Convection Cell

From Hall A Wiki
Revision as of 16:45, 23 October 2012 by Zwzhao (Talk | contribs) (oven drawing from Alan Gavalya and Susan Esp)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

file dir

from Al Tobias

As a first step toward A1N target development, it was agreed to produce two (2) prototype all glass cells of convection style. As a first step toward a prototype design, I am sending you some preliminary drawings. The first drawing contains nominal dimensions of our current convection cell. The second and third drawings show a cell with larger pumping chamber and larger target chamber, respectively.

1) Can a convection cell of dimensions shown in the first drawing fit into the JLab target oven? 2) What is the largest pumping chamber the current JLab target oven can handle? 3) We may want to extend the length of the transfer tubes, so that the distance between the pumping chamber and target chamber is greater than the amount shown in all of the attached drawings. What limit would be imposed given the current JLab target?

from Yi Qiang

The designs are pretty good to me. I'm thinking about another variation with narrower transfer tube for the wall effect. The inside dimension of the current oven which holds the pumping chamber is 6.4 inch (W) x 6.4 in (L) x 7.0 in (H). The side windows which allow laser passing through are 3.8 inch in diameter, they are horizontally centered on each side, and vertically, their centers are 2.625 inch above the bottom of the volume. So in principle it should be large enough to hold all your designs with maximum 2 inch pull-offs, but the room for the laser spot is really tight. Also, a new bottom piece needs to be designed to mount the new cells. More drawings of the current oven and the Transversity/d2n cells can be found here:

from Al Tobias

Attached is the final version of the 1st Convection cell (with string) we'd like to have constructed. It has smaller diameter transfer tubes and its "height" is slightly larger.

The 1-inch bulb was suggested by Gordon as a well defined and larger volume of gas on the transfer tube (TT) to perform pulse NMR. The bulb would allow us to test polarization measurement techniques on the TT in case any future designs of a target cell limit us the ability to do such measurements on the target chamber.

Scott used 1-inch spherical cells in his research.

As for the bulb, some of the motivations: a) to provide for more volume than the transfer tube alone: keep in mind that a TT cylinder with 6mm I.D. is much less volume then a sphere with 22mm I.D. b) allow for doing "Spin Echo" measurements: for various reasons, a spin echo may allow one to more consistently measure the NMR amplitude for polarization. For example, in cases where the FID signal is too short (decays too quickly due to field inhomogeneities ie. short T2*) it becomes difficult to fit the FID back to time zero to extract an amplitude. This is especially true when there is much RF "ring down" from the initial pulse that contaminates the beginning of the FID. The solution is to do a spin echo in which case one just measures the peak of the echo. To do a spin echo successfully, you want most of your gas in a region of good RF (H1) field homogeneity since the pulses used in a spin echo are important to flip the 3He spins properly. Gordon's idea is to use a solenoid twice the length of the bulb so most of the gas lies well inside a homogeneous RF field during pnmr measurements, with minimal gas within the TT and other edges of the solenoid where the field is less homogeneous.

In a typical spin echo, one first applies a pulse that tips the spins as much as 90-deg, just enough to get a large signal. During this time, the spins dephase. After a time t, one applies a 180-deg (pi) pulse causing the spins to rephase an additional time t later to give the spin echo. We simply measure the amplitude of the echo.

As for tube tolerance, I can specify an outer diameter on the drawing if you prefer. I spoke to Mike and asked him what he thinks the OD of the tube would be if I ask him to make an ID of 5 to 6mm. He said the OD would be about 8 to 9mm (ie. 1.5mm wall thickness).

from Mike Souza

This new design represents some major changes compared to the Transversity Experiment done several years ago.. Though this particular cell does not require precision target windows. The Pumping chamber is larger and more complex and each of the transfer tubes are significantly longer. All of this has to be made from 16 mm od. tubing.

Based on those parameters I am quoting as follows:

1 ea GE 180 Convection cells w/ standard wall windows $1,800

1 ea 3 -piece GE 180 String manifold graded to pyrex $220

For GE 180 Convection Cell as shown using precision blown windows There will be an additional $600 per cell depending upon window thickness and tolerances identical to the Transversity Cells

hi, Mike

Here are a couple more questions. 1. I think you made a convection cell for UVa sometime ago. Why do you compare this design to the regular Transversity Experiment cells instead? 2. Will the bulb on transfer tube be any problem? 3. what will be the tolerance of the bulb on transfer tube? 4. Can you confirm the tolerance for 5-6mm I.D. transfer tube is 8-9 O.D.? 5. How long will it take to make the cell?

hi, Zhiwen

1.) The convection cell I made was made of pyrex glass. Aside from the fact that the material cost is about about $5 per lbs verses $100 per lbs for GE 180, it is a very forgiving material compared to aluminosilicate. Pyrex has materials in all sizes. Though the cell was still re-blown, I could use 70 mm tubing to make the pumping chamber. The transfer tube was 12 mm. I simply heated 12 mm pyrex and re-blew the glass back to its size, similar for the target tube.

The cell also did not have inner lips (aka Souder rings) on the transfer tube seals to the pumping chamber.

So the main reason I used the Transversity Cells as a basis for the quote is because those cells had to be wholly fashioned from GE 180 and from 5/8" size tubing as the sole source for materials. It is very difficult to blow 3.6" o.d. uniform shaped pumping chamber w/ substantial wall thickness to withstand pressure, when it is made from a tube that is nearly x6 smaller. Every fraction of an inch over 3" is very difficult and yields are significantly reduced. This will be made even more difficult making the two "lip" seals.

2.) Honestly the bulb will not be that big of a problem. Far less then keeping the bends uniform and in line with the specifications.

3.) I expect I can make the bulb to the tolerance of plus or minus 0.5 mm.

4.) As I discussed w/ Al Tobias, I cannot gaurantee 100% uniformity. However, assuming there is a linear volume that each tube must have using the tolerance provided as the equation. I expect to fabricate tubes and measure the outside diameter w/ calipers and check their volume w/ deionized water before sealing them into place and I expect to fall within that range. Otherwise I will continue making tubing till I achieve those parameters.

5.) I am not sure how long this first cell will take. I'm hoping to learn a lot . There are a lot of set-ups. I hope to get it done in 20 hrs. If things go bad it could be a lot more then that, this material is extremely susceptible to cracking under thermal stress. So making the various parts to specs is one aspect. However, assembling them and sealing the parts together is the risky part. Since the target chamber is anchored to the transfer tube in two places, simply annealing a completed cell may pose problems.

As for delivery: I have scheduled vacation time beginning June 27th to July 11th. I will be attending a week long conference beginning 6/28 to 7/3 in the Washington DC area. When I return home on the 4th, I hope to use that time to work on the Convection cells. And would expect to deliver the 1st cell no later then 7/11 and the have enough parts ready to I can delivery the 2nd cell one week later.

2.) I don't know if I could make a 4.3" spherical. I tend to think 4" is as big as I could go in any consistent manner. Bear in mind the larger the cell the thicker it needs to be to survive pressures. Keeping wall consistency w/ all the weight of that sphere on such a small axis is the other problem. Can I make a 4.3" eventually I think it could be done. But having it come out with the quality and uniformity of the 3 inch sphere is most likely beyond my means.

3.) Zhao the problem is that at some point the pumping chamber needs to be sealed to the target tube at two points simultaneously. To make a proper seal I need an air tight situation so I can blow or inflate the glass. I'm not sure if you've blown glass before. But I have two holes on the target cell some 14" apart and now I am sealing the pumping chamber w/ two sidearms that have to center to center the holes. If I seal one I have a huge leak on the other side and if I simply weld w/o inflating the weld area the glass tube which is only 5 mm id. will close. And this is aluminosilicate . I think Al and Todd would have a grasp of what I am saying but. I'm going to be like a one-armed paper hanger making the twin seals.ANd If the bends are sloppy and things don't line up really good I won't be able to make corrections afterwards.

> So accordingly we will plan to make oven bottom piece with two holes 10mm in diameter and 2" apart. Do you think it's reasonable? It could be, If the specs allow me 8 to 9 mm on stems and I am allowed + or- 1 mm on the 2" center to center... If I can make a suggestion, why not send me a jig that exactly matches the holes used for your oven top? Something as simple as as strip of tin or brass w/ two holes in it? Then you're covered by saying "must be made to fit..."

convection cell tolerance

the drawing requires the distance between two centers of two transfer tubes 2" apart. tolerance is 1mm (0.04") according to Mike Souza. so it varies from 1.96" to 2.04"

the drawing requires each transfer tube is O.D. < 0.4". The actual size of Protovec-1 is 0.330" and 0.338". tolerance is 0.5mm (0.02") according to measurement on transversity cell. so it varies from 0.31" to 0.36"

So for two transfer tubes, the max outer distance is 2.4", the min inner distance is 1.6"

1. if the jlab oven bottom piece has two holes 0.625" in diameter and 2" apart. its outer distance is 2.625" and inner distance is 1.375". The min gap would be about 0.1" on each hole.

2. if the jlab oven bottom piece has two holes 0.7 in diameter and 2" apart. its outer distance is 2.7" and inner distance is 1.3". The min gap would be about 0.15" on each hole.

transversity cell tolerance

The three transversity cells have average transfer tube O.D 12.5mm=0.5" with tolerance less than 0.5mm (0.02").

The transversity cells don't have transfer tube O.D specified on the drawings.


Yi Zhang's measurement

Jie Liu's summary based on Yi's measurement

oven drawing from Alan Gavalya and Susan Esp

  • convection cell in transversity oven with new bottom piece

The oven bottom piece with two Tee is designed similarly to UVa's, but with large holes, diameter 0.625" comparing to 0.58"

the windows center to bottom is 3.15" while UVa's oven has about (3.325+3.5/2)"

At where the Provect-1 transfer tube connects to the oven bottom piece, it has inner distance 1.59" and outer 2.26", the left tube D 0.335", the right tube D 0.330", while the bottom piece has inner distance 1.376" and outer 2.624", both tube D 0.312" cell in oven Isometric view of the cell mount view of the cell mount Working Cell A1n 3.60 Dia Cell Mounting Plate Cell Central Retainer Cell Outer Retainer
  • transversity oven

"transversity target was angled at about 41 degrees to allow vertical pumping. so moving the target in line with the oven requires moving the support over with respect to the design position. We may not have space to do this. There are probably more considerations to be found in the future." (window 3.8", current) (window 3.13", outdated)

transversity oven photos