An interesting illuminator from an old photocopier

[TerraHertz]

Back in May I dismantled an old photocopier. Since then I've been busy with a bit of a workshop upgrade and other things, so assorted minor projects lay neglected. Now I'm done with that, and cleaning up the detritus of small jobs that accumulated in the last two months. One piece of junk that floated to the surface, was the scanner light unit from that photocopier. Discarding the plastic molding, this is what's left:

The tube seemed curious, unlike a standard fluorescent light.

There are no conductive penetrations through the glass! The inside is completely electrically isolated, and it works purely by capacitive effects on the contents.


The story and more pics: http://everist.org/NobLog/20140702_green_glow.htm

I'd never seen a light like that before. Now of course LEDs have superseded such things, but it's nice to see that for a while photocopiers contained something like a CRT.

[bktemp]

Interessting electrode design. I have never seen something before. Maybe it has a better life expectancy because no internal electrodes that degrade over time.
Except the odd electrodes, green flourescent tubes seem to be common on old photocopiers. I have some tubes laying around having normal electrodes but a similar phosphor coating, covering not the whole tube, but leaving a gap for the light to come out.

[SeanB]

That one is interesting, normally you have a standard fluorescent tube there with a slotted phosphor coating to get the light. Having a cathodeless tube that runs via Rf drive is quite an interesting variant, and not common. It probably has a fill gas of a penning mixture, as that will have the lowest excitation voltage, and likely there is a drop of mercury inside to provide the vapour that will provide the UV to excite the phosphor. Likely the end metallisation that is used as an amalgam reservoir.

[ovnr]

Neat.

This reminds me a bit of EEFLs, the weird stepchild of CCFLs. (That's External Electrode Fluorescent Lamp and Cold Cathode Fluorescent Lamp, by the way). For a few years before LEDs took over for LCD backlighting, they seemed to be making inroads in the market, without very much fanfare. They boast a bit longer lifetime and little edge damage over time, something CCFLs have large issues with (burnt phosphor + sputtered metal from the electrodes).

Here's a presentation on them. The voltage seems pretty similar to the tube you have - only real difference is that EEFLs are only coupled at the ends, not across the sides like your tube is.

[TerraHertz]

That one is interesting, normally you have a standard fluorescent tube there with a slotted phosphor coating to get the light. Having a cathodeless tube that runs via Rf drive is quite an interesting variant, and not common. It probably has a fill gas of a penning mixture, as that will have the lowest excitation voltage,

That part I agree with.

and likely there is a drop of mercury inside to provide the vapour that will provide the UV to excite the phosphor.

And this part not. So far I haven't thought of a way to be sure about this, but I don't think that the usual 'excited gas emits UV, which excites the phosphor' is the process happening here. There's definitely a gas inside, since in a small area at the wired end there is a slight 'squirming' of the light output from the phosphor, that's very typical of dynamic gas/ion instability. And yet at no point in the tube is there any visible indication of a gas glow. It looks like ALL the light output is from the phosphor, and very evenly across the surface.

One possibility that occurs to me is it's a purely electroluminescent effect - a total vacuum in the tube, and the phosphor excited directly by the field. But in that case you'd think the illumination would be brightest in the film on the wall between the electrodes, not uniform over the whole surface. Also why the little metalized strip at the end? Maybe...
In any case the light looks very uniform. So I think there's a gas, with just enough density for some collision scattering of electrons, to achieve the even light output.
Interestingly, holding strong magnets next to the tube has no visible effect at all.  Ha ha... while being very careful not to get 2300VAC through the fingers. There's an outer plastic film, but it's extremely thin.
The lack of effect from magnets suggests to me the electron mean free paths are short, and quite randomized by collisions. But I'm just guessing.

I can't think of any way to be totally sure which mechanism is correct. Can you?

Hmm... I could try cooling half the tube in liquid nitrogen. That should freeze out any mercury vapor.
Wish I had a spectrometer, that went up into the UV. Or rather, a working one. I have a couple of Shimadzu QV-50  Spectrophotometers for restoration, but they are a long way from working. Some degraded front-silvered mirrors, that I don't know how to restore. They are not flat, so need to be de-metaled and re-silvered.

Likely the end metallisation that is used as an amalgam reservoir.

If there's mercury...
My thought on the reason for it, is it forms a point of higher E-field intensity at the ends of the small strip, to ensure ionization begins. It definitely does not go all the way round inside; just stops with a short overlap inside the outer electrodes. If its purpose was as an amalgam mercury source, why not run it right around? Of course it could serve both purposes.

There might be one at the other end too, but I can't see there due to the glued-on fitting.

[rs20]

My first reaction when I saw this was also "this looks just like EL wire"! That seems like a rather large plate spacing for a capacitor to conduct enough AC current to ionize gas, but who knows. I guess you could destructively test it by drilling a hole in the tube...

[IanB]

So far I haven't thought of a way to be sure about this, but I don't think that the usual 'excited gas emits UV, which excites the phosphor' is the process happening here. There's definitely a gas inside, since in a small area at the wired end there is a slight 'squirming' of the light output from the phosphor, that's very typical of dynamic gas/ion instability. And yet at no point in the tube is there any visible indication of a gas glow. It looks like ALL the light output is from the phosphor, and very evenly across the surface.

If the excitation is UV, and the phosphor very bright, it is possible you wouldn't be able to see the UV (i.e. black light)?

[TerraHertz]

If the excitation is UV, and the phosphor very bright, it is possible you wouldn't be able to see the UV (i.e. black light)?

Not sure. In ordinary fluorescent lights you can look in through the small area near the end caps that isn't coated, and see the faint purplish mercury ions conduction path. In this tube, because the inner phosphor coating stops near the end before the outer electrodes do, there's some 'excited gas volume' visible without the bright phosphor behind it. Especially if looking diagonally into it, like in the pic below. But I can't see any glow at all in the tube volume. The photo isn't that great, it's much clearer to the eye.



Also, having a tube with raw UV output, in a consumer appliance would seem to be something safety regulations would forbid. Though glass attenuates UV, and the glass scanning bed of the copier might be enough to block the UV. The very thin glass of the tube itself probably isn't going to attenuate UV much, same as with the fluoros used in UV EPROM erasers that simply don't have the inner phosphor coating. Ha ha... which btw, as an experiment I once exposed the back of my hand to at close range for about 10 seconds. Don't try this, unless you want an _extreme_ case of sunburn.
I was never stupid enough to try looking at an eraser tube.

Hmm. Has anyone tried taking a pic of a running EPROM eraser tube with a digital camera? The UV shouldn't hurt the sensor, if it could even get through the glass lenses and filter layers?

Still have three old EPROM erasers, including a big home-made one whose safety interlock is easily defeated. I could put this tube next to an eraser tube, see if the UV excites the green phosphor. The unknown being, how much does the glass of the photocopier tube attenuate UV?
I only have one nice digital camera atm and can't afford to replace it. Anyone know if it's safe to expose it to hard UV?

[rs20]

Still have three old EPROM erasers, including a big home-made one whose safety interlock is easily defeated. I could put this tube next to an eraser tube, see if the UV excites the green phosphor. The unknown being, how much does the glass of the photocopier tube attenuate UV?
I only have one nice digital camera atm and can't afford to replace it. Anyone know if it's safe to expose it to hard UV?

I think the question to ask is whether this experiment would prove anything useful -- even if the phosphor does fluoresce in response to UV, that doesn't prove that that's the mode of operation of this tube. I feel like UV-excited phosphors and electric-field-excited phosphors aren't mutually exclusive at all.

[mikeselectricstuff]

I found one of these in an old scanner - surprised it isn't more common as removing metal from the envelope makes tubes more reliable and presumably cheaper. Maybe they couldn't make them work as well with the white needed for colour scanners.

ISTR it gave pretty bright flashes from static, and tesla coils.

[SeanB]

They could have doped the glass to absorb more UV, or have evaporated an inner coating that is a UV block. Possible they did not use mercury, but simply added a thin conductive coat to the top of the phosphor layer to increase the capacitance, with the backing providing the right amount for the other plate. Then they would not have needed any special filling, and could have used plain Ar or Ne just as a filling to keep the interior dry and non degrading.

My first Eprom eraser was very simple, one old mercury vapour lamp inner quartz capsule with a 20W lamp ballast to limit the current to keep it in a glow discharge only, with the starting probe still connected. It was enough to erase a single prom in around an hour, and was in a small box with a block of black antistatic foam to place the chip under the lamp. The foam was rather degraded after a few hours of use. The quartz capsule ran quite hot, but was not hot enough to melt the plastic box. I upgraded to a light box with 3 UV tubes later, that has a ZN1034E timer in it to control the lamp on time.

[TerraHertz]

I think the question to ask is whether this experiment would prove anything useful -- even if the phosphor does fluoresce in response to UV, that doesn't prove that that's the mode of operation of this tube. I feel like UV-excited phosphors and electric-field-excited phosphors aren't mutually exclusive at all.

That's true. Ever used a party blacklight in a room with several old CRT scopes? Some of the screens really glow. Others don't.
Also, even if there's NO green glow under UV, it still doesn't prove anything since the walls of the tube might (somehow) be UV-opaque.

But I went ahead and did the experiment anyway.
Result: Under intense UV, no sign of any green glow. See pics.

I'm thinking it's pretty unlikely the tube walls can be totally UV-opaque, so this strongly suggests the tube does not use a phosphor that requires UV excitation. That leaves direct excitation by the AC field, or excitation by electron impacts.

Incidentally, the old eraser in the pics really does have an interlock. I'm using the end of the small fluoro tube to hold down the microswitch lever that is normally depressed by inserting the tray.

[TerraHertz]

ISTR it gave pretty bright flashes from static, and tesla coils.

Ah ha! Thanks for reminding me. Back in my days of designing for arcade games, an invaluable 'stress test' device was a piezo spark lighter. Because that's what the customers would use to try and crash the machines, for uh... benefits.
I still have that. It's modified to have an adjustable spark gap, and two output wires - the 'body' and the high side.
Setup shown in 1st pic below.

Very interesting and pretty definitive result. It sure does light up the tube, but most significantly, there are distinct 'streamers' perpendicular to the tube electrodes. They're quite bright and well defined but very brief. Some of them are very sharp edged, but they vary - some diffuse, a few sharp. They occur many times for each spark button depress, but occur in random positions. As with all the light from this tube they appear to be on the phosphor surface, not in the volume of the tube. But being so small and fast, it's hard to be entirely sure.

To me they look like the result of Birkeland current filaments terminating on the phosphor and making it glow in lines. They're occurring in this setup due to the sharp risetime from the spark gap, and high impedance of the source. So the gas goes from insulating (no free electrons) then breaks down locally and a current filament forms. But the drive isn't sufficient to cause the ionization to spread much via electron collisions with gas atoms. The filament just dies away.

So now I'm about 90% convinced this thing really is working via electron collisions with the phosphor.

There's also a background of more even glow, that makes it hard to photograph the streaks, but I tried. The photos don't really do them justice.

It's pretty!

It also reminds me of a waveguide. Trying to think of a way to apply something like the 25KHz HV AC but just below the ionization threshold, PLUS a much high lower voltage RF signal at high enough frequency to get standing waves on the electrodes. Without either source blowing up the other one. Idea being to adjust things so the gas breaks down at the standing wave nodes, not the nulls.
Because it would be cool to be able to directly see standing waves.

Now I'm going to look for more of these old copiers. Having only ONE tube sucks. If I break it.

[SeanB]

What copier was it? I will look out for them then as well.

[bktemp]

I'm thinking it's pretty unlikely the tube walls can be totally UV-opaque, so this strongly suggests the tube does not use a phosphor that requires UV excitation. That leaves direct excitation by the AC field, or excitation by electron impacts.

I think you are wrong:
- I have a similar looking lamp but with electrodes. I can clearly see a bluish light from the area around the electrodes like in a normal fluorescent tube.
- A UV-C lamp does not produce the slightes amount of green night. Neither does a normal white flourescent tube give any light output when held next to the UV-C lamp. This confirms that even the thin glass wall absorbs the 254nm emitted by the mercury vapour completely. Otherwise a protective screen would be necessary around every fluorescent tube.
I have green tube with broken electrodes. If I can find it, I am going to open it to remove the glass.

[josem]

This NEC patent seems to be a good match:

http://www.google.com/patents/US5117160

Even suggests intended use for office automation equipment.

Perhaps with the striation improvement from HP:

http://www.google.com/patents/US5760541

This one from a spare parts seller looks like a simpler model of that: http://www.donberg.fr/catalogue/pieces_detachees_ordinateurs/cfx12ayg-36h.html

[TerraHertz]

I'm thinking it's pretty unlikely the tube walls can be totally UV-opaque, so this strongly suggests the tube does not use a phosphor that requires UV excitation. That leaves direct excitation by the AC field, or excitation by electron impacts.

I think you are wrong:
- I have a similar looking lamp but with electrodes. I can clearly see a bluish light from the area around the electrodes like in a normal fluorescent tube.

That's with mercury vapor. Spectrogram for mercury:

That's high pressure mercury, but much the same for low pressure. Lots of UV, some peaks in the visible range, so you can see a bluish glow in the gas.

- A UV-C lamp does not produce the slightes amount of green night. Neither does a normal white flourescent tube give any light output when held next to the UV-C lamp. This confirms that even the thin glass wall absorbs the 254nm emitted by the mercury vapour completely. Otherwise a protective screen would be necessary around every fluorescent tube.

However the phosphor layer is completely UV-opaque. No extra shield required. UV-fluoro tubes like in EPROM erasers are exactly the same except the phosphor is omitted.

josem Thanks, those are very interesting.
This is getting more and more fascinating. There are several details in those patents that are ... curious.
A minor detail is their term 'Xenone gas'. I can't find any reference suggesting this is not plain Xenon, so it's probably just patent-drone bullshit jargon.
Much more interestingly, there's something wrong with their description of the operating principle of these lamps.
They get it right for mercury fluorescent lamps, stating:
"Fluorescent lamps of the type typically used for household and office lighting have internal electrodes. Heating the electrodes causes thermoionic emission of electrons. Providing a high voltage from one electrode to the other causes electron flow between the electrodes. The electrons then excite mercury atoms. The excited mercury atoms release their acquired energy in the form of ultraviolet radiation. The ultraviolet radiation excites a phosphor coating on the interior of the lamp, resulting in emission of visible light."

But for these lamps, take this example statement:
"When an RF voltage of 20 to 100 kHz and 1 to 2 kV is applied to the two belt-shaped electrodes, discharge of xenone gas is generated in a discharge space in the bulb in a direction perpendicular to the bulb axis, and the phosphor film on the inner wall of the bulb is excited to emit light. At this time, since the discharge gas is xenone gas not containing a metal vapor such as mercury vapor, the phosphor film is efficiently excited by exciting light (147 nm) of xenone gas to realize a high luminous intensity."

What?! What is this 'exciting light (147 nm) of xenone gas'? Assuming they mean Xenon (and since it's an inert gas that doesn't like to bond with anything else including itself, what else could it be), then electrically excited xenon gas should produce the typical xenon spectra. Which is this:


Point is, Xenon has a lot of visible emission, not much UV. If there was Xenon in this tube and it was being excited to photon emission by the electron motion, it would be emitting visible light. More than UV. It would definitely be visible.
So that's problem #1 with the patent.

The next one is the 'striations' they mention. Yes, this tube has the serrated electrodes the 2nd patent mentions, as used to stop the striations wandering. But... think about it. If the tube is operating by current flow exciting the gas, which emits UV, and the UV excites the phosphor, how can striations in the bulk gas current produce fine striations in the light emitted by the phosphor? The gas ions would emit photons in all directions, so the effect should 'blur out' by the time the photons reach the phosphor on the walls.
Their description doesn't make sense.

On the other hand, where electrons impinge on the walls certainly can be confined to sharply defined regions, under some operating conditions.

I'm wondering if they deliberately obfuscated the mechanism, or if they simply made bad assumptions based on old style fluro-tube mechanism? It's very suggestive that they nowhere mention Birkelund filaments/currents, which are the fundamental process causing non-uniformity of the current through the gas. It's also true that plasma physics is very little taught anywhere, and has some very counter-intuitive behaviors. Perhaps it's possible that the people developing these tubes simply didn't understand how they work? Notice the amount of purely experimental fiddling they mention in the 2nd patent.


About the photocopier type. I was kicking myself that I didn't make a note of the model, and have since thrown out most of the case. But on checking, the case piece with the model sticker was still in the pile to be binned. The unit was:
Sharp Corp. Digital Copying Machine AR-201.
The inverter PCB says: USHIO PXZ170E1
There's no marking on the tube itself, but the attached connector has: USHIO UXFL-08YG37D

Oh, and dates in the machine's service card say it was in use from 2004 to 2009.

[bktemp]

I have searched for 147nm in xenon discharges but could not find anything usefull except this patent about flourescent materials:
http://www.google.com.ar/patents/EP0738311B1?cl=en
They talk about phospors for generating colored light using 145nm-185nm. At the end they describe an example lamp using xenon and it seems to be quite efficient at 40lm/W. So xenon must emit    somewhere in the UVC range to make this work.
147nm from xenon seems to te related to excimer emissions.

[TerraHertz]

Chase continues.

This: https://www.ushio.co.jp/documents/technology/lightedge/lightedge_27/ushio_le27-04.pdf
is all in japanese, but diagrams suggest it discusses this lamp type. Here's one showing the Xenon process:

How frustrating. Can't find an English version of that pdf. On their site they have _some_ in English, but not that one.
http://www.ushio.co.jp/jp/technology/lightedge/index.html
http://www.ushio.co.jp/en/technology/lightedge/index.html#paper_list   english list  le27-04 not there.
http://www.ushio.co.jp/jp/technology/lightedge/lightedge_27.html       japanese. Includes le27-04

The   le27-04 description translated by google:
 Chapter 4 Basic knowledge of each lamp PDF (1044KB) 4.1 xenon lamp 4.2 xenon flash lamp 4.3 ultra-high pressure mercury lamp 4.4 projector for ultra-high pressure mercury lamp (NSH) 4.5 capillary lamp (capillary type ultra-high pressure mercury lamp) 4.6 high-pressure mercury lamp 4.7 low-pressure mercury lamp 4.8 Mercury xenon lamp (Deep UV lamp) 4.9 metal halide lamp
And then relevant:
 - 4.10 external electrode fluorescent lamp gas ShikiNozomi
 - 4.11 dielectric barrier discharge excimer lamp

google: dielectric barrier discharge excimer lamp
 http://en.wikipedia.org/wiki/Dielectric_barrier_discharge
 http://www.coronalab.net/wxzl/plasma-16.pdf
 http://en.wikipedia.org/wiki/Excimer_lamp     Excimer lamp. Lists wavelengths for different gases.
           Xe2*   172nm   7.21eV  (same as in the Ushio diagram.)

172 nm is a lot different to 147nm.  Also I'd still like to know, if the phosphor is excited by far-UV photons from the gas discharge in the tube body, how come the phosphor surface can glow in fine lines? Sounds like multiple things going on here.

[josem]

The chemistry is way over my head but isn't this the same principle (excited high pressure Xeon gas generating UV which then glows phosphors) used in display cells of plasma TVs?

Also a search for "147nm resonant Xe* emission" turns out a lot of research papers that seems to point that is the wavelength.