Inside IBM's vomit-inducing, noise-free future chip lab There’s a bunker inside IBM Research's labs in Zurich which will make you want to throw up - at least according to its architect. It’s a white room - literally. The walls are covered in white foam-like, patterned tiles. But it’s a room that has been specially engineered to block out any interference from any source - noise, …
That thing looks seriously overbuilt. Do they have a locker and freezers with food and water for more than a year? Does the facility have its own generator? Guns and ammo? Magazines with articles about going off-grid?
As for the nausea inducing silence, the Ontario Science Center used to have an eerie soundproof passageway and I can vouch for the fact that real silence is compellingly strange. It would not surprise me if what he says about this is correct. It would not surprise me, though, if 'alien anal probe' found its way into his explanation.
Being Swiss, then, there is a very real possibility that there are guns and ammo.
If it's not the case now, then the change only came recently, that all citizens in Switzerland must have access to a bunker capable of surviving nuclear holocaust, and all men must serve in the militia and have access to an assault rifle during that time.
I met a Swiss guy a few years ago who was incredibly patriotic.
That's what I thought - you couldn't get a lockup/garage around here for that price.
My mind did, however, add "million" to the back end of the figure automatically. I tend to do that when big corporate science projects are involved. Which doesn't make sense, because we are then in the Merkin Billion Zone and no doubt that's how it would have been entered in the article.
A conundrum.
This post has been deleted by its author
... stroking his persian cat ...
No, what worried me most about his probable evil plotting is that it was called a "nantechnology center" in the text. That made my mind boggle, and required a stiff drink to unboggle.
Which meant that I started to worry his concrete air spring would spring a leak, so to say; just outside of the warranty period, leaving them sitting on solid bedrock. Not sure how they'd send it in during warranty cover either.
All in all, more questions than answers.
"clean room produces less than half the noise of any rival facility – 30dB versus 65dB.
Uh? 30dB ≠ ≈ half 65dB. And I'll ignore the confusingly placed dash.
Anyway, '65dB' relative to what. And does it refer to sound or EMR/RF? They are different you know.
Presumably, '65dB' refers to nominal/ambient sound level, if so then 30 dB down on that isn't much if the anechoic chamber is state of the art. And what frequency/range of frequencies does this apply to?
Perhaps 30dB versus 65dB actually means 'anechoic chambers typically attenuate 65dB and this one is better by another 30dB'. If so, then it would be truly a state-of-the-art facility.
If it's EMR/RF then again the figures make no sense. The EMR/RF attenuation of Faraday cages is measured in dB relative to background radiation and specified across bands of frequencies. So what are those frequencies?
If the Faraday cage attenuated some 65 dB down on background then this is good figure but it's hardly state of the art (unless that limit is determined by radiation from plant and equipment being used within the cage).
Incidentally, dB is a logarithmic ratio. Here's a very shortened precis from the Wiki defn.
The number of decibels is ten times the logarithm to base 10 of the ratio of the two power quantities.
This article left me not much the wiser.
Please clarify.
And I meant to add:
We have to ask: Wasn't that a bit risky putting the future of IBM's nano-technology research in the hands of an building-design novice?
Building designers are extremely unlikely to know anything about anechoic chambers, very low frequency vibration, EMR shielding (Faraday cages) and Helmholtz coils.
Specialist laboratories have always been designed and or specified by the researchers. Loertscher and other researchers OBVIOUSLY are the only ones qualified to attempt this work!
Ahhhhhh!!
Higher frequency screened rooms / chambers (good ones) can achieve very high attenuation. Low frequency screened rooms are a different issue - at 1s to 1000s of Hz changing magnetic fields need a lot of metal to reduce them effectively. The attenuation achieved by metal screening is very thickness-, and material- and frequency- and geometry-dependent below 10 kHz. I've just spent a fair bit of the last year and a half working on a similar screened room for an FEI Titan electron microscope and I really really don't find this article very clear or informative !
At very low frequency / near DC / DC a (e.g. Helmholtz or Maxwell coil) field bucking system is about the only option to get big attenuation unless you want to pay for a heck of a lot of mumetal - common steels at the odd few grand a tonne only give the odd single digits of dBs at reasonable thickness at near DC. However, a static unchanging DC field is an offset generator, not a blur generator. The earth's magnetic field contains components from DC up to quite elevated frequencies. Its changing field that is the biggest (electromagnetic) killer for e-m work IMHO.
In any of these rooms, the attenuation at mains frequency upwards is firstly limited by how much steel of what type is put into the walls, floor and ceiling. If you actively buck the AC fields, you have to decide where do you stop - mains with harmonics on it goes up and up in frequency.... At some point the control of a field buck system becomes too difficult. Cables and electric equipment systems in the facility generate a fair amount of field depending on the details of what is installed, how. Good detailing of the equipment and installation cuts the source field a lot.
There's an awful lot of numbers being bandied around in this article, but it doesn't leave things that clear at the end. Could we have some more science, some more engineering and a bit less gee-whizz., please. A proper detailed technical presentation would be great - if IBM or Dr Loertscher have published such a thing online can you link to it please ? Any open visit invites going ?
"IBM research staff member Emanuel Loertscher – the room’s architect – claims his clean room produces less than half the noise of any rival facility – 30dB versus 65dB."
That is NOT a halving of the noise levels. Decibels are logarithmic, whereby a 10 decibel difference represents a 10x difference in noise levels. A 35 decibel reduction in noise means that it is about 3000 times quieter.
One of the key items was maintaining a uniform temperature.
In this context that means uniform ant quiet, which is even trickier. The aircon is pretty special.
But note that surprising levels of vibration abatement can be achieved quite easily..
People were astonished when a few years after IBM developed the AFM someone built one in their garage.
They built it on a table loaded with sand sitting on 3 piles of old car tires laid out in a triangle (helps with the kinematics).
However running observations taking days calls for something much more serious, hence this puppy.
Thumbs up for an interesting article, but could have been a bit more informative.
What's the book?
One of the key items was maintaining a uniform temperature.
On the IBM link which RustyNailed kindly provided (above), I noted that IBM specifies:
Temperature stability: 0.1°C/1 h and 0.5° C/24 h
and I can attest that this is a pretty difficult task. I recall having to calibrate thermometers in a calorimeter and trying to get temperature stability to less than 0.1°C was both difficult and it took ages for the calorimeter to stabilize--and a calorimeter is not a room!
I note with interest that IBM only specify both hourly and diurnal temperature tolerances. In such an establishment I've have thought that a monthly and annual specification would also have been necessary. If the error is cumulative then longitudinal calibration errors could occur as a consequence of temperature drift.
0.5° C/24 h might be OK but not specifying monthly/yearly could mean much larger drifts over this time frame.
I wonder why longer times were not specified.
"What's the book?"
Don't remember, I was put onto it by a Usenet post but since Google's usenet search is FUBAR'd I can't find it.
IIRC it was written by a senior engineer with an instrument maker in the US. It started with how do make an acurate surface reference to begin with (you need 3 surface plates, some fine powder, a scraper and lots of patience) and went on from there. I think part of it was a case study about designing a Coordinate Measuring Machine, but I may be conflating 2 different books.
It dates from the late 80s or 90s as I referred someone else to it about 10 years ago.
Sorry I can't be more helpful.
The classic method for reducing disturbances is (essentially) to put a box in a box. Maintaining a temperature of 1 Deg C over the range of say 0-50C is tough, but doable. A box inside that could then maintain a temperature of say 1/50th of 1Deg. The principle applies to any disturbance.
I found RV Jones "Instruments & Experience" a fascinating read on instrument design.
Shame. I understand the Usenet problem, had that problem myself.
The classic method for reducing disturbances is (essentially) to put a box in a box.
Agreed, and although not spec'd, I'm sure IBM is aware of the problem. The box within the box is the correct soln., it's what I employed (a calorimeter within a calorimeter). Still, I found to achieve consistent drift below 0.1C required considerable attention to detail.
Isn't having a TEM in there instead of some "scanning probe" variant somewhat ironic given the facility is named after Binnig and Rohrer ?
I'm sure if my lab had been as well equipped as this, I would have got some better results. Would have kept wots-her-name from bursting in and turning the lights on/off whilst I was in the middle of an experiment.
In our building, with large open plan seating for staff, the FM guys put up some 'sound deadening' material on one of the long walls. When you walk past this wall, the lack of reflected sound does give a distinct uneasy feeling. Not quiet as bad as the effect quoted in the article but decidedly unnerving.
Back in the early 1990's, IBM built a "quiet room" at their facility in Boca Raton, FL.
It held the worlds record for the quietest spot on earth for a long time, until IBM moved from the site.
There was an Interstate highway, and railway and an airport, all within a mile of facility.
The fun thing to do was take a chair with arms and sit for awhile.
The arms kept you from falling out of the chair due to balance issues that came up.
After 10's of minutes, your hearing would adjust to the new quiet and soon enough you would be hearing your heart beat and the blood rushing in pulses through your head. Your breathing sounded like a some hollywood horror movie of rushing air.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
