Researchers such as these typically go about the Nat labs giving talks. Many times, they present what they've done, and where they are stuck. Oftentimes, the audience they speak to has sufficient expertise in their problems that brainstorming post-talk provides sufficient answers or at least promising directions of research.
The MIT tokomak guys were just here, and they had some really interesting challenges that the superconducting magnet specialists are able to address. The technical challenge they had was making a 20 tesla peak toroidal magnet for a stable plasma.. Not a big deal with the current generation of HTS. And, non soldered interconnects capable of 20 kiloamps or so..
Micro wire EDM is easily capable of doing this.
Here's a gear that's about 400 microns across as an example. I've papers on a spiral drill as well, but would have to look for the paper in the pile.
edit: just found a paper on a 25 micron end mill.
Fabrication of micro end mills by wire EDM and some micro cutting tests:Jiwang Yan, Kazuyoshi Uchida, Nobuhito Yoshihara and Tsunemoto Kuriyagawa in the
JOURNAL OF MICROMECHANICS AND MICROENGINEERING (caps not mine..)
edit: waittaminute. a number 80 drill is smaller than .35mm dia, a #97 is about .1mm. Watchmakers do this all the time..
John
Attachments
Last edited:
I used to work with someone who said the biggest technical jolt was going from doing some pA work to an interview at JET where they asked him how to measure leakage currents from the bus bars. when he asked the order of magnitude of the leakage they said 'about 200A'. Shifting ones brain the order of 10^15 on the fly often causes an audible clunk 🙂
Although not in your league, my eldest has a placement for next year at Daresbury. Daresbury Laboratory - Science and Technology Facilities Council They have some cool toys and I hope she has a lot of fun there.
Yes, I know. I have the bruises on my forehead to prove it. 😱
My league...your eldest...the jury is out on that, I've set the bar low.
My employer...perhaps not exactly, it looks like your eldest has gotten into a great facility.
Really great news. If she has any Q's arise w/r to supers, accel magnets, medical gantry's (ffag), PM me. I would be happy to help..
PS.. If she gets involved with a superconducting helical undulator there, you HAVE to have her contact me. They made a few fundamental errors on their proto that they did not understand.
John
Last edited:
We don't even bother with EDM for something as big as 0.25mm. The CNC lathe pushes the rod out of the chuck as the machining occurs so that it's supported close to where the action is.
BTW -- the sound of the gravity waves isnt a chirp, really. it is too low freq.
https://youtu.be/egfBaUdnAyQ
BTW2 - the Tokomat was designed and built and used at LLNL which then sent it to MIT when they were finished with it. I saw it at LLNL being built and used. From those studies LLNL went to a larger machine.... the one I was Sr EE technical coordinator on.
-RNM
https://youtu.be/egfBaUdnAyQ
BTW2 - the Tokomat was designed and built and used at LLNL which then sent it to MIT when they were finished with it. I saw it at LLNL being built and used. From those studies LLNL went to a larger machine.... the one I was Sr EE technical coordinator on.
-RNM
Last edited:
Apparently the physics is such that exceeding break-even requires scaling down the size of the vessel to roughly a meter cube and bringing center magfield up to 9 tesla. They gave a long and detailed physics explanation, complete with math, theory, and progress to date. Some in the audience understood it...while I was among them, I was not among them.. 😕
The big takomak's cannot get the magfield required for plasma stability. Even at a meter cubed, the super needs to support 20 tesla to get 9 central.
I believe that there are about 4 or 5 R and D groups poised to go over unity power with smaller devices.
For the LLNL unit, how did you quench detect the central solenoid at plasma initiation? The ITER guys came to me asking me about 500kohm resisistors capable of 100 kV ratings, I designed a co-wound tap detection scheme that eliminated the high voltages required and sped up the detection without the blanking period required.
John
Takomats? Takomaks? Tokamak? Or is this a different architecture that the googles doesn't know about? I'm assuming misspellings before misfusings.
Hey, I refused english as a second language when they recommended it freshman year.. 😉
Them:
Math...check, pretty good.
French...are you sure you took it in high school?
English...we give ESL courses here, you need it...trust us..
Me:
Huh?
John
At the time it was absurdly limited for the kind of decisions required, and of course the proponent knew that.
This is also at the root of super computing systems. We read only about their speed to number crunch as never being fast enough. There is a more important factor in super-computing apps and that is accuracy. A faster machine isnt always more accurate.
The math always is ahead of the machines to do the calcs. Most of the calc have to be simplified to get to fit into the machine in many ways. The numbers are so large or so small that the SOTA is only found in the calc at the last 1-2 digits. Which is also where the machine errors are greatest.
As the calc machines or super-computers get faster and more accurate, models can include more details for better outcome or predictions. Even so, the codes are so complex that these machines at thier speeds have to run for hours and sometimes over night to complete. So we can always use more speed.
THx-RNMarsh
Last edited:
As I'm sure the proponent also knew that determining the pencil's date of manufacture was absurd..
John
It was a con of course, but in the service of the truth---he really had prior art.
Right, I mean, it was more that the proponent was firing across the bow of the opponent.
And it is possible to date the paper with a bit of precision. But not when he wrote on it! Whatever happened to signed notebooks? Oh yeah the new law is first to file.
I saw very nice blog post (anonymous) but the guy claimed to have been the "driver" of the first Cray XMP at LLNL, so it must have been someone you knew. It was a long very intelligent post about relative computing power over the years. As one could expect there is no one answer. There are problems where indeed an off the shelf current PC is faster and those where the original Cray would still be faster. It has to do with many issues like on chip cache, memory footprint and memory speed and the fact that years ago huge problems eventually had to swap to slow disks or even tape.
- Status
- Not open for further replies.