The SHINE experiment - Measuring the lifespan of nixies

[AlexTsekenis]

It is my great pleasure to announce that the SHINE (SALTechips IN-13 Endurance) experiment is now online.

For the first time since the Cold War era we are able to rigorously measure the end of life of a nixie tube due to darkening of its glass envelope.
We are super nixited as we can now accurately answer the frequent question of the lifespan of a given nixie.

The experimental method relies upon a theoretical basis and custom precision instrumentation developed by SALTechips specifically for this purpose.
In summary, we are monitoring the optical absorbance demonstrated by the envelope of the tube as cathode material is being deposited by sputtering. The absorbance increases with time, thereby causing the tube to become dimmer. At some time a brightness threshold is reached which we treat as the end of life of the tube.
The method and experimental setup are described in detail here: http://www.saltechips.com/lab/
This time we are measuring the IN-13 bargraph tube that is of immediate interest to us, although the tools and techniques can be used on most glow discharge tubes.

We anticipate the experiment that is currently in progress to take a while.Data is plotted live as it is collected and we also have a live stream of the instrumentation screens available 24/7(follow link above or directly here: http://www.ustream.tv/channel/from-the-salte-lab). These are the same screens we use to record data! Nearly 7000 minutes into the experiment and we are already collecting valuable data.

The future of this method is bright as it is applied to measure and compare other nixie tubes!


Alex

[Dalibor Farný]

Hi Alex,

this is pretty interesting! I cant access the website, it is probably
unavailable now, but the stream is clear! Do You have any graph how
the absorbance rise during that those 7000 minutes?

Are You running it on specs current or on higher to make the process
faster? The sputtering is not directly proportional to current, so it
would be nice to have data for currents from datasheet..

Thanks!

Dalibor

2013/11/22 AlexTsekenis <alexts...@gmail.com>:

> --
> You received this message because you are subscribed to the Google Groups
> "neonixie-l" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to neonixie-l+...@googlegroups.com.
> To post to this group, send an email to neoni...@googlegroups.com.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/neonixie-l/b16aa111-40c9-49f9-b6c5-f0a5b4a788a0%40googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.



--
Dalibor Farny
http://dalibor.farny.cz

[AlexTsekenis]

Hi Dalibor,

Thanks. I'm surprised you cannot access the site, no issues here from multiple computers. Maybe try from the home page http://www.saltechips.com/ and clearing your cache. Please let us know if you are still having issues, there is a wealth of information on the site including equations, plots etc.

We have data for the first 7k minutes (6810 to be exact), but I would like to collect more before drawing any conclusions. The plan is to do curve fitting on the collected data points, thus suppressing noise and other artefacts. If you follow our theoretical working you will see that the light emitted is falling exponentially, so we expect the data points to fit such a curve: http://www.ibm.com/developerworks/rational/library/mar06/cantor/fig7.gif The sampling rate of the data published is currently slow (1 point per week). This Sunday is time for the second measurement. I will need to judge whether to increase or decrease the sample rate then.

Yes, you are right on current affecting sputtering rate. Also the gas pressure and composition as well as the electrode material may I add. This is mentioned in the description and an empirical equation is given. However the method we are using is independant from these factors - we are just interested in the rate that the glass becomes darker. So the experiment can be repeated for different currents and tube constructions to optimise the tube and operating point.

We are running it at about 4mA with a constant current sink. This makes the glow reach about 3/4 of the tube's length. In other words the tube is in the normal glow region and we are 1/4 away from abnormal glow. In the normal glow region increasing the current increases the area of cathode being covered by the glow. These conditions were selected to best replicate typical usage of this particular tube, thus giving the most accurate lifespan figure possible.

How is website access now?

Alex

[Dalibor Farný]

Hi Alex,

it still didnt work, are You sure that there are no firewall rules to
deny users from Europe? I tried it via US proxy and it works.

thanks!

Dalibor

2013/11/22 AlexTsekenis <alexts...@gmail.com>:

> https://groups.google.com/d/msgid/neonixie-l/732ecf75-a44a-4ffb-b877-9f3bfad6072b%40googlegroups.com.

[Dalibor Farný]

Hi Alex,

I think it will actually be something with settings of the firewall on
your server. The server doesnt respond to ping, from my IP and even
not from my server that is in a big czech telehouse.

however, I checked your site from the proxy and there is plenty good
information!

I was just reading about sputtering two days ago, (Weston, 1968),
there were mentioned some experiments for measuring the deposit
thickness, have You already read it?

Bye!

Dalibor

2013/11/22 AlexTsekenis <alexts...@gmail.com>:

> https://groups.google.com/d/msgid/neonixie-l/732ecf75-a44a-4ffb-b877-9f3bfad6072b%40googlegroups.com.

[Billy Watson]

I got onto all three sites with the greatest of ease. nice work and product.  I would like to pickup three for gift's

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CA%2BnkT5raxrTnFCjShmghV__kTonyM%2BMz_QMSewsoPF8V9HnxGw%40mail.gmail.com.

[AlexTsekenis]

Hi Dalibor,

I just got confirmation from Eastern Europe that they can access the page. Can I ask who is your ISP/host in Czech Republic so I can check with our host. Thanks in advance. The same goes for anyone else that might be having issues. Glad you have temporary access!

It seems our timing is good ;-) Certainly, Weston's book is idea for bed-time reading.The chapter on sputtering (ch 4) is informative of the state of research in those days, but it falls short if you are reading it to find out when the tube has reached its end of life. Here is why.

The theoretical models presented (section 4.3) are too low level. The researchers tried to create models almost from first principles. The sputtering process is very complex, so they inevitably resorted in number of assumptions and simplifications. As you go through the chapter you will see that the results agree under some conditions, but disagree on others.I think Weston mentions this somewhere.

In terms of experimental work most setups were at too low pressures to sustain a glow discharge (fractions of a torr). Consequently the back-diffusion of cathode material that is present in commercial nixie tubes was not present (as stated in top of page 138 and section 4.2.2) in those experiments. Additionally the ion beam setups look nothing like a nixie tube (mercury pool, charged collection target etc).

As I described in SHINE, the most relevant experiments were those from Stocker (reference in SHINE page, page 140 in Weston's book). Stocker used a cylindrical tube with a wire cathode and a cylindrical anode (much like an IN-13 tube!). However he used a filament light source inside the tube, whereas we are using the emission from the glow itself as the light source. We are not using internal light sources as anything carrying current or being at a potential will distort the fields inside the tube. Stocker ends up with the empirical equation 4.18 in Weston's book or 2 in SHINE page which you one can use. But the equation only gives the sputtering rate, not the loss of brightness. To put the final nail in the coffin, none of the above experiments or theories incorporates anti-darkening measures such as a mesh anode like the IN-13 uses!

We decided to measure brightness directly as at the end of the day that's what causes one to throw the tube away. Doing so automatically accounts for whatever processes might be taking place at an atomic level and any measures taken by the manufacturer to increase the lifetime. Using this method I cannot do a calculation and tell you how long a tube you made will last if you run it at 5 or 10mA. But I can measure it and then tell you with great confidence :-) hope this clarifies the distinction.

As long as chapter 4 (and other literature on sputtering) is concerned, my personal feeling is that technology declined before the leaders at the time had time to develop a unified model. What do you think?

Alex

[AlexTsekenis]

Thanks for the kind words aesops.tech. To keep this thread as illegible and technical as possible, could I invite you over to here so we can configure yours. Alternatively you can get in touch directly here, happy to hear from you :-)

Alex

[Dalibor Farný]

Hi Alex,

I can now access the website finally, interesting! About the Weston
book, You are right. It is too general and low level, there is no way
how to calculate actual sputtering rates using information provided
there. Also as You mentioned, they deal with low pressure mostly. But
there are some interesting information such as the experiment with one
shape of anode grid and different pressures. They found that the rate
of deposition of the metal on the glass at 20torr was much lower than
at 50torr, there is a relation between the shape (size) of the hole in
anode grid and pressure inside the tube.

I think it is better to spend a few hours setting up the real
experiment than bothering with theoretical calculations in this case.
There is too many variables - metal composition, anode cage shape,
purity of the gas and so..

I am definitely interested in your experiment as it is far the closest
to our needs!

> https://groups.google.com/d/msgid/neonixie-l/5b1ece2d-1b69-47a7-8980-66787cda3dcd%40googlegroups.com.

[AlexTsekenis]

Hi Dalibor,

Great, good to hear.

The next step (more like parallel step) would be to setup an experiment with a numerical nixie.
For reasons that I will elaborate on in a future post, I believe testing numerical nixies will be easier.

The experiment was picked up by Hack-a-Day, which was fortunate as a commenter (Tyler Laseter, I believe) had the idea of severely over-driving the tube under test to establish a relationship between darkening rate and current (Similar to Stocker's experiments). His suggestion was then to extrapolate to normal operating currents. As I explained to him as far as the IN-13 is concerned one cannot extrapolate to normal operation as any equations would be developed at completely different operating regimes. However, as a numerical nixie is already operated in the abnormal glow region maybe, just maybe, we can work out a relationship in a region not too far off from normal operation therefore not too inaccurate.

I think the promise of getting results in a shorter time for numerical nixies makes at least attempting this worthwhile. In that case 3-4 samples of tube X will be needed, preferably no mercury or other inhibitors to prove the method.  Even defective tubes with a single cathode working should work.

More coming.

Alex.