EMC Chamber build log

[PedroDaGr8]

This thread is badass. I never realized how much work went into making a chamber like this. Very impressive and thanks for posting it.

[Tabs]

...white polystyrene tiles are glued into shape and fitted on top of blue pyramidal absorbers. ...

Are those just impact protection, or do they actually serve an RF purpose?

The white tiles are made from polystyrene which Nick tells me is completely transparent to RF so they offer no technical/performance benefit to the chamber characteristics. They are added purely for protection and decorative purposes. They make the chamber brighter to work in and protect the tips of the foam absorbers.

Currently, all the pictures I've taken inside the chamber are lit with the actual lights that were intended to be installed in the chamber (I thought they were temporary work lights). They were plugged into a temporary extension chord because the chamber hasn't been wired by the electrician yet.

The lights were trac pro 70w SON lights which are sodium based and give a very yellow/orange light similar to street lights here in the UK. I asked Global to replace them with something whiter because it was too gloomy inside. Global don't normally use SON lights and were happy to replace them.

The final lights (metal halide) are still wired up using a temporary connection but give a much better light.
See the pictures below:

Tuesday 22nd April (After bank holiday)

Lights changed to metal halide, ferite tiles added to floor and pyramidal absorbers placed on top of those. Nick arrives to begin next stage of testing (All the pictures below are from Nick and show the test setup he is using. The equipment is his as well)
White tiles fitted on top of all side walls, with tiles missing in some areas (Global sent a full complement, but murphys law got us). Replacement tiles were sent down free of charge. Global had to deliver some of Nicks test gear so they sent the replacement tiles in the same shipment.
The picture is taken at the corner where the door is and is looking towards the back of the chamber where my antenna will be placed.

Here's another picture showing the same view. This time you can see a walk way on the right that leads up to the antenna area.


The ferrite tiles and pyramid absorbers on the floor are not fixed in place. This allows us to move the floor around.
Nicks tests will assume the floor is going to stay in this position. Global EMC call this an offset chamber which means the antenna, floor absorbers and EUT are not on the center line of the chamber. This allows us to have the walk way on the right. Nick will check to make sure we can establish a uniform field around the EUT.

Another view of the antenna area, taken from the EUT position.


Final picture showing one of Nicks setups. This time taken from the antenna area, looking towards the EUT area.

[lewis]

That looks awesome. What measurement setup are you using?

Are you likely to get interference from the ballasts in the halide luminaries?

[Tabs]

That looks awesome. What measurement setup are you using?

Are you likely to get interference from the ballasts in the halide luminaries?

I'll leave Nick to talk about his setup since I don't know it in enough detail.

In normal operation, the lights are switched off during emissions and immunity testing and no one will be inside the chamber. I've been inside chambers whilst doing mil-std-461 tests but this is when we start approaching the higher Ghz. By the time we hit 18Ghz the chamber doors were wide open.

Our plan is to to do conducted immunity, ESD, surge/bursts etc inside the chamber and for these types of tests, the lights shouldn't cause a problem.

[lewis]

In normal operation, the lights are switched off during emissions and immunity testing

Of course! Senior moment 

[gregariz]

It's interesting to see the varying types of constructions of chambers out there. Is this a common construction method in europe?

I've been involved in a couple chamber builds and most in the US that I have seen offered in recent years are constructed simply of about 1" double plated fibreboard. The panels and corners are then clamped together using a steel profile, with the joints stuffed with wool.

[Tabs]

It's interesting to see the varying types of constructions of chambers out there. Is this a common construction method in europe?

I've been involved in a couple chamber builds and most in the US that I have seen offered in recent years are constructed simply of about 1" double plated fibreboard. The panels and corners are then clamped together using a steel profile, with the joints stuffed with wool.

Again, this is probably one for Nick or Global EMC to answer. I've used different types of chambers before but this is the first time I've been involved in building one so I don't really have experience of other methods.
I'm assuming the fiberboards are plated with with something good enough to act as a shield and that the wool is what I was calling the EMC gasket?
I don't see anything wrong with this type of construction. Using a steel profile may make the chamber stronger than the way this one constructed (especially if using steel profiles on both sides of the plated fibreboard). Because my chamber is small its constructed using only 1 and half panels per row which are arranged like a brick wall pattern. This makes it very strong, especially when combined with the structural steel supports around the outside and not likely to sag/bow in the middle.
I've been told the the ceiling can support an additional 1 tonne of weight so this style of chamber construction is very solid. The stage 1 tests done by Nick to measure the shielding effectiveness showed the chamber was achieving between 115-120dB of shielding so the construction style is definitely good enough in terms of EMC performance.

I'm guessing the construction style is standardised across all of globals chambers. The chamber installer told me he's done loads of chambers for global across the UK and at least one in Switzerland. Can't say for sure if this style is specific to globals designs or if its normally done this way in Europe.
Global can answer that one best.

 

[Nick Smith]

Hi,

Firstly regarding the chamber construction using double sided boards - we call this a 'sandwich' system chamber. It's not very popular now in Europe or Asia but still seems to be the preferred system in the US. I've tested many of these and many of the 'pan' type rooms (like this one). I can categorically state that the pan-construction rooms are capable of higher shielding performance, however the sandwich systems are usually good enough for most purposes. I've no idea why the pan system hasn't really caught on in the US yet. The sandwich construction is a bit more flexible because it can be cut on site if required whereas the pan system has to be made correctly in the factory before shipping to site.
The 'wool' mentioned is usually bronze wool - this is similar to steel wool - it's commonly used in the leisure marine industry because it doesn't rust, but it's also useful for stuffing the corner joints in sandwich rooms!

As far as the chamber testing goes (apologies for the poor photos BTW - I've got loads of better ones but seem to have given Tabs all of the blurry ones) - we have basically performed two tests, one is a 'Normalized Site Attenuation (NSA) test and the other is a Field Uniformity.

The NSA test essentially measures the loss between a pair of antennas and compares the measured loss with the theoretically calculated loss for free-space. This is in accordance with the standard CISPR 16-1-4. In this chamber, we treated it like a Fully Anechoic Room (FAR) which has absorber on the floor rather than a semi-anechoic chamber which has a reflective ground-plane and uses a height-scanning antenna mast.
In this case, we actually used the 'Reference Site Method', which means that the antenna pair were first measured on an extremely good reference site (the National Physical Laboratory (NPL) in the UK) and we then do a direct comparison with this as the reference site and measure the difference between the two sites.
This is the test where you can see a pair of antennas - one antenna is a miniature biconical antenna, which is used as a transmit antenna and we use a biconical or log-periodic antenna as a receive antenna. The measurement is repeated at a number of positions defining a test volume inside the chamber where later on the Equipment Under Test (EUT) (i.e. Tabs' equipment) will be located.

The second test is a Field Uniformity test in accordance with the standard IEC 61000-4-3. This uses a transmit antenna at one end of the chamber and an isotropic field probe to measure the field at the EUT end of the chamber. The field is measured at 16 points which describe a test plane of 1.5m x 1.5m (W x H), and we then compute the variation in the field across this plane. To be honest, in this chamber the 1.5m x 1.5m plane is larger than really required, because it is a compact chamber and the EUT is relatively small (i.e. much less than 1.5m^2), but in order to prove the chamber performance we measured over the whole 1.5m plane anyway.

As always, I'm happy to answer any questions on the testing process.

Regarding the lighting - as far as I am aware, there are no significant emissions from the metal-halide lights that have been installed. Fluorescent lights can definitely be problematic as can some of the newer LED lights, so it's good to be aware of potential problems because they are not all good. There should be no problem with the lights installed in this chamber even if they are left on during testing. It's simple enough to check the emissions anyway if required.

the white polystyrene tiles (Styrofoam in US speak??) are really for decoration only - they are pretty much transparent to RF, so have no effect on the chamber performance. They do however help to reflect light and make the chamber a brighter workspace. They also look quite nice and most of the chamber suppliers do offer these nowadays however they are really only for cosmetic purposes.

I haven't worked out how to post pictures here yet, but if somebody can explain how, the I have loads of non-blurry images of the testing if anybody is interested...

[gregariz]

Hi,

Firstly regarding the chamber construction using double sided boards - we call this a 'sandwich' system chamber. It's not very popular now in Europe or Asia but still seems to be the preferred system in the US. I've tested many of these and many of the 'pan' type rooms (like this one). I can categorically state that the pan-construction rooms are capable of higher shielding performance, however the sandwich systems are usually good enough for most purposes. I've no idea why the pan system hasn't really caught on in the US yet. The sandwich construction is a bit more flexible because it can be cut on site if required whereas the pan system has to be made correctly in the factory before shipping to site.
The 'wool' mentioned is usually bronze wool - this is similar to steel wool - it's commonly used in the leisure marine industry because it doesn't rust, but it's also useful for stuffing the corner joints in sandwich rooms!

Thanks for that. Not having seen a chamber go up in europe I had not seen that style before. It certainly looks the goods. I suspect though the sandwich construction does go up quicker. I saw a 12x8x12 go up in a day and a 24x12x12 go up in 2 days. However these were standard kit sizes I believe so there was no cutting on site. And that was just the box, no fitout.

[lewis]

Hi Nick, would love to see some more pics. Click reply, then click "Attachments and other options" under where you would usually write your reply. Under 'Attach' -> choose file... etc.

[T3sl4co1l]

the white polystyrene tiles (Styrofoam in US speak??) are really for decoration only - they are pretty much transparent to RF, so have no effect on the chamber performance. They do however help to reflect light and make the chamber a brighter workspace. They also look quite nice and most of the chamber suppliers do offer these nowadays however they are really only for cosmetic purposes.

Transparent?  Are you kidding?  Those damn things will vastly increase the ambient radiation, especially in the 400THz+ band!

... 

Tim

[mikeselectricstuff]

In normal operation, the lights are switched off during emissions and immunity testing

Of course! Senior moment 

I wouldn't have thought halide would be the best choice for lights you want to be turning on & off multiple times - warmup time & restrike issues?

[Nick Smith]

Thanks - hopefully these pictures will work:

[Nick Smith]

Transparent?  Are you kidding?  Those damn things will vastly increase the ambient radiation, especially in the 400THz+ band!

... 

Tim

They don't increase the ambient radiation, they simply reflect whatever is already there. There are no significant sources of Terahertz emissions in the chamber to be worried about, so the reflections at those frequencies are no great concern.

[T3sl4co1l]

If that were true at HF as well, all those pricey tiles and foamy things wouldn't be necessary.

[SeanB]

Metal halide lights with a magnetic autotransformer ballast will not have much electric noise over 1kHz and it will mostly be low harmonics of the line frequency, caused by the non linear resistance of the arc in the lamp. There will be a small amount of wideband noise generated by the lamp itself, but most will not make it out of the lamp housing if it is metal. There will be a larger noise generated during the lamp startup as the ignitor generates a very high voltage to start the lamp, but this will stop after about 30s as the lamp warms up. If you want to leave the lamp in the chamber you might want to open it and place a ferrite bead on each lamp lead inside the case, close to the lamp holder. Otherwise you just have the lamp power cord very long and a hook on the outside to hold it, and a power point to connect the lamp so that you can simply carry it in when needed, and take it out still lit and leave it outside during testing, so you do not have the 3-10 minute restrike time of the lamp.

[Tabs]

In normal operation, the lights are switched off during emissions and immunity testing

Of course! Senior moment 

I wouldn't have thought halide would be the best choice for lights you want to be turning on & off multiple times - warmup time & restrike issues?

Nick and SeanB are saying its not an issue if the lights were one so I guess its better to leave them on if they've been turned on already.
Didn't think about lifetime of the metal halide lights if they were switched in short intervals. Thanks Mike, I would've got through a lot of these if you hadn't asked the question.
Would the effect be visible? (similar to the black stuff that develops at the ends of fluorescent tubes? or would I have burnt out the starter?)
 
Would be interesting to do an emissions test with & without the lights on and see how much wideband noise is added.

[lewis]

Another dumb question - how is the shield grounded? Did you need to dig in earth rods all around the chamber and use big beefy straps, or is bonding back to the consumer unit with a bit of 10mm² adequate?

[Nick Smith]

Another dumb question - how is the shield grounded? Did you need to dig in earth rods all around the chamber and use big beefy straps, or is bonding back to the consumer unit with a bit of 10mm² adequate?

There is no need for a special ground for RF purposes. Some people feel the need to put in a special earth rod for their chamber but it makes no difference to it's performance. A faraday cage will still work whether it is grounded or not and a typical ground wire is a pretty high impedance at RF frequencies anyway.

It is important to have a good safety ground though - the power filters have big Y-Class capacitors between the lines and ground so there is significant earth leakage - certainly more than enough to take out an earth-leakage trip (RCD).

[SeanB]

Metal halide lamps fail by having a rising internal pressure as they age, and this translates into a rising arc voltage until it eventually is too high for the ballast to maintain the lamp current, so it fails by cycling on and off continuously. You will see a lamp with black and coloured ends near the internal arc tube terminals as normal, it is merely metal halide and metal deposits, and indicates that the lamp has only been run for a short time then turned off. Run it for an hour or more and it will go mostly away, but any lamp that has been used will develop this to some extent. Even a brand new lamp might have some from testing during production.

The standard failure mode is cycling for a long time, then bursting of the inner arc tube. Please note that if the lamp is used in a fixture with a missing tempered glass cover, or is used in an open fixture, it must be a blast protected lamp, which has an inner guard of quartz glass surrounding the arc tube to contain the blast of the arc tube failing. If you use a regular unprotected lamp there and the arc tube ruptures then it will blow up the outer envelope as well, showering the area with red hot glass. If you use a fitting with a broken glass front and the lamp gets wet it will also explode.

[Pat Pending]

That's a very interesting build - thanks for posting.
With a lot of foam in there, what precautions can you take for fire?

[Nick Smith]

That's a very interesting build - thanks for posting.
With a lot of foam in there, what precautions can you take for fire?

The foam used in the absorbers is treated with fire retardants and meets the requirements of NRL 8093, which is the US Naval Research Laboratories standard for urethane foam absorbers.
You do still have pay attention to what you are doing though because it is still possible to burn the absorbers even though they meet the NRL requirements. You need to particularly careful when using very high power RF amplifiers and high gain horn antennas, since this can cause heating inside the absorbers however this shouldn't be a concern in most chambers for commercial EMC measurements.

[SeanB]

The absorbers I worked with for RADAR testing did tend to cook themselves with the first few hours of running them when new, but afterwards they worked well, even though they were rather brittle then, as opposed to them being flexible like foam sponge when new. Those will probably get a little brittle with time as the plasticisers evaporate from them, but they will still work so long as they are mostly intact, a few chipped edges here and there are not going to make much difference.

[Tabs]

Been off work the last few days.
Here's some pics taken by my colleague:

Timber frame going up to separate the corridor from the chamber control area followed by plaster board and skim.
Once this work is complete, the electrician will wire the chamber to its proper supply. We'll probably call Nick in again once its all cleaned.

[Precipice]

This is fascinating, thanks for posting.
I'm currently reassembling a rather larger third-hand chamber from the same company - the new stuff looks far, far easier to build! (Still, what did I expect for £200 on ebay, buyer collects

2 questions - anyone got a source for that dogbone-shaped gasket strip? I was going to reuse the nickel felt gasket material I scavenged from the last build (which looked like it had already been reused), but the crushable dogbone stuff looks like it'll be easier, and might soak up uneven-ness a bit better. (The structure of my chamber is galvanised steel sheets on a wood and ply frame, with gasketted aluminium closure plates covering the joints, so I need the gasket to bite into both galvanised steel and aluminium). I've got all the panels up, so the closure plates are next, and, since it's 2500 self-drilling screws, I'd rather not have to do it twice because I chose the wrong gasket...

Second question - those polystyrene cone-covers, are they a GS thing? They're tempting, since some of my cones are much, much bigger (no ferrites, so the cones are trying to work down to a much lower frequency), and I've got cones (and raised walkways) on the floor. Stopping the cones getting damaged, making the chamber less of a black hole for light as well as RF, and reducing the number of things I lose in the cone field on the floor, all sound like good things.

Steve