Shielding the Weiner PSU

7 May 2013

Kiril Marinov

ASTeC, MaRS, DL

1

Concept: a box with a pair of “chimneys”

Bin

max<<<B0 540X620X280 AISI 1010, 5 mm thick steel

Air outtake, inner diameter 60 mm, outer 80 mm, 60 mm high

Air intake, inner diameter 60 mm, outer 80 mm, 60 mm high

Top

Design considerationsThe air in/outtakes serve several purposes:

• additional shielding of the “hole” areas, as will be shown• access for cabling (4+1 sets of cables needed)• Bottom-to-top air convection.• Possibility to install additional fans in a well-shielded area (1-

3mT).• Supplying air from a compressor should also be an option.

Heat losses need to be quantified as they affect the design of the shield. Different people I have spoken to have quoted different values (from a 100W to 400W-1200W).

Access to the PSU is needed for servicing. As will be shown the box is capable of shielding the PSU, but additional mechanical (opening and closing) and electrical engineering (cooling, cabling) effort is needed. Mechanically the box needs to be designed in such a way that opening and closing it does not affect the shielding performance.

The position of the air intake/outtake pair does not need to be in the centre (this is chosen for symmetry only). A second pair could be a added, if needed. There is a 50 mm distance between the PSU walls and the shield in each direction. This can be increased, but within reason. Larger shielded volumes require thicker walls...

May not fit in a standard rack.

Method

The shield is placed between the poles of a window-frame magnet generating uniform field pointing along Z of 30 mT strength. The shield has been rotated around its three main axes and in each configuration the field in the shielded region has been assessed. The PSU is believed to be able to tolerate 25 mT.

In all the subsequent plots the external field is always along the Z axis.

External field parallel to the air in/outtake axis

B(X=0, Y,Z)

External field parallel to the air in/outtake axis II

B(X, Y=0,Z)

Flux density is highest on the sides that are parallel to the external field, as should be expected. In the areas outside the shield, adjacent to those sides the field is lowered as well.

External field perpendicular to the air in/outtake axis

8.8 cm steel

B(X=0, Y,Z)

External field perpendicular to the air in/outtake axis II

B(X, Y,Z=0)

External field perpendicular to the air in/outtake axis III

B(X, Y,Z=0)

This should be the same distribution as the one shown on the previous slide, only rotated 90 degrees along Z.

External field perpendicular to the air in/outtake axis IV

B(X, Y,Z=0)

The same as on the previous slide but the two “chimneys” have been removed. The effect is easy to see.

External field perpendicular to the air in/outtake axis V

B(X, Y=0,Z)

External field perpendicular to the air in/outtake axis VI

B(X, Y=0,Z)

The same as on the previous slide but the two “chimneys” have been removed

SummaryThree orientations of the proposed structure with respect to the external field have been considered.

With an external field of >30 mT the field inside the shielded area is well below the target value of 25 mT. The shield seems to work. Better shielding can be achieved with thicker walls and longer chimneys.

When all the other issues - cooling, if it turns out that the amount of heat generated is high, cabling, easy access without compromising the shield - are taken on board this may no longer be the case. Replacing the units with new ones (MARATHON range) should be considered as well.

If a decision is made to keep the present units and go ahead with building the PSU shields then a prototype should be built and tested in R9 – the sooner, the better. This will help with eliminating a number of uncertainties, e.g. model validation, effect of welding on the magnetic properties of the steel, etc.

I would like to acknowledge useful discussions with Ian, John, Craig, Melissa and Paul.