At the start of 2023. NLnet decided to sponsor the development of BB3-CM4.

As we mostly designing HW and we need to comply with EMC standards and regulations we decided to run some in house
experiments to find or design affordable opensource equipment for pre-compliance measurements.

Chek it out?

Make sure you read my prevews article, as there are lots of info in that one

Starting up!

So now that we have equipment, let's find the best house appliance that could be used as a testing chamber.

After some tests, I have concluded that a microwave oven will be best to block external signals.

That is also logical, as you do not want signals to escape outside the oven - you want to keep them in.


Then I searched around the internet for EMC microwave testing, and even found something!

So let’s get one microwave from the recycling yard!


As I have worked for years (around 15y) as repair technician I really like to repair and reuse!

As my good friend is near an electronic recycling yard and he is often there to check for something he can re-use or repair.

I asked him if he could find one microwave oven that does not need to work.

And the next day the microwave oven was delivered!

I have removed all parts ( as we just need the box ) and did a detailed cleaning.

Adding up!

I wanted to have two antennas inside, but also power supply. 

First antenna goes on top!

For power, I have decided to go with USB as that will be good enough for this experiment

I have used aluminum tape to seal all the holes I have used to distribute power!

If you read the first article I have linked you will get the idea of what two antennas are used for.

The one antenna would be used for sending known signals, with that signal activated we use MAX HOLD on the second one to get the strength of the signal on each frequency. 

Then we replace the sending antenna with the real device and do HOLD MAX recordings.

If recorded signals from the device compared with transmitting antenna signals are lower on every frequency, then our device is emitting under that limit.

Testing if ULX3S is powering UP.

ULX4M as DUT (device unter test)

Here is one picture with second antenna.

First you do meassurements with just emitting antenna, then you add DUT and take meassurements again - second antenna can be used as holder :)

Initial measurements

After extensive testing of many devices, I have found that Raspberry Pi 1 is a great device that can be used instead of the first antenna.

I have read on the internet that it barely passes EMC testing. Some people even reported that they failed. 

So that was a good device for comparison with the ULX3S ( old versions).

RP1 was running some version of Linux, as most noise is generated when you use SDRAM.

ULX3S was running PAWSv2 it was using SDRAM.

ON the picture is showed some kind of chamber calibration.

So firstI took long MAX HOLD on empty chamber (yellow trace).

Then I did meassurements of first, and second device.

After that I just select SUBSTRACT trace1 on trace 2 and 3.

That ignore all the noise that is in the chamber before device is powered up!

Once you do substract TinySA makes a mess with scale, but we do not even know te scale, we just need to compare two devices and confirm that one is better or not!

I am also using pre-amplifier here, and TinySA does not know that so even first empty chamber is out of scale...

Microwave conclusion

So you can get decent results with microwave tests plus probe tests to locate what is out of the specs.

But we know microwaves are built in a way to reflect all signals - so we will get lots of reflections and our measurements will not be as good as we want.

DIY EMC chamber

So maybe we could build a small EMC chamber from the start with some non reflective material.

After extensive research, I have found some foil that could be great for our purpose ( on one side it has reflective - probably aluminum part, and on the other side it has some magnetic absorbing material ).

As I was looking into the datasheet I found one picture that shows how to measure how blocking material is.

As that was clumsy to do without any holders

I decided to create open holders that will hold two probes and have some spacing where we put material.

One probe is used for sending and the second will pick.

And now we can test how blocky different materials are.


After testing many materials I had around in my house I have concluded that this special foil is the best choice, and second best choice is just a simple aluminum tape!

So I have decided to combine those two!

Shiny is not always blocky!

What I have also discovered that shiny "aluminum" cake cary material is not blocky at all!

The Box

Next is to build a box.

I first used OpenSCAD with a great box plugin.

Just set assemble to false and you will get dissasembled box


Then I exported the OpenSCAD Model to DXF - and imported it into the laser cutting tool LightBurn

LightBurn is not open-source but it is still a great tool with a great price so it is worth supporting ( if you have a laser ).

Isolation is the key!

Once the box was cut and engraved I needed to isolate it.

First I will use just ALU tape, as that is available everywhere and it is easy to get.

Then I will do measurements.

After that, I will add special foil.

This was first cut so it was still not isolated as I did not know this shiny is not blocky!

Antenna is not fixed!

As antenna was not fixed, I have created (in OpenSCAD) antenna fixer :)

It is also good to mention that those fois can be bought in two different types, here is link to one of the types.

It depends on what you need but they have sticky parts on different sides, so if you want to stick that foil make sure you buy the right version as the magnetic side needs to be from inside of the box - signals from our device first need to hit that magnet so that we do not get any reflections.

Close sesame!

As doors needed to be frequently opened and closed I have added magnets, so now I can push doors and close the chamber.

Internal decoration

Internal view with the special foil in place!

Foil is still not glued, as we have plan for a new chamber, and I will try to re-use parts - specially foil as much as I can...



I did a few measurements, but the microwave oven measurements were not on the same graph.

Luckily for getting both measurements, I used EEZ Studio, but export/import from the graph was unavailable. So I contacted Envox and in a day Martin ( main developer ) added JSON EXPORT/IMPORT

With the new version, I could export data, and add all measurements on one graph.

The results are here: You can see that we can use a microwave oven for measurements, but our custom DIY box is even better than a super closed microwave oven!

Sweeping from 10MHz - 3GHz

Just to add quick sweeping test.

I have put one TinySA ultra inside of the box and it is doing sweeping test at -40dBm - I did that 3 times with 3 different ranges

Second one that is recording signal is set to HOLD MAX and it is connected over pre-amplifier.

Yellow line is HOLD MAX before emiting device starts to sweeping.

First range is from 10MHz to 1GHz - so I would be confident on meassurements that starts from 100MHz as bellow some FM radio is still making true the box...

Ignore small drop lines, TinySA is scanning to slow to get everyting in two passes - I would need to take at least 3 passes for 1GHz range

One pass from 1GHz to 2GHz - looks perfect!

One pass from 2GHz to 3GHz - my wifi router makes into the box!

Router is one meter away from the box, and my phone is really close to the box.

But I would say if you move a little bit away from the signal that range could also be usefull!


Project link


Radiona/intergalaktik discord:

Envox discord:

Many thanks to NLnet for sponsoring our work!

By: Goran Mahovlić

This project is funded by NLnet

This project was funded through the NGI0 Entrust Fund, a fund established by NLnet with financial support from the European Commission's Next Generation Internet programme, under the aegis of DG Communications Networks, Content and Technology under grant agreement No 101069594.

© Intergalaktik