Dear All,
Here is the summary about handling pumped Oxford superconducting magnets.
Thanks to all who respnded to my post: Janusz Koscielniak, Jeoff Armstrong,
John Davidson, Tara Sprules, Steen Nielbo, Hamish Grant, Beverly
Ostrowski, Nick Wilson,
and Andy Shepheard.
Some short info and list of identified facilities with similar
instruments is posted here:
http://nmrwiki.org/wiki/index.php?title=Pumped_NMR_magnets_from_Oxford_Instruments
Please feel free to add to that list and comment.
Thank you!
Evgeny.
1) Background
2) Overview of cooling system
3) Needle valves
3.1) Regulation of Needle Valves.
4) "Fridge flooding" (and the prevention of)
5) Monitoring the magnet system.
6) The proper way of isolating magnet from the pumping system.
7) A shortcut procedure for briefly disconnecting the pumping system.
9) Pump electromechanical valves.
10) Pump power supply.
11) Other problems that have been occuring in the past.
12) Original post.
1) Background.
Coil of pumped superconducting magnets is cooled to ~2.18K so that it can
sustain higher current and thus produce the higher magnetic field.
Cooling below 4.2K is achieved by pumping continuously pumping helium with the
vacuum pumps. Oxford Instruments has built ~40 such systems in the world.
At 3-3.5K magnet will quench.
So it is very important to follow the procedures of handling the magnet
carefully as small mis-steps can force the magnet to warm up.
Generally there are two parts that need to be closely watched:
(1) needle valves, (2) pumping station.
Stability of magnet temperature. Someone formerly with Oxford once
turned pump off for 12 hours with no change in coil temperature
(I guess this assumes that he closed the needle valves - not sure -
second hand information) Someone else told Geoff Armstrong that
coil won't warm up for several days without pumping if NV's are well closed.
Some info about Helium:
At ambient pressure <sup>4</sup>He boils at 4.2K.
<sup>4</sup>He starts transitioning to superfluid state below 2.1768K
(lambda point).
When temperature approaches that point heat capacity of Helium increases
so cooling He below that point by evaporation is very hard.
(But temperatures of about 1K are obtainable at a cost of evaporating
about 50% of liquid.
This is how 1K-pots work)
2) Overview of the cooling system
Magnet has two chambers separated by spring loaded valves.
The lower chamber contains the magnet coil immersed in the liquid He
at around 2.18K,
while the upper chamber contains 4K He. Both valves are permanently closed –
they would open only during quench. So the Helium in the lower tank is
trapped permanently.
The system for cooling the lower chamber starts at the block
with the two needle valves (NV's) located at the bottom of the upper chamber.
The NVs regulate the inflow of liquid He from the outer 4K chamber.
The metered flow of He enters the piece of tubing called refrigeration loop
where He expands into the vacuum created by a rotary vacuum pump and
cools to ~ 2.18K.
Finally the vacuum pump is connected to the refrigeration loop via long
large diameter corrugated tube.
3) Needle valves.
The two needle valves are connected in parallel for redundancy.
They are controlled by the knob at the end of about 1 meter long shaft.
A trace amount of solidified gases trapped on the surface of valve
block can make the valve feel 'sticky'. This effect is compounded by the flexing
shaft when torque is applied making regulation feel jumpy.
The two valves are slightly different - their flow rates are
calibrated at installation.
(Hopefully you have a record of that in the documentation).
Needle valves need to be exercised every 6 months
(close one valve and open another one) so that they don't get
stuck in one position.
If they are set incorrectly the magnet will eventually warm up.
If the valves are closed - refrigerator won't work. If on the contrary -
they are too open - liquid He will fill ("flood") the loop
and will start warming up the inner chamber.
3.1) Regulation of needle valves.
Adjusting these needle valves takes some practical skill.
All valves have right standard screw thread ("clockwise close"
"counter-clockwise open") and it takes more than fifty
turns to travel the full range ( useful positions are within 2-3 turns max from
fully closed position one turn opens a lot already).
Needle valve control has some hysteresis: once valve knob is turned
some way, slight change in resistance in turning will be noticed.
One small adjustment like that is done at a time,
then one must monitor fridge pressure.
However, pressure reading alone is not enough for correct adjustment.
Sufficient flow of Hellium should be allowed to keep temperature constant.
(Pressure value will depend on the pump model)
The desired flow should be set while watching pressure not to come any
close to 40 mbar - better well below this value.
If it seems impossible - either pump is too weak or there
is something wrong with the magnet.
The biggest disaster apart from spontaneous quench, is when moisture
gets inside pumping line. It will freeze in its coldest spot making
it impossible for the helium gas to flow through the fridge.
There is no way of fixing other than warming up the magnet (a very
costly affair).
4) "Fridge flooding"
This situation is created when fridge loop fills with 4K liquid He.
This can happen when pumping stops, but needle valves are kept open
long enough (see above) or needle valves being open too much.
If that happens, temperature of He inside the lower chamber will start rising
and will start slowly warming up the coil.
Fridge flooding is not immediately dangerous to the magnet.
It may take a couple of days for the coil to warm up to the alarm level.
After 12 hour increase of temperature will be noticeable.
All liquid that has flooded fridge will have to removed
before cooling action can be resumed. Pumping all this liquid will take
a lot of pumping effort. For such occasion older versions of maget
came with one normally used RV5 pump and one RV12(higher capacity)
service/backup pump.
If the fridge is flooded with 4K liquid He, both needle valves
should be closed, both pumps turned on until pressure reaches stable
lowest value.
The lowest achievable pressure is <1 mbar if the needle valves don'l leak.
Pumping to this level of pressure assures that the fridge is completely emptied
of the liquid.
5) Monitoring the magnet system.
Monitoring equipment consists of the monitoring PC and three old
fashioned flow meters.
Fridge flow meter is connected to the output of the vacuum pump. Two
other meters measure
boil off through the "conventional" exhausts of N2 and He.
The flow information is very important. In interest of saving helium the
flow should be kept as small as possible for keeping temperatures constant.
Pressure inside fridge loop is displayed on the PC monitor as
"manostat pressure".
This pressure should be always interpreted in a context of flow gauge reading.
Coil temperature (as well as temperature of some other parts) is
monitored by the PC.
Alarm is set to 2.5K
6) The proper way of isolating magnet from the pumping system.
The surest way to protect magnet with the pumps off -
completely close the needle valves, pump for several minutes till the
lowest pressure
is reached, then close the ball valve. Now pumps can be safely disconnected,
but coil temperature should still be continuously watched.
7) A shortcut procedure for briefly disconnecting the pumping system.
Given all the hassle that goes into regulation of needle valves,
a modified procedure can be followed if pumping is to be interrupted
only briefly
(i.e. several minutes).
The short procedure is the following:
* if the pumping system needs to be open to ambient pressure
- close the ball valve on magnet leg
* close pump valves
* monitor fridge pressure to make sure it does not start rising rapidly
* perform pump replacement, reset circuit breaker, etc.
* once pumping is restored confirm that the pressure is dropping
* monitor pressure and coil tempereature periodically for a couple of days
If this procedure cannot be performed quickly enough so that the
fridge loop pressure rises above 40 mbar, it will fill up with liquid He
and will stop working as a refrigerator. So this procedure is only advised
when you are sure that a longer downtime won't be necessary.
8) The Pump Trolley.
Pump trolley provides some protection against pump failures:
* a relay trips when pump current drops to zero
* pump has thermal protection - it stops on overheating
Pump failure trips it's cirquit breaker and triggers
automatic switchover to another pump.
Pump trolley is not monitoring system pressure.
Circuit breakers inside the pump control box look like the ones
you might find in your garage. If handle is down - it's tripped.
Just lift the handle and it will reset.
To access them the pump control box needs to be turned off -
by turning the big red knob on the side by 90 degrees.
Then open side door with key or skrewdriver, reset tripped breaker.
Alarm wont stop on until the breaker is reset.
Pumps are working fairly hard even though they are pumping inert gas.
Working pressure is ~18 mbar (x100 above ultimate pressure)
I've measured working temperature of 68.9C - which is very close to
70C - maximum temperature allowed by the manufacturer.
Oil should be changed every 3000 hours of continuous operation.
It is recommended to rebuild the pump once a year, and discard
it after maximum two refurbishing cycles.
9) Pump electromechanical valves.
Remember to keep both valves in "Open" position in the
normal operation. Backup pump won't start if it's valve's control
is in "closed" state. Secondly, leaving valve knob in
"open" state does not mean that the valve is actually open -
it will only open when pump is on for 30 seconds.
So when both valve controls are in "open" state with one pump in
"standby" and another one in "on" state -
only the valve that is being pumped will open.
10) Pump power supply.
Use appropriate UPS (whose VA specification well
exceeds power consumption of the pump motor).
If possible - plug UPS to line backed up by some emergency generator.
We have "tripp-lite smart pro net plus 3000" - supposedly 2.4kW -
should pull two RV5 pumps for 20 minutes )
11) Other problems that have been occuring in the past.
Monitor, the PC computer, pumping station, etc.
Replacing the monitoring station may take several months.
12) Original post.
....This might concern only those who have a Pumped Oxford Magnet. (we
have 18.8T/2.2K/63mm bore magnet; with Edwards dual RV5 pump platform)
But I hope that those who do have similar system find some time to
read through this.
As I found out from my own experience it is nearly impossible to
obtain technical information and guidance from those few
engineers who know this system. And it is quite positively impossible
to get it in the timely manner.
So I think it would be great to establish a network for those of us
who maintain a system similar to this one.
About a month and a half ago we had a small but inconvinient incident:
One of the pumps backing the Helium fridge stopped running because of
a tripped cirquit breaker inside the pump control box.
That's a minor problem - just open the box and reset the breaker. Also
there is a second backup pump - so it's not a critical problem.
The issue was that (1) pump control box manual had no troubleshooting
section nor a cirquit diagram supplied -
provided documentation did not even explain how to turn the box off
and nothing useful about the breakers.
(2) oxford magnet manual gives identical instructions for opening and
closing needle valve - an obvious mistake.
also the manual warns of possible manget quench if valve is left at
wrong setting and strongly suggests to receive proper training
before adjusting the valve.
(3) according to the magnet manual I was supposed to temporarily close
that valve in order to replace pump and I had to reset the cirquit
breaker.
It took five days for the manufacturer to find the cirquit diagram for
the pump box
and I've never heard anything back about the valve.
For all that time we had no emergency backup pump so we had to closely
watch the one that was on. Since I did not have instructions on using
magnet valves,
I did not want to risk stopping pumps by doing something wrong with
the pump control box (which as I already mentioned lacked the complete
manual).
That's about it.
We still don't have any instructions on how to regulate the valve
- I had one suggestion to temporarily close the ball valve without
touching the needle valve when changing pumps
- but that is not what the directions in the magnet book say. One day
we might have to regulate the fridge valve
and by then we may not be able to get any instruction at all.
So perhaps if you have a similar setup we could exchange contact
information and stay in touch just in case.
Hopefully we can close the gaps in the documentation too.
Thank you.
Evgeny.
--
Evgeny Fadeev, Ph.D.
Director, BioMolecular Spectroscopy Facility
1212 Natural Sciences 1
University of California Irvine
Irvine, CA 92697
telephone: 949-824-5842
www.physics.uci.edu/~biomolenmr
nmrwiki.org
--
Evgeny Fadeev, Ph.D.
Director, BioMolecular Spectroscopy Facility
1212 Natural Sciences 1
University of California Irvine
Irvine, CA 92697
telephone: 949-824-5842
www.physics.uci.edu/~biomolenmr
nmrwiki.org
Received on Thu May 08 2008 - 06:39:45 MST