DIY semi-anechoic chamber thoughts & ideas

[zebra50]

Hi!

I wanted to build some kind of anechoic box for comparing microphone response, and am canvassing thoughts and opinions. I wondered if anyone else had been down this road with success, or otherwise. Anyway, here's my basic thinking so far.

1. Size & shape of the box:

I have a decent amount of space in my workshop, which is under our studio live room, and is about 5.5 metres x 12 metres - currently has concrete floor and brick walls. Ambient noise is pretty low, unless the studio is in action, in which case I wouldn't be making measurements.

. I could probably make a room-within-a-room, up to 2 meters high, and around 3 meters x 5 meters. But obviously small is better for me in terms of cost and space utilisation. If I can get it do double up as a 'dry' room for recording then I might be able to allocate more space.

Reading around I understand that bigger is better in terms of controlling low frequencies.

The video here shows Thiersch's chamber - (jump to about 18 minutes and 10 seconds). It looks a bit over a meter wide but long and thin.

http://www.erf.de/index.php?content_item=137&node=68#68/137

Is rectangular the way to go? Or would something asymmetric help fight the standing waves?


2. Construction

My basic plan was to make a wooden frame then a couple of feet thick with wire mesh either side and filling with roofing insulation. Then I'd cover with fabric on the inside to keep and line with thick acoustic tiles.

I can't decide what to do on the outside of the box - whether to use plasterboard, wood, or just fabric again and let the bottom end go through the 'walls' into the outer room.


3. Measurement

Speaker & amp with a flat(ish) response. Certified measurement mic for comparison. Swept sine wave / impulse response type measurements.

I guess I'd install whatever I have lying around and then upgrade when I know what I need. I'd imagine that a concentric monitor or just a single speaker would be best to avoid phase issues.

One big concern is positioning of the thing under test - I guess the box would need to be profiled to find the area of flattest response.

Well, that's about where I'm up to. Any thoughts or shared experiences gratefully received.

Cheers!

Stewart

 

[zayance]

Nice video of M.Tiersh, too bad i don't understand German  , his anechoic chamber looks nice, and easy to make, i wonder also about the rectangular form, interesting post, i'll have to check back.

 

[zayance]

He wasn't talking technical actually, he was talking about the fact that after the wall in germany, he was in big trouble, no orders nothing and until 1993 were everything started back, and at one point in the video, he shows one microphone headbasket,  he says to be the future membrane he's working on, that's will be a revolution from whatt he says, or something like that, my cousin from germany just translated, just to let you know, if you were curious like me.

 

[abbey road d enfer]

Hi!

I wanted to build some kind of anechoic box for comparing microphone response, and am canvassing thoughts and opinions. I wondered if anyone else had been down this road with success, or otherwise. Anyway, here's my basic thinking so far.

1. Size & shape of the box:

I have a decent amount of space in my workshop, which is under our studio live room, and is about 5.5 metres x 12 metres - currently has concrete floor and brick walls. Ambient noise is pretty low, unless the studio is in action, in which case I wouldn't be making measurements.

. I could probably make a room-within-a-room, up to 2 meters high, and around 3 meters x 5 meters. But obviously small is better for me in terms of cost and space utilisation. If I can get it do double up as a 'dry' room for recording then I might be able to allocate more space.

Reading around I understand that bigger is better in terms of controlling low frequencies.

The video here shows Thiersch's chamber - (jump to about 18 minutes and 10 seconds). It looks a bit over a meter wide but long and thin.

http://www.erf.de/index.php?content_item=137&node=68#68/137

Is rectangular the way to go? Or would something asymmetric help fight the standing waves?


2. Construction

My basic plan was to make a wooden frame then a couple of feet thick with wire mesh either side and filling with roofing insulation. Then I'd cover with fabric on the inside to keep and line with thick acoustic tiles.

I can't decide what to do on the outside of the box - whether to use plasterboard, wood, or just fabric again and let the bottom end go through the 'walls' into the outer room.


3. Measurement

Speaker & amp with a flat(ish) response. Certified measurement mic for comparison. Swept sine wave / impulse response type measurements.

I guess I'd install whatever I have lying around and then upgrade when I know what I need. I'd imagine that a concentric monitor or just a single speaker would be best to avoid phase issues.

One big concern is positioning of the thing under test - I guess the box would need to be profiled to find the area of flattest response.

Well, that's about where I'm up to. Any thoughts or shared experiences gratefully received.

Cheers!

Stewart

I have some experience regarding measurement of microphones (having been the distributor for NTI, I went to all the training courses, and also for E-V and Audio Technica). I'm still involved in a lot of measurements on loudspeakers.
I've experimented with all sorts of boxes intended for proper QC of loudspeakers when they get off the line.
I have access to a proper anechoic chamber that's roughly 4x5x7 meters. It is only certified down to 90Hz. All the boxes I have experimented with would only have been certified down to 500Hz. The size of Thiersch's chamber makes it accurate down to 2-300Hz only.
And it's not a big problem, because what you want to do is differential evaluation. You won't be capable of producing absolute frequency response graphs, you'll have to apply a correction curve your measurements.
And you can do very close proximity measurements for the lowest frequencies, which you can't do for loudspeakers because of the sheer size of them, and their less controlled directivity.
I would recommend a symmetrical shape, although non parallel walls seems a good choice. The reason is the directivity pattern may be twisted by non parallel walls.
Regarding construction, you don't need insulation, you want absorption. Since it is almost impossible to have LF absorption given the basic parameters, your idea to "let the bottom end go through the 'walls' into the outer room" is probably a good one. You may try using some Helmholtz resonators to tame resonances, once you assess them. Again, this is made possible by the small size of the space that needs to be linearized.
Reference loudspeakers are quite expensive and somewhat limited (most of them use a compromised sized element, heavily electronically processed to extend the frequency response).
I have definitely excluded their use (again in light of not needing absolute measurements) in favour of using double measurements, one with a 12' woofer (60Hz-2kHz range) and a tweeter (500-20kHz), and some post-processing of the results.
Positioning of the DUT has never been a major problem for me; I've always found relatively easy to find an area where the response is coherent and continuous.
If you want to do directivity curves, you may want to provide a way of rotating the mic from the outside of the iso box.
Hope this helps

Regards

 

[zebra50]

Thanks Abbey. That's a big help and lets me know my thinking is not completely stupid.

Comparative, rather than absolute is all I hope to get, and indeed all that I need. I like the woofer & tweeter idea. I'll probably start out with speakers from my scrap pile and experiment until I have a better idea of what I want. Current plan is to build a long thin tapering room, with the long axis being the measurement axis and lots of absorption especially at the long ends

Rotating mic mount is a good idea for polar patterns - I'll worry about that once I've got the basics sorted out.

 

[PRR]

A classic technique is "outdoors". Bury a pretty-flat speaker face-up flush with the ground, hold the mike up with a 10-foot pole. Use pivot and string to get directivity. Take a break when airplanes fly over.

That would work better here than back in my old house a half mile from the interstate highway. The rumble never stopped. Also the neighbors would complain about test-tones.

 

[QUEEF BAG]

+1  for outdoors if feasable

that's a good way to test to a lower frequency. 

no low freq. reflections from padded side walls or nodes from what ever
shaped room.

if it's not, venting sounds like a good idea

 

[zebra50]

I think I heard this referred to as 'the redneck's anechoic chamber'! Not sure outside is going to work with our climate unless I pick my mic testing days very carefully.

Yorkshire is wet and windy, which the ribbon & condenser mics may not appreciate. Might get a few weeks in summer when it's dry enough but not too dusty. At least there is no one to upset but the cows.

 

[Marik]

Hey Stewart,

What software you gonna use? Does it have room/speaker correction? What is the reference speaker?
To build the "real" full size anechoic is not a trivial task and you will need to spend quite a bit for wedging/damping, to fight reflections, so most of all, this is a matter of budget and time. Don't forget, that wide band speaker will be also of significant cost.
On the other hand, you can make a (cheap) small sized box, say down to 500Hz, and measure the rest outside. This will greatly reduce the cost, and the speaker itself will be much simpler task.
Something like this:

http://www.parts-express.com/pe/showdetl.cfm?Partnumber=264-848

used with elctronic FR correction (like for example DEQ2496) will be quite sufficient.

On the other hand, you can invest into something like TEF25, which will do it for you in an ordinary room over MOST of the bandwidth, with most of the medium quality speakers. But again, still for the very low freq. precise measurements you will need to go outside, anyway. IMO, some $2.5K-$3K for the TEF will be much cheaper solution (and in the end, might give more precise measurements). It has some real cool features, including polar plots, phase characteristics, etc. etc.

Best, M

 

[zebra50]

Hi Marik,

Thanks for your input. Really this is something I expect to build and rebuild over several months or even years as I get to understand the parameters better. Sometimes you have to build something once before you can design it! I can spend a bit of money on it if needed, slowly over a few months. Recycling some of the building materials can help keep the cost down and I do have some friends who can help me with the necessary building work for the cost of beer and sandwiches.

At the moment I'm using Fuzzmeasure for my amplifier frequency responses, and it does have microphone calibration and the ability to save and recall multiple calibration profiles, which is nice. However my evaluation copy has just expired so I need to make a decision on that front pretty soon.

I was expecting to do polar plots by hand, in the old fashioned way.

And I haven't decided on a speaker yet so thanks for that suggestion.

Cheers!

 

[Marik]

At the moment I'm using Fuzzmeasure for my amplifier frequency responses, and it does have microphone calibration and the ability to save and recall multiple calibration profiles, which is nice. However my evaluation copy has just expired so I need to make a decision on that front pretty soon.

Ah, you are on Mac  8)

Here are a few ideas for you:

 

[abbey road d enfer]

On the other hand, you can invest into something like TEF25, which will do it for you in an ordinary room over MOST of the bandwidth, with most of the medium quality speakers. But again, still for the very low freq. precise measurements you will need to go outside, anyway. IMO, some $2.5K-$3K for the TEF will be much cheaper solution (and in the end, might give more precise measurements). It has some real cool features, including polar plots, phase characteristics, etc. etc.

Best, M

Are you a TEF25 user? I am, since TEF10, then 12, 20 and now 25.
It has the same limitations regarding low frequencies than any other measurement system (MLS, Room Tools,...). The advantages are those inherent to any time-windowed system, the reflections can be eliminated as long as their are outside the duration of the lowest period. It means that reflections can be eliminated as long as their are not too close to the direct arrival. In a typical room where you can put the DUT at a 2 meter distance from any wall, the first reflection will be about 10msec late. Windowing for 10 msec guarantees accuracy only down to 200Hz. And all the TEF's I have used suffer from a repeatability problem. For some unknown reason, the calibration settings seem to disappear after some time! For this reason, we run a calibration cycle before each set of measurements.
However, I must say that, although we have actively looked for an alternative solution, we haven't found one yet.

 

[zayance]

Wow impressive pictures.

 

[thermionic]

A bit OT, but do any of you guys know about microphone tunnels? Apparently there's only one left in the world, @ Neumann, Berlin. EMI Hayes had one, but it got broken up a few years back. One esteemed engineer tried to save it, but it would've costed thousands to move and occupy his entire garden (which probably wouldn't have been very wife friendly). I heard about this at an AES lecture years ago, so my memory's not more specific.

Justin

 

[bockaudio]

Schoeps has a tunnel as well. More complicated to run than it appears on the surface.

 

[abbey road d enfer]

Please pardon my almost-infinite ignorance, but what is a microphone tunnel?
Is it the orifice where the microphone larvae are expulsed from the womb of the queen microphone? :

 

[zebra50]

Is it the orifice where the microphone larvae are expulsed from the womb of the queen microphone? :

No, that's here

But joking aside, thanks to everyone for the input and inspiration photos - why would you want to put a tank in there?

 

[zayance]

I believe the less noise makes it the perfect wheapon to sneack around and blast you  .

 

[Rossi]

Please pardon my almost-infinite ignorance, but what is a microphone tunnel?

I'd like to know, too.
I visited Neumann Berlin about a year ago, and I did not see a tunnel. Their measuring chamber looked pretty much standard.

 

[zayance]

Thanks for claryfing this Rossi, so it's possible, i can understand that real situtation reflects more, seem to be more logical for me, good to know i'll have to try that for myself in some time....

PS: By the way nice article, very instructive and complete, i like your style, all articles of products on magazines are not that described...