Rasnik for ATLAS

Rasnik is a wide range, high precision alignment monitoring system, developed at Nikhef.

Sensiflex has brought the system to the commercial market. Nikhef participates in and collaborates with Sensiflex. While the Rasnik system for Atlas uses a dedicated read-out system, the Sensiflex systems use commercial available sensors, read out via USB or Ethernet.
Rasgap is equivalent to the first USB version of a Sensiflex system.

Sensiflex offers some standard products, primairy intended to monitor tunnel segments and construction sites. Systems for bridges are under investigation. But also smaller scale high accuracy systems are possible with this principle.

go to the Sensiflex homepage

RasGap monitors the gap (expantion joint) between the H-building and the newer N-building of Nikhef. We noticed that the gap increased with time the past ten years or so. Therefor it is an interesting testing location. The system is also used to compare various software versions in terms of speed, noise etc..

Rasgap is mounted near room H147a (second floor) in the Nikhef building.
On the left one sees the illuminated coded mask. The lens sits in the middle and the sensor is on the right. The white plate below covers the gap. The black box on top with the green Leds is the processor. It performs the reconstruction(s) and buffers the results. Every hour this data is copied to a central database. This data (X, Y and Z) can be plotted to analyse the results. Later on this year it can be accessed using a browser. In this case the X is the vertical movement and Y the horizontal movement between the buildings. Z is the distance between the buildings or size of the dilatation joint. This caused the crack in the floor carpetting.

The X direction (vertical) is relatively stable. The Y direction moves a lot more, mainly due to temperature changes, where the buildings react a bit different. The same goes for Z. One can nicely distinguish working days and weekends, especially in Z.

When you take a running average of the data, you can also plot a running rms. Using the right parameters one can reduce the daily pattern and thus look at small effects in more detail. Doing so, you see three peeks during the day: when people arrive, when they go for lunch and when they go home. First we thought this was due to turbulence and we closed the ends of the cover. Turbulence is the biggest enemy of this system. This had no effect. Then we turned on the monitor screen continiously instead of working hours only. That didn't help either. Then the system was completely covered to block ambient light. Still the same effect was seen.

One can see the results of three types of reconstruction sofware running on the same picture. They rather disagree in Z. The Foam version is very noisy. The Rasnik version is the first reconstruction software, written by Frank Linde some 15 years agoo. The Sefo version uses the same principle as Foam, based on Fourier transformations, but was completely rewritten. One can also see some activity on saturday, the right most peak.

We concluded it must be mechanical noise introduced by people passing by. This seemed unlikely, since we only take a measurement every 30 seconds. So we did a controlled test with people standing at different points in the corridor.

At every point, one to seven, nine people stood still for two minutes. The measurement rate was increased to one every ten seconds. After that people walked around randomly in the same area for two minutes.

Watch out for the different scales. The large difference in Z still has to be solved. But when the system is calibrated it doesn't matter much. So roughly we can say that the vertical movement is 0.5 micron per 100 kg. The horizontal movement is "as usual" and not correlated to the vertical. The random walking did not show a noticeable effect. So, the rms increase shows groups of people passing by.

you can go to :

the NIKHEF Rasnik overview page.
the NIKHEF home page.

You can mail

suggestions for this page to: Henk Groenstege.

updated: april 2012