try to remove all series resistors if possible, e.g. by using many resistors in parallel in a circle (like a star), or maybe by replacing 0805 with 1210, since they have around 5 times the power handling capability. Also resistors can tolerate high power for short periods, so a large quantity may not be needed (e.g. with cooling or some heatsinking), for example instead of 10 size 0805 parallel, there could be just 3 size 1210 in parallel as below, allowing the length to be very short. Anyway your solution may be fine too.

    image

     

    Here's some random google images searching for dummy load.. someones star configuration for a dummy load (not attenuator, but the principle would be the same):

    image

     

    And this should be avoided except for perhaps HF at a stretch:

    image

  • in reply to shabaz

    Hi Shabaz,

     

    I decided to redesign my dummy load / attenuator in the following way:

     

    image

    R1 consists of 40 100 ohm resistors. 4 in series and 10 in parallel.

    R2 consists of 9 100 ohm resistors in parallel.

     

    R3 and R4 are single SMD resistors.

     

    Now the attenuator should be symmetrical with 50 ohm on the input and the output.

    Unfortunatelly the damping ratio is now a little bit odd. It is -22.36 dB. But this should not be a big problem.

     

    Bernhard

  • in reply to bernhardmayer

    Hi Bernhard,

     

    For testing, a spectrum analyzer, or an SDR (e.g. RTL-SDR dongle) can be used too, provided you have a way to generate a signal (e.g. some wireless chip dev-boards have test modes that can do this). See here for some ideas:

    Molex 2.4GHz / 5GHz Antenna Kit - Review

  • in reply to shabaz

    Hi shabaz,

     

    thank you for your answer and the links. They are both very interesting.

     

    Unfortunatelly I do not have a network analyzer and at the moment I do not have access to one. So I cannot do any measurements and can only try to make the design as good as possible. One key element was to use cheap and easy available parts.

     

    The inductance is hopefully low as every path is a straight line and the magnetic field of adjacent paths should cancel itself.

     

    Placing resistors ontop of another is no option for me as every single resistors needs its space to dissipate the heat it generates.

     

    I also tried to regard the design under the statement that every trace longer than one tenth of the wavelength has to be regarded as transmission line. That also means that every trace that is shorter than one tenth will work anyway. In my design the path from the input SMA connector to ground is about 20 mm long. So everything with a wavelength longer than 200 mm should be fine. This means frequencies below 1.5 GHz. Higher frequencies have to be measured.

     

    Regarding the attenuator: You are right. I made it wrong.

    I wanted to make an attenuator but deigned it as a dummy load and now its asymmetrical. I will redesign it.

     

    Thank you

    Bernhard

  • Hi Bernhard,

     

    I don't think this will operate in GHz region. You can see the difference the construction makes, here (there is a small BNC there, that only works to 100MHz for instance): Which 50 ohm Loads are any good? Finding out with the FPC1500 Vector Network Analyzer

    (There is my (awful) home-made construction there, which cannot work beyond HF, but there is also a commercial one (BNC terminator) there which is poor beyond a few hundred MHz (mainly because it is not intended for radio, it is intended for LAN termination).

     

    So from that experience, I believe there is a large amount of inductance in the design, due to the chains of resistors. Is it a dummy load, or an attenuator? I think it's an attenuator, but in this case the best way is to use this website (it is awesome): https://www.changpuak.ch/electronics/calc_02.php

    Using that topology (pi or T), you'll get a perfect match for source and load (albeit still only in theory, since it still depends on construction).

    The topology you have will result in a mismatch at the end marked 'output', so a symmetrical design is preferred.

    In your case, if you need to attenuate a lot, then perhaps an option is to do it in stages, since it gets awkward to attenuate more than (say) 30dB in one stage. Also, another option is to use parallel resistors (e.g. stacked on top of each other) rather than serial strings of resistors.