The paper refers to tubes that are "specifically engineered for this type of display". I followed up on the citations and got here (
). The application note states "until recently, the high peak currents required caused ionization to spread beyond the numeral to the leads and the pins. This extraneous glow within the tubes resulted in a generally poor aesthetic appearance of the display". It doesn't make any mention of any other deleterious effects (though that doesn't mean there aren't any).
The note goes on to explain that the high peak currents are required to maintain the brightness of the display.
The Z5600M data sheet specifies a continuous current of around 1mA and a pulse current of around 10mA. I might assume(!), therefore, that these tubes could also be safely driven by a multiplexed method that adjusted the current to maintain brightness,
Power supply voltage, V 200 -
Anode current, at pulsed voltage: (50 Hz), mA (median) 2 to 4
Anode current, at rectified voltage: mA 4 to 8
So. All things considered, I now side with Ty (is that your name? Apologies if not) when it comes to IN-18s, or anything that does not explicitly specify a higher pulse current.
Subjectively, I have not noticed any difference between my multiplexed IN-14 displays and my direct drive IN-14 displays, in terms of brightness, flicker or glow leakage. I suspect that if the current is limited to that stated in the data sheets, there is not a problem. I will let Ian comment on whether or not that is the case with his current (no pun intended) hardware.
It would be an interesting experiment to run the IN-1s with a variety of activation schemes to test for glow leakage and for tube lifetime:
Plain direct drive.
Dimmable direct drive.
Multiplexed and staying within current limits.
Multiplexed and not staying within current limits.
Dimmable/Multiplexed and staying within current limits.
Dimmable/Multiplexed and not staying within current limits.
Regarding the post directly above... you dug up something useful that I've never seen before: some real citations regarding multiplex, and most interesting (to me because it sort of backs up one of my assertions) is the paper in which they state "specifically engineered for this type of display". Good finds!
I agree with you that the Z5600M is a good candidate for successful multiplexing, given that they include a spec for pulsed current and it's higher than the continuous current spec. Stands to reason they had multiplexing in mind when they came up with the spec. I would imagine this holds true for any data sheet being read in its original language that gives a pulse current value that is substantially greater than the continuous current value.
I suspect that the IN-18 datasheet has lost something in translation but not being a Russian speaker I have no proof - only an ASSumption. I'm assuming that "Anode current, at pulsed voltage: (50 Hz), mA (median) 2 to 4" is talking about poorly filtered half wave DC that ripples severely (pulsates?) at mains frequency. Is it too far fetched to think the datasheet dates back to a time when the Nixies might well be fed from a selenium rectifier or even a vacuum tube rectifier?
Yes, my name is Ty, the Eeberfest part is a made up word I use a lot on the Internet. Eeberfest = a slightly less harsh (and highly technical) term for a clusterfuck.
Writing style, okay no problem. There is so much pertinent info missing when discussing things in text as opposed to face-to-face that it's hard for me to gauge attitude and intent sometimes.
In the video clip I was seeing flicker watching it on a desktop PC using the embedded YouTube player. I tried it later on a phone, again using the embedded player, and the flicker was still there but much less noticeable. Watching it now directly on YouTube seems the same as watching on the embedded player.
Look into it later when the dust is clearing off the crater.
1 year 7 months ago - 1 year 7 months ago#7522by judge
There is an interesting discussion going on over on neonixie-l about IN-18 failures on Spectrum kits. In it, Manuel Azevedo observes that older spectrum kits apparently heal failing tubes in newer spectrum kits. I guess there is an optimum current to use here that balances cathode poising with sputtering. It also makes me wonder if a better way of performing ACP would be to boost the current to all the tubes for a while (while displaying all the digits). Of course, I have no idea what would be involved in making a current controlled power supply, so this is easy for me to say!
I finally read about getters while I was making sure I had some idea of what I was saying in this post. Now I understand a bit more about what Dalibor was doing in his making a Nixie tube video. I also found another wonderful article about the complexities of tube design
. If you ever wondered what all that stuff inside the tube was for, it is worth a read.
from the same site that describes the theory of operation in terms that I can understand!
1 year 4 months ago - 1 year 4 months ago#7679by Ian
I don't see it as "bashing" at all, and I think it's totally on topic: It's a natural discussion topic because lots of people use the Spectrum as their "reference" implementation of a large format direct drive clock, and you can only compare what you have in common and where there is a large enough sample size to gain some information with a fair level of significance. Name me one other clock where the majority of people could say "oh yeah, I have one of those, and it does that as well/doesn't do that any more/never did that".
It comes down to the fact that if there are problems (blue spot, poisoning), you can't achieve much without comparison with the experience of others.
If the Supertex HV chips were not designed for Nixies, perhaps there is a problem that needs solving. If the old Spectrum doesn't poison and the new one does, (an observation, but not based on data at this point), what is the Cause? Do we have to collect more data to see that there really is a difference between old and new Spectrum? Has something happened to the HV? Is it the HV, or is it something that changed betweem the old and the new Spectrum? Is it just the Spectrum or do other HV direct drive clocks also have this problem? Is the problem in the tubes? If there is a problem, can it be mitigated?
Many questions. Few answers, and I'm glad that others are thinking about it as well, because perhaps with the introduction of a Zener or a resistor or something we're all good again, and that would only serve to feedback to an improvement in the Spectrum, so everyone wins in the end.
Last Edit: 1 year 4 months ago by Ian. Reason: typo
Of interest - all of the early and later IN-18 Spectrums use the same output drive. PV shows (6) revs to their builds. Up through v5, the anode resistors were 10K. The v6 kit uses 8.2K anode resistors but all use the same output chips and circuits. There were several firmware updates. It seems unlikely that an older version of the same clock would "heal" tubes from a newer version of the same circuit unless there is something else is going on, like maybe a problem in a specific clock unrelated to the design. Especially if the anode resistors on the latest version are slightly lower resistance. So I question the reason for the observation from the other forum as being related to output circuitry. Maybe if it had been between two different designs of clocks with different output drives, I could draw some conclusions. I think that the conclusion that the Spectrum has issues is unfounded.
I do have a clock that has demonstrated instances of cathode poisoning in 3.5 years of continuous operation - the IN-18 Blue Dream. I've had (2) total tube failures due to leakage and (3) instances of cathode poisoning in different tubes that have been subsequently cured by high current operation with an external power supply. This has occurred over several date codes. My observation is that the Blue Dream has a brighter display than the Spectrum and the output circuitry of the Blue Dream uses a totally different output circuit using discrete devices. Based on the experience that I have had with the Blue Dream, I would say that the Spectrum clock and the driving method is easier on the IN-18 tubes. Unfortunately, I can only measure the drive voltages at the output stage but can't give you any definitive information on the current waveforms in the Blue Dream. I've attached a typical section of the Blue Dream output on a couple tubes.
I still love the Blue Dream but I probably would not have another one until the sockets change to something more gentle on the pins. As far as the Cathode poisoning goes, I will heal the tubes when needed but would like to understand more about why this particular clock design seems to be more prone to the problem. There is nothing in the Blue Dream configuration that allows me to adjust the brightness except to run it in a dimmed mode 24/7.
Last Edit: 1 year 4 months ago by jwalton_mn. Reason: Unnecessary comment in the title.
Regarding the assertion about early Spectrum vs later Spectrum kits - I give a red flag to the claim. I have both and the drive circuity is essentially the same. The only evident difference is that the latest version actually uses 8.2K anode resistors vs. 10K anode resistors in the early version. This would imply that the later versions might actually cure the early version cathode poisoned tubes, not vice-versa. The control element that has to be established between the two is that the features in firmware, such as dim or off period and the frequency of cathode poisoning protection need to be identical to have a valid comparison. Using the schematics from PV's web site (Thanks!) you can see that the output devices and connections are the same. I've only been running the Spectrum 18 for about 2 1/2 years and have not had any issues with tubes. If you want to compare to the Blue Dream IN-18, I have had one running for about 3 1/2 years and have had issues with five of six tubes in this clock. (2 seal failures and 3 cathode poisoning).
Regarding the sputtering that causes cathode poisoning: It would be interesting to actually see a graph that could demonstrate whether the sputtering increases linearly or exponentially with current. That could be useful in trying to understand whether pulsing at higher currents for shorter times is any more or less detrimental than a continuous sputtering generated at lower currents.
I'll jump in, as I've been mentioned in this thread too.
As I mentioned in the neonixie group - and as proved by Jeff - I have both models the same release (Rev 6). But the older clock is using Rev 5 10K resistors, not Rev 6 8k2 resistors.
I didn't build the older clock - I bought it as a ready-built clock off eBay from a user called joemorphia.
But here's a side-by-side comparison of both clocks:
On the left side, with the acrylic case, you can see the older clock - it's quite brighter than the newer unit.
They are both set for 173.1V and here's the readings I got at each resistor:
I was out on a two week vacations. During this time all my clocks were off. When I returned, on the new clock, the leftmost tube has now poisoning on several digits: 2, 3, 5, 6, 8 and 0. Curiously all on digits, it's on the same position.
Here's a small video:
I will, again, remove this tube and put it on the old unit and see if restores back to live.
But I want above all to say that we're talking about one unit where I see this problem: there's not enough data to say accurately what's going on!
All I see so far is a single Spectrum unit that seems to work not so well.