Are the isolation tables using nitrogen or compressed air supply the best? What do you think about the active isolation platforms available from Herzan and Accurion that runs simply on electricity?
Can you explain a bit more? You mean hanging from the ceiling with four ropes !! My lab does not have that much height .. nor is it possible to do any damage to the ceiling.
Dear all, a quite cheap AND effective way is to mount the AFM instrument on a marble base the size of which just large enough so that the instrument fits on. This hard and massive base (with extremely high resonance frequency) can easily be equipped with 4 hooks (two on two opposing sides), and you can connect the AFM-base with the ceiling using 4 elastic straps with these hooks. These straps used e.g. for baggage racks or camping purposes are exremely soft, thus they have low resonance frequencies, and by combining these two resonating systems with two substantially different Eigenfrequencies you can isolate the vibrations from the lab and the AFM - so you end up with the performance of your AFM system.
By the way, in many cases I have experienced that acoustic noise is a serious problem for AFM - so some providers also offer systems that are acoustically isolated ... Best regards to all of you, Dirk
Thank you everyone for your answers and proper justification. What I can conclude is that the hanging marble from ceiling is the best solution. The AFM I plan on buying is Cervantes Fullmode AFM from Nanotec Electronica (http://www.nanotec.es/products/cervantes.php).
Acoustic noise is not an issue because I shall be the only person in the room. However, I am a bit concerned about the air conditioning which shall be at a distance of 2 metres from the AFM.
Let me add, that in case of hanging scanning head solution you may need also some supporting mechanism to avoid the swinging during the manipulation. It should allow to release the suspension system smoothly in order to avoid unwanted oscillations. On the other hand, I could test the home-made, cheap solution based on the massive base supported on tennis balls, with few concrete 20kg blocks with elastomer spacers between. Such a "sandwich" was quite effective and allowed to obtain the noise level below +/-2 angstroms while the trucks and trams are moving on the road 40meters from the lab (no better location available).
Thank you so much Prof. Sikora for this valuable information. It would be kind of you if you could explain a bit more about the tennis ball supported concrete block.
The construction is relatively simple: you start with some rubber (I'd suggest some thick, complex structure instead of flat thin layer), than you put 3-5 layers of concrete blocks stacked with few millimeter thick elastomer layers (the lateral size of the structure depends on the size of the scanning head and other equipment you want to place on the table). On the top you place few tennis balls. You need to stabilize them, otherwise whole table will be unstable. Then you place the granite (or other rigid, heavy material) table, which will be the base for your equipment. You may plan to put some commercial damping system (active, or -K one) in the future if necessary, therefore your table shouldn't be small. 50 cm x 50 cm should be enough.
Let me confirm that air conditioning can be a serious problem. It is a problem for vibrations but also for temperature stability (most air conditioning does not blow continuous flow but have ON/OFF cycles regulated by a thermostat). If your air conditioning exhaust is only 2 meter from your AFM I would consider to either redirect it or to make the exhaust diffuse. In my current lab, the exhaust is "plugged" to a long cloth tube which allow the air to go through in a very diffuse way. This is a inexpensive and very efficient solution. Have a look here for an example:
Whilst working with Heini Rohrer and Christoph Gerber, I learnt that when a hanging suspension elongates 25cm under the combined weight of AFM & support platform, the system oscillates with a main resonance frequency of aprox. 1 s-1. This can be calculated and measured.
My background lies not in your equipment but in sound and vibration. I have never worked with AFM but I have been involved with the vibroacoustic design of a Nanolab where there were some SEMs, EBLs etc and a Synchrotron machine, both at Lund University. I have worked also with lots of vibration isolation cases for offshore piping and machinery which can be quite challenging due to limited space and the fact that offshore constructions though weighing a lot can be seen as a high performance lightweight building.
I started out to write a shorter piece but got carried away as I enjoy the topic.
For what it is worth here are some ramblings of mine, they relate also to clean room environments. Some of it may be useful even if not much of it specifically relates to AFM
There are three kinds of pressure (http://qringtech.com/2010/09/15/wave-steepening-increase-peak-pressure-piping-pumps/ ), Line Pressure (LP, Ambient), Kinetic Pressure (KP, vind) and pulsation (sound).
These are different beasts who act differently.
In a clean room environment, one need to cater for all of them as one is dealing with a 'closed' volume.
LP varies whenever staff enter/exist the clean room and when vacuum units open.
A remedy for the former can be a revolving door.
A remedy for the latter is to have a stiff internal pipe in which to vent units before lids open.
Also, one may use simple vents where a pipe end slit at an angle and loaded with a membrane lifts the lid to remove much of the LP variation. This trick is commonly used in cars as it otherwise would be diffiucult to close the doors.
There is KP (vind) in a clean room as it very much ties in with its core functionality.
KP does provide a wind load.
It is a weak force but sufficient to knock out sensitive machinery when exposed to the sensor part of, e.g. a EBL machine.
The remedy is to allow wind only at the time of loading/unloading samples and to block wind once the sample has been loaded.
Pulsation (which, as opposed to sound, covers also frequencies below 20 Hz) matters due to
HVAC systems tend to have its fundamental resonance at a few Hz.
Apart from a tiring work environment, this resonance does affect vibration isolation performance when air springs are involved, see below.
Fan noise providing harmonic excitation.
Human speech has a tendency to jump from soft spoken with an A-weighted Sound Pressure Level (SPL) of about 55 dBA re 2E-5 Pa to declamatory SPL of 80 dBA re 2E-5 Pa when the background SPL is above, say ~50 dBA re 2E-5 Pa.
Declamatory speech levels are sufficiently strong to knock out sensitive machines.
It is therefore good to keep HVAC system noise low.
This directly translates to having sufficient wall-roof height.
Good HVAC outlets that quiet tend to require space as one slowly expands the airflow.
It is prudent to install so called hygienic sound absorbers.
The objection is that these add surface to clean which drives up operating cost.
The remedy is to fold them in a way that does not obstruct flow and pack the area below the clean room floor with absorbers.
An absorber is just a well designed pressure drop and, hence, can be designed from any material and work at 'any' frequency.
For the absorber to function, it needs to stand still such that pulsation flow can oscillate through its pressure drop.
A low frequency absorber can therefore be designed using microperforated metal or glass that is rigidly attached to the building structure.
Vibration isolation can be provided by active isolation, passive isolation or a combination of both.
Disclaimer. I work with passive isolation.
That said, I have been involved with projects using active systems though I have not been in charge of this part.
Active isolation
Uses filters to negate unwanted vibration, sound etc..
Filters require time for tuning and settling.
The more effective the filter, the narrower the filter bandwidth, the longer the settling and tuning time.
As rules of thumb
Continuous source(s) where a phase reference(s) can be provided allow ample time for tuning and settling, i.e. active systems work very well and usually provides about 40 dB (100x) reduction.
Continuous random source(s) and slowly varying sources without phase reference usually procide about 20 dB (10x) reduction.
Short duration sources that can be predicted and foreseen can likely be handled using active systems. I have no experienece from such situations.
Short duration sources that cannot be predicted or foreseen require wide filters that tune fast. The active system performance then depends strongly with the source characteristics but isolation tends to land in the ballpark 5-10 dB (~2-3x).
An active system needs a support. As we cannot avoid Newton's 3rd law that an acting force at an interface must have an equal and opposing force, we land in the situation that the active system may improve matters in domain A (the isolated zone) while adding vibration/sound to the adjoining zone B.
This situation can become aggravated when a series of active systems are placed next to each other as crosstalk between them may remove their usability. In such situations, one must use a multichannel active system rather than a multiple of single channel active systems.
Active systems need one channel per degree of freedom (dof) it caters for. A hard mechanical connection involves six 6 dofs, X, Y, Z, RX, RY, RZ vibration. Omitting any of these overlooks transmission paths.
Passive vibration isolation
Can be provided by suspended or hung systems.
Suspended systems involve buckling loads for the spring elements.
Suspended systems become statically unstable when one starts to explore maximum performance.
Hung systems can provide superior isolation as one makes good use of pendulum dofs, as buckling is avoided and as centre of gravity can be kept low.
As rules of thumb
Metal springs and rubber isolators work down to a vertical direction resonance of about 3 Hz.
Systems hung from precompressed metal springs can get a vertical direction resonance slightly below 1 Hz as spring length start to become long. Note that long springs have anti-resonance at fairly low frequency.
Airsprings can achieve very low resonance, in particular when using coupling volumes.
The larger the airspring, the lower is its resonance frequency.
Smaller airsprings tend to land at ~3 Hz and larger airsprings ~1 Hz.
Comment. My personal record is 0.2 Hz and yes, this system used coupling volumes and was statically unstable and its tilt had to be computer controlled.
Suspending elements that are allowed to buckle provide a negative spring stiffness.
If instability is handled, such systems may provide high performance and vertical direction resonance down to 0.5 Hz.
Problems with passive systems
Any isolator will have anti-resonance at which frequency it becomes very stiff and transmits.
A remedy to anti-resonance is to use multistage isolation systems.
However, multi-stage systems only work well for stationary and slowly varying sources as they involve a mechanical filter that needs time to tune in and settle.
Airsprings are sensitive to pressure variation as it provides a lift that is relative to the surrounding pressure. Main influences come from LP, pulsation and leakage.
Leakage in the airspring system must therefore be minimal.
HVAC acoustic resonance
Strong infrasound sources (harbours, large open pools, large diesel machinery, steam outlets, etc)
Combined passive/active systems
I am just guessing here as I have yet to design such a system.
My guess is that this is where the maximum performance would be found.
The passive system would do the rough work and clean up the situation such that a simpler active system may provide high performance and avoid cross talk or overlook transfer paths.
The active system may then handle also internal sources.
Rigid body modes
Provides acoustic short circuiting of high performance vibration isolation systems.
Are the strongest sound radiators and hence, the strongest to sense sound.
There is strong acoustic coupling when there is half an acoustic wavelength between front/back of a rigid object as it then forms an acoustic dipole.
Similarly, half an acoustic wavelength across the face forms an acoustic quadrupole.
Vibration requirements
The best is to know the workings of your important machines inside out.
However, this is not always possible.
Machine requirements may be provided by machine suppliers.
These sometimes provide the lower envelope of all known installations of this machine and, hence, may form a rather conservative design criterion.
There exists generic criteria, the so called Vibration Criteria (VC-A to VC-G).
VC values rely on vibration velocity while some machines are displacement sensitive and other respond to acceleration.
A Catch 22 situation
Once you isolate enough, you become sensitive to internal sources and to sources involving transport power/of media to/from the machine.
Media can be handled with pulsation bottles.
Power cabling is handled using blocking masses.
Internal sources are hard to do much about. One my add losses to the isolators at the cost of isolation. It might work to use active systems to handle internal sources.
Soil properties may matter greatly
Rayleigh waves act along the surface and are easy to break with a ditch.
Shear waves may exist and travel long distances
Love waves are tricky as these need a stiff soil boundary, a free upper boundary and a slab to exist.
Once Love waves exist, they are easy to excite in the vertical direction and they respond in the lateral direction.
The deeper the soil, the lower the resonance frequencies involved, see below.
Building
Usually, the higher up, the worse, as building motion is imposed by external loads.
Walking is made with a repetiion rate of 2-4 Hz and with harmonics thereof.
Most suspended vibration isolation setups tend to have its rigid body resonances in the ballpark 3 Hz to 8 Hz, which is unfortunate.
Floors should therefore have its first resonance well above 12 Hz.
When possible,
Separate lab from walking loads by transferring these down to ground and barring this, into the inplane direction of major structural elements of the house.
This advice in particular applies to stairs where walking down applies far greater loads than walking up.
The plates you walk on in a clean room can provide string impulsive loads is they are unevenly supported.
Should you isolate the slab/foundation for your most precious machine form the building or integrate it with the building?
This is a pest or cholera decision.
The former maximizes the slab's ability to pickup vibration from the soil while the latter enables trasnfer of high frequency structureborne vibration.
The better way if soil vibration is an issue, is to have as large and stiff a house body as possible and to isolate the machine from the floor, i.e. have a long solid house body and an isolated slab floating on top of this floor. .
Sources
Trains and traffic are obvious sources.
Speed bumps and driving on uneven, e.g. icy, roads worsen the situation.
Road trains are worse than trucks separated by distance.
That said, the strongest vibration and pulsation source usually is the people using the lab, so called cultural noise.
A short story - I once measured vibration at an AFM.
It happened to be placed in the basement and such that one side was sitting on the lab side with the other side sitting on the building side.
The lab was isolated from the rest of the building which was an ordinary office building surrounded by major streets with mixed traffic, albeit, placed on Swedish bedrock.
Taking measurements on both sides, it was, much to the amusement of the lab manager, quite clear that the better position was on the ordinary building side.
Why? - with X dedicated researchers stomping around in the lab it was way noisier on this side. That said, late at night, i.e. when everyone had gone home, the best performance would likely have been found on the lab side.
Practical ways
Developing a good lab culture can go a long way, at least if it is a small lab with limited user circulation.
Usage strategy may help as well.
If you have lots of users, section them into A, B and C category.
Educate those in need of performance.
Bunch long time users with high performance demands into simultaneous lab use - a bit like having ladies only at the swim club.
The lab does not need and rarely is able to provide top performance 24/7.
Usually, the best performance is found at nighttime with good weather conditions.
Measure relevant parameters 24/7 to know your facilities and tell the people in need of maximum performance when they can get it.
Here are some general ramblings of mine on the topic of vibration isolation
Here is some work that was made on the Nanolab. I did not know it at the time, but the problems I picked up using FE simulation and later verified using measurements are caused by Love waves.