IN-9 tube gas discrepancies and cathode poisoning issues

[Alex]

Hi guys,

I have been stock piling bar-graph tubes for a while now due to their relatively cheap price and potential for interesting looking clocks or other animated displays. 

Due to a few being smashed in a recent shipment I decided to sit down and work through testing all 450 or so IN-9's that I have to see what they are like.

Now, the first interesting point is that it seems that the white topped IN-9 can be either Neon or Argon, it seems most of mine are Argon which is quite annoying as the neon is a much richer red! I actually paid a bit more for a couple of purple topped argon tubes assuming them to be a lovely purple colour hence I was disappointed to find they are identical to about 70% of my conventional IN-9's (and badly cathode poisoned)! 

This brings me to the second point and a quick question, has anyone got any advice on getting these things to behave a bit better, some have fairly epic cathode poisoning with the glow starting in the middle or snapping to / hugging the other end or making large jumps up the tube. Are these things usually quite awful or have I just got bad batches? I have tried burning some in at upto about 40mA which has resolved some of the minor issues but on some seems to make them even more keen to glow from the other end...

I have a few hundred IN-13's to test as well soon and am just hoping they are not as bad as these have turned out to be...

Any advice on these tubes would be appreciated, hopefully a 64 channel VU meter will follow before long ;-)

- Alex

[Grahame Marsh]

Hi Alex

I've only used the IN13 (the one with the primer electrode) and they all
did as you described initially (despite the primer). I wrote the
microcontroller code to repeatedly sweep the tube current up and down
over about a second and left a batch of 8 running continuously. I think
it was several days before the "worst" tubes started to behave
correctly. Rinse and repeat until my purchase (a few dozen only) were
nearly all working. I have a couple that I couldn't unstick. I only
used a modest overcurrent at the top end. Since then the seven I have
in service (a "Jon Ellis" giant 7 segment clock) have operated without
fault. I don't know if the sweeping was an effective method or whether
it was just the exposure time.

http://www.sgitheach.org.uk/ss.html

Grahame

> --
> You received this message because you are subscribed to the Google
> Groups "neonixie-l" group.
> To post to this group, send an email to neoni...@googlegroups.com.
> To unsubscribe from this group, send email to
> neonixie-l+...@googlegroups.com.
> To view this discussion on the web, visit
> https://groups.google.com/d/msg/neonixie-l/-/W7oabYSL9SgJ.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

[Tim]

Hi Alex,

First things first, I love IN-9 tubes and am in the process of making an audio spectrum analyser with them.  I have both the orange and violet versions.  My violet ones having the purple top and do illuminate with a very purple glow :o)  These tubes are trouble and very hard to make them consistently illuminate from the bottom as they do not have the extra electrode the IN-13s have.  But there is a way!

I would be very surprised if your problem is being caused by cathode poisoning, it is more likely to be because of the power supply you are using.  These tubes do not like being run off nice clean DC from a SMPS. One must consider that when these tube were first being used (decades before I was alive) SMPS were not used to any extent (if at all?).  They were powering the tubes from good old half wave rectified DC.  I personally power the tubes from a mains isolation transformer (I think it came out of a bathroom shaver socket) and a single 1N4007 rectifier diode.  I don't think I have ever had one misbehave being powered in this way.  Give it a go and see if this solves your problems. 

Don't forget that the IN-9 are very power hungry compared with that of the IN-13.

Hope this helps,

Tim

[Alex]

Cheers for the replies guys, I suspect that with a bit of running in most of the tubes I have labeled "Slight poisoning" will come right... It is strange that your purple topped tubes are distinctly different from white topped, maybe I got lucky with my white topped but I defiantly prefer the true neon IN-9 glow...

My test rig could not be better for the tubes:

UK Mains -> Isolating TX -> Variac -> Full Bridge -> Large reservoir cap -> 25K W/W Pot -> 1K 10W Fixed R -> mA Meter -> Tube.

Defiantly no switching noise!

I will leave a few of the worst offenders on burn in for a half day and see if they wake up a bit better....

- Alex

[Tim]

That is very interesting that you are not using a SMPS and still having the trouble.  I have never tried the tubes with a  full bridge only a half, I wonder if the difference between 50Hz and 100Hz makes any difference, I cant see why it would.

You say you have a large reservoir cap, effectively giving you smooth DC.  I dont think these tubes like smooth DC hence the SMPS being trouble. Try removing the cap and see what happens.

Hope this helps,

Tim 

[Jon]

I spent quite a bit of time with IN9 and IN13 making my one digit 7 segment clock (http://youtu.be/mQ1567EFCY0). Someone was also asking about them over on tubeclockdb.com, and this is how I summarised my experience there:

 

My recollection is that a batch of neon IN-9 tubes would generally divide into thirds when I was selecting tubes for the clock. One third would be A grade tubes - pretty homogeneous in performance; one third would be Grade B stuff - definitely functional, but with little hiccups, sticky points or perhaps incomplete removal of poisoning - quite usable, but not good for the clock where you want all the segments to track each other reasonably. The final third of tubes had gross issues and I chucked them. Argon IN-9 were worse than neon (fewer grade A), and the IN-13 were almost all grade A with very little cathode poisoning.

 

The para above refers to selection of tubes after a burning in process at around 2x full scale current to remove cathode poisoning which is universal on these tubes. So you see, even after pretty extensive clean-up, there's still a high attrition rate. Interesting you say that you have argon-filled tubes with a white top - all the argon ones I ever had came with purple tops.

 

To Tim's point about what you feed them, I only ever worked with clean DC from a SMPSU, and that was fine on the good tubes. I'd be more inclined to blame the tubes - IN-9 are quite sensitive to rate of change of current. If you want to optimise your chances of keeping the glow in the 'proper' place, then you need to limit dI/dt, which may prove somewhat difficult for your VU application if the signal is swinging around on some ahem, banging music. IN-13 would likely give you better stability, though they're not as fun (too dim, orange only).

 

Jon.

[Dekatron42]

Here is a link to the first VU-meter that I came across several years ago: http://m.bareille.free.fr/vu-in13/vumeter_in13.htm#vu4301 ,I built a few of the ones with the That2252 IC. It works very well and unlike other people here on the forum I never had any problems with either IN-9 and IN-13 back then, they all ran perfectly after the initial burn-in.

 

I always do the burn-in process on one tube at a time and oversee the current/voltage over a few hours or days until the tube works as it should. With the batch of IN-9's and IN-13's that I bought back then, close to five years ago now, I did not have to run the burn-in procedure for more than 24 hours. It seems that I was very lucky with those that I bought compared to other people.

 

/Martin

[Alex]

Tim, spot on!

I finished testing and had a nice pile of mis-behaving tubes that would "snatch" to the other end on the first glimpse of FSD and jump about during ascent and descent. I de-soldered one leg of my reservoir cap so they have a nice 100Hz ripply supply and they all, instantly, work perfectly! The tubes are slightly (only slightly) dimmer and the tops develop a hazy arrow instead of a sharp end to the bar but they slide up and down smoothly and don't seem to mis-behave at all.

I am still, however, convinced that most of mine are some argon-neon mix which is a bit annoying as the few that are true bright orange neon are lovely...

Cheers guys,

- Alex

[Alex]

Also, anyone thinking of audio VU meters, check out the MSGEQ7 chip if you have not already, i have a dozen or so from this seller : http://www.ebay.co.uk/itm/270847592314 (not related) and used some in a work project (audio -> DMX) and they worked wonders, nice stable DC level out for reading by a micro ADC. There is another (harder to find) MSI chip that has all 7 channels broken out so in theory a 7 channel VU without a micro would be easy, that chip and some op-amp / high voltage transistor glue. A micro + DAC's would allow some nice winamp swooshey effects in between songs though as well :-)

- Alex

[Tim]

Glad I could help :o)

The next test will be to see how much the channel set to operate at a frequency of 100Hz will behave!  Depending on the Q factor of the filters you are using you may need to go one side or the other of 100Hz by a few Hz.

Tim

[Michel]

I bought a couple of IN-13 tubes a few weeks ago for my pending clock project. After reading this, I thought I better hookup a couple just to see what problems I can expect :-). Initially the tube would only light up to about 80% of full length at a current of 4.7mA (SMPSU). I then pulsed (by hand) the tube on/off say 5 seconds on, 3 seconds off. Every time the tube would light about 1 millimeter more than the time before until it reached the full length.

Sometimes the primer cathode really helps and sometimes it doesn't and the beam moves up the tube. I assume that using the 100Hz rectified mains will solve that problem, but I haven't tried that yet.

Michel

[Alex]

Hi guys,

I would like to contribute a few points to the discussion.

In contrast to the IN-13, the IN-9 does not have an auxiliary cathode to anker the glow discharge. This we all know.
However it does have a zirconium bump/spike at the bottom of the main cathode which is otherwise made out of molybdenum. The combination of Neon/Zirconium gives a lower breakdown voltage than Neon/Molybdenum, everything else being the same. The purpose remains the same: the glow to start at the bottom of the tube.

For the auxiliary cathode or zirconium spike methods to be effective, the voltage must be applied with a limited rate of increase. In other words, turn up the voltage slowly. This gives the physics in the IN-9 tube (have a read here http://www.saltechips.com/products/thermneon/theory.html) enough time to strike at the bottom of the tube before the applied voltage reaches the threshold where the rest of the cathode (molybdenum) strikes. This could be why Tim is finding the tubes to strike more reliably with half-wave rectified DC - the voltage rises slowly. I certainly don't think the tubes have developed the sort of nostalgia we have for neon, but for their original power supplies! :-)

Assuming there is a glow going at the bottom of the tube, that still does not guarantee that the bargraph will not 'break' if a fast rise in current (i.e bar length) is allowed. By increasing the current the glow will try to cover more of the cathode surface. The glow at the bottom of the tube is providing ions locally, thus giving the glow a preference to continue growing from there if enough time is given (back to formative lag). If the current rises very quickly, distant positions on the cathode become 'disconnected' and will strike independently. 

At this point I must mention that argon-filled tubes have a shorter formative lag than purely neon tubes. This could be why Jon has had better luck with neon IN-9s.

I experienced this first hand when working with the IN-13. Even if the aux cathode was active, the bargraph column would break when I wanted to 'instantly' move from the bottom to the top of the tube. The first measure taken was in hardware, by implementing an RC lowpass filter on the signal that controls the bargraph length. This slows down the signal to the tube's driver, and also forms part of a DAC. Have a look at the implementation on page 4 here http://www.saltechips.com/docs/thermneon-assembly-manual-v1.1-web.pdf
The second measure was in software, the bargraph was incremented in thousands of small, fast steps, rather than in one big step. This eliminated the column breaking up issue.

Regarding cleaning by cathodic sputtering, I found there was significant variation in a batch of 20 IN-13s. At the rated 5-ish mA, a few would light up up to the middle, a few all the way to the top and variations thereof. I was fortunate enough to come across and salvage a PS325 programmable HV power supply, so I used that to pulse the tubes at 100% overvoltage and overcurrent (280 V, 10mA) for 10ms every 1s. The current was limited to 10mA. I used a higher voltage to minimise the formative lag and a pulsing scheme to prevent overheating of the cathode. The current was chosen on a guestimate basis, the physics of cathodic sputtering are above my head. It took a minute or less for each tube. I am aware these tubes were heavily sputtered at the factory to give a high purity cathode surface - crucial for linearity. Briefly connecting two pieces of bare wire will work just as fine, but either put a current limiting resistor in series or wear goggles.

Hope you find this useful.

Alex.

[threeneurons]

Wow, thank you, for that analysis ! I guess the short answer, is change the current slowly, in particular, from the OFF to ON. Slow being a few milliseconds, since these thing tend to strike in the few 10s of microseconds. Maybe, a bit longer for those bargraphs.

[start end]

With DC high voltage power supply, the plasma didn't anchor to the bottom. With 50Hz square wave pulse current, the tube worked correctly.

https://youtu.be/mwD-WWELFqY

[Dekatron42]

You can find a PDF-file on the IN-9 & IN-13, from a Russian book, with the original Russian and a translation made with Google Translate here: https://drive.google.com/open?id=0B2XLFMydwkBgc2lhSEk5WXgxLXc in the TEMP folder here at the forum.

In most designs that I've seen in Russian books the IN-9 is almost always driven by half-wave rectified voltages and is usually used to measure an AC mains voltage or similar voltages, just like the first schematics shows in the document. The IN-13 is mostly used where you want some precision and also when you need better control of the ignition of the tube.

/Martin

[start end]

Thank you, Martin. The PDF-file is very useful.

Dekatron42於 2017年4月30日星期日 UTC+8上午2時33分27秒寫道：