Archived - On the way to a microtron, testing of the magnet system

[Rapp Instruments]

Hi,
today I finished the magnet system and made some tests. The pole pieces have a diameter of 145 mm with a gap of 26 mm.
A hall sensor fixed to the moving arm of a XY-recorder allowed to make a map of the field with a computer and AD- and DA -converters. The map shows the rather uniform field over the area of the pol pieces. The map was taken with a coil current of 2 ampere.
For running the microtron I need a field between 100 and 200 mT. Because I have not the right power supply until now I made a test with a charged electrolytic capacitor (20000 µF, 40V), with 40V 10 A a field of 270 mT is generated. In this pulsed mode ever higher fields seems possible, perhaps even some 800 mT for a small cyclotron.
Now I'm going to build the chamber

Thomas

[Larry Upjohn]

Rapp;
did you build an algorithm to scan the hall device over the magnet or did you use a manual method reading the A to D's and then step the XY position with the D to A output. Just wondering as it is an elegant method either way. Best wishes as I avidly watch for your next installment! Larry Upjohn.

[Doug Coulter]

You do some very nice work, and apparently do it pretty fast as well!

Have you considered using NeFeB magnets along with this? I built a small test jig, intended to someday be a cyclotron, with them and easily get ~8k gauss in a 1/4" gap, as measured by a hall effect magnetometer, and like yours, it's pretty uniform. I used 1.5" diameter by 0.5" thick magnets at each pole piece, and used 1008 alloy soft iron for the yokes for mine, welded and then copper plated. You could also have some windings for fine tuning, but in this case you'd save a lot of power and heat.
To make magnetic losses minimal, I used 1" by 2" iron pole pieces on the outside of the yoke.

Here's a pic of my little version while I was trying to measure the field via NMR (which didn't work well).

The screws on top are for taking it apart, not holding it together, which it does fine on its own, eg they are threaded into the top piece to push it off of the rest.

For higher fields, you could use better grade magnets (NeFeB comes in several grades, and there are other types of magnets that are bigger but have much higher curie temperatures), or use bigger ones and then tapered pole pieces. And then there's always the wire windings.

The main disadvantage of the new magnets is the low curie temperature, I've ruined a few that way.

[Chris Bradley]

Thomas,

If you're after only 100mT then you might be interested to visit the construction I posted on viewtopic.php?f=15&t=7189#p51095

I would suggest the biggest benefit for the assembly I built is uniformity of field within the volume. However, clearly it is a fixed field magnitude and won't serve if you want to adjust the magnetic field strength.

[Rapp Instruments]

Hi Larry,
I used a small programm written in Borlands Delphi to scan the area and to store the hall values. I like this pascal language to write all kind of applications.

I used low carb. steel, the german name is st37, I think it's similar to the american 1018 steel. With my 9300 MHz frequency I need, depending on the orbit energy, 168 mT (255 keV), 112 mT (170 keV) or 84 mT with 127 keV.
I thought about using perm. magnets at least for biasing, the rather low fields should be no problem even with bigger gap. I need the 25 mm gap because the X-band R-100 waveguide has this width. But for flexibilty I decided to try my luck with a old fashioned coil. It's true, I waste a lot of power, but I planned to build a switched current source. With a on-time of 0.5 sec and an increased rep. rate of the magnetron (300 Hz seems possible) I should get enough electrons for a good measuring statistic and the magnet get not to hot.

Hi Dough,
a very interesting approach to measure the NMR, why you wasn't satisfied with the result? I think for NMR you should use a modulation coil at the Magnet pol pieces for sweeping over the resonance line. Or did you sweep the rf-frequency.

Hi Chris,
I have read about your measurements when you posted them. In fact your post initiated the idea to try to build a microtron myself

[Doug Coulter]

The NMR approach didn't work well because I wasn't even sure what octave I should be in at the time -- I didn't want to make too many assumptions, and it turned out that was wise. I got barely detectable absorption peaks at two frequencies an octave apart. This is mainly because my signal generator was aged and harmonic laden and a bit unstable (since fixed by getting a new one -- the 1938 tech was ready for a rest). Yes, I swept the frequency, as the field was what I was trying to measure and didn't want uncertainty there. In the end, I bought a magnetometer from a source ChrisB found for me, and it works pretty doggone well. It has a bunch of other uses too, like finding out when I need to pull out a degausser for something.

On the other hand, the field vs magnet vs gap web page he found for me gave WAY wrong answers, even though the magnets measured to spec with the magnetometer -- go figure. The web page he found was wildly optimistic about the realized field in the gap -- 12kg vs the about 8 I measured.

I did this structure as it would fit through the door on my main tank (note rounded corners on the pole pieces to make this so) -- this means I don't need as long a gap, as big magnets etc, because the gap isn't full of vacuum chamber walls and so on -- the whole thing is inside. This brings up cooling issues for the magnets, which I planned to solve via water cooling through some holes in the magnet yoke structure.

Good luck on your project. Whatcha gonna use it for when it works? I fantasized early on about making a linac with a microwave oven magnetron at 2.45 ghz, but never did get around to it. It was good motivation to learn more about RF modes in waveguides, though, which has turned around and paid for itself working with the new ion source I'm getting ready to put into service. Knowledge is never dumb to acquire.