Science/Engineering/Technical forums

I like your "poor man's DRO" made out of those digital calipers. I did spring for a "for real" DRO for my milling machine, from dropros.com. The lathe I get by with the traditional dials.

Yeah, the calipers thing is working out well - mounting one end via magnet on the ways turns out to be good, other than the mess with magnets and steel incidental razor blades. It's already saved the caliper head a couple times by sliding when travel is exceeded; it's a one foot range thing.
Just generally improving things in the shop - I got rid of 15 mils play in the threading simply by putting a real thrust bearing in the end of the drive screw....stuff like that.

Make the place more worth heating and hanging out in during upcoming winter - in between trips to the hospitals.

Now, would anyone in their right mind build an adapter so they could replace a camlock chuck with an R8 collet? Maybe I'm not the sharpest fork in the toaster, but I did, because I have them and it'd be a mere few hundred extra bucks to get some other type (C or ER) collets and ??? mounting system for this machine. The inside of the spindle starts at 1.725 and tapers. I cut the taper on the adapter a little faster so it can swivel ever so slightly, though it centers on a fast 60 deg taper at the front. This way, by moving the back end of the draw bolt (Which I'm still putting together, it's nicely threaded and I used thread wires for the first time) if I want, or centering it with a bushing on the back if I want. Because I can, I don't have an obvious use for that, but was off about a degree on that taper that I had no way to measure internally back in that deep. Bug == feature? Yes, I know the oil level is low - it leaks, slowly, and I put really thick stuff in there so the gears stay wet even so. It's a pain, I haven't found the leak yet, but the stuff does show up on the floor, eventually.

All done. Indicates around .0003" off center. Probably about the thickness of that spot of corrosion on the collet...perhaps just a kiss with steel wool in the right spot. It was all done in one chucking, so I know my adapter is concentric...maybe the spindle itself? I'll have to test! I had to use a 1/2" milling bit to indicate, because no 1/2" steel stock I have is all that round. I could have (and eventually will) try with a 3/8", as I have a project involving that size of centerless ground O1 tool steel - that broach to make holes for screw-in panel mount BNC connectors. The one I made from hardware store steel worked fine for a few holes (my case harden was both thin and sloppy), but...I'm going to make a good one and by golly, solve that issue for good and all. The female BNC's I have in bulk are a not-quite 3/8" hole but with two flats ~ 5/16" apart to keep them from turning. Of course, they aren't exact size, so I'll mic 10 or so before finalizing the broach dimensions...

I like BNC's as connectors for data aq and the like, but doggone it, those ones that unscrew on each use I can do without, and I can also do without the panel mount ones that need 4 tiny accurate holes as well. It's one of those "I'm on a mission" sort of things - I've suffered long enough.

O think what I'm going to do to get my measurement on ion reactivity (Whatever you want to call what happens when a bunch go by the electrode) - is going to be something like:
Make a pentode output stage with the highest possible output impedance, driven off a high output sweep generator I have. I'll make another of those nifty high bandwidth HV probes I did awhile back, which will be the only load other than the tank stuff.
As it sweeps, I can take data on amplitude and phase, and do that with and without ions, as well as compare with just a capacitive load that matches the apparatus if I should want to (which would ignore nonlinear things and parasitic inductances, but those shouldn't really be very present).

I can calibrate the output impedance which will no doubt change with both frequency and amplitude with various other loads to see how much attenuation and phase shift those cause.
Some math (gotta love computers) and bingo - I should be able to see at least a strong hint of what I'm looking for.

Bill brought me one of these a few years ago, and I've been guarding it: https://www.radiomuseum.org/tubes/tube_715c.html
There's even data for it. These were used for pulsing big radar stuff before other ways were found to get that job done.


Any comments by the RF competent welcomed...the expected effect is megohm level reactance over most of the range of interest, but could go to lower if I get geometries right and have some luck...

Noodling the load Z of a plasma. At the pressures that seem best for fusion, expect a 4 megohm DC resistance paralleled with capacitance, I'm pretty sure it's not going to be an inductive load.

That's pretty close to what I've measured at "good for Farnsworth" speeds and feeds - around 2 megs in my lashup - close enough given all the other variables ( like my oddball ion source stuff).
That's with DC and the usual stuff going on we all know about. Or I think we all know that most ions in a Farnsworth are generated by electrons that were emitted as secondaries from the grid via other ions hitting it, and are generated near the bottom of the potential well, which is one reason it takes what it takes to get noticeable fusion at all - most of what we put in is simply wasted - as the emitted electrons speed up far from the grid, the ionization cross section goes down - the peak is at ~~~ 100ev.

What I'm looking for here, though, is to start very low pressure - all ions generated with an ion generator starting around e-5 or less mbar, and working up, looking at reactance.
(that huge magnet based ionizer I made for this is super effective even at very low pressures)
While you can't accelerate a neutral plasma with an E field...you CAN accelerate a polarized one in an E field with a gradient. (similar to an iron bar in an H field).

We know you can manage to trap ions in a purely E field setup - quadrupole mass specs work this way, keeping some, and rejecting other e/m ratios. There are also just plain traps, and we know they work and we know the math.
However, at higher pressures, that would be better for fusion, this plan - unmodified - is too simple to work. Once you have a lot of ions, you also have electrons, even if you don't intentionally emit them - the charge yanks them off the walls - to paraphrase a bible saying "electrons you always have with you". In fact, a huge fraction of the losses in a Farnsworth setup are from electrons emitted from the grid via ion impacts there, which are simply accelerated to the tank walls, which heat up where they hit - a thermal camera even shows that they are hitting where the "rays" of the poisser end.

Here, I'm on a little bit of a different track. Ok, we're going to have a plasma (probably not even neutral, but net negative as we have in a Farnsworth setup) - maybe it's not even full ionization, but more like rydberg atoms - I don't care for the moment. Anything polarized will do for what I'm looking at. If we can get all the + facing one way, and the minus end facing the other, we have the electrostatic equivalent of bar magnets...

Those can be accelerated and managed in similar fashion to ions, but at far higher density (at least in my theory). You need more than a simple E field, you need a gradient, like you would
with iron bars or magnets, so what, we already have one via our geometry.

But like any ion trap, to get the to and fro motion such that things repeatedly go through the middle, and turn around and come back before they hit the tank walls - we need to get our drive
right. To do that, since there is no math out there yet for this impure situation - we need to measure how what we're putting in results in motion of ions, which will show up as a reactance.
Starting at super low pressures and densities, where conventional ion traps work, we can find a solution to the known equations (mathieu flavor stuff). Now let's sweep the pressure/density parameter - which will begin to affect things dramatically around e-4 mbar if my mass specs are any indication - and look at how we might add more predictive terms to the existing math that make some sense with what we know is going on (or will learn) - due to adding more stuff in there. If we can get to some sort of better prediction than "ions start to all get lost and the resonance smears way out" which is about all you can tell from a mass spec plot - perhaps my "incident" really did happen. And maybe I can reproduce it. I think it was likely real.
Reproducing it would be really neat. I'm going to go for it, health permitting. I sure have picked up lots of cancers...perhaps there's a relation there. Whatever, it seems worth doing, but to
get anywhere faster than "just try everything and every guess" = which hasn't been working out, perhaps a real parameter sweep with meaningful measurements would be a better path than what's been tried so far.

Thus I need to make what amounts to a VNA for yea, multi-meg level reactances, with the additional requirement that I can do this at varying AC+DC voltages such that I can start in a place
where the simpler known math surely works, to account for any scaling factors from our somewhat different geometry and so on. And then sweep up the ion density (becoming more of a plasma density as you just can't have that much single polarity charge flying around without it picking up electrons via it's own attraction for them - CERN has this issue with their beams already at only some 10's of ma).

I expect this to be a bit difficult, since at the roughly predicted speeds and feeds - kv and mhz - just getting the signal into the tank is going to result in a lot of other reactance due to the air around things, or .... that's plasma ball in air kinds of levels and it's going to take some doing.

At any rate, I'm starting with building a custom data aq for that because it's comparatively easy and it's my expertise anyway - something to do while I try to figure out how to make the data I want to aq. I've got some pretty cool stuff going as prototype, and am now in the process of building the "real" hardware for that so I can have a nice thing to finish the code for.

You probably picked up cancers like I have, just by virtue of being old (like I am) and having crummy genetic code for resisting cancer (like I do). This November one is getting cut off my head behind my ear. As long as we catch them early enough, we can keep going.

There's no telling about causes of cancer, to be sure. I don't have the family history of some of this, but indeed, I did a lot of things known to cause DNA damage "in the day" so...it's just a factor in a guess - I did use qualifiers on purpose, and it's not the only factor in that guess. I'm hopeful I will continue to be "hard to kill" as has been the case so far, I'm not dwelling on it much, though the medical system scares me a little bit.

Example is the pulmo guy wanting to "go right in" to look at the thing in the lung, ordering a further PET scan to get a better look at it - and in the process having to have a heart monitor put on by the cardiologist guy (due to the main complaint) removed to get the picture, which then shows other issues down in my gut - so far, it's like astronomy - every new observational capability has shown new stuff, and it's all bad (in this field, they're not looking for good stuff)...The specialists all not talking (despite in some cases sharing the same office suite - literally the same receptionists) and all calling all this stuff suddenly a major issue when I went in for chest pain in the first place...it's a little daunting and I'm having to be my own patient advocate, using a little more ... forcefulness ... than I think should be required, but such is the deal one sometimes gets in life, and I'm not the first, I'm sure. But yeah, having the cardio guy tell me to tell the pulmo guy to wait because what he wants to do might kill me on the table due to the heart issues - sigh.
At least the stuff cut off my skin has not returned, as far as I know (I do have family history on that one).

You are certainly correct in that simple old age is a factor, and IMO, not completely a bad one. I say that as I've watched progress in the field my whole life, and having all this stuff show bad all of a sudden not that many years ago would be a certain, short term, death sentence - now a lot of people survive these things. Finding them early is better, but wasn't enough most of the time in the past. So - living long enough for medicine to progress has been a good thing!

Meanwhile, and considerably more fun - I've spent the past few good days improving my lathe. A feature I didn't pay for or get (horrible fright ... not a surprise ) was adjustable gibs on the main saddle. This showed up as the saddle kicking back toward the operator a bit, sliding up on the inverted V part of the ways, when sudden pressure was applied back toward the operator by the tool entering the cut - very visible at the start of threading, or in the current case, attempting to make a broach out of semi-annealed O1 tool steel. I have some video of it, it was visible.
Here,you can see the whole main saddle jump back when the tool enters the cut while threading a barrel. No blood, no foul on this project, but this is the one I got video of, so...here it is.
https://youtu.be/RmNKIjr6LG8?t=54
That jump is what I think I've just fixed by modifying the saddle.

I downloaded a bunch of manuals for this generic copy of South bend item - Everyone sells a version it seems, but all others have some way of holding the saddle down to the ways and not allowing this shift under load that causes the saddle to "climb the V" slope.
So I tore the thing down - a challenge as none of the manuals I did manage to find actually show enough detail to make it obvious how to get the thing apart without bending stuff or really taking it to tiny pieces....
And added some stuff to prevent this motion. Where other versions had an adjustment - say "4 setscrews under the saddle bearing on the bottoms of the ways" kind of thing - here, those were omitted, and there was 20+ mils of motion allowed in the vertical plane on the front V, and 100 mils of horizontal motion towards the operator allowed by lack of any stop other than "size" on the back L of the bed ways. Here they just left a gap when they processed the casting so things wouldn't bind, and were done with it. Now I've added pads and setscrews to set the clearances on both front and back, all adjustable in case it seems I'm over-controlling things - and in limited tests (just got this done some hours ago) finding that they did do a decent job machining the basic ways, and with all the adjustments now added, and tight, the main saddle moves nicely with no binding...and it's probably going to be as if I'd bought the roughly thousand dollar more expensive version from Jet now. I used brass where things slide on cast iron and there are now neat little tracks of brass deposited from a few back and forth passes testing the adjustments, and it feels "right" so I'm hopeful - I'll get some real testing in on some now ruined stock - O1 can be hard to cut - it wants to rub until a lot of force is applied, then the tool digs in - before I wreck any more of that stuff - the same operations in 12l14 or even 1018 were no problem. The poor man's DRO I'd added to the Z axis of the mill made it possible to create nice accurate pockets for some bits of brass I had left over from an old diffusion pump project...seems the shop is getting to critical mass on tools and stock, and that's happy-making. I plan to put a link here to a post I have yet to make on all that - I got pictures, it really happened...Lathe improvement project: http://www.coultersmithing.com/forums/v ... =35&t=1175.

Still have that new-to-me old man issue with having to get up in the night - dunno if this gets posting times correct, but for me, it's 4:40 AM after having slept since midnight...waiting to get tired again so I can go back to bed. Gheesh. Seems better than adding another drug or drugs to the lengthening list, just wait awhile and get sleepy again.

Yeah, if I get up in the middle of the night and can't get back to sleep, I'll go down to the office and do boring paperwork I've been putting off. That gets me sleepy again real quick.

Not to pile on, but that's an old man problem I share with you, and can commiserate with.

These days I sleep in three hour watches, but I'm never quite asleep. A dog can bark, a fan in a computer can change pitch...

Bah.

Hey, if we tough old birds can't commiserate a little and share tips - what's the point? No piling on detected.
I did this on a lousy day after a tough last night - on a raspberry pi running "twister" which is a fancy respin of debian to support retro pi, dosbox and old time gaming kinds of stuff - which I have yet to use for that, but it looks cool and is generally smooth. I overclocked it to 2 ghz (with a cooling tower) and use it on low energy days...a lot of things don't need the monster I recently built, like developing this toy for machine shop use.

This just calculates speeds for me. Give it any two of diameter, feet/min, rpm, and it'll give you the third one.
http://www.coultersmithing.com/forums/v ... 7027#p7027

Looks like this: