The Trials and Rewards of EMI / EMC

[NukerDoggie]

I'm involved in pre-compliance EMC work for a prototype medical device I am developing. I wanted to log my progress here for the benefit of anyone else who has similar interests and requirements. I started on the Rigol DSA815 Spectrum Analyzer thread and moved the EMC stuff here to a new thread. I hope some persons will find this beneficial.

In my pre-compliance EMC efforts regarding my prototype medical device I must accurately measure the EMI being radiated from the DUT. To do that I need a good spectrum analyzer, an ultra-wide-band listening antenna and an RF isolation cage so as to block out virtually all RF but that coming from the DUT.

This post has to do with the RF isolation chamber. I have been experimenting with such a chamber and here are my findings, for the benefit of anyone else with similar needs.

I have a portable battery-powered transmitter-receiver pair mfg by Linx Technologies that operates at about 430 Mhz. These are being used as test assets to demonstrate when and if I have achieved a desired measure of RF isolation. I place the receiver in each of my experimental chambers and try to ping it with the transmitter from only a few feet away. If I can, then I have not achieved the level of RF isolation I am seeking. The receiver is extremely sensitive - able to receive signals at levels down to -118Dbm - which equates to little more than 1 picowatt - a very tiny signal level!

I have successfully tested this transmitter-receiver pair at a distance in excess of 2500 feet of separation from each other, with the standard whip antennas, just to illustrate how very sensitive the receiver really is. If I can get this receiver to 'go dead' within an RF isolation chamber then I have really accomplished something.

I have succeeded in doing just that. Here are the details:

I used 10 mil aluminum foil to construct a foil box with a lip all around where the lid clamps on, and a lid itself. The box itself has no holes or voids of any kind, except of course the open top where the lid fits onto it. The lid is clamped onto the box lip all around using large binder clips like you get at Staples office supplies. The box is a cube of about 18 inches square, but the shape doesn't matter. I use 10 to 12 binder clips to secure the lid in place all around, after placing the receiver inside the box. The homemade Faraday cage is not grounded. It need not be in order to snuff out RF radiation.

This cage demonstrates virtually 100% RF isolation at the 430 Mhz frequency. The transmitter is unable to reach the receiver at all. However, if I remove only a couple of binder clips from the lid and wiggle the lid edge just a little, then the signal gets thru. And I mean only a very, very slight gap being allowed between the box and lid.

But with all clamps in place, RF isolation is 'perfect'. No significant gaps exist for RF to leak into the cage, and the clamps also help ensure that box and lid are electrically One Conductor - very important for any Faraday cage.

 I was very pleased to discover that this very high level of RF isolation can be achieved so easily with materials that are cheap. Now, in practice, my isolation chamber won't be made of 10 mil foil, but rather of 0.090" or 0.125" aluminum plate, so as to make a chamber with good structural integrity. But you don't have to spend a lot of money - you just have to be very careful in the design and construction of your Faraday chamber.

Soon - a post on my EMC UWB antenna that is now under construction.

[HLA-27b]

Very interesting. Thanks for sharing.

Now, for the sake of curiosity, what would have happened if you had a power cord coming out of the box?

[jahonen]

Note that box shielding (faraday cage) is only one part of the problem. You want the inside chamber walls to be non-reflective, so that the reflections from the walls won't disturb your readings. You would need ferrite tile absorbers and graphite foam pyramids on the inside walls. Also, standard practice is to have the distance from the antenna and DUT of 3 or 10 meters depending on the DUT size, so it measures more like farfield. The problem with a small box is that you'll be measuring mostly reactive nearfield and get lot of false alarms.

However, you might want to look up a TEM (or GTEM) cell. That is basically a huge piece of a terminated coaxial, where the DUT is put between outer shield and inner conductor.

Regards,
Janne

[nctnico]

I agree with Janne. You'll need a large room with proper shielding and absorbing materials in order to get useful measurements. The materials are insanely expensive. I really think you're better off renting a chamber for half a day to see which frequencies exceed the limit and use near field probes to find the causes of emissions in your own lab.

Take a look at this website:
http://www.rainfordemc.com/

[NukerDoggie]

Very good points in the previous two posts regarding false vs accurate measurements. It is insights like those that you simply cannot put a price tag upon!

Keep in mind, however, that I want to measure BOTH the near field and the far field emissions - the near field levels are important for me since this device will be in intimate contact with the patient's skin.

I am most likely, therefore, to take your advice on renting a standard test chamber AND to also at the same time continue building my own smaller chamber as well. I do need to seriously consider the inner reflections issue, too. And those ferrite tiles and RF-snuffing foams are ridiculously expensive, as you point out. There may be much cheaper alternatives, perhaps. I am looking into fine copper gauze padding that is attached to the inner walls and lid surfaces.

Also, what effect would grounding the box to a true earth ground have upon reflections from the inner surfaces? What if the DUT ground, Faraday box and true earth ground were all made common? Would that sink away the RF striking the inner walls of the box?

And the issue of a small port for power and test leads adds to the complexity too. But there are solutions for such things too.

[Neilm]

I assume that you have copies of the standards you have to meet as these should detail the test methodologies and set ups.

If the DUT is in contact with a persons skin I believe that there is a recognized mounting method that will will represent the affects of EMC when pressed against skin. I have to confess I am not too familiar with these as I don't work with this sort of equipment. I know that small GTEM cells are available for hire or purchase. These range from very small units that can stand on the desk to very large units that take up a whole room and you can walk inside them.

However, if you want to use it to measure the emissions you will need some other equipment. Most standards relate the emissions to the typical results from an Open Air Test Site (OATS). A cell like a GTEM requires some very complex maths to do this. The DUT is tested in three mutually perpendicular axies over the frequencies of interest. The three runs are then scaled to account for the different performance of the cell at different frequencies and then correlated to produce the final results. Fortunately, there are computer programs that will do this for you so if you are looking to hire a GTEM the company will probably have these for hire as well.

If you are planning on doing any immunity tests remember that you will want the area where the DUT is mounted to be as close to uniform as possible. The last thing you want it to have two points in the DUT that have drastically different field strengths.

I would suggest (and have repeatedly) going to the compliance club at http://www.compliance-club.com/. There are many articles on good EMC design.

Yours

Neil

[Construct]

Sounds like you're off to a good start. As you already discovered, the hardest part is sealing the seams. My rule of thumb was to place a fastener every 1/4 wavelength or closer at each seam. If using spring-loaded fingers or similar, you'll want them even closer due to the imperfect sealing.

I am most likely, therefore, to take your advice on renting a standard test chamber AND to also at the same time continue building my own smaller chamber as well. I do need to seriously consider the inner reflections issue, too. And those ferrite tiles and RF-snuffing foams are ridiculously expensive, as you point out. There may be much cheaper alternatives, perhaps. I am looking into fine copper gauze padding that is attached to the inner walls and lid surfaces.

Copper gauze won't help you. You need a very lossy material to dissipate the RF energy. Otherwise it will bounce around internally and potentially cause havoc in your DUT or skew your results significantly.

Take a look at the white papers here (you'll have to register) for some decent information: http://www.eccosorb.com/resource-white-papers.htm

And the issue of a small port for power and test leads adds to the complexity too. But there are solutions for such things too.

There are quite a few solutions out there, but you'll have to be careful about your application. If you're not careful, the cables will conduct the RF energy out of the chamber and re-radiate it. This is even more likely without anything inside to absorb and dissipate the RF energy. Several companies make filtered connectors specifically for these applications, but you'll have to find one with enough bandwidth for whatever communication protocol you're using.

[NukerDoggie]

Yes, I see that metal gauze is a non-starter for reducing reflections inside the cage.

Man, the various RF-absorbing materials are expensive! Tempted to try urethane foam resin with fine powdered iron added (I can get both real cheap) but the problem of even distribution of the iron powder is probably a blocking issue. Besides, iron powder is unlikely to manifest the same RF-absorbing properties of ferrite. It might just make a wonderful reflector, or snuff magnetic fields while doing little to RF fields.

Ferrite tiles look really good for my app (<= 1Ghz) but the price is way out of reach.

I've got the 'rent-a-faraday-room' blues!

[SeanB]

I used to use a ferrite loaded epoxy in a microwave dummy load. Basically a tapered steel waveguide with the thin end of the taper welded shut, and filled with the iron filing/ferrite/epoxy mix. You could get it to glow red hot with a good magnetron, but the epoxy would not survive this smoke test. The portable test chamber had a carbon loaded waffle foam inside, much the same black foam that you get components in, just really big thick sheets. When they were replaced ( fell apart) the new ones would burn white at the tips initially as they were burnt in. was only a 2kw magnetron, but I never did servicing on the ground based radar units.

[Neilm]

Ferrite tiles look really good for my app (<= 1Ghz) but the price is way out of reach.

We were having problems at work with our GTEM and ended up having to put some ferrite tiles on the floor to sort out the problem. Each one tile (100mm x 100mm) was about £5 and we needed 1.5 sq meters.

Neil

[G7PSK]

Copper gauze won't help you. You need a very lossy material to dissipate the RF energy. Otherwise it will bounce around internally and potentially cause havoc in your DUT or skew your results significantly.

This is not a topic I know much about but some years ago I used sintered stainless steel in a filter on an essential oil still, this sort of material ranges from 2 to 200 microns size filtration I would expect that it is very loss'y electrically. The type of material I am referring to can be seen on the link.  http://www.purolator-facet.com/porous.htm

[NukerDoggie]

I'm building my own EMC antenna for use with my new Rigol DSA815-TG (which is to be delivered to me on Wednesday). Now, understand that I don't intend to end up with a certified EMC antenna like those you'd expect to pay $2k or $3k for. But, I want an UWB antenna with reasonable performance which I can use in my own lab, inside my own Faraday cage, to get a good feel for how good or bad my medical DUT is with regard to EMI.

I was inspired by this youtube video:

 

I've chosen to build a solid biconical antenna, but made from .016" copper sheet rather than from aluminum mesh as the one in the video. I'm also making cones with a larger diameter (about 17") in hopes of snagging a bit lower frequencies. I'd like to have an antenna usable from 20Mhz to around 2Ghz.

The cone half-angles will be about 45 degrees, and I'm hoping to achieve down to near 50 ohms impedance so as to avoid having to do any impedance matching - but time will tell on that.

I expect to have this prototype antenna completed within a couple of weeks or so. I'll post a photo or two when I'm done with it. It should weigh in at about 7lbs total weight (the copper weighs about 4.5 lbs. The rest of the weight will mostly be the acrylic support frame. The cost? Including shipping for the materials, about $90 for the copper, $100 for the acrylic sheet and rods (the darn UPS on that was $35 alone!), and another $25 for misc stuff like fittings, etc.

So for about 200 bucks I could end up with an antenna that isn't too much below the expensive ones in performance, if I'm very careful. How will I know how good my antenna really is, though? Off to the lab with it! I'd have to have it professionally analyzed - maybe at a university or maybe at the Colorado School of Mines? I'm only about 100 miles from there.

[NukerDoggie]

Well, I've finally completed the construction of my biconical antenna. See the attached photo. I added a second photo attachment since the angled perspective of the original photo makes the antenna look a little squished vertically.

The biconical is 16 inches in diameter and 16 inches in overall height. Its cones are constructed of 0.016 copper sheet. If I've done my job right, it should be close to 50 ohms. The cone angles are about 95 degrees, and the lead-in is low-loss 400-series 50 ohm coax with a TNC connector embedded into the apex of the top cone. The center connector continues down to connect to the bottom cone. Distance between the two cones is about 4 mm. There is no impedance-matching balun at this point - I'll have to see if one is needed. low-mid freq and wideband ferrites are installed at both ends of the coax lead-in.

Now comes the hard part - finding out how good it is, fixing its shortcomings and calibrating it. I'm going to need a lot of help here - so I hope you'll offer all the suggestions you can.

Many thanks in advance!

[MartinX]

I have been working for a number of years using a rather primitive set up for EMC testing and I think getting accurate level measurements without a proper absorber lined chamber and a calibrated antenna is not a realistic goal, achieving it can be a task more expensive and time consuming than developing the product you want to test, however the test gear you have constructed so far can be very useful and save a lot of money, this is how I do it:

Go to a real test facility and pay for it, there is no avoiding it, their measurement will always be better than yours, the trick is to minimize the time you need, a pre compliance test covering a lot of things can be done in two hours. Being prepared is key, read the standards and make sure you understand how the different tests are done and prepare every cable and other thing that may be needed in advance so you can walk in there and set up your test object and be ready in a couple of minutes. Find out what kind of equipment the lab is using in advance and ask them how long they think the measurements you want to do will take, modern test receivers are significantly faster than old ones so a more expensive lab per hour may get more done than a cheaper lab using older equipment. Make sure they understand you are not working for a mega corporation with a big wallet. If they drag their feet around or spend a lot of time searching for lost coaxial cables politely complain about it and they will extend your test time. Don´t forget to ask what they charge for a test report, sometimes the cost is ridiculous, if so take your own notes and just ask them to print the curves.

Now lets say there was an emission problem around 120MHz, a peak 8db above the limit.

At home repeat the measurements you can using your own equipment, meticulously document how you did it, it is important that it is possible to replicate the measurement you made exactly, making some sort of fixture is a good idea. Your measurement will probably be a bit different compared to the lab but that is not critical, watch out for really large dips or peaks though, if they are close to the frequency you are having problems in, try to reposition the antenna to get them out of the way. If the shielded box proves unworkable try making a measurement out in the open, if you work in a neighbourhood with a lot of transmitters nearby masking the frequency you are trying to measure consider other places you could go to.

Now you have to do something about the 120MHz peak, after making changes to the test object you repeat the well documented measurement you made earlier and the peak is now 15db lower, excellent! There is a good chance that the 120MHz peak will measure 15db lower in the proper lab even if the absolute level in your and their measurement is not the same. This is not a method without weaknesses so there is no guarantee of success but I have never seen a reduced level in my "primitive" A to B comparisons turning out to be nothing or an increase. The ability to do all the fiddly fault finding and experimenting without paying big money for lab time will make your investment pay for itself I am sure.

[free_electron]

Here's my two cents on these kinds of experiments : useless.
For two reasons
1) there are standards and regulations pertaining to what you are designing and the tests it must pass. The tests are very detailed and tell you exact
Y what chamber, how large , damping , what receivers etc.... Any substitution invalidates the test and the testresult. You may have stellar results with you paperclip annex tinfoil hat. Stick it in the real
 Testchamber when you go for certification and there's a 90% chance it will fail.

Nearfield and far field testing is different. You may pass one but fail the other. Tweak the design retest and now the scenario reverses.

You also need to know the working environment. If your medical device is for implant you will need to test  in a body model...

In othere words : mucking around woth homebrew is a sure path to test failure.

I was in singapore a couple of years ago for a drive that failed test in far field.
It failed in the chamber with the antenna at 3 meters height and 4 meter off the device at a precise angle... Good luck with you 10 cubic inch testchamber....
This was a 2.5 inch laptop drive... There was a spur at 3 MHz that was 4 dB over limit.. Turned out to be ringing in one of the switching regulator...

The testsubject is placed on a rotational table and slowly spun. The antenna goes up and down end back and forth and they essentially make a 3d field cloud. The full test is over 4 hours...

We had the chamber for 48 hours... Including two operators to run the test... Total cost 8k$...
After the first run we recorded the position of the drive on the table and the antenna postion. Then a few drives were modded with different filtercaps and tweaked regulators and test was rerun for one position only. This still takes a good 20 minutes. Setup in chamber . Close chamber. ,test, swap drive out , reset setup etc. the two best candidates were then subjected to the full suite.

EMC testing is not something to be done lightly.
Get in touch with rohde and schwarz and amplifier research. They can get you the right stuff.

And then you have only covered emission... What about susceptibility....

Once you know the frequencys that are over limit you can trace with sniffer probes. Easy to make yourself.
Get some rigid coax , solder an sma at the end strip the other end, make a little coil using magnet wire wound around a mechanical pencil refill . Two or three turns is fine. Solder one end to the sleeve the other to the center and slide a bit of copper sleeve over the coil solder the sleeve shut , put a dot of hot melt glue in the tip. You now have a very sensitive pickup.
But.. Beware of magnetic fields !

Bob pease had an article on how to make good sniffer probes.

[Short Circuit]

Yes and no f_e; There's no way of diong any absolute measurements using some homebrew setup indeed.
But same applies to a real EMC receiver with 10k$ antenna's in a non calibrated environment,
and a calibrated chamber is expensive, really really expensive...

Personally, I have had fairly good results from a low cost spectrum anaylyzer and a fairly low cost antenna
set. After having done a real (pre)compliance test, you have a EMI "footprint" of the device, which can be
used to correlate the homebrew measurements to the real thing. Quite usefull to get a coarse idea of the
impact of various circuit modifications on the results.

Susceptibility is another story, but much less relevant for most electronics. Unless you're talking about
safety stuff, life support, anthing where circuit reliability is of any concern beyond warranty claims, I would
not worry about it at all. Apart from that, it's quite hard to fail on immunity. Never had any problems with
that, even with industrial and automotive levels.

[free_electron]

The original poster talks about medical stuff... You can bet it will have to go through susceptibility...
Cant have a bloodglucosemeter go wonky because the cellphone nearby is receiving an email and sends out a couple of chirps to the tower ...

I agree that you can do some sniffing work to get an idea.. And you can get away woth an analyzer and a good antenna , but to do real 'homework' you need a directional antenna , a turntable and a shielded cage... Preferrably an anechoic one ( anechoic for rf that is.. Sticking it in a recording studio doesn't work )

And the first thingyou need is the relevant regulations to know how much energy in which frequency ranges ( it's not flat , but a stepped plot ).

[NukerDoggie]

f_e,

"Useless" is a pretty strong word - but I think I know the context in which you used it and your motives in doing so. I know you have a great deal of experience, and I respect that. I'd much rather have everyone here give me the unvarnished truth as they each see it - I can benefit the most that way.

These most recent posts were just what I was hoping for, therefore.

I would like to reiterate that my homebrew approach at this point in the game is only intended as a precursor to the time and expense I absolutely know I will have to incur in the certification labs. I don't want to go to the lab largely ignorant of what the emissions footprint of my device is. I want to minimize the trips to the lab. I want to make the most of this pre-compliance phase, and educate myself as much as possible, but not step over the line from self-education to self-delusion.

I see the time and expense now, before my first trip to the lab, as time and money well spent, as long as I always keep a realistic and scientifically logical perspective on what I can accomplish in my own lab. When I walk into the certification lab I want to know I've got a device that is very much nearer to, rather than farther from, compliance. The only way I can know that is by taking matters into my own hands now to the extent that is scientifically possible.

There will always be surprises - but I want to do all I can to make them fewer and smaller.

I really appreciate all the advice and insight from all you who are much more experienced in these matters.

[NukerDoggie]

I think the first thing I'd like to do as respects my new biconical antenna is to find a way to 'calibrate' it against something fairly reliable. Please critique me on this:

I have the Rigol DSA-815TG, so I'd like to use the Tracking Generator to transmit to the biconical, which will be connected to the SA inputs. I would place both antennas (biconical and whatever transmitting antenna I decide to use) inside a faraday box to eliminate all but the TG's signal.

The transmitting antenna has to be wideband, of course, and has to have a known, documented 'signature' across the 100khz to 1.5ghz range. What I end up with on the SA display is a 'signature' for my biconical - and I learn how good or how bad it is at various frequencies.

If my Faraday box is good, as far as blocking out interfering RF, but not so good as far as snuffing internal reflections, I can live with that for the time being. It is a problem I can address a little later. First, I need to produce a signature for my biconical, even if it's somewhat messy due to internal reflections. I can then tweak the antenna to improve the signature, to try to get it to match that of the transmitting antenna more closely.

As for the transmitting antenna - I could rent one. I could research the possibility of building one, if I could find detailed plans/designs that would give reasonable confidence of obtaining a desired signature in the finished product. I simply need a wideband antenna that will have a characteristic signature for the frequency range I'm interested in - up to 1.5 ghz.

I'm posting this here in case anyone knows of a simple transmitting antenna I could use or build, one that would have a characteristic signature I can count on if I follow the design precisely. It doesn't have to be one that has a perfectly flat response - it just has to be useable across the freq. range and has to have a known signature. I can then calibrate my biconical to it, and create a rough conversion table to convert my biconical readings to actual values, based on the reference signature of the transmitting antenna.

Then, once I have a biconical that is 'calibrated', I can place several already-certified compliant devices (one at a time, of course) that are similar to my device, under test, take readings, make a reference chart, and then place my device under test. I can then determine how my device compares to known certified, compliant devices.

That won't be a guarantee of compliance for my device, but it will be a strong indicator, and should give me clues as to where I need to focus attention on reducing emissions. For each mitigation effort, I can re-loop thru the tests to see the effects.

But I have to start with calibration of my biconical - not trying to make it like the ones found in the cert labs - that is way beyond my capabilities. But I need to establish a reliable signature for it in the range up to 1.5 ghz so I know, at each frequency, what its response is relative to something fairly trustworthy.

[MartinX]

Trying to calibrate a homebrew antenna with another homebrew antenna in a shielded box is not going to work, the number of unknowns just went through the roof. All you can hope to achieve in your home made lab is comparing two measurements, the absolute levels you measure will never be accurate.

I would not bother trying to calibrate that antenna at all, there is no point in knowing the antenna correction factor is 5db in some frequency, the error in your measurement set up could easily be 30db or more.

At work I have a 10m2 shielded room to work in, no absorbers, this works better than I initially thought. We have had this set up for almost ten yeas, and by comparing our measurements to real lab measurements of some 15 products trough the years our estimates are now not totally off, sometimes the correlation is spot on but that is just luck.

I was in singapore a couple of years ago for a drive that failed test in far field.
It failed in the chamber with the antenna at 3 meters height and 4 meter off the device at a precise angle... Good luck with you 10 cubic inch testchamber....
This was a 2.5 inch laptop drive... There was a spur at 3 MHz that was 4 dB over limit.. Turned out to be ringing in one of the switching regulator...

I assume the spur was at 3GHz if it was that directional, anyway picking it up in a shielded box is not that far fetched, without absorbers the signal will  be reflected bouncing around reaching the antenna eventually, kind of like the stirred mode chambers with rotating reflectors in the corners but you have to do the stirring yourself moving the antenna and test object around. I am doing pre pre compliance tests sometimes, if I pick up nothing then I move the antenna and test object around a couple of times and measure again, still nothing then probably all is good. If I get a spur 10db over the limit, crap, difficult situation, is it worth it to try to get rid of it or am I chasing ghosts? I have never had an unseen signal popping up in the real pre compliance tests but I have been chasing ghosts sometimes. In these kind of tests if you see nothing then nothing is probably emitted, if you see something, investigate if it is easy to rid of, if it is, do it, if it is difficult or expensive to get rid of  then you just have to do a better measurement at a lab to see what you are dealing with.

[Rufus]

I think the first thing I'd like to do as respects my new biconical antenna is to find a way to 'calibrate' it against something fairly reliable.

You are screwed really and are going to have to pay for use of proper test facilities. MartinX has the best advice here.

You are basically hiring the test chamber, equipment, and an engineer for a day or half day and it isn't horrendously expensive. If you haggle/plead and are flexible you might get some deals when they are light on work.

They ought to have calibrated broadband noise emitters to check their equipment and would probably let you calibrate your antenna (and SA) against them.

This chamber (DUT obscured) with remote control rotating platform, wind up/down H/V log periodic, the equipment outside and an engineer goes for around US$1500 a day for pre-compliance work.

Like MartinX says be well prepared and smart about get the most from your money. There is no harm checking for and tying to fix anything that looks like it might be a problem with a completely uncalibrated lash up in advance.

[free_electron]

I would focus on doing a good design first. After all , the testing is only an indicator of how good the design is. If it fails you 'll still have to do the work to make the device properly ...

Control your edges , trap everything you can in transmission lines. Avoid impedance mismatches and use planes properly. That will get you 90% there.
points of attention :
- fast clocks
- switching regulators especially with synchronous rectifiers.. there is such a thing as switching too fast ... sometimes injecting a 0.1 ohm resistor in the tank capacitor path dampens peak currents enough so they don't radiate that hard..
- anything digital...
- anything that can be catalogued as 'rf' or containing 'rf' ... RF = radiating frequencies ...
- if your system has signals leaving the confined space : ferrite beads and proper i/o mechanisms

[G7PSK]

Instead of messing around with tin foil to make a Faraday cage why not use an old microwave oven stripped out. I use one to house a large dummy load.

[NukerDoggie]

I'm not messing around with tin foil. I merely conducted an experiment to see how thick the material needs to be in order to be quite effective.

As a matter of fact I have constructed an "RF Coffin", where all RF dies except the kind I want to 'come awake at night' like Bala Lugosi.

It's a brand new large aluminum truck tool box (24 x 24 x 72), seamlessly welded, made from 0.060 aluminum sheet. The lid gasket has been replaced - it's not rubber, but multi-layer fine pure copper mesh spiral-wrapped around a very durable foam center. It is seamless, running all around the lid where it mates with the box. The mating surfaces of the lid and box have been finely sanded to remove oxidation, and a good layer of anti-oxidant designed for copper-aluminum connections has been coated on all these mating surfaces. When the lid closes it compresses the copper gasket about 40%, and seals out all unwanted RF.

Total cost - $300 for the box, $32 for 100 feet of 6 inch wide copper gauze, $10 for the dense foam gasket and $7 for the anti-oxidant paste.

This RF Coffin is large enough to hold my new biconical antenna along with just about any DUT I choose to place in there with it. Much, much better than an old microwave oven stripped out.

Bela would be proud!

[SeanB]

Good idea and implementation, as the microwave would only work if you replaced the door seal with the copper foil, as in a microwave the seal for RF is a quarter wave dipole in the slot, so that it is a high impedance to the 2.4GHz microwaves only. It will block WIFI, not GSM calls if you have ever placed a cellphone inside and called it.