My Newest Nixie Clock

84 views
Skip to first unread message

Tom Katt

unread,
Jul 29, 2026, 7:54:41 PM (yesterday) Jul 29
to neonixie-l
Retro_Nixie.jpg

Still needs some final polishing, but since it worked when I plugged it in after stuffing everything into the case, I couldn't help myself and took a birthday photo ;-)

I recently got myself a small Carvera Air benchtop cnc machine and decided to try making a pcb board via the milling-etch method, where the trace isolation lines are milled out using a fine engraving bit.  This actually worked quite well, though there is a limit to how fine you can get the traces - so far it seems like 0.75mm is my limit while still being able to solder.  That actually works fine, given my eyes and hands aren't as good as they used to be lol - I'm pretty much a through-hole guy anyway, so this is perfect for me.  I can't take credit for the case - I found a model online that was originally 6 digits IN-12 tubes and modified it for my 4 x B5991 and 2 x IN-17 (shout out to Nicholas for helping procure the IN-17's) and printed it on my 3D printer.  It's amazing what technology offers for the home hobbyist these days!

I milled the pcb for the tube display board, and while I ordinarily build my own controllers I cheated this time because I had to know if those $15 shipped Chinese clock boards were any good - tell you what, they are actually pretty nice lol.  Direct drive using ULN2803 chips that are fed by 74595 serial-parallel converters, so no multiplexing (I prefer direct drive for tube life).  They also have several good options built in, including anti-cathode poisoning routines that flip through all 10 digits on an adjustable periodic basis if desired and 'night mode' that disables the HV at user configurable hours so the tubes don't run all night when nobody is awake to see them.  It would definitely cost me more than $15 for the parts, let alone the time and cost of making the board itself (though that option just got better as mentioned).  Still - if you want a cheap turn-key solution, you could do far worse than these things.  This isn't the seller, but here is an example of the control board I used: Universal Nixie Clock Control Board 4/6-Digit Remote

Anywho, I'm going to button this up and find it a home now ;-)

Martin Černý

unread,
3:56 AM (18 hours ago) 3:56 AM
to neonixie-l
Looking great, want to make a variation of this design one day as well. As a side note, the ULN2803 (or more commonly ULN2003) used in these boards is not a good choice for nixies. It does work and it's a cheap reliable option but the performance is sub optimal. Because it's clamped only to 47V, the unlit cathodes are operating very close to strike voltage, this results in "foggy", less crisp numbers and unreliable blanking. What's great is there's SN75468 chip which is pin-to-pin compatible with ULN2003 and you can use higher clamping voltage (70V or more) for best performance and reliable blanking.

Dne čtvrtek 30. července 2026 v 1:54:41 UTC+2 uživatel Tom Katt napsal:

Tom Katt

unread,
7:53 AM (14 hours ago) 7:53 AM
to neonixie-l
You know, I was amazed they worked at all but then figured they must know something I don't because I don't think I would have even tried it in the first place lol.  But that makes complete sense and I guess you DO always get what you pay for.

Thankfully, mine seems to be working pretty well (at least for now).  I just had to know how they managed to do things so cheap...   My small stash of 74141's remain guarded for the next project lol.

Thanks for the comment!

gregebert

unread,
8:11 AM (14 hours ago) 8:11 AM
to neonixie-l
10+ years ago another member on this forum recommended the HV5530, and I've made several clocks with it and have been very happy and zero problems or failures (OK, I did toast one due to my own stupidity, but that doesn't count...). The other plus is that this was the final nudge I needed to go all-in on surface-mount parts, using thru-hole only when no SMT options exist.

Tom Katt

unread,
9:16 AM (13 hours ago) 9:16 AM
to neonixie-l
I'd love to work with smd, but I think with my hands and eyes those days are behind me now lol.  More and more parts are smd only now, or through hole is becoming 'vintage' and getting a price premium as a result.

Martin Černý

unread,
9:38 AM (13 hours ago) 9:38 AM
to neonixie-l
Reminds me I wanted to write a blog post about all the drivers and methods to run the nixies. There's so many fun options, even if we're just talking Microchip drivers, there's quite a few for different use cases.
Dne čtvrtek 30. července 2026 v 14:11:57 UTC+2 uživatel gregebert napsal:

Tom Katt

unread,
9:49 AM (13 hours ago) 9:49 AM
to neonixie-l
This got me curious - how have I gone all these years and just realized that even my trusty 74141's only have a 60 volt clamping voltage?  I must have built 7 or 8 clocks with the 74141 chips and their Russian clones...  Though I've seen many comments suggesting the Russian K155ID1 chips handle higher voltages?  Not sure if that is true.

But I've never done much with blanking - and most have been in the 170 volt supply range, which limits the issue?

The SN75468 seems like a good option - combined with SN74595 they only require a few i/o pins to manage many digits, and the SN75468 appears to still be easily obtained.

Thanks again!

Tom Katt

unread,
10:05 AM (12 hours ago) 10:05 AM
to neonixie-l
Forgot to show the pcb I milled on my Carvera Air benchtop cnc - this is after tinning:

PCB_TIN.jpg

Martin Černý

unread,
10:23 AM (12 hours ago) 10:23 AM
to neonixie-l
I don't know where this notion that russian chips handle higher voltages comes from, it's total bogus unless someone corrects me with hard data. If anything, they are less reliable in my experience... Both have 60V clamping voltage, so it doesn't really matter what the supply voltage is. If you run it too high, like above 190V, the OFF outputs will light up (200V - 60V = 140V, sustaining voltage for most tubes). 74141 works fine with 170V, but even at that voltage, with all outputs OFF, the tube can dimly light up. These drivers are not really designed to blank all digits, it's kind of a afterthought. If you want 100% reliable cathode side blanking, the clamping voltage needs to be 70V or higher (for 170V supply). It's worth mentioning that blanking the 74141 will not damage anything and it's not out of spec, it can introduce visual imperfections.

SN75468 is really the most versatile option, it works perfectly with both direct drive and muxed drive, not too expensive and it's in stock with most suppliers. I prefer it for most projects.

Dne čtvrtek 30. července 2026 v 15:49:19 UTC+2 uživatel Tom Katt napsal:

Tom Katt

unread,
10:44 AM (12 hours ago) 10:44 AM
to neonixie-l
Information like this is why I love this group ;-)   And no disrespect to Russian members, but this would be the first time that a cold war cloned chip exceeded the original US version lol.  So that makes sense to me as well.

Mitch

unread,
10:51 AM (11 hours ago) 10:51 AM
to neonixie-l
Nice work. I take the lazy way out, ordering boards from JLCPCB.

I've used the through hole HV5530 since the Jeff Thomas, NixiChron days. They never fail, and I won't multiplex for reasons discussed.

This time around I used the SMD version. It's really not that hard to solder, using a little flux and solder wick.

Mitch

Martin Černý

unread,
1:32 PM (9 hours ago) 1:32 PM
to neonixie-l
I know this is getting way off topic but I have to respond to this :D There's no reason to think, or at least I haven't seen any reliable data to suggest, that muxed nixies last less than driven directly. If anything, it's possible to drive the nixies much more efficiently and multiply their life, if done right. But there's alot to that, worth a whole another discussion, lol.

Dne čtvrtek 30. července 2026 v 16:51:29 UTC+2 uživatel Mitch napsal:

Nicholas Stock

unread,
2:09 PM (8 hours ago) 2:09 PM
to neoni...@googlegroups.com
Indeed. This seems to be a perennial topic on the forum.... I've had
multiplexed and direct drive clocks going for years and years and haven't
noticed any tube lifetime differences.... (I know, not a definitive study
by any means!). You're probably aware of the attached from Burroughs
explicitly outlining multiplexing for their tubes...

On Thu, Jul 30, 2026 at 10:32 AM Martin Černý <mce...@gmail.com> wrote:

> I know this is getting way off topic but I have to respond to this :D
> There's no reason to think, or at least I haven't seen any reliable data to
> suggest, that muxed nixies last less than driven directly. If anything,
> it's possible to drive the nixies much more efficiently and multiply their
> life, if done right. But there's alot to that, worth a whole another
> discussion, lol.
>
> Dne čtvrtek 30. července 2026 v 16:51:29 UTC+2 uživatel Mitch napsal:
>
>> Nice work. I take the lazy way out, ordering boards from JLCPCB.
>>
>> I've used the through hole HV5530 since the Jeff Thomas, NixiChron days.
>> They never fail, and I won't multiplex for reasons discussed.
>>
>> This time around I used the SMD version. It's really not that hard to
>> solder, using a little flux and solder wick.
>> Zev-7 Mini — World's Smallest 7-Digit Nixie Clock — Angry Electrons
>> <https://angryelectrons.com/clocks/zev-7-mini/>
>>
>> Mitch
>>
>> On Thursday, July 30, 2026 at 10:44:57 AM UTC-4 Tom Katt wrote:
>>
>>> Information like this is why I love this group ;-) And no disrespect
>>> to Russian members, but this would be the first time that a cold war cloned
>>> chip exceeded the original US version lol. So that makes sense to me as
>>> well.
>>>
>>>
>>> On Thursday, July 30, 2026 at 10:23:33 AM UTC-4 Martin Černý wrote:
>>>
>>> I don't know where this notion that russian chips handle higher voltages
>>> comes from, it's total bogus unless someone corrects me with hard data. If
>>> anything, they are less reliable in my experience... Both have 60V clamping
>>> voltage, so it doesn't really matter what the supply voltage is. If you run
>>> it too high, like above 190V, the OFF outputs will light up (200V - 60V =
>>> 140V, sustaining voltage for most tubes). 74141 works fine with 170V, but
>>> even at that voltage, with all outputs OFF, the tube can dimly light up.
>>> These drivers are not really designed to blank all digits, it's kind of a
>>> afterthought. If you want 100% reliable cathode side blanking, the clamping
>>> voltage needs to be 70V or higher (for 170V supply). It's worth mentioning
>>> that blanking the 74141 will not damage anything and it's not out of spec,
>>> it can introduce visual imperfections.
>>>
>>> SN75468 is really the most versatile option, it works perfectly with
>>> both direct drive and muxed drive, not too expensive and it's in stock with
>>> most suppliers. I prefer it for most projects.
>>>
>>> --
> 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 view this discussion, visit
> https://groups.google.com/d/msgid/neonixie-l/6ad3d7a3-60b3-49db-ad27-c745804453a0n%40googlegroups.com
> <https://groups.google.com/d/msgid/neonixie-l/6ad3d7a3-60b3-49db-ad27-c745804453a0n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
Burroughs Application Note.pdf

Nicholas Stock

unread,
2:12 PM (8 hours ago) 2:12 PM
to neoni...@googlegroups.com
What on earth did my confuser just do?

Anyway, here's what the original text said..

"Indeed, this seems to be a perennial discussion on the forum. I've had (like many others here) both direct drive and multiplexed clocks running for over 10 years and haven't seen any difference in tube lifetimes. Not a definitive study by any means, but the attached note from Burroughs is probably something most have seen.

On Thu, Jul 30, 2026 at 11:08 AM Nicholas Stock <nick...@gmail.com> wrote:
Indeed. This seems to be a perennial topic on the forum.... I've had multiplexed and direct drive clocks going for years and years and haven't noticed any tube lifetime differences.... (I know, not a definitive study by any means!). You're probably aware of the attached from Burroughs explicitly outlining multiplexing for their tubes...

On Thu, Jul 30, 2026 at 10:32 AM Martin Černý <mce...@gmail.com> wrote:
I know this is getting way off topic but I have to respond to this :D There's no reason to think, or at least I haven't seen any reliable data to suggest, that muxed nixies last less than driven directly. If anything, it's possible to drive the nixies much more efficiently and multiply their life, if done right. But there's alot to that, worth a whole another discussion, lol.

Dne čtvrtek 30. července 2026 v 16:51:29 UTC+2 uživatel Mitch napsal:
Nice work. I take the lazy way out, ordering boards from JLCPCB.

I've used the through hole HV5530 since the Jeff Thomas, NixiChron days. They never fail, and I won't multiplex for reasons discussed.

This time around I used the SMD version. It's really not that hard to solder, using a little flux and solder wick.

Mitch

On Thursday, July 30, 2026 at 10:44:57 AM UTC-4 Tom Katt wrote:
Information like this is why I love this group ;-)   And no disrespect to Russian members, but this would be the first time that a cold war cloned chip exceeded the original US version lol.  So that makes sense to me as well.


On Thursday, July 30, 2026 at 10:23:33 AM UTC-4 Martin Černý wrote:
I don't know where this notion that russian chips handle higher voltages comes from, it's total bogus unless someone corrects me with hard data. If anything, they are less reliable in my experience... Both have 60V clamping voltage, so it doesn't really matter what the supply voltage is. If you run it too high, like above 190V, the OFF outputs will light up (200V - 60V = 140V, sustaining voltage for most tubes). 74141 works fine with 170V, but even at that voltage, with all outputs OFF, the tube can dimly light up. These drivers are not really designed to blank all digits, it's kind of a afterthought. If you want 100% reliable cathode side blanking, the clamping voltage needs to be 70V or higher (for 170V supply). It's worth mentioning that blanking the 74141 will not damage anything and it's not out of spec, it can introduce visual imperfections.

SN75468 is really the most versatile option, it works perfectly with both direct drive and muxed drive, not too expensive and it's in stock with most suppliers. I prefer it for most projects.

--
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.

Tom Katt

unread,
2:52 PM (7 hours ago) 2:52 PM
to neonixie-l
Well, I obviously fell for the old mother's tales lol.  But I'm used to being wrong and learning new things lol.

That said - I did think that some tubes were explicitly designed for multiplexed operation.  The Burroughs app note suggests that their B57xx and B58xx series tubes were in that class.  I'm guessing that perhaps this also applies to my B5991 tubes (at least by series numbering).  Or does multiplexing apply to all tubes with modern circuitry?

PS - thanks for helping with the IN-17's...  they look great in my new clock ;-)

Martin Černý

unread,
3:51 PM (6 hours ago) 3:51 PM
to neonixie-l
Multiplexing works with most if not all mercury doped tubes (also known as long-life). I'd assume it will work with non-doped tubes as well, haven't tried it (could be too slow) but if you run it at like 1.2x the recommended current (which is still within spec), 1/6 duty, you could probably significantly extend their short life at the cost at running them dimmer, so there's potential there. The Burroughs document does mention tubes specifically designed for muxing but it also mentions 5% duty at 15mA, which is a rather crazy number. So they are made for muxing 20+ tubes, they need to be faster and sustain 5x+ the normal current. In a nixie clock, you have 1/4 or 1/6 duty and like 2x the normal current max which, when driven moderately is still within spec. I guess what I'm saying is that muxing in terms of nixie clocks can barely be called muxing in terms of stressing the tube :) It's more like PWM dimming all things considered. :)

Dne čtvrtek 30. července 2026 v 20:52:26 UTC+2 uživatel Tom Katt napsal:
Reply all
Reply to author
Forward
0 new messages