Nixie tube clock... without the nixie tubes!

[linux-works]

<soapbox>

btw, the reason I had to mount that dc/dc upverter upside-down is that the designer of that module did not -care- about being .1 hole space standard.  sigh.  I emailed him, told him how inconvenient it was to deal with his module that way and his attitude was 'just design your own green board and use my pin spacing for when you mount my module'.

sigh.  he just does not get it.

folks, if you -can- follow the .1" hole spacing standard when designing DIY modules for sale, please do.  some of us still do appreciate being able to perf-board a proto or one-off build, and if you are not .1 compliant it causes us much grief.

</soapbox>

[cprobertson1]

nothing beats real nixies.  don't try to simulate it, just buy the tubes and build your own clock.  not hard.  here's mine

Nice work! I approve

Aye, the problem is the initial cost of the nixie tubes (I'll probably end up forking out on them the next time I see a decent deal for a bunch of them online: need to wait until I can afford some throwaway cash first though xD

[c4757p]

Nixies occasionally go for quite cheap. I've got rather a bunch of them that I was planning to use - never got around to it - and paid a few bucks for them all IIRC. Be patient, and don't insist on having one of the two or three extra-popular parts.

[kwass]

I did no work on the psu; just bought a module from fleabay and it works great:

http://www.ebay.com/itm/140964710736

$15 each - hard to complain about that.  I give mine 5v from usb power and it works super!  mine is almost a year old, now, and still chugging along.

You can buy directly from Taylor and they sell a vertical mount 0.1" pin spaced version:

http://www.tayloredge.com/storefront/SmartNixie/PSU/index.html

I've used this in a couple of my Nixie clocks with good results.

[linux-works]

I did buy a taylor dc/dc, as well.  yes, there are vertical mount ones.  I liked the stability of the horiz one, though.  just wish he could have moved a few pins over a bit

there are in14 and in16 tubes that are the size I used.  I used the ones that have the proper 5 numeral.  I think that's the in16.  around $60 for a set of 6 from russia or ukraine (for some reason, most are from ukraine).  shipping is about a month ;(

lots still around.  I don't know how long the supplies will last.  wonder if anyone has any idea at all how many are left in the various seller's hands.  do we have years of hobby clock building still left?  what if someone snarfs up most of the market?  they are not being made anymore, that's for sure.

also, on the really large tubes, I've heard about a blue glow that you want to avoid; and that often means using US 74141 chips instead of the russian versions.  and those chips are long gone from the parts channels; even used its hard to find the US chips.  and the large (nice!) tubes are hundreds of dollars for a set of 6.  but you can still buy them and I've seen at least 10 sellers on ebay with them for sale.

[macboy]

With your access to a glass blowing friend you could make your own "real" nixies.  Or if high voltages are a real problem, look into making your own "vacuum fluorescent" displays in a nixie configuration.  Those were used in automotive applications in the 70s, using 12V mains.  I don't believe they stepped the voltage up much, if any.  Or as a final thought, the science experiment/surplus market has trimmable electro-luminescent sheets that could be cut into numerals and packaged in a tube.  This option would have stacking problems since the lumens per unit area would force you to make the numbers "thick"".

VFDs in the 70's!? Yes. Today? Yes. My 2009 Subaru has beautiful white-on-black VFDs for its audio system and clock/trip-computer displays. In the dash, the odometer is a cheap looking black-on-grey LCD. What a shame. VFD may be an older display technology, but it is still relatively widely used, especially in harsh environments like cars. VFD's can be lit with as little as 12 V on the grid and anodes, depending on construction, but 20 V - 40 V or higher is more common.

[nickds1]

Did you mux them? A lot of "nixie clock" designs floating around do that, and it abuses the hell out of the cathodes. Drive them direct.

Absolutely not so - many of the 100s of different types of nixies were explicitly designed for muxing - Burroughs, National and others detail the methods in their technical documentation - it's not complicated.

Also, most nixies have a small amount of mercury in them to limit cathode poisoning and sputter. Most such nixies have a designed lifespan in the order of 200,000 hours, or nearly 23 years of continuous running.

There are reasons why a nixie may not last that long - it may be old, it might not have Hg in it, it might be damaged, or, most importantly, you may be over-driving it - overdriving causes an exponential shorting of a nixie's life.

FWIW, "end of life" for a nixie is not well defined - Weston says that when 50% of the tube is sivered or a digit being poisoned , its time to call quits, as would cathode failure due to sputtering.

Some people are making new nixies - its VERY complex , time consuming & expensive - have a look at Dalibor Farný's site at http://dalibor.farny.cz/ - its a tricky process - I, myself, have made neon letters and artwork - there are many many skills needed, none trivial, many dangerous, one certainly potentially lethal and all require extensive study & practice

Any questions about nixies, just ask

Cheers,

Nick (owner/moderator of the Google! NEONIXIE-L group at https://groups.google.com/forum/#!forum/neonixie-l)

[cprobertson1]

@nickds1: Useful advice! Aye, I was discussing with my glassblowing friend over the weekend, he can definitely make me an outer shell for one: and might even get away with putting wires through it - making it airtight and pressurising it with a particular gas mixture would be hard - and he's not sure how his glass would respond to heat (reckons we'll need to get a hard glass like borosilicate or specially treat it if there's going to be any real heat (otherwise it becomes weak and brittle - and depending on if it's been destressed may simply shatter - particularly under pressure)

Looks like making my own might be out of my reach =/

@linux-works I was thinking of making my own power supply (mostly because I want to try making one ) - I'll have a good look around for chips, and if worst comes to worse I can probably hack something together.


---Acrylic arrived today - will get pictures up later once I put a jig to hold it all together, but in my test it looked decent - though obviously no replacement for a nixie (still looks cool though)

---Still waiting for LED filaments - I'm hoping these will show some promise!

---Still waiting on a good deal for Nixie tubes. Theres a pack of 24 for £120 odds, +20 shipping from Ukraine - tempted but that's a lot of money


Pics coming soon once I get a jig together for testing

[cprobertson1]

I just got a bonus at work. I've already spent it. 24x IN-14 Nixie tubes for £5 each - apparently new (from Ukraine! Wonder how much that's going to cost me!)|

Apparently they can be multiplexed - at least according to the datasheet. Not sure I will though, direct drive seems simpler - still looking around for ICs to drive them: I'm tempted (I think I said this already? Hmmm...) to use a shift register feeding through high voltage transistors and simply switching between which cathode to drive with the register (disconnecting the anode with another transistor to avoid it flickering through the other digits as the data is fed in.) - you know, I am almost certain I have said this before... because somebody suggested an IC for me... Here we are, "Tiltit" suggested the HV5812 shift register - I'll have a gander for that when I get home!

Anyway - I'm aiming to make two clocks now (WHAT!?) one with nixie tubes and one with fake nixie tubes (this way I can practice making my own power supply with the fake-nixie solution ) - and who knows, maybe I'll find a decent alternative on my journey - so I'll keep you updated on my designs as they progress!

Going to give the acrylic-number another go tonight: got a new router burr to make the digits with - see if I can't get a jig together for holding them in place + stopping light going from one plate to the other. I'll need to use 5mm LEDs (to light 1mm acrylic plates... yeah) - so I'll only be able to light 3 or 4 of the plates individually but it will be enough to give us an idea if it's feasible. The final design will hopefully use 1005 or 1608 LEDs at the bottom (and maybe the top) - separated by an opaque "comb" (to stop the light from one acrylic sheet polluting the other). Going to mount this in a glass tube with an open bottom and drive each LED with a common cathode/anode. It will give more of a "floating sign" effect than I'd like, BUT, I have an interesting plan: mounting it inside some caging reminiscent of a real Nixie Tube (which will hopefully hide the corners of the acrylic, and the black or foil background will help with visibility)

Its not a real substitute, but for cheapskates or people with too much time on their hands it may just be enough (Cost per Tube approx £3 - but you need to buy materials for more than 1 tube) to be worth the bother of making one Plus I don't think they'll look too bad. Should have pictures by the end of the week - maybe even tonight if it works (just depends on how much energy I have left after a day at work )

[nickds1]

IN-14s are an "acquired taste" - the "5" digit is actually the same cut-out as the "2", but flipped and inverted - very odd - many people really dislike it. No idea why the Russians did this - we doubt that it would have saved much money, printing double the number of twos and no fives... personally, I never liked them and have never used them. Just checked my stock - out of the 7,500+ nixies I currently hold, only 9 are IN-14s

The IN-16 is slightly smaller and has the correct "5", like most nixies. I have over 1,200 of those at the moment...

The problem with the Supertex drivers is getting hold of them - in the UK, the usual suspects (RS/Farnell/Rapid etc.) often don't have them, and sampling is not easy. Depending on the CPU you are using, you may need level shifters as well. Having said that, they are a good chip and widely used in many designs.

Direct drive is simpler, but those Supertex chips aren't cheap

There are several latching shift-registers/drivers that will do which don't need extra HV cathode drivers, i.e. they handle that themselves. If you factor in the voltage drop across the nixie tube (typically around 120V), then you really only need an 80V-tolerant driver to be safe - 180-200V is completely unnecessary...

[cprobertson1]

Hmmm - I could have sworn I replied while I was at work...

@nickds1 - aye, Linuxworks mentioned that - according to the wikipedia, there are a few soviet designs that do it - apparently to reduce manufacturing costs. Meh, to be honest, looking at pictures online I personally don't it looks too bad - adds flavour if anything [I'll no doubt start a flamewar saying things like that!]

Some pics of the sidelit acrylic - which actually didn't look too bad! Decent light (even in a lit room - though not the best of course!), good viewing angle, and it will look great once it's properly mounted inside a wire cage/glass tube designed to mimic a nixie (it actually looks alright on it's own as well xD)

is it a nixie tube? no - but it is pretty easy to make - and though it has it's obvious shortcomings, when combined with a glass tube and some wires I think it may actually look pretty decent. Pictures were taken under different lighting conditions, and the photography is terrible - so I apologise in advance - gives some idea of what I can do with it though. These were 0.9mm thick with a 0.4mm etching on them (+/- 0.1mm because there's backlash in the jig I was using to cut them) and the number of plates doesnt affect visibility much (I tested using the three numbered plates + 8 blank plates which I couldn't be bothered etching)

Anyway - catch you folks late! and thanks for all the help! With any luck I'll get an actual prototype made up before next week - depends on how busy I am

Digit 2


Digit 3


Digit 1


What the plates look like

[linux-works]

if you want me to laser etch and cut some acrylic, as a test for you, I can do that.  might be a bit easier and more accurate than cutting directly with a friction tool or router.

(I tend to hang out at my local techshop at least once a week, anyway, so I could probably cut a few samples for you if you want, when I'm there next).

[kwass]

Apparently they can be multiplexed - at least according to the datasheet. Not sure I will though, direct drive seems simpler - still looking around for ICs to drive them:

I've built a few Nixie clocks and all are multiplexed, much easier, I think.  I use a single 74141 for the cathodes and high voltage optocoupler, TLP627 (available in single, dual or quad packages) for the anodes with just a single current limiting resistor.  I've settled on this circuit design as it has a very low parts count, inexpensive and is very reliable.

[Alex Eisenhut]

With your access to a glass blowing friend you could make your own "real" nixies.  Or if high voltages are a real problem, look into making your own "vacuum fluorescent" displays in a nixie configuration.  Those were used in automotive applications in the 70s, using 12V mains.  I don't believe they stepped the voltage up much, if any.  Or as a final thought, the science experiment/surplus market has trimmable electro-luminescent sheets that could be cut into numerals and packaged in a tube.  This option would have stacking problems since the lumens per unit area would force you to make the numbers "thick"".

VFDs in the 70's!? Yes. Today? Yes. My 2009 Subaru has beautiful white-on-black VFDs for its audio system and clock/trip-computer displays. In the dash, the odometer is a cheap looking black-on-grey LCD. What a shame. VFD may be an older display technology, but it is still relatively widely used, especially in harsh environments like cars. VFD's can be lit with as little as 12 V on the grid and anodes, depending on construction, but 20 V - 40 V or higher is more common.

1966:



http://www.bigblockmopar.nl/2010/12/chrysler-electroluminescent-dashboard-lighting/

[nickds1]

With your access to a glass blowing friend you could make your own "real" nixies.  Or if high voltages are a real problem, look into making your own "vacuum fluorescent" displays in a nixie configuration.  Those were used in automotive applications in the 70s, using 12V mains.  I don't believe they stepped the voltage up much, if any.  Or as a final thought, the science experiment/surplus market has trimmable electro-luminescent sheets that could be cut into numerals and packaged in a tube.  This option would have stacking problems since the lumens per unit area would force you to make the numbers "thick"".

VFDs in the 70's!? Yes. Today? Yes. My 2009 Subaru has beautiful white-on-black VFDs for its audio system and clock/trip-computer displays. In the dash, the odometer is a cheap looking black-on-grey LCD. What a shame. VFD may be an older display technology, but it is still relatively widely used, especially in harsh environments like cars. VFD's can be lit with as little as 12 V on the grid and anodes, depending on construction, but 20 V - 40 V or higher is more common.

1966:
...

http://www.bigblockmopar.nl/2010/12/chrysler-electroluminescent-dashboard-lighting/

Electro-luminescent (EL) is not the same as VFD - completely different technology:

VFDs are essentially triodes where the anode is coated with phosphor and you have a directly heated cathode (the heater) and a control grid.

EL displays utilise an optical/physical property of the material being used whereby given the correct stimulus, the material emits light.

Interestingly, for years, experimenters have been using VFD displays as triodes - some work quite well into low RF frequencies. More interestingly, Noritake (probably the world's largest producer of VFDs) in association with Korg (the audio effects people), have recently started to make novel audio triodes on their VFD production lines:

Many of us will have used Noritake VFDs, and as any fule kno, VFDs can be used as triodes, so maybe Noritake decided to cash in on this or just try to see where it'll end up. They've called it the "Nutube 6P1" (*) - the VFD heritage is obvious for all to see - basically, they seem to have been able to use exactly the same manufacturing technology, specifically in the use of new long-life cathode coatings, and also the same form factor & encapsulation as a traditional VFD. Great idea.

They're claiming 30,000 hours of continuous use, only 2% of the power consumption of a traditional triode (e.g. 12AX7) and very small size - 30% of the size of a "conventional vacuum tube", however big that might be. Also claiming "rich in harmonics"...

Homepage: http://www.korg.com/us/news/2015/012212/ also see http://www.synthtopia.com/content/2015/01/29/korg-noritake-introduce-futuristic-nutube-vacuum-tubes/

Can't find a datasheet ATM.

(*) For some reason, when I first read the name "Nutube", I was taken aback as I read it as "Nut Lube". I need to get out more

[cprobertson1]

if you want me to laser etch and cut some acrylic, as a test for you, I can do that.  might be a bit easier and more accurate than cutting directly with a friction tool or router.

(I tend to hang out at my local techshop at least once a week, anyway, so I could probably cut a few samples for you if you want, when I'm there next).

That'd actually be incredibly useful! I'll let you know if I take my tests to the next stage or not - I'm going to do a by-hand mockup this weekend hopefully; if it looks decent I'll give you a shout! We have CNC mills at work I can make use of if it comes to it - should be able to get hold of a 5mm V-Bit or 3mm End Mill. A series of <1mm holes in the shape of a digit doesn't look too bad either!

Anyway, I'll PM next week if I need it - I'll reimburse you somehow

@Kwass - ooh, thanks for the tip! I'll see if I can't get my circuit diagram reviewed by the community before I put the clock together just to make sure I'm not doing anything daft I don't usually work with multiple voltages at once - can't beat a challenge!

@nickds1 + Alex Eisenhut - thanks for all teh info! I do enjoy a good read... followed by a wiki trawl...

[mikerj]

I think the routed/engraved line for acrylic digits needs to be a lot thinner, under 1 mm wide would be good I suspect.  As it stands, the digits are noticeably obscuring each other.  A small PCB router could probably do a good job here.

The "dotted" lines in Jurgen Grau's edge lit nixies seems to work pretty well.

[klr5205]

Re-create the nixie tube construction using electroluminescent (EL) wire!

Probably best suited to fairly large characters, since I don't think you can create extremely tight radii.

[linux-works]

I think the routed/engraved line for acrylic digits needs to be a lot thinner, under 1 mm wide would be good I suspect.  As it stands, the digits are noticeably obscuring each other.  A small PCB router could probably do a good job here.

The "dotted" lines in Jurgen Grau's edge lit nixies seems to work pretty well.

even though this is not about real nixies, I'm getting curious about this, now   I wonder how good it can look when its well executed.

next time I go to the laser cutter place, I'll try some varieties of continuous lines and dotted lines.

[cprobertson1]

Just finished up a run of experiments: mixed results to be fair. Hoping to improve with time though, going to try drilling out a series of 1mm holes in the shape of a letter next time instead of just using a hand router as I did with all but one digit here.

Brightness is a problem - can't see the digits if there's strong light (fair enough, I expected that - I was going to mount the acrylic platter against a black background anyway)

I've encountered several problems - and came up with several good practices:


-Good viewing angle (up to 60-70 degrees I reckon)

-Low visibility in bright light (to be expected really - you can see it, but particularly with the overhead halogen bulb it just gets drowned out by all the white light - going to try putting against a reflective or black background to see if that helps any)

-Unable to see digits near back of the platter (see photo).

-Etching gives decent brightness

-Drilling a series of 1mm holes (spaced between 0.5 and 2mm apart (from the center of the drill) gives better brightness than etching (need to be careful not to distort the acrylic plate if you're doing this! Drill against a hard material and make sure it's clamped down!

-Going to try using thinner Acrylic - acrylic must be thin (I'm using 1mm: even less, ideally 0.5mm would give a good combination of fidelity and visibility.

-Care needs to be taken when cutting and sanding edges; if the acrylic isn't cut cleanly it doesn't let a lot of light through the edges (which is where our lights are coming from!)

-Probably going to need to light it from the top as well as the bottom: or use angled lights to ensure good lighting near the top of the plate.

-Putting 30V through an LED because you had the wrong preset on your bench supply breaks the LED. Who'd have thought.

Note all pictures are taken under medium lighting conditions (1x cold cathode overhead work lamp + 1x fluorescent at other end of desk - the white balance is all over the place as well because of the visible LED at the bottom (I just used a single LED with a 1mm hole drilled in the end and some tape to try and occlude some of the ambient light)


2: drilled - 5 plates (no background)


3: etched - 5 plates (no background)


7: etched: 10 plates (black background: not light isn't focused on that plate properly, 8 is also slightly visible (can be avoided by better LED placement)


9: etched: 10 plates (seen behind 8 plates) - thickness of the plates is causing problems. Problem may be improved with drilled numbering instead of etching (as light from rear plates may travel along holes in front of it: we shall see!)


2: etched (before I tried drilling it): 10 plates (black background): note 3 is also partly visible due to bad LED Placement


3: etched: 10 plates (black background):note 2 is also partly visible due to bad LED Placement

[cprobertson1]

Does anybody know what the part number for these LED Filaments (or similar) is?

I cannot find datasheets or reliable suppliers for them I might try making my own by soldering 0603 or 1005 LEDs to a wire (or two wires) and coating it all in resin: that way I can make them in arbitrary lengths; we'll see how badly that turns out - but I figured I'd ask around as I'm struggling to find them anywhere excepting weird chinese parketplaces.

[kwass]

Does anybody know what the part number for these LED Filaments (or similar) is?

How about this one:  http://www.runlite.cn/userfiles/5cjerp755q19k1413258994.pdf

[codeboy2k]

Does anybody know what the part number for these LED Filaments (or similar) is?

I cannot find datasheets or reliable suppliers for them I might try making my own by soldering 0603 or 1005 LEDs to a wire (or two wires) and coating it all in resin: that way I can make them in arbitrary lengths; we'll see how badly that turns out - but I figured I'd ask around as I'm struggling to find them anywhere excepting weird chinese parketplaces.

It would be interesting to break one apart... maybe it's been said already, but I think these are strings of chip LEDs (blue?) in series to get the high voltage, like 60V - 80V and then the tube is white phosphor coated, just like the white LEDs.

[cprobertson1]

It would be interesting to break one apart... maybe it's been said already, but I think these are strings of chip LEDs (blue?) in series to get the high voltage, like 60V - 80V and then the tube is white phosphor coated, just like the white LEDs.

Definitely interesting! Probably one of the first things I'll do with them

If they look good when mounted in a glass jar in a numeric-display shape I was going to try making a simpler version using a string of SMD LEDs and soldering them (in opposite directions) across two strands of copper wire (like a ladder) - varnishing it all and then twisting it into a helix before coating the entire lot in a transluscent resin (to act as a diffuser - might even try adding some aluminium oxide to aid in the diffusion) (this would perforce put them in a parallel configuration which is not something I'd usually do - I would much prefer to put them in serial, which is unfortunately more complex as it'd involve soldering them all onto one wire and then trimming off the wire between the two pads on the same LED to prevent a short. I'm also not too sure how much strain an SMT

If it works, cool, if it doesnt, meh it was only £1 I wasted (and I'll have a bucketload of SMD LEDs for other projects left over anyway)

Anyway - still waiting for my LED filaments to arrive, and the R&D into the sidelit acrylic is going well - going to try drilling holes next: will probably have another prototype by tomorrow!

[cprobertson1]

Tried a the through-hole design: while it doesn't look too bad, it's not bright enough for any real daylight usage... and doesn't look particularly nixie-ish =/ Oh well! Dem's the breaks!




And yes, that is a spice jar it's living in

It a decent view when the diffuser and hex-grid aren't mounted - I might keep these in mind for future experiments!

Here's a link to this series of experiments

Im going to (temporarily) abandon this line of development while I try playing with LED matrices and diffusers - if I get the masking process right I might actually be able to make quite sharp-looking glowing numerals.

I'll need to wait for my SMT LEDs to arrive - and I've also a motor controller to build (just a simple NE555 with a pot for the duty cycle running through a transistor linked to a ON-OFF-ON switch to control motor direction - nothing too fancy) - so it might be next weekend before I post anything else up!