Re: Helping Nixie Tubes Fire in a Darkened Room

[gregebert]

UV light source, such as from an LED ? 

On Wednesday, November 29, 2023 at 6:43:48 PM UTC-8 Jeff Walton wrote:

I've recently come across a situation where I have some tubes in a clock that are being directly driven and are having trouble starting when the room is darkened but light right up when a room light is turned on.  These particular tubes were probably intended for use in a calculator.  They are seven segment neon MG-17G tubes.  Once the tubes have any of the segments lit, there is really no issue with the performance.  It's when the tubes go completely dark if a space is used while scrolling a message or lighting a dash on and off to emulate a colon.  I'm wondering if others have found any particular tricks to help convince tubes to light up.  There is no "baselighting" and the HV is ~172v.  I'm considering increasing the HV by 10-15v but don't want to over drive the tubes.  Short of putting a radioactive source in the vicinity, are there other things that anyone has had any luck with?

Jeff

[Jon]

I think the best approach would be to increase the HV. If you look at the discussions of glow physics in Weston or Acton for example, the lag in striking a glow from dark (which is what we're talking about here) is reduced by using an anode voltage materially above the threshold striking voltage. Overdrive is a matter of tube current rather than anode voltage per se, so a simple compensatory tweak to the current-limiting resistor value will get you back to the same tube current as now. The cost of the change will be a little extra power dissipated in the current-limiting resistors, but that should be immaterial in most circumstances.

Jon.

[Paul Andrews]

Are they also dimmed using an LDR? If so, turn dimming off - if you have access to the source code, you could turn dimming off briefly when the clock turns on.

[Tidak Ada]

Why  not a trial with UV-LED’s 365 nm is a useful wavelength and you don’t need to use then at full power.

You can use them in a pulsed mode only during refresh of the display to avoid too much O₃ production. Iyt is not as hazardous as the use of ß-radiating isotopes as used in rigger tubes.

 

eric

--
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 on the web, visit https://groups.google.com/d/msgid/neonixie-l/ee14735e-1e40-4c6b-a815-77e670e1c89cn%40googlegroups.com.

[John]

I had a radio (can't remember the type,) that had a small neon bulb under the chassis ...and NE2 perhaps. I think it was used as a cheap and dirty voltage reference .   The thing was, when the chassis was flat on the bench in operating position, the bulb would never fire. When I tipped it on its edge and the bench light shined brightly on the undersides, it would fire fine as apparently the bench light induced just enough additional excitation to allow it to light.

It took a while to figure that out.

John

[Jeff Walton]

Not dimmed.  Tried to do a momentary hit on all segments to "wake" up the tube and it helped some but the extra flash is distracting.

[Jeff Walton]

Have some 365 and 395nm LED chips on order to experiment with.  Will try some things when I get them.  Hopefully they will help and I can hide them under the tube.  I saw some 245nm LEDs but super expensive and larger.

[Nicholas Stock]

May I suggest a small piece of uranium ore placed next to the clock?

LOL.

Just kidding Jeff! I like Jon's idea, up the voltage/series resistor.

--

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 on the web, visit https://groups.google.com/d/msgid/neonixie-l/69cbf16c-56c2-4785-b4d2-ab0f6bcb76e1n%40googlegroups.com.

[Jeff Walton]

I can try to modify the Omnixie HV module if higher voltage shows promise.  The module will be a little bit of a challenge as it has really small parts and I have a rework gun that might be on the large side and melt everything around the target resistor.

[Jeff Walton]

Give me the glow, Nick!

[gregebert]

I've never had problems with tubes failing to ignite, perhaps because I use higher-than-minimum supply voltages, typically around +200VDC. My first clock uses 340V.

[liam bartosiewicz]

If I remember correctly, there were some tubes manufactured that had a small external radioactive source to make them more reliable. It wouldn’t be very practical, or safe for that matter, but a radioactive nixie clock does sound pretty cool.

On Nov 30, 2023, at 1:43 PM, Nicholas Stock <nick...@gmail.com> wrote:



To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAOX%2BRHKPrgPAGPbF1%2B2d0hNNp8Vk%3DXGe5OP5V2PropW0Oy7Xnw%40mail.gmail.com.

[Nicholas Stock]

Some tubes were doped with Kr85. They have little radiation signs on them.... 122P224, Paul has a nice write up here (https://www.nixies.us/bwg_gallery/122p224/). Half life is 10 years, so a lot of decay since manufacture....

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/233BA8F1-AF6E-4CC8-B499-2A9ED404EC37%40gmail.com.

[Robert G. Schaffrath]

It would an interesting idea but the glass would block the ionizing radiation (alpha and probably most beta particles from daughter elements in the ore).

I have the same issue with B-7971's in a dark room. Unless I shine a flashlight at them, it takes them a bit of time to turn on from a cold start.

[Paul Andrews]

You could have a high value resistor permanently connected between the cathodes and ground so there is always a minimum current flowing through them?

[Mac Doktor]

On Nov 30, 2023, at 4:42 PM, Jeff Walton <jwalt...@gmail.com> wrote:

Have some 365 and 395nm LED chips on order to experiment with.  Will try some things when I get them.  Hopefully they will help and I can hide them under the tube.  I saw some 245nm LEDs but super expensive and larger.

I'd stay away from 245nm. It's very dangerous plus the glass is opaque at that wavelength. 365nm can be hazardous as well. Some of these UV LEDs are really bright and you can't tell.

Terry Bowman, KA4HJH
"The Mac Doctor"

https://www.astarcloseup.com

"If only you could see what I've seen with your eyes."—Roy Batty, Blade Runner

[Jeff Walton]

I've done some testing with common 3mm UV LEDs that are in the 395nM range.  Running them at ~5mA provides a dim purple color light that when placed under or in front of the tube actually makes the reliability of the ignition very stable.  Even in a fully darkened room, the segments and decimal points light reliably.  

This particular tube (MG-17G) has segments with no anode screen in front of the tube, so they seem to behave a little differently than a regular nixie.  I also found that if I placed fingers on the sides of the tubes while operating, the segments lit more reliably.  I think that I will go for an LED under each tube but powered just enough to make the tubes work better and not to try to supply any visible base lighting.  I will try 365nM LED's when they arrive and may use them if there is less noticeable, visible light.

Thanks for the suggestions!

Jeff

[Tidak Ada]

Hi Jeff, good to hear you are content with my suggestion.

Two remarks: I  bought at a time small 365nm transmission filter from a seller at either AliExpress or Banggood. That reduces the amount of visible light considerabele. Second, make that the UV flash only during a short period around the firing of the of the nixie segments. That will reduce the production of ozone considerable.

Cheers,

               eric

Verstuurd vanaf mijn iPhone

Op 3 dec. 2023 om 08:09 heeft Jeff Walton <jwalt...@gmail.com> het volgende geschreven:



--

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 on the web, visit https://groups.google.com/d/msgid/neonixie-l/cfde3460-1737-4506-a58e-a1a6401ca7cen%40googlegroups.com.

[Christian Riise Wagner]

You won't generate ozone with 365nm. You'd need shorter wavelength UVC for that.

[Tidak Ada]

The problem is that a 365nm  LED has a much wider spectrum, definitely producing ozone. I have a floodlight with a 365nm COB-LED that I can smell. OK It’s a 10W chip, but despite that.

Further, it’s fine to reduce the LED’s burning time to avoid their light and color. That’s the main reason to start  with only pulsing during transition of the figures. I think the UV pulse may be short enough that it will be not visible

 

The UV trick works also fine with dekatrons and other trigger tubes.

 

Cheers,

               eric

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/51995e71-f9b2-48c1-addf-fba4f10327e8n%40googlegroups.com.

[Jeff Walton]

Here is a closeup of the tube with a UV light source assisting...   395nM  LEDs.  Waiting for samples to test 365nM.

[Adrian Godwin]

Here's another approach, not really a viable retrofit to a nixie unless an extra electrode fixed to the glass might work ?

https://youtu.be/y2_dTEXptZQ?t=227

 

--

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 on the web, visit https://groups.google.com/d/msgid/neonixie-l/04812860-adfe-487b-8eb0-82d572e13df8n%40googlegroups.com.

[gregebert]

885-890nm ? That's infrared. If the LEDs are always-on, then I'm glad you're using IR instead of UV because it's much less harmful (perhaps harmless ?) to materials and humans/pets, etc.

BTW, those are really interesting tubes. I dont think I've ever seen tubes with a solid/opaque anode.

On Monday, February 19, 2024 at 5:19:12 PM UTC-8 Jeff Walton wrote:

I posted this a couple months ago regarding the MG-17G display tubes that wouldn’t fire reliably in the dark.  The final resolution for it was to mount some 885-890nM UV LED’s under the tubes.  A completely updated display board with LEDs driven from a fixed supply and a transistor on an LDR so that they LEDs only turn on in a darkened room.  This 885nM wavelength seemed to work better than the 865mM and was less visible than the 905nM LEDs.  I wasn’t trying for a visible underlighting effect on the tubes, so they are being driven at about 2mA so that there is sufficient light without excessive glow.

 

The clock is a Richard Scales design and this is the only one (so far) with the “keep alive” LEDs.  The tubes do not have a keep alive cathode, so this method works.  Otherwise, any single segment or decimal point would not fire reliably.  Now the tubes light dependably.  Thanks to Richard for the new board and a lot of back and forth to get this working right!

 

The original PCB:

 

The new PCB:

 

The display in operation:

 

The case in Place with LEDs on:

 

The finished clock:

 

Once we got past the issues with dark operation, the clock performs very well!

 

Regards,

 

 

 

Jeff Walton

---

jwalt...@gmail.com
jwalt...@msn.com 
 (952) 943-2064  Home
(612) 865-5560  Cell

 

From: neoni...@googlegroups.com <neoni...@googlegroups.com> On Behalf Of Jeff Walton
Sent: Wednesday, November 29, 2023 8:44 PM
To: neonixie-l <neoni...@googlegroups.com>
Subject: [neonixie-l] Helping Nixie Tubes Fire in a Darkened Room

 

I've recently come across a situation where I have some tubes in a clock that are being directly driven and are having trouble starting when the room is darkened but light right up when a room light is turned on.  These particular tubes were probably intended for use in a calculator.  They are seven segment neon MG-17G tubes.  Once the tubes have any of the segments lit, there is really no issue with the performance.  It's when the tubes go completely dark if a space is used while scrolling a message or lighting a dash on and off to emulate a colon.  I'm wondering if others have found any particular tricks to help convince tubes to light up.  There is no "baselighting" and the HV is ~172v.  I'm considering increasing the HV by 10-15v but don't want to over drive the tubes.  Short of putting a radioactive source in the vicinity, are there other things that anyone has had any luck with?

 

Jeff

--

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 on the web, visit https://groups.google.com/d/msgid/neonixie-l/d5713f02-ec1d-493d-b12f-644fb080ee71n%40googlegroups.com.

[Jeff Walton]

Hi Greg, 

The LEDs were 385nM, not 885nM.  

Clearly the LEDs were UV.  I initially thought the 365nM would be more effective but they were not.  The 365nM parts might have also been less desirable from a safety standpoint and were also quite a bit more expensive.  The 405nM UV LEDs are less expensive and more plentiful but seem to have more visible light spill. 

I wish that I could edit the original post, but no such luck!

Jeff 

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/bb3b4ba6-8597-48c9-af56-c0a9e3de97c7n%40googlegroups.com.

[gregebert]

No problem; amazing what you can do with 500nm.....

Since they actually are UV LEDs, you may want to tweak the software so they can be turned off after a few seconds if your clock is direct-drive. If it's multiplexed, you probably need the UV on continuously in the dark (would be very informative to run an experiment here to find out....).

[Jeff Walton]

Repost with correction to the LED information: 

I posted this question a couple months ago regarding the MG-17G display tubes that wouldn’t fire reliably in the dark.  The final resolution for it was to mount some 385-390nM UV LED’s under the tubes.  A completely updated display board with LEDs driven from a fixed supply and a transistor on an LDR so that they LEDs only turn on in a darkened room.  This 385nM wavelength seemed to work better than the 365mM and was less visible than the 405nM LEDs.  I wasn’t trying for a visible underlighting effect on the tubes, so they are being driven at about 2mA so that there is sufficient light without excessive glow.  I also tried 10-15v increased HV voltage but it did not really help.

 

The clock is a Richard Scales design and this is the only one (so far) with the “keep alive” LEDs.  The tubes do not have a keep alive cathode, so this method works.  Otherwise, any single segment or decimal point would not fire reliably.  Now the tubes light dependably.  Thanks to Richard for the new board and a lot of back and forth to get this working right!

The original PCB with no provision for adding LEDs:

 

 The new PCB (unpowered) with 3mm LEDs added to aid dark starting:

 

 

The display in operation (with UV LEDs active):  The camera sees more light than I do…

 

 

The acrylic case in place with LEDs on:  The LEDs are mush less apparent with the top acrylic cover – 5mm thick

 

The finished clock (with added LDR to run UV underlighting only when in darkened surroundings):  The blue is from another clock and the blue silicon mat below.

 Once we got past the issues with dark operation, the clock performs very well!  Thanks to everyone from the group that made suggestions! 

Jeff

[Jeff Walton]

The tubes are direct drive but the issue is that the tubes which display only a dash (instead of a colon) are cycling on and off every second, so the LEDs would be constantly turning on and off.  It is only when one segment or DP is energized that the dark turn-on is an issue, so it would be too noticible to be doing that with the LEDs.  We did try to send millisecond hits to all of the tube segments on each illumination of the dashes.  The thinking was that if it was brief, it might not be noticible.  In practice, the other segments had to be powered long enough to actually ionize the gas and they do not turn on and off fast enough to avoid a very visible (and annoying) flash.  

[newxito]

The clock looks great!

It seems we have the same PCB manufacturer. Only 402 orders so far? I'm already at 418 :-)

[Richard Scales]

That board was from weeks ago!!!

I had no idea things had got that serious.

My last order was No. 428 

... I need help....

- Richard