Page 1 of 1

cyclotron electromagnet - 1 Tesla achieved

Posted: Sun May 13, 2018 4:47 pm
by Chris Mullins
Well, the embarrassment of a naked pole in our cyclotron magnet is finally over - the second coil is wound and installed. Winding 900 feet of 1/4" ACR copper tubing isn't fun, but far worse is sliding 900 feet of fiberglass sleeving over it, a couple of inches at a time (it was truly a family affair). There's more info in general on the coil construction here: https://mullinscyclotron.weebly.com/magnet-coils.html, but below are some pics for this second coil and the finished electromagnet.

Unlike the first coil, I embedded two K-type thermocouples in the center of the winding:
IMG_8616.JPG
After two weeks of evening and weekend sleeving and winding, the coil was done - 900 feet, 18" diameter, around 80 pounds, 0.5 ohms, 269 turns:
IMG_8620.JPG
The pole and magnet top piece weigh around 280 pounds, so time to pull out the Harbor Freight shop crane:
IMG_8637.JPG
That round plate resting on the bottom coil will be used to hold the upper coil onto the top piece - that's 0.190" aluminum plate.

After a few false starts, my son and I ended up resting the coil across two saw horses (same two used to wind it), and lowering the pole piece down, so it pushes out the wooden form in the center of the coil:
IMG_8644.JPG


After it's in place, the coil is attached with 1" aluminum square tubing and 1/4" threaded rod. Now it can be picked up as one unit:
IMG_8646.JPG
Next, the top piece, along with pole and coil are lowered on the frame, and the coil is wired up. With our normal 1.42" gap between poles, it's reading around 820 mT at full power:
IMG_8651.JPG
Each coil is around 0.5 ohms at room temperature, so the set of both coils is just around 1 ohm. Full power is around 55V at 55A:
IMG_8652.JPG
By inserting a 0.5" pole piece extension (8" diameter, 1018 CR steel), the gap is reduced to 0.92" With that smaller gap, I got just over 1 Tesla:
IMG_8653.JPG
On paper, I should get around 1 Tesla with the 1.42" gap, at 55A. With one coil, I was getting around 470 mT, so doubling that would be 940 mT. Instead, I'm getting around 820 mT, so the "efficiency" (ratio of actual vs predicted field strength) is around 90% at half power, and 80% at full power (3 kW).

820 mT is just barely enough to get past 160 keV, or p-B11 fusion. I was hoping for closer to 900-950 mT, to give some margin. Otherwise the protons at the desired energy will be very close to the pole edge, and approaching the edge of the chamber itself. I may have to replace the power supply with something beefier to provide some margin.

Total coil inductance is difficult to measure directly. My LCR meter only goes to 100 Hz, not low enough to get the "DC" inductance. Readings at 100, 120, and 1000 Hz are 34.3, 32.9, and 14.6 mH, respectively. A (very) rough L/R time constant measurement going from zero to 12 amps gives around 80 mH.

Next step for the magnet: either cooling, or bulking up the power supply. Will be a while though, now it's time to circle back to the cyclotron itself to make further improvements.

Re: cyclotron electromagnet - 1 Tesla achieved

Posted: Mon May 14, 2018 2:25 am
by Rich Feldman
Great work there. Do you now have dirty gloves in papa, mama, and baby bear sizes?

Regarding your inductance measurements, I'm guessing that eddy currents in the steel will be hugely significant at 60 Hz or even 10 Hz. What's the skin effect depth at 60 Hz in solid steel? When you apply voltage to the coil, and the current has stabilized, I think it still takes a while for the magnetization to penetrate all the way to the pole axis. The coil ampere-turns are opposed by eddy current ampere-turns in the steel (along the entire flux path). I haven't personally run the numbers. If your gaussmeter doesn't have an analog output, maybe a video of the digital display could answer the settling-time question.

By coincidence, today I was thinking about a second "temporary" winding for my 3 inch magnet. That after assembling the whole yoke for the first time, this weekend. 1 tesla should be easy with no air gap. With 1 inch gap, and the orange extension-cord coil, the "theoretical" drive requirement is on the order of 100 A at 100 V. The wire can handle it long enough for a flux density measurement, but it would be easier to get the electricity if it were a 50 amp problem.

Re: cyclotron electromagnet - 1 Tesla achieved

Posted: Mon May 14, 2018 9:39 pm
by Chris Mullins
Rich,

Good question on skin depth. I'm not sure about 1018 steel, but iron is about 80 mils at 1 Hz.

My gaussmeter doesn't have an analog or digital output, so I have to video a test run and transcribe the numbers later. I ran a test by putting a (much abused) resistive shunt in series with the coils, so I can measure current with a voltmeter next to the gaussmeter. I then went from zero to half scale on the power supply. Here's the setup:
depth.png
The left voltmeter is reading current at 10 mV/A. The right meter is my gaussmeter. In separate tests, both seem to respond to full scale changes and also refresh the screen in about 0.5 seconds. Probably should have used a real stopwatch instead of my timer.

Here's a step change from off, to half strength (454 mT), and back to zero:

Code: Select all

seconds	Amps	mT
12	0.18	15
13	22.49	819
14	28.54	450
15	28.52	451
16	28.50	452
17	28.50	453
18	28.50	453
19	28.50	453
20	28.50	453
21	28.49	454
22	28.49	454
23	28.49	454
24	28.49	454
25	28.41	454
26	4.36	137
27	0.45	38
28	0.17	19
29	0.18	16
30	0.18	15
I'm not sure what that reading of 819 mT is - some sort of anomaly from the meter. The field seems to reach almost full strength within a second or two, but takes several more seconds to reach the final value of 454 mT. The power supply is in constant current mode, and reaches steady state several seconds earlier. It also takes a few seconds longer for the field to decay back to steady state with no current (around 15 mT).

Here's the exciting video of the test: https://www.youtube.com/watch?v=ofALoFKxFtU

I get an extra digit of resolution under 200 mT, so I repeated it with a lower current:

Code: Select all

seconds	Amps	mT
17	0.18	15.8
18	12.43	125.6
19	11.41	192.0
20	11.42	193.1
21	11.42	193.3
22	11.42	193.5
23	11.42	193.6
24	11.42	193.6
25	11.42	193.6
26	11.42	193.6
27	11.42	193.6
28	11.42	193.6
29	11.42	193.6
30	11.20	193.6
31	0.97	89.3
32	0.17	18.6
33	0.17	17.2
34	0.17	17.0
35	0.17	16.8
36	0.18	16.6
37	0.18	16.5
38	0.17	16.5
39	0.17	16.5
40	0.17	16.4
41	0.18	16.4
Here the effect is more noticeable. The current is stable at 11.42A after 2 seconds, but it takes 3 more seconds for the field to reach final value of 193.6 mT. When the current switches off, it takes even longer - around 9 seconds. For the record, here's the video:
https://youtu.be/odahbSQCHl0


In these tests, I wasn't breaking the current with a switch - I was setting the source to zero amps.
The effective source impedance from the power supply when sourcing current is much lower than the diode/MOV protection network that the magnetic field discharges into with a step change to zero - perhaps that's some of the difference between a step change up, and down. Would be interesting to try a physical switch.