1. Detector requirement:

Spatial Resolution:

• What is the required resolution in the spectrometer?

I assume the question is about position resolution in the drift direction. There is no explicit requirement on this. The requirement is on single event sensitivity (SES), which is a combination of momentum resolution and pattern recognition mistakes. For hit position resolution below 250μm, momentum resolution is dominated by material. This can be interpreted as <250μm position resolution, but the more stringent requirement comes from pattern recognition ... that is, to minimize mistakes.

With the issue being mistakes, no standard measure of resolution is a good figure of merit. We emphasize getting the outliers by the following process.

o Let GEANT generate clusters

o Simple drift model for these clusters, cross checked with GARFIELD

o Simple electronics model, cross checked with SPICE

o Apply threshold

We of course don't accept the results blindly. Using this procedure, we generate resolution versus drift distance plots and compare against published ATLAS results. We also compare simulated to real Fe⁵⁵ for pulse shape parameters, particularly rise time.

There is also a requirement specific to the electronics: <0.5nsec for drift time (time relative to the accelerator clock); and <0.1nsec for difference in time for the same hit measured at the two ends of the straw. These requirements are driven by the desire to make the digitizer's contribution to the resolution negligible. See document 3879.

• What is the simulated and measured resolution (with prototype)?

See above regarding resolution in the drift direction.

For "time division" -- measuring positon along the straw based on timing -- noise is the dominant issue. We still use GEANT, GARFIELD, and SPICE (as above) to reproduce variations due on cluster statistics, but we also rely on measured resolution to set effective noise. See document 2310.

Occupancy:

• What is the maximum total and local (Hz/cm²) hit rate in the straw during data taking and beam splash?

Beam flash (electrons) gives a ~150nsec burst at 52MHz at the start of each microbunch. Muon related hits are ~0.2MHz over ~1200nsec every microbunch.

• What is the expected total charge/cm?

Simulation as of mid‑2013, with the standard proton for that time, and for gas gain of 50K, showed the highest dose as 1.5C/cm… see document 2535.

• Is aging an issue?

We have tested to an accumulated charge of 1C/cm with no observable aging. Depending on simulation model updates and proton absorber redesigns, this test may need to be extended.

Cross talk and noise behavior

• Has the cross-talk been measured? If yes what is the level cross-talk?

Cross-talk is primarily an issue for large proton pulses causing hits in neighboring straws. This was measured using a proton beam at LBL. At nominal gain and threshold the rate (fraction of time neighbor sees a hit) is of order 5%; the actual value varies with straw, a phenomenon which needs further investigation. See document 5777.

• Has the final front-end boards been used?

No. The basic design, including connectors, is what we expect to be final. However, due to the geometry of the prototype vs an actual panel, the boards are not final.

2. Operation conditions

Counting gas:

• What are the main arguments behind the choice of Ar/CO2 (80/20)?

I assume the argon part requires no explanation. CO₂ was selected for several reasons.

CO₂ is nonflammable, simplifying storage and distribution. In addition, in the event a straw ruptures, the vacuum pumps would immediately be exposed to the drift gas. A non-flammable gas eliminates many safety issues associated with pumping and venting the drift gas.

CO₂ abundance and unique phase diagram makes it easy to separate and purify. Thus high purity CO₂ is cheap compared with comparable purity hydrocarbons. Further, the absence of any H atoms in the drift gas inhibits polymer growth. All these factors add up to low aging with CO₂.

What is the nominal high-voltage and what is the gas gain?

Between 1350V and 1400V, and gas gain ~50K

3. Material and coating of the cathode foil

Cathode material: As cathode material a 36μ thick PET (2x15+6μ glue) has been chosen

Not quite. There are two layers of 6.25μ PET plus adhesive. Based on measured weight per unit length, the effective thickness including adhesive is 15μm

• How and where is the base material produced (coating)?

Vapor deposition by Sheldahl, aka MuTek (Minnesota).

• What are the QC steps on the base material?

Sheldahl performs adhesion and resistivity tests and provides us copies of the results.

• Are the two sides electrically connected and if yes how?

In the base material, the two metallization layers are not connected. We can chose to connect them at the straw ends with silver epoxy, but the data to date indicate this is unnecessary. See document 5331.

• Has any studies concerning effects on discharges on the straws been carried out and is there any risk of increased leak/permeation?

See document 1808. Those studies used a 2nF capacitor charged to 2kV. A single spark blows away the metallization in a sub-millimeter region. The Mylar remained intact. This will have some, but not much, effect on diffusion. However, ~10 sparks in the same spot would blow a hole through the Mylar, a far more serious problem.

Based in part on these tests, we have reduced the HV blocking capacitors to 200pF per side but have not yet repeated the discharge tests.

• Has any large surface irradiation tests been carried out on the straws in order to look for Malter effects?

With both inside and outside surfaces metalized, there is no risk of Malter effect.

Straw gluing:

• Given the axial tension applied during the detector assembly, has (glue) adhesion to the straw wall been made?

Yes. The straws ends are epoxied into a gas manifold.

• Which glue is proposed?

Epon 815 is our selected resin. Curing agent is under consideration.

Permeation: The coating of the straws

• Are there any studies made on the base material concerning permeation? How big is the contribution from the glue joint between the films?

Such a study was attempted, but did not yield reliable results. Given lack of resources, we focused on measuring the time dependent leak rate of completed straws. Since this was found to be well within our requirements, we did not return to studying the various contributions.

4. Straw production and QC

Straw winding:

• Is the winding process defined?

Yes, although details are proprietary and have not been revealed to us.

• How is the cleanliness guaranteed and is the production lubricant free? (In case aging is issue)?

Both vendors interested in the job made it clear that production will not be done in a cleanroom, and they need some form of lubricant to get the straw off the mandrel. However, they agreed to use the same lubricant for all runs and our aging tests were done with such straws. No statistically significant aging was observed.

• Where are the straws produced? How is the production quality ensured?

PPG, in New Jersey. We explicitly test all straws for leaks as well as end-to-end resistance. We implicitly test size since the parts would not fit if the straw was the wrong size.

Quality insurance:

• How is the leak tightness measured and what is the specification?

We will test each straw for leaks using a CO₂ sensor; see document 5256. This method have been crosschecked against the more fundamental leak rate into vacuum; see document 5246.

The specification, agreed to with the Muon Beamline group, is 7ccm on the tracker overall. Based on ~100 straws, we estimate the straw leak contribution will be <10% of this limit. Other sources of leaks are yet to be measured, but the straws themselves are not a problem.

• What is the expected failure rate (major leak or straw failure)?

Since we are able to disconnect straws individually from high voltage, no form of straw failure is a big issue except for developing a major leak. We have pressurized straws to 60 psi for a few minutes, as well as kept some pressurized to 15psi for 120 days with no failure. Straw creep tests have been ongoing for ~3 continuous years. See document 2633.

These are all low statistics tests, however. To allow major leaks without having to stop data taking, we will send gas to each station separately. This allows us to turn off a station with a leak while continuing to take data with the rest.

• What is the experience from prototypes in terms of failure/major leak?

Kinking a straw does not destroy it, nor leave any noticeable crease when the straw is pressurized. However, during high pressure tests (see above), one straw popped at a kink location at 70psi. We have also found less catastrophic leak problems at 15psig after a straw is kinked.

Based on this experience, we keep straws in a protective tube from the time they are leak tested to the time they are put into the panel, except for a short time when they are cut to length

5. Detector construction and QC

• How is the pretension applied and the straightness guaranteed?

Our previous prototypes used dead weights to apply 700g tension. We expect to use load cells and motors on the next assembly fixture.

The 700g tension was selected to overcome end affects so the straw is straight. See document 1343.

• What is the sag for the horizontal 1200 mm long straws?

At the time of installation at 700g, ~100μm. Tension drops over time; our current best fit model extrapolates to 280g at infinite time, giving a sag of 250μm.

• How is the straightness of the straws checked during assembly and what is the specification?

There is an implicit requirement and test: the straws must not touch each other. Beyond that, we rely on sufficient tension to keep the straw straight.

• How is the leak tightness verified before installation

Besides leak testing of each straw, each panel is tested in a vacuum vessel before being assembled into planes.

6. Electronics and HV

• What is the modularity of the high voltage?

~200 HV channels available but we have not decided how best to distribute this. Note, however, straws can be individually disconnected from HV by remotely blowing a fuse.

• If there is a problem with FE electronics how is it repaired?

Barring catastrophic failure preventing a large section from taking data, we would continue data taking till the next repair period. At that time, the faulty plane would be removed, and the appropriate panel replaced with a spare. The defective panel would be repaired at a later time.

• Has the components inside the gas manifolds been tested for aging (if aging is an issue)

All material except for the active elements (chips) have been irradiated by other experiments to higher levels than we expect and so have in affect been tested.

Besides direct aging of the components, they may outgas and cause aging of the straws. All components of the final detector, except electronics, were used in the aging prototupe. The geometry of the parts have changed substantially, but not the material.

• How is the cooling of the electronics done?

Electronics are mounted to an aluminum baseplate which in turn is cooled by a SUVA (2-phase refrigerant) cooling loop.

• What is the estimated radiation levels for the FE-electronics? Is it an issue?

Radiation is primarily low energy neutrons. Expected dose is 2×10¹⁴n/cm². We plan to test the full FE chain, both for material damage and SEUs, at a neutron facility at Berkeley.

6. Operation of the detector

• How is the sudden leak in the straw handled?

o How many straws have to be turned off in case of sudden failure during operation?

We can shut of one plane: 576 straws, or 1/36^th of the detector.

o Can a straw be replaced?

Yes, but not easily. We would more likely cut the straw out and plug the holes in the manifold.

• How is the T0 found?

We have two approaches: use t₀ from the calorimeter, or fit it directly from the track. Our normal fitting, which assumes an electron, simply includes t₀ as a fit parameter. For PID, where we want also to get velocity, a local t₀ is calculated using "doublet" hits: two hits in the same plane.