Difference between revisions of "Minutes-9-23-2010"

From GlueXWiki
Jump to: navigation, search
(→‎Engineering)
 
(6 intermediate revisions by the same user not shown)
Line 18: Line 18:
  
 
= Minutes =
 
= Minutes =
−
Participants: Eugene, Fernando, Beni, Bill, Chris, David, Simon, Gerard (on the phone), and Lubomir.
+
Participants: Eugene, Fernando, Beni, Bill, Chris, David, Simon, Mark, Gerard (on the phone), and Lubomir.
  
 
It was a long (2.5 hours) but productive meeting, half of the time discussing the FDC requirements for the electronins, and the other half - the wire stringing procedure.
 
It was a long (2.5 hours) but productive meeting, half of the time discussing the FDC requirements for the electronins, and the other half - the wire stringing procedure.
Line 40: Line 40:
  
 
- fADC125 mapping: the 2.6mV/fC card output should map onto 80% of the full scale of the fADC125. Additional decrease of the range down to 60% is acceptable if it is needed
 
- fADC125 mapping: the 2.6mV/fC card output should map onto 80% of the full scale of the fADC125. Additional decrease of the range down to 60% is acceptable if it is needed
−
to match the CDC requirements for the fADC125.
+
to match the CDC requirements for the fADC125.  
 
   
 
   
 
- Gain for the anode cards: with the present gain of 0.6mV/fC and gas gain of 8x10^4, the wire signals saturate. Since the cards will be used in discriminator mode,  
 
- Gain for the anode cards: with the present gain of 0.6mV/fC and gas gain of 8x10^4, the wire signals saturate. Since the cards will be used in discriminator mode,  
 
the saturation doesn't affect the time information that will be recorded with F1TDC. If there are other saturation effects, like increased recovery time, then we need to
 
the saturation doesn't affect the time information that will be recorded with F1TDC. If there are other saturation effects, like increased recovery time, then we need to
−
reduce the gain by about a factor of two. Again, this will be problematic with the current version of the card since 0.59mV/fC is the lowest possible gain.
+
reduce the gain by about a factor of two. Again, this will be problematic with the current version of the card since 0.59mV/fC is the lowest possible gain. Gerard will look into that.
  
−
<!--
+
- Other fADC125 requirements: Lubomir investigated the effect of the peaking time on the resolution (page 517). Eugene: very small effect and it is not worth optimizing.
 +
Gerard expects the resolution with fADC125 to be closer to the one with discriminator and has doubts about this result. For the next meeting, Lubomir will prepare additional plots using other interpolation methods. Gerard: the common noise subtraction is not possible in the FPGA that collects the information from all the channels, because the zero suppression should be done before that in the FPGAs for the individual channels. Lubomir and Gerard will discuss other possibilities for the noise reduction. 
  
−
- Fernando prepared two pre-amps in discriminator mode, together with the cables and ECL translator. Beni is testing them. Fernando wants us to answer in one week whether they are acceptable. There's some delay with the delivery of the PCBs. They had to be here on Monday, but will have them next week. The contract for the PCB stuffing is ready submitted to the procurement. Will have the first 2x6 PCBs (1st article) within two weeks and after the approval the rest (26x6) will be done in 4 weeks. We discussed how we will store the PCBs: in boxes, wrapped and taped. Fernando is also working on the procurement of the HV mini cables.
 
  
 
== Engineering ==
 
== Engineering ==
  
−
- Cathode foils: Roger got an answer from Advanced Circuits that they can't produce the cathode foils because of the big sizes. Roger is still waiting for an independent estimate from them. He submitted to the procurement answers of a second set of questions from them regarding details of the attempts to find another source. Ross Small from the procurement states that "at this point approval of a sole source is not a sure thing" and he suggested advertising it in the FedBizOps site for 21 days. We discussed what are the pros and cons of such advertising. Pros: no longer sole source and it takes only 21 days. Cons: if some vendors decide to bid they have to prove they can do it and go through the same stages (first article production) as with Allflex which will take a lot of time.
+
- Preparation for wire stringing: the holes in the granite tables are being drilled now and all the initial parts to be attached to the table are ready. David has been working
 +
on the movement system and the encoder. There are several options to connect the stepper motor to a new computer, one of them would require a $2K adapter.  
  
−
- Some of the fixtures for the wire stringing were made in the machine shop. Tomorrow they will drill holes in the granite table in 126, that will be used for stringing. Starting next week Bill will have more time to work on the stringing. David brought the position measurement system from UVA on Friday 08/27. It turned out the motion carriage doesn't have enough travel to cover our chamber (970mm). Last week David disassembled the carriage with the stepper motor and the control from the FNAL wire winding machine and got it working with a very old computer! Bill prefers to have a ready system (stepper motor with the control and the software) that works with the new computers. Beni has a small power stepper motor and we will see if we can use it. Bill wanted to know the status of the tension measurement system at UVA. Lubomir will contact them after the meeting. Bill and Simon prefer to have system like in IUCF: starting vibrations mechanically and optically looking for resonance. Eugen asked about the procedure of repairing the wires if they don't meet the tension specifications. It was already done here by Brian many times.
+
- Bill presented his plan for aligning the wires to the soldering pads by using a precise squiggle motor (see links above). After adjusting each wire, it will be soldered at the same time. In this scheme the wires will be glued later. Eugene made a point that this is a new procedure that requires time for R&D and we should use a procedure
−
 
+
demonstrated to work. Instead of touching the wires, at UVA they applied pressure on the frame locally from below to adjust the vertical position of the board,
−
== Full-scale prototype  ==
+
glued the wires and then soldered them. Everybody agreed that our frames are different and this procedure may not work for us. Still, important is first to glue the wires and then
 +
to solder them to avoid tension on the soldering. The problem is that when gluing, the wires must be at the right positions and this can't be done for a group of wires with the
 +
squiggle motor. For that one needs a comb or individual weighs on the wires from the inner side of the chamber. During the discussion people exchanged other interesting ideas. At the end we agreed all this requires prototyping and we can start with something simple: after applying glue and before it hardens, manually adjusting the wire position (as much as possible) while looking at a microscope.
  
−
- Lubomir studied the chamber resolution with Garfield with two gas mixtures (40/60 and 90/10) and simulating also different electronics; see above links to the FDC log book. Apart from the resolutions itself, plotted are also the distributions of the differences, actual minus reconstructed position. Eugene pointed out they are asymmetrical and the uncertainties are big. Simon will use these distributions in the GEANT simulations to investigate the effect on the tracking. When comparing different gas mixtures (page 513 in FDC logbook), one can see that 90/10 has wider range with resolution below 200microns. The improvement for small distances (up to 2mm) comes from the smaller averaged drift velocity with 90/10, while in the 4-5mm region where the field is highly non-uniform the independence of the drift velocity from the field with 90/10 gas mixtures removes the tails in the distributions that we have for the 40/60 mixture. Simulating the discriminator mode (pre-amp in discriminator mode plus high resolution TDC) and the fADC125 without the shaper (that introduces additional integration of 40ns) at the same threshold, gives advantages to the discriminator. If, however, one can reduce the threshold (e.g. by common noise subtraction) from 30mV to 15mV than the fADC is compatible to the discriminator. Simon pointed that without shaper one still has to take into account the cables. Lubomir will update the simulations changing the peaking time from 11ns (pre-amp only) to 14ns (pre-amp plus cable). Still, comparing the resolutions with 11ns and 40ns peaking time the preference is for the smaller peaking time. We discussed whether we can reduce the preamp-cable-fADC125 peaking time to 14ns (question for Gerard) and whether it has to be coordinated with the CDC people. Fernando explained there will be two versions of the fADC125: for FDC and for CDC. The reason is the difference in the dynamic range and the gain of the preamps. Eugene and other advocated to have the same fADC125 for the two detectors.
+
== Full-scale prototype  ==
  
−
- Beni is working on the top and middle chamber with the new discriminator card and an analog card connected to the wires. He observes oscillations when the cards are connected to the central wires. Fernando will look at the problem tomorrow. Now Beni is doing tests with the last 24 wires. The reason for the lost Ar bottles is the regulator which at higher temperatures (when there's direct sunlight on the bottle) opens a relive valve. The regulator was replaced but we don't have more spares.
+
- Beni is testing the miidle and top chambers comparing the new method, pre-amp in discriminator mode, with the old one with an external discriminator. There were some noise problems especially in the middle chamber. The source of the noise: the CAEN HV supply used for the scintillators.  
  
−
-->
+
- Simon showed his new results (pages 516 and 519) for the momentum resolution including more realistic Garfield simulations of the drift times. The momentum resolution gets worse
 +
(page 516) compared to the standard 200 micron position resolution that was used before. However, Simon demonstrated that the knowledge of the error distributions can be
 +
used to improve the momentum resolution (page 519).

Latest revision as of 10:19, 24 September 2010

September 23, 2010 FDC meeting

Tentative Agenda

  1. Production
    • Clean room update
  2. Electronics
  3. Engineering
    • Cathode foil: sole source procurement status (Fernando)
    • Wire stringing: status and plans (Bill, David) [[1]], [[2]],[[3]]
  4. Full-scale prototype
  5. Other


Minutes

Participants: Eugene, Fernando, Beni, Bill, Chris, David, Simon, Mark, Gerard (on the phone), and Lubomir.

It was a long (2.5 hours) but productive meeting, half of the time discussing the FDC requirements for the electronins, and the other half - the wire stringing procedure.

Production

- Only good news about Blue Crab, procurement will finalize the contract with AdvanceTEC for the construction of the clean room by the end of the month. The AC for the whole area is being installed now.

Electronics

- Fernando: we have now all the PCB boards (except maybe one set, out of 28, that has to be checked). The contract with the assembly company is awarded.

- Fernando: according to the procurement, the sole source for the cathode foils is out of question. Therefore, Fernando submitted the specifications to be posted on the FedBiz site with a deadline for possible offers of Oct. 13.

- Lubomir started preparing a document explaining the FDC requirements for the pre-amplifiers and flash ADC125. It is based on the tests of the full-scale prototype including with fADC125, and on Garfield simulations. This document linked above contains plots and more details that were discussed.

- Gain of the cathode cards: present value of 2.6mV/fC will work for the FDC. To be safe we specify gain range of 2.0 - 3.2. The problem is that in the current pre-amp design, the next possible gain below 2.6 is 0.91. Lower gains might be needed if the cards generate with 2.6mV/fC. The final decision about the gain can can be made after building and testing one full package of 6 chambers.

- fADC125 mapping: the 2.6mV/fC card output should map onto 80% of the full scale of the fADC125. Additional decrease of the range down to 60% is acceptable if it is needed to match the CDC requirements for the fADC125.

- Gain for the anode cards: with the present gain of 0.6mV/fC and gas gain of 8x10^4, the wire signals saturate. Since the cards will be used in discriminator mode, the saturation doesn't affect the time information that will be recorded with F1TDC. If there are other saturation effects, like increased recovery time, then we need to reduce the gain by about a factor of two. Again, this will be problematic with the current version of the card since 0.59mV/fC is the lowest possible gain. Gerard will look into that.

- Other fADC125 requirements: Lubomir investigated the effect of the peaking time on the resolution (page 517). Eugene: very small effect and it is not worth optimizing. Gerard expects the resolution with fADC125 to be closer to the one with discriminator and has doubts about this result. For the next meeting, Lubomir will prepare additional plots using other interpolation methods. Gerard: the common noise subtraction is not possible in the FPGA that collects the information from all the channels, because the zero suppression should be done before that in the FPGAs for the individual channels. Lubomir and Gerard will discuss other possibilities for the noise reduction.


Engineering

- Preparation for wire stringing: the holes in the granite tables are being drilled now and all the initial parts to be attached to the table are ready. David has been working on the movement system and the encoder. There are several options to connect the stepper motor to a new computer, one of them would require a $2K adapter.

- Bill presented his plan for aligning the wires to the soldering pads by using a precise squiggle motor (see links above). After adjusting each wire, it will be soldered at the same time. In this scheme the wires will be glued later. Eugene made a point that this is a new procedure that requires time for R&D and we should use a procedure demonstrated to work. Instead of touching the wires, at UVA they applied pressure on the frame locally from below to adjust the vertical position of the board, glued the wires and then soldered them. Everybody agreed that our frames are different and this procedure may not work for us. Still, important is first to glue the wires and then to solder them to avoid tension on the soldering. The problem is that when gluing, the wires must be at the right positions and this can't be done for a group of wires with the squiggle motor. For that one needs a comb or individual weighs on the wires from the inner side of the chamber. During the discussion people exchanged other interesting ideas. At the end we agreed all this requires prototyping and we can start with something simple: after applying glue and before it hardens, manually adjusting the wire position (as much as possible) while looking at a microscope.

Full-scale prototype

- Beni is testing the miidle and top chambers comparing the new method, pre-amp in discriminator mode, with the old one with an external discriminator. There were some noise problems especially in the middle chamber. The source of the noise: the CAEN HV supply used for the scintillators.

- Simon showed his new results (pages 516 and 519) for the momentum resolution including more realistic Garfield simulations of the drift times. The momentum resolution gets worse (page 516) compared to the standard 200 micron position resolution that was used before. However, Simon demonstrated that the knowledge of the error distributions can be used to improve the momentum resolution (page 519).

Retrieved from "https://halldweb.jlab.org/wiki/index.php?title=Minutes-9-23-2010&oldid=22447"