Pulsed DC vs direct DC in cathode poisioned

[Trumpeter]

Hello gents,

Some of you may remember the issues I have had with cathode poisioning on my in18 clock. I have been using a heathkit ip17 to repair the cathodes which has had some success, though the positioning has creeped back in most cases.

One thing I have noticed is the tubes seem to light better/more evenly with the dc current from the ip17 than the PSU in the clock. It has had me wondering if there is something to the pulsating dc used in these modern clocks to control brightness, perhaps this can cause some tubes to develop poisioning more quickly? I ask because I can have a tube run for hours on the ip17 at 5mA then on the clock at the same current the same digit is poisioned within an hour or less. Why wold this occur?

[Jim_Z]

A few questions:

1.) How do you know its 5mA in the clock and how did you measure it? Peak or Average?

2.) What is the voltage to produce 5mA in each case?

3.) What is the waveform associated clock?

5.) Who manufactures the clock? Have you discussed with Him/Her?

6.) What Voltage/Current/Time did you use to "heal" the digits? 

7.) Any consistency with the in-18 tubes that are failures WRT manufacturing date? 

Jim

ka7gzr

[Trumpeter]

I'll do my best, I am not an electrical engineer, just a hobbiest so my answers may be frustrating to some.

1) Measured between the annode and anode resistor.

2) On the clock I dont know, I'd have to lift the resistor again and check. On my heathkit 5mA is about 140ish volts. I have a 10k potentiometer installed between the PSU and annode.

3) No idea, I don't have a scope.

5) I have discussed this issue with the creator. Chalked up to in18 being prone to poisioning. I'd prefer not to mention the design. My intention is not to call into question the design, rather discuss pulsating vs continuous dc.

6) Healing was done 10mA at about 149v for 30 min-2 hours depending. I expect their ability to heal will diminish.

7) These are all 1981 tubes. Some are failing, some are chugging along seemingly without issue so far. (Clock running since January 2017)

[Instrument Resources of America]

How are you measuring the current when the tube is in the CLOCK?? Ira.

[Trumpeter]

If I lift the annode resistor I can take a measurement in line between the resistor and where it is soldered to the board no? Maybe I'm doing this wrong?

[Instrument Resources of America]

No need to lift the anode resistor. Just place a scope across it, and
measure the voltage pulses, and then calculate the pulse current. Ira.

[Luka C]

@Ira, I think he already mentioned that he does not own a scope. If we knew the model of the clock, maybe there would be someone who knows such data from the firmware of the clock. Btw, if the tubes are NOS and have been running since January 2017, then there is something seriously wrong with the mentioned clock? Does it have anti-cathode poisoning routines activated and how often?

[Instrument Resources of America]

Sorry!!! I guess that I'm not paying attention!! ;>( Ira.

[Trumpeter]

Slot machine effect every 10min. Tubes were NOS but I wonder if I got a bad batch. I guess I should buy a scope, any reccomendations for an economical model that will get the job done?

[Terry S]

If the tubes show signs of poisoning within hours, something is seriously wrong with the tubes, not the clock.

I have a Nixichron that ran for many years, and some of the tubes started to show poisoning symptoms very gradually over time. Jeff sent me info on a hardware update, as well as advice on how to de-poison the tubes. It took a tad of rework and a few hours total time. The rework lowered the anode resistor value and raised the high voltage, so the tubes ran a little "hotter", but the poisoning has not returned at all. And I completed that rework at least five years ago. The clock is running strong and I expect it will for many more years. I run it 7 hours/day, and not at full brightness.

I did have one tube fail... what looked like poisoning was in reality a slow gas leak. Probably a result of my removing the tubes from rigid sockets.

The Nixichron has no anti-poisoning routine. I've run across numerous clocks that do, but some don't up the brightness to max during the routine. So I find that of questionable value. Others do up the brightness... don't put those clocks in your bedroom.

Terry

[Trumpeter]

I'm not sure what could be wrong with the clock that would cause the cathode poisioning. I have measured 5mA to the annodes which is in spec. The hv PSU is cranking outhe 170v. What else should I check?

[Terry S]

I'd suggest your local Craigslist -- any Tektronix scope from the last 2 decades would be a good bet. I'm fond of the TDS series, but you need the bench space, or a cart. The newer units are LCD screens and only require a few inches of depth on your bench.

Here's a particularly good buy in my metro area: https://minneapolis.craigslist.org/hnp/ele/6151008545.html

$200 for a 4 channel 400 MHz scope. Overkill for what most hobbyists need, but a great value. You will spend some more money on probes, but you do not need to buy the expensive Tek probes.

Terry

[Paul Andrews]

Maybe you just have a bad wall wart. It would be a cheap and easy test to buy another. Not all wall warts are created equal. They can be poorly regulated, noisy and what it says on the label can be an outright lie. I know you have tested the current, but I still think it would be worth a try.

[JohnK]

err, watch the potential there. Scope earthed/grounded? Scope floating? HV
supply earth/gnd referenced?

Maybe a different technique for a non-EE?

John K

> --
> 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/d8b3ae86-d0c9-8980-8108-c9ca01a12e70%40HUGHES.NET.
> For more options, visit https://groups.google.com/d/optout.
>

[Trumpeter]

I'll give another wall wart a try. I hadn't thought of that.

On the subject of scopes and grounding, is there some reading I should do? I do like to be cautious around these high voltages.

[Luka C]

@terry, I agree, but it seems like his tubes got poisoned over three months or so as far as I understand. Could be a slow gas leak from a broken seal, but I'm not sure he could have healed them (and have them working for some time clean of cathode poisoning without raising the voltage) if that was the case. Gas leak typically irreversibly progresses until the tube completely fails.

[Trumpeter]

Luka that is why I am investigating further. If it was the tubes loosing gas why amy I able to heal them?

[Instrument Resources of America]

I fergit that we may be working with non EE's Ira.

[Scott McInness]

The scrolling date and temperature, if you have them enabled, effectively function as an anti-poisoning routine on the Nixichron. I received mine in October 2011, installed some 9102 date-code nixies and it’s run 24x7 ever since, except for the odd power failure, with no signs of poisoning at all. Other than the scrolling date and temperature, mine only dims from 21:00-07:00 but I can’t remember what my wake and sleep brightness are set to.

 

Cheers,

Scott.

 

From: 'Terry S' via neonixie-l [mailto:neoni...@googlegroups.com]
Sent: Tuesday, 30 May 2017 11:47 PM
To: neonixie-l
Subject: Re: [neonixie-l] Pulsed DC vs direct DC in cathode poisioned

 

If the tubes show signs of poisoning within hours, something is seriously wrong with the tubes, not the clock.

--

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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/a2daa453-783b-443e-b908-8812724d541b%40googlegroups.com.

[Tomasz Kowalczyk]

There is no "economical" model of a true oscilloscope. You might use some sort of audio card oscilloscope, but with such voltages you have to be very careful not to burn your PC - for example, use a voltage divider made of 330k and 1k resistors (1k between GND and audio card input and 330k from audio card input to HV). This + some software (there are many audio card oscilloscope programs) should do the work for frequencies from audio range (20Hz-20kHz). I would also place two standard diodes in series, anode to audio input and cathode to ground - cheap and rather reliable way to produce a ~1,4V voltage limit.

Datasheet of IN-18 is hard to understand for me - I can read cyrillic, but barely understand language alone, so I make alot of assumptions from my native, also slavic language - but as far as I can read, allowed continous current is 4-7mA and average current for pulsed operation should be 2-4mA (specified for 50Hz pulses). If the current you have measured in your clock is a true average value of 5mA, then it is out of specs.

Can you check clock PSU voltage, anode resistor value, anode voltage drop (avg) and number of tubes used in clock? I bet doubling anode resistor value would fix the problem, as this starts looking like it isn't a problem with pulsed vs direct operation, but a design which is out of datasheet values for the tube. This will of course change the brightness, so you might want to use some other resistor - 30% bigger value than the original should be also okay.

Normally pulsed operation doesn't do anything bad to the tube, unless it is out of specs (too high or too low current*). Many smaller tubes in later designs were multiplexed, for example calculators - I've seen displays of one east german and one russian calculators and they both were multiplexed (Z574M and IN-14 tubes).

*I am not sure if too low current can hurt a nixie, but I believe it might increase sputtering - I have no data to back it up, it is just my strong feeling.

[Jeff Walton]

Nixie tubes that are run with too low driving current are very susceptible to cathode poisoning.  Glowing cathodes sputter.  The sputtering is what causes the gas to glow.  The reason for running within spec range is that when the cathodes are properly driven, the cathode that is operating will be driven hard enough so that it will drive off impurities that have been deposited on the surface of the cathode.  Driving the cathode at too low current will not allow the normal operation to drive off cumulative deposits from sputtering of other cathodes. 

 

These cathode deposits can occur from:

·         Long term storage when impurities in the tube will contaminate the cathode surfaces.  This is why tubes that are NOS and in boxes since 70’s and 80’s will often need to be run for some time in order to get them to completely glow on all cathodes.

·         Running at too low current means that the cathodes cannot properly “clean” themselves from deposits of adjacent cathode sputtering.  Higher currents cause higher excitation of the gas and “heating” of the cathode in operation.  The higher excitation and heat cause the contaminants to be driven off of the surface of the cathode in operation. 

 

Running a tube at currents higher than the normal ratings will cause more overall sputtering but cathodes can clean themselves better.  That is why higher currents are used to “de-poison” or “clean” a cathode which does not glow completely due to the contamination on the surface from sputtering or impurities in the gas.  The downside to high current operation is that the tube life will be shortened significantly if driven too hard for long periods of time.

 

The conclusion is that reducing the drive current below the tube spec is NOT the way to prevent cathode poisoning, it will actually cause the problem. 

 

Also, pulsed current drive requires higher current levels during the pulse than continuous drive to achieve the same brightness.  The debate is in the physics of operation and whether the pulsed current times duty cycle has an equivalent effect to constant current of the same computed value.  There is a case to be made that a pulsed tube actually is less susceptible to cathode poisoning but there is insufficient data regarding the long term life of the tube for pulsed vs constant current.  I have read conflicting opinions over the usable life of the same nixie using the different drive techniques for operation. 

 

My experience is that a nixie tube operated in spec, with good seals AND mercury doping can run for more than 30 years, as evidenced by a clock that I built in the early 70’s which ran continuously for over 35 years without tube failure. The brightness was lower after 35 years and some of that was attributed to silvering on the inside of the glass envelope after many years of operation. 

 

 

 

 

 

From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of Tomasz Kowalczyk
Sent: Thursday, June 1, 2017 3:45 AM
To: neonixie-l
Subject: Re: [neonixie-l] Pulsed DC vs direct DC in cathode poisioned

 

There is no "economical" model of a true oscilloscope. You might use some sort of audio card oscilloscope, but with such voltages you have to be very careful not to burn your PC - for example, use a voltage divider made of 330k and 1k resistors (1k between GND and audio card input and 330k from audio card input to HV). This + some software (there are many audio card oscilloscope programs) should do the work for frequencies from audio range (20Hz-20kHz). I would also place two standard diodes in series, anode to audio input and cathode to ground - cheap and rather reliable way to produce a ~1,4V voltage limit.

--

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 email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/0e39db9b-b7f3-4370-8f70-0c22c72d5981%40googlegroups.com.

[Trumpeter]

I replaced the wall wart with a salvaged laptop supply that is rated at 12v 2amps. Maybe this will help.

I did a lot of depoisioning over the last few days and the tubes (six in total) are mostly clear now so it will be interesting to see if they start poisioning again. They are all 1981 dated tubes purchased as nos.

To reinterate:

HV supply is putting out 170v
5mA after the annode resistors measured with a multimeter.

I have not taken a voltage before and after the anode resistor (10k) but perhaps I should?

I have an entire box of 1988 and 1991 dated tubes at my disposal but I'd rather not simply replace these expensive tubes especially if my problem may be with the clock, for fear of ruining another batch.

[gregebert]

I will add that:

1. Not all nixies will respond to depoisoning attempts. I have a few Burroughs 5031's (very early nixie) that are hopelessly gone.

2. It takes weeks to  depoison at normal current.

[Luka C]

@greg, I might add that I have actually healed the IN-18 tubes (which I have purchased with cathode poisoning for a discount)  in less than a day per tube, but increased the current like 50% above maximum specified in the datasheet...at that point, the tubes run warm (not hot that you'd burn your finger on them), but I guess it won't kill them if run for less than a day.