Back to the NIKA2 main page

Second intervention on dichroic and Ar2 shims to try compensating 2mm-1mm focus difference witnessed since the March 2026 intervention

Work Plan

Extracted from July 15 regular NIKA2 meetings with few more details:

• The plan for the intervention is the following: day 1 we open the cryostat and dismount the dichroic ; night 1 we let the dichroic relax ; day 2 we put back the dichroic membrane in its mount, measure its possible residual distortion, and mount it back in the cryostat with shims for Ar2 that depend on the measurement done, then close the cryostat and start vacuum tests ; day 3 continue pumping and start cooling down.
• From the 2 experiences of cool down after measurement on dichroic (1st NIKA2+ cool down in March 2025, and cool down after shims intervention in March 2026) we experienced that the distortion stayed almost unchanged at 1st cool down after opening. In fact in March 2026 it move from 2.0 mm measured just before closing the cryostat, to 1.5 mm according to the Ar3-Ar2=0.85 mm focus difference at beginning of run 80. Based on this empirical knowledge, depending on the residual distortion after having put the membrane back in its mount there's 2 possibilities:

  • 1) If the distortion is < 0.5 mm we can put back the dichroic in the same orientation as before and insert the shims that compensate exactly for the measured distortion. The chance that the situation after cool down is worse than currently, is negligible.

  • 2) If the distortion is > 0.5 mm, since we can't reduce more the distance between last lens and Ar2, the idea would be to put back the dichroic flipped compared to the present orientation so that the focus shift is not toward the lens but further away from the original position of Ar2, and we insert the shims that compensate for this new focus shift. The chance that the situation after cool down is worse than the current situation is higher than the previous case, because we are not totally sure that the dichroic will stay distorted in the direction that is opposed to the direction it distorted until now. The chances that it reverses its direction of distortion during cool down seem pretty small though. And if the distortion reduces a bit as we expect, we don't know by how much it will exactly.

• On day 2 of the intervention, before mounting back the dichroic in the cryostat and inserting shims, we can organize a meeting, and describe the situation to take a common decision.

Preliminary work before September 14th, 2026, and staff for the intervention

• NIKA2 warmed-up, and at ambient pressure
• Shims at the telescope.
• Tool box for opening NIKA2 cryostat.
• Torque wrench at the telescope.

Intervention team: Martino Calvo, Benedetta Kalemi, Samuel Leclercq, David John

All the team at the telescope on Monday 14th, to prepare everything for the intervention on Tuesday morning.

September 15th, Tuesday

Opening NIKA2:

Opening of the cryostat quite efficient, done in a bit more than 2 hours.
We measure then the distortion of the dichroic. Side note: on the last picture of the group below one can see very well the inner structures of the dichroic in the reflection of the ruler lightened in green.

==> The dichroic is distorted by 2.5 mm, a bit more than the distortion deduced from the Ar3-Ar2 = 1.2 mm focus difference obtained in the latest focus test done before warming up.

We dismount completely the dichroic and let the free membrane relax for 7 hours.

After the 7 hours, the membrane has taken the chips distortion it always had naturally when free of any constraint. Putting it back in the rings and screwing them with a 0.2 N*m torque (intermediate between the 2 previous values), we measure again the residual distortions.

==> The membrane didn't relax to a flatter surface! The distortion stays practically unchanged.
==> No need to wait longer, it's clear it won't evolve much more. So we decide to flip it and introduce shims that would compensate for the expected distortion at cold temperature.
Base on previous experience where the membrane flattened by a bit more than 0.5 mm, Samuel do Zemax simulation supposing a 1.5 downward distortion at cold temperature.
==> Zemax simulation says we have to introduce 17 mm shims thickness on Array 2 mount to compensate for the 1.5 mm deep dichroic distortion. So we mount back the flipped dichroic, measure it again as a double check, and put the corresponding group of 3 shims (1 + 0.5 + 0.2 mm thicknesses) on array 2 mount.

The measurements with the dichroic back in place and distortion downward (vs upward previously) gives and amplitude of ~2.0 mm. But this doesn't change the decision made about the shims.
End of day 1.

September 16th, Wednesday

We started measuring again the dichroic distortion before closing. We distortion is still 2.0 mm deep, nothing moved during the night. So we start the process of closing the cryostat.

After 4 hours of work, the cryostat is closed and back in place. Note: the 11th picture above shows the fixation device of the lower part of the "tirant" (tie rod / strut) on the 50K part of the nose (M7-M8 assembly); to see the upper part on the 300K cryostat body, look at the 3rd picture of the cryostat opening of the intervention made in march.
We start pumping just before lunch. Unfortunately the leak detector that Dave had send for revision many weeks ago, and that should have been returned to IRAM a few weeks before the intervention, has still not been returned by the contractor.

After lunch the pumping rate is visibly slower than after the previous interventions. It might be a leak, or degassing from the additional super insulation we put during the closing process, or it might be that the auxiliary pumping unit we put to help the cryostat pumping system is less efficient than the leak detector we normally use? Too bad that we don't have the leak detector. It should arrive tomorrow morning. In the meantime we tried reviving a very old spare leak detector, but this was unsuccessful; this unit is broken. Nothing much we can do. We just let the pumping system running.

End of day 2.

September 17th, Thursday

Given the pressure decrease rate overnight, and after having mounted a bigger turbo pump, without much success, it's clear we have a leak. At 9:30 the repaired ASM leak detector arrived at the telescope (thanks to the IRAM Granada staff who picked it up at the depot where the manufacturer had sent it, and drove it up at the telescope as soon as possible). We crane it up in the receiver cabin, mount it on NIKA2 cryostat, and as soon as the vacuum is low enough, we start searching for the leak. We start with the worst place the leak could be: the interface between the 300K body and the base plate to the cryostat... Alas, the leak is there! We try to tighten the screws and add clamps in the hope of applying enough pressure to the two parts to seal the leak. But it doesn't work. The time to do all that it's about 11:45 when we finally decide to open the cryostat again. In principle, Benedetta, Martino and Samuel, should leave in 30 minutes to be at the Malaga airport in time for their trip back home. Having done the check-in online, they decide to postpone their departure by 1 hour to help Dave and Marcos opening the cryostat at least up to the "tirants" and let them dismount these delicate parts with some guidance that would help when closing again the cryostat after having hopefully found and fix the leak. 13:30, 1st "tirant" dismounted, Benedetta, Martino and Samuel leave the telescope hoping that there will be no traffic jam on the road to Malaga airport, and confident that Dave and Marcos, with possibly the help of Victor, will manage to fix the leak and close again the cryostat.

Dave and Marcos found the leak. It was caused by a small piece of torn super-insulation that was crossing the O-ring on the cryostat base plate. It must have moved there when we closed the 300K body, in the last millisecond before it touched the base plate, pushed by the displacement of air (the location was clear and being watched when moving the 300K body!). They clear all the joints again, and carefully close again the cryostat.

Pumping to try again to put the cryostat under vacuum at the end of the day.

September 18th, Friday

Leak check successful. The rate of pressure drop in the cryostat corresponds to that previously observed in the absence of a leak.
Vacuum reached and NIKA2 is cooling over the weekend, and beginning of the following week.

END of the 2nd intervention mission.