Rejuvenating Nixies; Some Comments / Photos

[VK5RC]

As a bit of a Nixie (and other display tubes) fan, some of the digits in one of my Nixie clock in some positions were fading, I believe the term "Cathode Poisoning" is used. See 1st photo below.
After reading the page of Dieter, at Tube Tester http://www.tube-tester.com/sites/nixie/different/cathode%20poisoning/cathode-poisoning.htm , I thought I didn't have much to lose.
I have a Keysight AC6802,a  very useful device (both AC variable V and freq -to 400Hz AND DC PSU) which has some current limitation/protection. Unfortunately the lowest current limit is 0.2A.
Carefully connecting a Nixie and slowly increasing the DC voltage was not a happy experience, letting out the smoke on one of the daughter PCB upon which my Nixies sit.
I decided to build a current limiting supply using a LR645 10V regulator, but it can handle up to 400V in or so and if TO220 type mounted to a heatsink can dissipate 2-3W and still not blow itself up. see photo below
The current limiting resistor is switchable to 1mA steps up to 6 or so the slowly increasing steps.
This allowed me to use the AC6802A as the DC source and current limit via the LR645.

[VK5RC]

The problem is that the AC6802 is a noisy beast, and rejuvenating Nixies a SLOW process, each digit of each tube if done PROPERLY can take 1hour.
The fan was driving me crazy, I was also attaching a device capable of delivering 800W when a far less powerful source would suffice. I came across a small switch mode booster from lumos, which goes pretty cheap for ~USD 13 see photo below. It is quiet acoustically, I have not tested it electrically.

The comments I would make, I am left with the idea to try and get the current as high as possible in Dieters advised 2-10x the advised current range on the data sheet for the Nixie. For my russian IN14 the data sheet Imax is 2 1/2 mA or so, I would try to get to 6-8mA if I could.
The limit I found was if the current was raised too quickly the 'lead in' wire would start to glow, but if you just waited a few minutes, often you could 'sneak it up' to the 'goal' levels. The best result was when I left one digit on and went for lunch, an hour or so did a really good job.
The photos below are of the Nixie above in early rejuvenation and the end result after about 20mins, I think I probably could have got better result by 'cooking' it for longer but is does take a LOT of time.

[VK5RC]

Below is the restored clock, I cheated when I got my Nixie clock and used PV Electronics kit QTC with IN14 tubes as there are quite a few of IN14s around and the clock has some nice features built in e.g. GPS lock, 'slot machine' switching on of the Nixies after each minute ti try and prolong Nixie life and the small daughter boards for each Nixie.  so pulling each Nixie was not at all difficult (10pin header) . 

In Summary Nixie re-conditioning as suggested by Dieter (see above link) worked well and I would recommend a 200V or so DC source and making a current limiting supply using an LR645. I was able to restore all of my tubes successfully although it is not a quick process. If you do this please be careful re the potentially lethal DC voltages.

[bitseeker]

Thanks for sharing your success story with rejuvinating your Nixies and the before, during, and after photos.

[FlyingHacker]

Yes, thanks for sharing. I have read about this, but never attempted it.

[VK5RC]

I am always happy to share my successes! ,  the others get buried HiHi.
Seriously I have learnt a lot from eev blog contributors,  so I try to put a bit back.  My technical understanding is modest and I am having a bit of fun restoring older gear,  about to get my first tube radio,  a yaesu ft101ee,    so getting a bit of experience with the higher voltages is important.

[CJay]

Silly question and maybe requires a thread of its own but cathode poisoning in Nixies to my understanding is due to running them on DC, would it be a better idea to work out a way to power them from AC or have I missed a fundamental mode of operation that requires DC?

[SeanB]

Running them on AC would stop the poisoning, but the front grid, which is the positive and normally does not glow, would swamp the light from the digit selected by its glow when negative.

[VK5RC]

I've not seen anyone actually identify what is the contaminant/s on the cathode,  it does,  to me,  appear to be heavier than neon as the bottom segment of the least used digits seem to be the worst.  Often the tens of minutes or seconds esp 6,7,8,9.
Full reversal AC has problems but I am not sure if even cyclic DC ie rectified AC would make much difference to the accumulation of the compounds on the cathodes.
Most people suggest the 'slot machine " switching on of all of the digits on a regular basis.  Dieter' s (and possibly others)  technique of ' burning it off' is often mentioned,  my experience supports this. 
Perhaps a technique that might be considered is to have an intermittent period (perhaps a minute per day) of high current 'slot machine'  cycling of the digits again as a preventative.  PITA to try to organise.

[Cubdriver]

Nice writeup, Rob    - tagged for future reference.

-Pat

[tautech]

Nice writeup, Rob    - tagged for future reference.

-Pat

Linked in this sticky:
https://www.eevblog.com/forum/repair/vintageclassic-renovation-techniques/msg922140/#msg922140

[bitseeker]

Most people suggest the 'slot machine " switching on of all of the digits on a regular basis.  Dieter' s (and possibly others)  technique of ' burning it off' is often mentioned,  my experience supports this. 
Perhaps a technique that might be considered is to have an intermittent period (perhaps a minute per day) of high current 'slot machine'  cycling of the digits again as a preventative.  PITA to try to organise.

This Arduino-based Nixie clock controller project uses the slot machine method to periodically give all cathodes a chance to light up (I think the interval was 10 minutes). There's also a manually enabled mode to rejuvenate tubes by burning off the cruft.

https://github.com/isparkes/ArdunixNix6

[VK5RC]

One of the Nixies that was rejuvenated has slowly declined again, in the same section of the number 5. Nothing to lose so I had another go.
Note this is with a "slot machine" running all the numbers ever few minutes or so.
The same technique of running it with 2-4x the usual current worked again - how long for I don't know.