Doing more thinking on the superconducting magrid feedthrough. Rather than have a bend between the magrid and a centered feedthrough, it would probably be easier to make a custom conflat blank with an off center 2.75″ pass though welded on. Like this:
The part would looks something like this (via MDC):
Shout out to Anthony Clark from the UK for his donation of $10! Thanks Anthony!
YES! Check this business:
Here it is with the color levels adjusted so you can see more:
This is really really exciting.
First thing I learned is you need some gas in that chamber to start a plasma. I started at pressures around 1 e-6 torr and got nothing. What you really need is pressures above 1 e-3 torr. My gauge doesn’t work in this range, so I was flying blind.
I had to install a valve to leak air into the chamber to keep the pressure where I needed it:
Just playing with this device for 10 minutes gave me more of an intuition for plasma than most of the reading I’ve done on the topic.
I’m now one small step away from first fusion. FUCK YEA.
Spent the last two days throughly grounding the entire system.
We start by attaching a ground line directly to the overhead sprinkler system. First we scrape off the paint:
Then we attach this purpose made ground clip:
Here it is installed:
We are using 12 AWG insulated stranded copper. In addition to this pipe ground, we are also using outlet ground from two different circuits:
Ground each piece of rack equipment, and the rack itself:
Ground several points on the chamber, sled, and pump:
Here is the high voltage connection point with high voltage divider inside glass insulation:
I’ve been doing some brainstorming on the standoff for the superconducting magrid. This is a messy problem! You need a cryogenic feedthrough that is also a high voltage standoff. Then you need to pass in the YBCO superconducting cables, and wiring for the persistent switch.
Yesterday I realized that we can get most of the way there by welding together two off the shelf components:
This diagram refrences parts 9812107 and 9611005 from insulatorseal which is a subsidiary of MDC Vacuum. I’ve sent this drawing to insulatorseal for a quote. One problem is that part 9611005 is only rated up to 6kV so we will need a custom variant to get to the 10kV to 40kV range.
The idea with this setup is that the high voltage could come through a standard HV feedthrough and connect to the insulated tip of this feedthrough via a connecting wire:
So the approach I used to secure the inner grid just wasn’t cutting it: The inner grid was not tightly secured. So I took a page from Andrew Seltzman’s liquid cooled grid by using telescoping ceramic tubes:
I had to file the OD 12.8mm tube with a diamond file to get it to fit inside the next larger tube.
Inside it’s wired together with wire nuts and teflon coated aircraft grade 24 AWG wire:
Note: we ended up using a shorter and thinner stretch of wire than pictured here, but this illustrates how it’s wired.
Here we see it all assembled:
During previous tests we noticed that the outer grid was not mechanically secure in the chamber. Today I added a support to the outer grid:
Now it fits snugly in the chamber. Keeping the alignment correct was important.
Also we received the high voltage power supply back from Glassman. It’s been converted to a negative potential for use with a fusor. We got a nice shock mounted shipping rack:
I’m getting everything ready for first plasma.
Created a new STL of the superconducting magrid using a tolerance of 0.01mm. It looks much smoother:
If you looks closely you can see the faceting. Also took some time to refactor the code that generates these shapes. I added a temp folder and a parts folder to keep the main directory free of trash. Now the resulting STL and PNG filenames include the hash tag of the current git HEAD so you can connect an STL file to the code that generated it. You can find the branch I’m working on here.
A few days ago I send the STL files for the superconducting magrid to Prometal for fabrication. They are having problems with my STL files – overlapping triangles; the red area will not clean up:
I have no idea how to fix this.
UPDATE:
Spoke with Mike at Prometal. He suggests trying again with higher resolution triangles. I’ll do that today.
Today we ran the first test putting current through the superconducting magnet. First step is wiring up an AC connection for the DC power supply (0-100VDC, 0-120 Amps):
I learned how to crimp. It’s easy, but you need the right crimping tool. This power supply uses 240V single phase on the AC input side. I include these rather mundane steps because they matter. They need to be done for the project to move forward; they need to be done well; they often have their own challenges, “right ways”, tricks and considerations. I learn more than I expect from the simple stuff like connecting a power supply.
So here is the superconducting magnet soldered to a 50 amp twisted copper cable for main power, and the persistent switch with it’s nicrome heater wires:
On the other side we have the probe for the magnetometer secured with Kapton tape:
Here is the whole setup with a high wattage power supply for the main superconducting current and a low wattage power supply to power the heater coil. Dewar flask below.
Magnetometer check:
OH. Speaking of this magnetometer: I found out today that it drops -42 gauss when you dip it in liquid nitrogen. At first I though my gauss meter was broken from falling off the table!
So here is how it went:
First we lowered the experimental apparatus into the dewar of liquid nitrogen. It took several minutes for the nitrogen to stop boiling.
Next we turned on the heater up to 1 amp, based on previous experimentation. From there we incrementally raised the amperage, waiting about 1 minute between steps. The good news is the superconducting magnet seems to work. Here are the magnetometer readings for each amperage (adjusted for the -42 G drop from cryogenic):
5 A: 58 G
10 A: 126 G
15 A: 183 G
20 A: 232 G
So far so good!
Next we tried putting the magnet into persistent mode by turning off the heater allowing the persistent switch to become superconducting again.
Upon turning off the heater, the magnetometer jumped from 220 G to 115 G with no change in amperage to the magnet.
From here I lowered the amperage to zero over 10 seconds. The magnetic field returned to baseline. In other words FAILURE – the current did not persist.
Stuart suggest the problem may be that we are leaving the power supply connected, and it’s acting as a resistive load. We can certainly test this by adding a circuit breaker at the power supply.
So… not a complete success, but we have moved the game forward. I learned a lot today. I plan to retool this experiment, get some fresh LN2 and try it again.
