Doug's latest feedthrough work

Things at the limits.

Doug's latest feedthrough work

Postby Rex Allers » Mon Jan 30, 2017 9:30 pm

In the "keep alive" thread Doug posted about his latest work on HV feed-through. This post was intended as a reply to that (hence the intro stuff) but that thread is locked and most of this probably fits better in this section.
Reference to that post: viewtopic.php?f=42&t=980#p5839
___________________________
Doug,

Sorry to hear you had some health issues but very good to learn you are OK now.

You mentioned missing recent posts. Maybe you know this, but on the BBS main page, near the top, you can click on "View active topics" to see all the recent posts. Default period is the last week but you can adjust it up to a year if one hasn't visited the forum in a while. That's the method I always use when checking a phpBB forum.

It was good to see the description of your latest HV feed-through. I'm very slowly working toward building a fusor. The chamber I plan to use is cylindrical, so I plan to employ a lot of the techniques you developed: grid style, and feed-through being the biggies, so far.

I wonder if I might trouble you with a few more detailed question about this feed-through.

-- You said the Cu rod was 14" x 12.75". I assume that was a typo and should have been 1/4" x 12.75".

-- On the aluminum end cap, about how long did you make the section that fits inside the glass tube?

-- I think the BN rod extends past the end of the glass tube. If so, about how far past the end of the pyrex?

-- (You did mention the Cu rod extending past the BN by 1/4". Just noting that I saw that.)

-- I'm curious if you did anything on the Cu / aluminum connection to allow for venting to aid any slow virtual leaks when evacuated? Maybe that's just not really somthing worth worrying about.

-- When you have the feed-through in its operating position, about how far does the glass project into the chamber beyond the inside wall? And, accordingly, about how much glass between the coupler and the aluminum end cap on the outside?

I had already been planning to make a feed-through based on your earlier work. I decided to go with 1.25" OD borosilicate tubing to keep the BN diameter requirement to 1". The ID of the glass is on the order of 0.93". So I got the tubing and a CF/compression fitting for 1.25". I haven't had the guts to order BN yet.

Being reminded, I just checked McMaster's BN page again. The 1" dia x 12" rod is $368 but I see they have 1" square x 12" for $302 -- $66 less. They both require lathe work. The effort and scaryness of knocking the corners off is probably justified at $66 worth. (I'm not paying myself per hour.)

-- I've never tried to cut big thick pyrex tubing and the pieces I've got need to be cut. Do you do that in a lathe with a diamond wheel? Any tips on technique?

Thanks for all the sharing you have been doing over the years. As you can see, I'm planning to steal extensively from your work.
Rex Allers
 
Posts: 28
Joined: Mon Sep 05, 2016 8:22 am
Location: San Jose CA

Re: Doug's latest feedthrough work

Postby Rex Allers » Mon Jan 30, 2017 9:37 pm

While I was typing, I see Doug added another post to that "keep alive" thread that answered a few of my questions.

Hope this doesn't scatter the information around too much on the forum.
Rex Allers
 
Posts: 28
Joined: Mon Sep 05, 2016 8:22 am
Location: San Jose CA

Re: Doug's latest feedthrough work

Postby Doug Coulter » Tue Jan 31, 2017 10:31 am

Rex, to use the unread posts feature you have to log in. I was busy. When you see the same name as last poster on a forum every time you check it *usually* means nothing's happened.

Yeah, I fixed some typos over there. I have the Cu sticking out around 1/4" which is completely arbitrary, zero might be better. But I will note that the back end of the grid must be spaced out from the end, at least if the end of the BN is square, as experiments here have shown (note back end of grid has its own extension). The optimal amount is unknown at this time, as is the ideal shape for the protruding end of the BN (likely they interact) or its length. I try to keep system openings to a minimum here - this will be the first time that's been needed in around 2 years. It's real clean in there.

On earlier designs, yes, I did the Kurt Lesker thing and had paths to help move trapped gas out of threads etc. In real life, with a reasonably tight fit on the BN, its hygroscopicity is the controlling factor,
(unless you want to buy the next grade or two up at 5x that price, and yes, I asked the guys McMaster gets it from) along with the slow diffusion of the internal water down its length and across its radius. You're stuck with that and you might as well just wait. It's a week or two to perfect. You can run some in the meantime, which speeds all that up somewhat (heat).
Note that the upside of all this is that it takes awhile for the BN to re-absorb much water. So if you bring the tank up with some inert gas and work quick, it's not going to make you wait a long time again.

I wouldn't even try turning a square one round, but that's me and my only 10 years or so of lathe work. This stuff is soft and weak, and an interrupted cut...just not gonna go there. I may not be paying myself, but I am paying for the stuff, at least the part Bill doesn't pay for. Heck, the drill for the center was serious bucks FWIW - foot long drills that are straight are not cheap either. The good news is this stuff puts near-zero wear on tooling. It's so slick it's hard to clamp.

As to protrusion, I think it's going to depend on scales of things, the idea mainly being to keep the glass out of danger, and to shape the net field as desired, which is kinda hard to instrument - or for that matter, know exactly what's desired (especially when RF is involved along with the dielectric constants of various things - including the plasma at its existing density and position moment by moment). Feel free to post well documented results and give some back, eh?

The exposed end of the BN seems to "find a happy place" as regards charge buildup on the surface. Whether that's a happy place for max fusion is yet to be determined. But if your gear fails, you can't do ANY fusion...gotta start somewhere.

Note that this ion grid, while real similar to the main one (except it's a lot smaller) is out in the big 14" dia x 26" tall tank, not in the middle or anything like that, though I usually try to point one of the "beams" towards the main grid in the 6"x6" sidearm. The point of this one is that it's the most efficient (by a huge factor) ion source yet discovered, and it works due to the Paschen's law effect of a lower voltage being able to maintain a discharge in the larger dimensions of the tank. For an ion source that works at the record lowest pressure, the microwave one I've described is better, but it not only uses around a kilowatt, but also makes only microamps worth of ions. This one makes tens of milliamps for under 200 watts, and provides a control electrode. It's a fairly big deal.

You might want to just generally do some homework on my youtube page.
Here's a start: https://www.youtube.com/watch?v=Us7Ma230huY which shows the ion grid as a control grid in a slightly different lashup.
Or:
https://www.youtube.com/watch?v=HiDUJajLBRw
Showing what happens when there's not a great fit between the BN and the glass - lengthwise arcs. Which is why the piece I'm using now was in the scrap box, it was too small for 1.5" pyrex tubing we use for the main FT.
For each of those videos there are maybe 2-3 others talking about the details. There ain't no free in education, though I paid most of the rent, you gotta do the rest. A lot are linked here *someplace* but organization that makes sense is an ongoing struggle and hindsight is hard to apply once an edifice is in place...

Somewhere there is also a video on how I cut glass. I made a toolpost grinder for my lathe, though my first one was nothing but an adapter for a dremel tool, which worked "well enough".
It's not as necessary for just cutting pyrex - the old score and snap seems to work up to around 3/4" pretty well....but if you want to diamond grind chamfer the ends, you need something like a grinder (use water as the lube/coolant). Annealing the glass is a good move, somewhere I have profiles (Kohl on materials for vacuum) on how to do this. Else it may shatter in use, and with a turbo pump, that's bad ju-ju. I use diamond tools from "Horrible Fright" - it's about the one thing they don't mess up too badly, the price is right. Sometimes it takes a little work to make the wheel or chamfer bit run true....

As for scattering info...organization is, as I mentioned, a bear. This could have originally gone into either HV or Vacuum, and surely not water cooler. Sigh,,,I used it as kind of "yes, I'm alive and this kinda proves it's really me posting". Not a great excuse...or the best example, but "other concerns applied".
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Doug's latest feedthrough work

Postby Doug Coulter » Tue Jan 31, 2017 12:52 pm

I suggest measuring the ID of your pyrex tube first. My thick-wall 1.5" OD was 1.115" inside. 1" BN will not work, which is why I even had this piece left over for the smaller one. Reading is Fundamental ;~) Or viewing. See second video above for what happens if you try that.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Doug's latest feedthrough work

Postby Doug Coulter » Tue Jan 31, 2017 3:23 pm

OK, here's the latest on this one (ion grid feed through). I changed the "stick out" of the copper part to near-zero, since it "didn't look right" and we know that things that aren't the grid sitting at HV in there just attract ions and waste power (as well as sputtering onto insulators). Since this grid design has plenty of offset already, and this rod is "tunable" by threading in and out of the cap...this looked right. The back half of the grid is bad enough, and I've thought I might cut vees between the rods, but you know, the real thing that gets hot is the center screw where the axial beam hits. In this case that's stainless steel, not perhaps the best choice, but it won't weld to the copper. I've used Ti, but then you have even less thermal conductivity to take away heat. It's a tough spot. At least the graphite is a good black body and radiates heat really well by itself.

While I did round over the corners and polish the end cap, this isn't the full corona ball treatment (it's not long enough for one thing). This will be at relatively low (for me) voltages and inside a pipe inside a pipe inside copper mesh that's grounded. It should do, we'll find out, of course.

The next step is to clean it as perfectly as I can manage (large ultrasonic cleaner for the not-BN parts), and put it together for real. Should I need it, "*Doug's magic vacuum grease" will lube the end cap O rings, but I may not need it. One definitely does not grease the O ring in the tubing coupler! It has enough creep issues already - and here I show an end view of the fancy clamp for the glass, which also centers the outermost insulating pipe and holds some of the weight.

*Doug's magic vacuum grease is a mix of beeswax (pure, from McMaster) and some low vapor pressure pump oil (Diffoil-20). Mostly wax, it's a solid at room temperature. Comes in handy now and then.
Doug's magic sealant is often the blue permatex gasket maker - this is a non-acetic acid curing silicone and has very low vapor pressure. If you're more patient and don't need this amount of flex, elmer's glue is even better(!). Sure, there are some epoxies *nearly* as good, but it's almost impossible to mix them perfectly enough for the results to not be horrible and outgas for months.

*Credit John Futter for that elmers tip. John also mentions that you should remove the BN from your tooling at first opportunity, and he's right. If on your lathe ways, your tailstock will soon lose the will to clamp. This stuff is really slippery, and becomes more so over time. It makes graphite look like silicon carbide.

Here's what it looks like now - that rightmost O ring just cushions the glass from the end cap as nothing is ever a perfect force distribution as is. Not strictly required, but this isn't a fun place to fix if things chip or break.
Fitup.JPG
Will be cleaned and assembled and installed next.


Glass treatment info here.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Doug's latest feedthrough work

Postby Rex Allers » Wed Feb 01, 2017 5:26 pm

Doug, you can use the "View active topics" link on the main Board Index page without logging on to the board. When checking here, and in the Fusor.net board (both phpBB) I use that to see the recent activity. I never log on unless I want to make a post.

Thanks for the replies. The additions you made in the "keep alive" as I was writing my post starting this thread cleared up a lot of my questions. The addition stuff here is great too. I have looked at your videos over time. I guess I forgot that one of yours included cutting glass tubing. I think I've seen several showing that general diamond disk method from different youtube posters and forgot you were one of them.

I have a better quality "dremel" (a Proxxon) that I made a holder for to fit in my lathe toolpost and some cheap ~2" diamond disks for it. I haven't tried cutting thick tubing yet. I did try cutting the end off a test tube and that chipped a bit at the end of the cut, but hopefully thicker glass will work better. But then, I think the Proxxon with the 2" wheel might not have enough clearance to cut in the middle of a heavy-wall tube. I also got some cheap 4" diamond disks that fit in my small angle grinder. I probably should come up with a stable toolpost holder for that and see how that works. So far, I haven't justified the cost or effort of a proper tool post grinder.

You said I should measure the ID of my glass tubing. I did. I mentioned the ID in my previous post. I went with 1.25 OD and it has about .93 ID; specifically the ID was about .925 min and .935 max at the ends where I could measure. There's a little bit of an internal lip at the very end, but I plan to grind that off before using. I picked the 1.25" because I thought it would be big enough for any voltage I'm likely to put through it and because the ID was less than 1" so BN rod would be a good bit cheaper than if I needed 1.25" BN.

You are probably right about trying to turn square BN to round. I think I would do it by chucking the BN deep in the headstock with a 4-jaw and cutting round with light cuts and many iterations of pulling the BN further out, but that would be tedious. Also I need to end up with about .93" so not too much slop from the 1" requiring lots of fiddling to keep it centered in a 4-jaw chuck. I guess just chucking BN for turning is probably a delicate operation, not to crush it. So, probably not worth the hassle and potential disaster. In fact I think I should do some tests on my lathe and make sure its not turning tapers. I don't usually turn long things.

I think I've now worked out all the lengths of pieces needed to make a feedthrough of my own. I picked up that you now have the Cu rod about flush with the end of the BN. And I also think I have accumulated materials and tools to make grids like yours.

Pretty sure I saw it here before, but good to see again the reminder to cover, or at least clean the ways while turning BN to keep from over-lubing them.

On the o-ring grooves, are you getting the smooth surface required just with a nicely shaped and sharpened HHS tool bit? (No polishing or something like emory paper to finish?) You didn't mention anything like that so I assume just the tool makes a clean-enough cut in the aluminum.

You mentioned pumping out the moisture and anything else over about a two week period. Vacuum is pretty new to me. Did you run the pumps continuously over that time? Seems like a challenging amount of 24-7 power from you solar electricity system.

The style of grid you came up with in a cylindrical chamber always made intuitive sense to me so I've been heading in that direction. I've got parts for a cylindrical chamber that will be a little bigger than most amateur IEC-style fusors but I don't know if it is big enough to support something like what I think you are doing with the second control grid. Anyway, I'm sure my first step should be just to do fusion with a single grid and learn to control and monitor things before attempting anything more complex.

I'm becoming pretty slow at executing things that I start at this stage of my life, but I intend to get there eventually. I do keep slowly moving forward. I'll post status later when I get to the point of actually having something working.

Thanks again for sharing so many details. Hope you continue to make progress and I look forward to hearing any new reports.
Rex Allers
 
Posts: 28
Joined: Mon Sep 05, 2016 8:22 am
Location: San Jose CA

Re: Doug's latest feedthrough work

Postby Doug Coulter » Wed Feb 01, 2017 10:56 pm

I get slight chipping on the thick pyrex, but just go slow and use plenty of water. I run the lathe at 50 rpm and the grinder "fast" - around 10k rpm or so. I follow the cutting with some horrible fright 60 degree diamond chamfer bits in and outside, moving them around a bit to reduce the grooving (leave some flat end). Then flame polish on the lathe, letting it cool while going maybe 200 rpm for some time. Don't go overboard on the flame polishing and reduce the ID, the idea is just to smooth the grooves in the chamfering and make it easier to slide an O ring past it without nicking it up.

It's rare I get a good enough cut on the hand ground tool to just use as is, even though the scratches are in the "right" direction. Just prying the o rings off while I'm testing to see if the fit is just right scratches it some. I polish the whole piece on a normal wheel with medium coarse polish (the grey stuff), then with a little wheel on the dremel in the grooves till it looks right. Judgement call...Maybe if I ground better tools and had a better lathe, but I don't and since I don't do this often, it's worth just getting it right by hand when required.

Which is actually a big point of this exercise. After around 6 *years* of trying stuff that "should have worked" - I have this design now and it finally just plain works.
It's an extremely big deal and a prerequisite to all other progress. Not only do all the other designs fail pretty quick, before that they inject little sputs and arcs and mess up your measurements and adjustments on the way to just plain failing in some mode. These just work.

We'll see about the Cu extension length. It's adjustable on this one, and I chamfered the end of the BN too this time. (always trying new stuff). On the one known to work great, it's about 1/4" and the BN is flat on the end, and much larger in diameter than the grid, FWIW.

McMaster sells some long drills for boring these. That's really the big challenge, getting a straight hole through the center at both ends. I often drill partway from either end before going all the way through and letting the drill enlarge the slightly off holes a little in the middle so a rod will slide in easily. This stuff doesn't wear tools...so that bit will last as long as you keep the rust off.

I don't worry too much about crushing it in the chuck, but I suppose you could. More danger is that you have it slip - and the chuck cuts it...and it likes to slip, so you just keep an eye on it and go slow. It'd be pretty easy to break it in half if you weren't careful and put too much side force in the middle.

I've been running my turbo far past it's expected hour life, now around 6 years, more or less continuously, though not always at full speed. I have the backing pump rigged off a solid state relay to kick on when the turbo draws over some setpoint of wattage to keep to speed. Outgassing, being slow once the first couple hours are done, doesn't need full speed on the turbo or a real sensitive setpoint on the backup pump...you just have to do that by ear. If you don't have an oil free backup pump and check valving, you have to be real careful with that to keep oil from being sucked backward into the system.
It IS a pain, one reason I don't put gassy things in there much - or even open the door often, and when I do, I use one of argon, nitrogen, welding gas mix, helium, to bring it up to 1 bar, work fast, then back to vacuum. This saves endless waiting (vacuum is all about patience when you get to the -7 or even -9 millibar). Every minute in shop air is an hour or more getting rid of that junk. Anything that puts heat/energy into the system helps that happen faster. Running the fusor in some clean dry gas - argon say if you don't want to waste D - speeds things up lots. BN is particularly slow as the water has to come out the long way and most of it you don't get very hot. Maybe you can find a reference on "molecular vs viscous flow" somewhere, Kurt J Lesker used to discuss this really well in their used-to-be-free catalog, most of which is just online now AFAIK. TL;DR - once things get into the molecular flow range (long mean free path) vacuum doesn't "suck" - you are waiting for particles to find their way into the pump inlet randomly - short fat and shiny is the rule there. And making them fly around more via heat or whatever speeds all that up enormously. (the binder in the BN is where that comes from, they ship it sealed with desiccant but by the time you get it into the tank it's too late)

FWIW, that turbo draws only about 20-30w (when the vacuum is good) so I just keep it going. The power draw is a very good indicator of how good the vacuum is at the moment, and that's why hooking it up so the backing pump, which is huge drain (About 1 kw) only runs a few seconds every few minutes.

Vacuum will teach you patience and how hard is is to find tiny leaks, and for that matter tell whether it's one of those or just something giving off gas. Just a bit of rust on something or the wrong metal in there will look like a leak. Get a good gage! We like PKR-251 which is a combined pirani and ion gage, there are a lot of them out there.

Yes, our second grid (which is what I'm doing here) takes advantage of something very odd in my setup - that tank we got surplus. You should do OK with just one in a big cylinder (we did), and if it's not a standard size, you'll have a bit of fun finding out the right diameter and length of the grid that works best there, and even I don't claim to have filled up that search space or even close. In the 6" world, going smaller - down to 3/4" from 1.25" was a huge improvement, and I made a 1/2" one (with the same length/diameter ratios) but haven't tried it yet. 8 rods has been working for me, but I need to make a new indexer to go to 6 for the smaller stuff.

You would think you could just make it any length and all the beams be sheets, but some experiments here show that making say, 50% longer just draws more power and makes more heat, but not more fusion...haven't tried the entire possible space yet on that one. I still wind up with a "kernel" poisser in the exact center in all axes on mine. When I make them too long, they just get hotter, waste more power, and depending on what they're made of, might melt and sag. Now I'm using tungsten (tig rods) and graphite (mcmaster) and that's best so far.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Doug's latest feedthrough work

Postby Rex Allers » Thu Feb 02, 2017 2:58 am

Lots of good real-world details in that description. I think a lot of that will give me a good feel when I get to starting up. Curious what big, power hungry backing pump you are using on your big chamber?

I think I've followed most of the evolution in your grid design. As I said, I like what you've done. It makes sense and has proven itself in the results. I was planning to start with about 3/4" diameter.
My chamber is 7.8" ID and about 16" long. First step is just to do some fusion and I probably have lots of things to do before I even get to trying. If I get there, I was thinking down the road, maybe I could put a smaller tube (4" dia x 6 to 8" long?) inside for the shell around the main grid and try adding a control grid in the full-size half of the tank. Not sure if my tank is big enough for that splitting to work at all but maybe something along those lines is worth a try.

I have a small turbo I got pretty cheap. I've only spun it up once with the input capped off and no vacuum gauge on it. From the drive monitor, it spun up to full speed but it made a lot of whine noise. Not sure if it is healthy. I need to do more testing but I might need to shop for a better and bigger high-vac pump too.

Unrelated...
I'm not sure about this technology stuff. It can't be trusted. I'm in San Jose. part of tech mecca "Silicon Valley". My internet connection is a screaming 5 Mb. Yesterday my ISP was down for about 6 hrs., a DoS attack, I think. Today my power went down for 6 hrs because it was fairly windy this afternoon. That was the third power failure in two weeks.
Arrgh!
Rex Allers
 
Posts: 28
Joined: Mon Sep 05, 2016 8:22 am
Location: San Jose CA

Re: Doug's latest feedthrough work

Postby Doug Coulter » Thu Feb 02, 2017 11:00 am

Here it is, more or less ready to install.
I had to make up another batch of "grease" as I temporarily lost my other jar. About 60:20 beeswax:diffoil.
I didn't wind up using that last cushioning O ring as my "pull it out of my butt" machining put the second seal/locating O ring too far back, and the chamfer in the glass let the 3rd o ring eject the glass from over the second. Probably don't need two there anymore since the BN supports things fine, but habit...And half the chance of a leak I'd have to fix.
Note that the seal O rings are pretty squished in there tight...that's one reason to handle the glass well, and use a little lube - maybe 5 milligrams worth (which should also help the O ring to Al seal). In theory, none would be better, but this seems to work in all my setups. No grease on the tubing coupler (other than lubing the threads for the nut).
IonGrid.JPG
Ready to install, I hope.


The whole point of this board is to share practical useful knowledge. So when you get something new, the idea is you share back, same idea as GPL.
And give credit where due, that's why I insist on real names here.

I don't trust that government owning tech biz either, or not much. Why am I off the grid, again? At least if my power fails, I know whose fault it was. So far, I got one heck of a lot of 9's uptime - many more than they do. And you live on the left side of the country - worst case for a lot of things (Frank Zappa's "Fresh Flakes" comes to mind). Jerry seems to manage some way (and really well at that), but it seems the exception that the real competent doers are over there these days.

I'll see if I can answer some of your other questions in forums where the answers fit better and link them back here.
Vacuum technique for this sort of thing
Thoughts on sizing and gas.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Doug's latest feedthrough work

Postby Doug Coulter » Thu Feb 02, 2017 3:20 pm

Installed! Looking good. More pix and post mortem of the old one in a link.
Yes, I've already run a discharge on this and the pressure's coming down nicely.
Installed.JPG
Slipped right in through the side door.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Next

Return to High Voltage, High Power

Who is online

Users browsing this forum: No registered users and 6 guests

cron