Langmuir Probe Creation and Design

For posts specifically relating to fusor design, construction, and operation.
bk8509a
Posts: 40
Joined: Sat Oct 31, 2009 8:47 pm
Real name:

Re: Langmuir Probe Creation and Design

Post by bk8509a » Fri Apr 23, 2010 12:54 pm

Chris, the probe is a triple probe and only needs current if you want to determine electron density. I have no electron saturation current, ion saturation current, or any graphs that I need pick data off of. I get a straight number out, every half second.

I'm using equation 16 of the paper to do all my calcs, which is non-solvable with algebraic methods. I end up making LabView do Newton–Raphson on it.

An average data run is attached in Excel.

The article i'm using for my triple probe is attached. I couldn't post the pdf (1.3 MB)....

The theory is completely different that regular Langmuir probe theory.
Attachments
pdf4.png
pdf3.png
pdf2.png
pdf1.png
LangmuirRun.xlsx
(133.04 KiB) Downloaded 108 times

User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Re: Langmuir Probe Creation and Design

Post by Chris Bradley » Fri Apr 23, 2010 2:47 pm

Thanks for posting up the paper, though my computer was struggling with those png images. (png?!) Are there any more pages?

I converted to pdf so I can view it... maybe it'll also be useful for others in pdf... and will study on it further, later..

view.php?bn=fusor_files&bn=fusor_files&key=1272034039 (BROKEN LINK)

bk8509a
Posts: 40
Joined: Sat Oct 31, 2009 8:47 pm
Real name:

Re: Langmuir Probe Creation and Design

Post by bk8509a » Fri Apr 23, 2010 3:08 pm

Chris,

The whole paper is on the files forum now.

User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 7:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Langmuir Probe Creation and Design

Post by Doug Coulter » Fri Apr 23, 2010 6:01 pm

I've done a very tiny amount of probing myself, nothing so pretty yet though.

I would suggest a couple possible improvements though (though I may be talking out of the wrong orifice since I've not duped what you did ;~)

One is that I have some of that 4 bore ceramic, and it's electrically pretty leaky, particularly when warm. This will affect a number of things, including the electric field far from the tip -- which may distort your measurements a good bit. Quartz is many orders of magnitude better for this kind of thing -- but I doubt you find it in a 4 bore tube, so it'd have to be larger...dunno which effect would be most important here. I'm thinking more of conduction to the plasma outside the main tube along the sides, not so much wire to wire leakage, but both would be fairly high once warm for the tiny signals involved.

Next, I'd use something other than tungsten -- something with a higher work function that won't have big errors due to secondary emissions. Ti perhaps -- it's cheap, anyway (McMaster of course).

Third, nothing beats being able to move the thing while it's working. A wobble stick or a push-pull twist adapter with this on the end would teach you a ton in very little time. It's just not the same having to tear down the system to put it in a new place -- the conditions won't otherwise be the same, and you might actually be able to see any leakage field of this distorting the visible plasma, which itself will tell you a lot diagnostically. As in, is my probe changing the thing I'm trying to measure too much?

I am lucky to have both a wobble stick and a push-pull twist for 2.75 CF here and they've sure taught me a lot -- fast. Even with a dumb pinhole camera for charged particles and X rays there were some surprising results as the thing looked from different angles, which I am still digesting, but it is definitely nicer to have things to ponder than not to have a clue!

For probe kinds of things, I just run the wires down the motion stick to another feedthrough for the signals....in my case I can do this "far from the action" as my fusor is a cylinder running in a 6" ID sidearm off a 14" diameter main tank, which helps for things like that. Plenty of room for other junk in the main tank that doesn't affect the main fusor operation much if at all.

Just my $.02.
Why guess when you can know? Measure!

User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 11:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: Langmuir Probe Creation and Design

Post by Carl Willis » Fri Apr 23, 2010 6:22 pm

Hi Brian,

This is a very good experiment. I haven't had time to digest it fully, but it's a well-documented effort that adds a new technique to the community repertoire, and the results appear to be very consistent between runs. Consistency is a challenge with fusors. As to what it means, or might mean, I have to consider later if I might have any observations on that front. In the meantime, I appreciate your contribution here and keep up the good work.

-Carl
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277

DaveC
Posts: 2346
Joined: Sat Jun 30, 2001 5:13 am
Real name:

Re: Langmuir Probe Creation and Design

Post by DaveC » Sat Apr 24, 2010 12:26 am

Hi Brian -

Very interesting results, to say the least. Also thanks for posting the paper.

Do you have a schematic of your measurement setup? One question the hopped off the page of the paper, was spawned by the phrase " floating" , which I loosely translate as meaning "isolated", as in not connected to anything at ground potential.

While an easy concept in theory, my own experience has been that this can be hard to accomplish experimentally. With electrodes surrounded by plasma, it seems hard at first glance to be able to even say the apparatus could be floating.

Experimental process here is very important.

Nonetheless, the very consistency of your data, suggests something real is going on.

I'd be delighted to see how you were able to do this. It's a very nice technique.

Best regards,

Dave Cooper

User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Re: Langmuir Probe Creation and Design

Post by Chris Bradley » Sat Apr 24, 2010 8:56 am

Thanks for uploading the whole paper. Not read it in depth yet, but my main question(s) arising so far is the envelope of applicability of this technique. I wasn't quite sure if I read it right but it seemed that there was an argument given over why the probes should be close together, but does this suggest, perhaps, that there is a limited range of applicable plasma parameters for the technique?

In the case of the fusor, in the position you are holding these things, I'd imagine the charge density to be sufficiently low such that the size of the probe sheaths are comparable with the probe separation. In fact, they might be very much of the same order, such that your measurements are actually some function of sheath thickness up until the overall charge density increases [due to more gross input power] quicker than the electron temperature increases, and the sheath thickness might then decrease.

bk8509a
Posts: 40
Joined: Sat Oct 31, 2009 8:47 pm
Real name:

Re: Langmuir Probe Creation and Design

Post by bk8509a » Sat Apr 24, 2010 5:00 pm

In advance, sorry for the long post. I will try to comment on all the great advice and then I'll talk to you about the newest problem that has come up.

Dustin,

I think you might be on to something about how I can't detect the whizzing by electrons. I have new data to refute that though.

Doug,

I looked for Ti rods on McMaster, the smallest they get is .062, which is .02 too big. I could use wire and straighten it, or try to find rods in a tinier diameter.

I've also looked into controlled movement of the whole device. I bought a 3'' bellows last week that I will eventually attach. These are just the fixed intro runs.

Carl,

The more I look at this data, I start to think, "It consistently means something, something that I have no clue about". I appreciate your praise of the experiment, as this is my precursor to grad school at Wisconsin Madison. Hopefully I can thrive there.

Dave,

I'm going to have my senior thesis on this website soon which will give diagrams, schematics, and other facts such as construction and experimental procedure on here. These are all important, I agree.

Also, in plasma physics, "floating plasma potential" means just hanging out in the plasma and taking on any potential that the plasma takes on. You cant get a true floating "ground" in plasma.

Chris,

Applicability is a huge concern of mine, as you will see from my next few pieces of data. Also, I'm doing the calcs on the Debye length and finding out that the density would have to be extremely low in these areas, with these temperatures, in order to have the sheaths not cross. You make a valid point.

NEW DATA:

Alright, so yesterday I decided that it would be a good idea to test to higher voltages of Vd3, to see if the temperature converged. This meant I had to use a different power supply than I used for the first couple of measurements. I got some results that are making me question certain parts of the experiment.

For the first series of measurements I used a Instek GPS-3030DD:

http://www.metrictest.com/product_info. ... -3030DD(N)

to apply Vd3. For the second series I used a OLD (but functional and accurate) tube power supply the Kepco ABC 200M:

http://www.photomachining.com/inventory ... 3102-L.JPG

The good thing is that the data taken with the GPS-3030DD was consistent with previous trials. But when the same Vd3 was applied to the Kepco 200M, the graph was similar, but not the same! Which power supply is telling the truth? What could be going on? To be honest, I feel that the tube is correct. The reason for this is I expect a RAPID increase in electron temperature as voltage goes up, the Kepco gives me this. I'm assuming those bumps in the beginning of all the graphs are just due to interference with the grounded grid.

There is a wild discrepancy between the two power supplies here (See picture 1). I know some of you are experts on this, so I expect some insight about this.

The new power supply made some new curves, all which are consistent (See Pics 1-5).

The comparison of these curves is shown in picture 6. The problem is, the more I increase Vd3 the higher the electron temperature goes. My high voltage supply should be making the electrons get fast, not the probe's voltage!

ANALYZING THE DATA:

As you can see in pictures 6, the curves maintain the bumps that were shown with the old power supply, but they have a high tail end. The tail end makes me think that the Kepco Power supply is giving an accurate reading. These electrons should be zooming around at -20000 V.

CONCLUSIONS ABOUT THE PROBE THUSFAR:

I believe its working, but there is some underlying error with the equipment such as the power supply. How do I know this? See picture 8 and 9. The curves are exponential and on track with the only calibration curve in the Chen Paper. Its definitely working.

PROBLEMS TO BE SOLVED:
-Figure out which power supply is working correctly.
-Firgure out why there is a discrepancy between power supplies.
-Understand why the bumps appear in all taken data (outer grid lensing?)
-Determine why the electron temperature goes up as Vd3 goes up in all taken data.

NEXT STEP:

I'm sticking this probe in the center of the fusor next. The plasma should be more uniform, allowing me to make some deep conclusions about whats going on outside of the outer grid, and if my probe is working.

Sorry for the long post, it helps me figure out whats going wrong/right, and helps you help me.

-BK
Attachments
calibrate1.jpg
calibrate.jpg
Comparison.jpg
50Kepco.jpg
40Kepco.jpg
30Kepco.jpg
20Kepco.jpg
PowerSupplyDiscrepancy.jpg
10Kepco.jpg

User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 7:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Langmuir Probe Creation and Design

Post by Doug Coulter » Sat Apr 24, 2010 6:08 pm

FYI, McMaster calls it "wire" below that size, and they go down to .004". I use the .032" stuff to make grids sometimes, and some times also the 20 mil thick plate for endcaps on the cylinder grids. I use the pure (not deliberately alloyed) grades always for things like this, they sputter very little and live long.

McMaster is a little strange that way -- for large clear mica sheets, you don't search for mica, you look for "india glass" instead....this kind of thing abounds there.

Whether it's a big effect or not for this (it has been here for probes) the alumina 4 core they have has roughly 10e4 times the leakage current as quartz does at a given temperature. That's a big one. It's even worse than 7740 (or K-33) pyrex glass, which is the most conductive glass in common use.

Ceramic has shown me enough leakage to wonder about an in-tank preamp on homebrew ion gages, it's pretty bad even when there's a lot thicker piece and a longer leakage path. Then I got into quartz...no problems now. The tiny multi bore ceramics are sold for thermocouple uses, where a fairly large amount of conductivity doesn't make any difference. I am using that for type C thermocouples with no problems, for example.

Very small quartz tubing, single core, is available at www.quartz.com/quartz.html -- down to .5mm ID by 1 mm OD, though you probably would want something a bit more thickwall for better insulation/standoff voltages. At those sizes, a bundle would fit in a larger piece of tubing which would still be pretty small. I like this outfit, and anybody who wants to go a group purchase from them can contact me, as I need to build an order to get to their discount threshold. I'm using it primarily for feedthroughs (tubing coupler or just glue through a hole with Hysol), as it has better performance for fusors than the commercial ones -- by far, and it's not just conductivity for that -- its comparative less trouble of being reduced by hot H, low dielectric constant, and low RF losses all at once that make it superior to the ceramics.

Keep those results coming! I get bored with this place when we don't do real experiments.

I'm pondering the paper and your results now, as they might dovetail with some of mine nicely.

But I'll wait to understand before comment (at least this once ;~).
Why guess when you can know? Measure!

User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Re: Langmuir Probe Creation and Design

Post by Chris Bradley » Sat Apr 24, 2010 6:16 pm

Brian Kelleher wrote:
> Alright, so yesterday I decided that it would be a good idea to test to higher voltages of Vd3, to see if the temperature converged. This meant I had to use a different power supply than I used for the first couple of measurements. I got some results that are making me question certain parts of the experiment.

I can imagine a host of issue that might lead to such effects; switch mode power supply effects, internal impedance, current feedback into the power supply from the plasma [current seepage through the earth connections] etc..

Unless there are issues I'm not seeing here, I would suggest as an initial check you stick a very large capacitor across the supply outputs and measure V across the cap with a trusted volt meter.

Not sure I agree with you about the Kepco supply. Remember you are likely measuring the state of the background medium, not the state of the beams. As far as I would presume, that'd mean you're only gonna measure a real think 'slice' of energy, so to speak, off the top of the fast electrons hammering their way outwards. I would tend to trust the linearity of results from your Instek, which generally extrapolate back to 0eV=0V drive (a know point!) from the higher values. Looks to me that, at that physical location, your readings from the higher drive potentials is showing a [background electron energy]:[beam electron energy] of around 1:1000 (given that the electrons where you are measuring is likely around a half-way up the drive potential).

Post Reply