INS-1 and IN-3 Nixie bulb panel mount holder

[Sparkst3r]

I have a bit of a sticking point with my project. It's a reasonably(I say reasonably, relays are expensive even those from china) turing complete computer built with relays.

As you may have guessed, the displays and indicators are nixies of some form or another - IN-12B, IN-3 and INS-1s.
I'm comfortable in the way I'm going to mount the IN-12Bs on a panel by using sockets however the accompanying decimal points which use IN-3s have long flexible non socket-able leads. I really don't want to solder them onto a board mainly because I can't bear to cut the leads off of the poor things but it's mainly in case they need to be replaced in the future. The binary displays use INS-1s for front facing panel mounting and have the same problem.

I've thought about LED panel holders for the INS-1s but I can't find any 7mm holders that would fit them. Plus the IN-3s need to be vertical and stand up from a base.

They are used in nixie clocks very frequently so I wonder if any of you have come up with a solution to this problem.
It's a darn shame they never made sockets for them in the past. And I haven't got the equipment to make bakelite to make a nice looking socket either.

Thanks for any suggestions!

[Paul Parry]

With the INS-3 I usually put them inside a little test tube, and fill with clear resin. The pins I attach to a 2 way IDC header and you can then plug and unplug them - the lower part of the test tube is usually hidden away inside the casing of the clock so you don't see that part, just the top with the INS-3 in there. For the INS-1 I sometimes make a coil of wire that the tube slides into and use that to support the tube between a couple of Nixies, but you can only really get away with that on a Steampunk type clock as it fits with the design. Have a look at rubber grommets for mounting INS-1's you get them in various sizes and they're designed to stop wires chaffing when passing through metal sheet - they will do a great job.

However, you may well be changing out INS-3's pretty often, they are not the most reliable of things. After a few months they will start to flicker, occasionally you get one or two that work perfectly for ever, but I stopped using them for clocks, and just use regular neon indicators.

How far along are you with the computer build?

Cheers,

Paul

 

 

[David Forbes]

The wire leads on these often fit in standard IC socket strips. Cutting
the leads short to fit the socket is not a crime, as the tube will wear
out in this machine, so its future use is not of interest.

The top "nipple" end of the tube may be held by a horizontal plate with
a small grommet in it. This was standard practice back then.

On 4/1/15 3:38 PM, Sparkst3r wrote:
> however the accompanying decimal points which use IN-3s
> have long flexible non socket-able leads. I really don't want to solder
> them onto a board mainly because I can't bear to cut the leads off of
> the poor things but it's mainly in case they need to be replaced in the
> future. The binary displays use INS-1s for front facing panel mounting
> and have the same problem.
>
> I've thought about LED panel holders for the INS-1s but I can't find any
> 7mm holders that would fit them. Plus the IN-3s need to be vertical and
> stand up from a base.
>

--
David Forbes, Tucson AZ

[Sparkst3r]

Thanks for the suggestions guys!

The computer is intended to look like it came out of the early 20th century
with visible wires, interconnects in wooden cabinets!
For a brief example, a module has a deep brown stained wooden facade with
an etched brass control panel filled with plenty of buttons, toggle
switches and neon.
As for the progress on the functionality, it's slow but steady. Most of the
ALU functions are either built or designed and ready to be built along with
the registers and some of the RAM. It should be capable of about as much as
the little man computer by the first working prototype. I'd quite like to
make it capable of controlling a stack but that may be quite far into the
future.

I have indeed noticed the IN-3s start to flicker after a while, in fact one
or two were flickery straight out of the box. If they are as unreliable as
Paul suggests I may buy a second box.



I do intend to go for a 'steampunk'-esque theme so the wire supports may be
an appropriate solution for the IN-3s, though I'll have to experiment to
get it to look nice. A quick test proved the wire is rigid enough to hold
the pinched end of the tube and spiral down over the wires like a spring to
hide or at least obscure them. My local electronics shop sells grommets
with strain reliefs cheaply and they fit over the nipples of the tubes and
obviously can be mounted in a hole drilled into the brass ceiling of the
cavity in the cabinet for the tubes.

Paul you may be onto something with those grommets to hold the ins-1s, The
6mm grommets I mentioned with the strain reliefs hold them really snuggly!
Thanks for that suggestion!

I have a few old Cambion DIP sockets that I got for pennies when a local
factory went bust that seem to work l but sadly I don't have enough for all
the bulbs and they cost a fortune on ebay. The modern ic sockets are a bit
brutal on anything you put into them so I much prefer the cambions for old
delicate ICs, they're fantastic little sockets, it's a shame they're out of
production now!

[image: http://www.surplussalespa.com/catalog/images/703-4051-01-0316.JPG]

On to some good news!

I tried using screw terminals and that proved to be rather successful. Also
as it turns out, corrugated cardboard is really handy. I'll try and get a
picture later of them lit up if I can put a temporary base together by then.



Thanks to both of you for the help!

[gregebert]

If you can use NE-2H bulbs instead, they have a rated lifetime of 25,000 hours when operated at the specified current.

My gut instinct tells me you will get better lifetime operating neon bulbs from a balanced AC source vs. DC (assuming identical RMS currents) because both electrodes are used, and any impurities inside the bulb would likely be distributed on both electrodes for AC, rather than on a single electrode for DC. Same for sputtering.

My big clock has 306 NE2HT bulbs and has been in service for over 2 years now; zero bulb failures. At any given time, only 14 of 306 bulbs are on, but in the past 2 years there is a total of 250,000 "bulb-hours" with no failures. I'm excluding 2 that were dead at first power-on.

[Sparkst3r]

I know of bulbs in projects my father built that are older than myself(16, that equates to well over 140,000 hours) that are running today with just a small flicker and barely any sputtering on the glass.

My only quibble with using AC neons is that they have 2 electrodes that glow rather than one making it look more like an equals sign than a solid glow of a decimal point.