12V UV Germicidal Light Bulb

[Computer Nerd Kev]

A while ago I bought a pair of these 12V 3W UV-C light bulbs to use to
build an EPROM eraser:
https://www.aliexpress.com/item/3W-E17-UV-Germicidal-Light-Bulb-Intermediate-3-Watt-UV-C-Replacement-Bulb-for-GG1000-1100/32665735988.html
- http://tinyurl.com/h45emmj

There are more details in this listing:
https://www.aliexpress.com/item/3-Watt-UV-E17-Replacement-bulb-UV-C-Replacement-Bulb-for-GG1000-1100-Air-Sanitizers/32665910312.html
- http://tinyurl.com/m5rqmg9

SHORT STORY:
^^^^^^^^^^^^
One blew on 15VDC (did nothing on a lower voltage), the other blew
on 12VAC (did nothing on 9VAC). How are you meant to power them?

LONG STORY:
^^^^^^^^^^^
In my ignorance, I assumed these would work like normal incandecent or
halogen bulbs, so I mounted a bulb in a tin with a hole in one end
covered in layers of hot glue to act as a hopefully eye-safe viewing
window (I only view it indirectly as an added precaution). Then, unsure
of whether the voltage specification was AC or DC (and frankly, assuming
that it didn't matter), I slowly raised the DC voltage from my bench
supply.

There was no light, and until I reached about 15VDC only a few tens of
milliamps current. Then the current suddenly surged to amps (still no
light) and even though I made a hasty attempt to cut the power, I found
that one of the two vertically stretched filaments had blown.

DC clearly wasn't working out, so I tried AC, starting low with a 9VAC
plugpack and the backup globe (multimeter in series to measure current).
This gave me the same "no light, a few tens of milliamps" condition, so
I moved up to a 12VAC plugpack (I don't have a variac unfortunately).

Now I finally got light and decent current (although around .6A
of it, so more than 3 Watts). As it ran, the blue light got darker and
apparantly less visible, seemingly as the globe warmed up. The current
also slowly increased to .88A after two minutes, then dropped back to
.81mA by 2m 30s with the light still looking good. But current rose
again to .92A, then it seemed to go into thermal runaway at 3min 20
and rapidly spiked to over 1.2A before I could cut the power. Inspection
revealed that one of the filaments had now blown in this bulb as well.

I tested one of the half blown bulbs on another 12VAC plugpack, this
time only rated at 500mA instead of the 1.7A of the one I used before.
I could fill a page with the weird behavior that it showed in regards
to both light an current consumption, but the blue light never became
very dark, and the data on two EPROMs I put in was unaffected (checksums
matched) after forty minutes. Current varied between .25A and .4A during
operation.

So I clearly don't know the right way to drive these bulbs. Does someone
know how they're meant to be powered, or should I just give up and go
with one of the flourescent UV bulbs (which are unfortunately awkwardly
long for the small device I had in mind).

P.S. I know about the $20 EPROM erasers sold from China, but I have a
strict policy of not buying mains powered devices direct from
China. Especially devices that I'll be tempted to leave unattended
during use.

--
__ __
#_ < |\| |< _#

[Rheilly Phoull]

Well the top url states 10-12vdc, did you try reversing the polarity ?

[Computer Nerd Kev]

Rheilly Phoull <rhe...@bigslong.com> wrote:
> On 6/5/17 9:31 am, Computer Nerd Kev wrote:
>> A while ago I bought a pair of these 12V 3W UV-C light bulbs to use to
>> build an EPROM eraser:
>> https://www.aliexpress.com/item/3W-E17-UV-Germicidal-Light-Bulb-Intermediate-3-Watt-UV-C-Replacement-Bulb-for-GG1000-1100/32665735988.html
>> - http://tinyurl.com/h45emmj
>>
>> There are more details in this listing:
>> https://www.aliexpress.com/item/3-Watt-UV-E17-Replacement-bulb-UV-C-Replacement-Bulb-for-GG1000-1100-Air-Sanitizers/32665910312.html
>> - http://tinyurl.com/m5rqmg9
>>
>> SHORT STORY:
>> ^^^^^^^^^^^^
>> One blew on 15VDC (did nothing on a lower voltage), the other blew
>> on 12VAC (did nothing on 9VAC). How are you meant to power them?
>>

> Well the top url states 10-12vdc,

I can't find it in the top link, but I notice now that when you
select the "One Bulb" option on the other listing, the image
that's shown does say "DC10-12V". I don't think I noticed that
before.

> did you try reversing the polarity ?

I didn't make a note of trying that, but I tried just now with
the half-blown bulb that still makes a fairly bright blue light
on 12VAC (but doesn't erase EPROMS). At 11VDC there's no light
with either polarity, except that I notice now a dim red glow
which is presumably the filament getting hot.

The 11VDC current is now .24A, perhaps because some of the blown
filament is shorting something, but the current is identical in
either polarity.

[FMurtz]

could someone get one working and see if it will cure loca glue?

[Andy Wood]

The bloke in the link below seems to be talking about the same sort of
lamps. He suggests an AC voltage of at least 24V with a series
resistor to keep the current around 300mA.

http://russellsrandomthoughts.blogspot.com.au/2013/05/the-gtl3-bulb-simple-and-inexpensive.html

Andy Wood
woo...@trap.ozemail.com.au

[Computer Nerd Kev]

On 06 May 2017, Andy Wood wrote:

>
> The bloke in the link below seems to be talking about the
> same sort of lamps. He suggests an AC voltage of at least
> 24V with a series resistor to keep the current around
> 300mA.
>
> http://russellsrandomthoughts.blogspot.com.au/2013/05/the-gt
> l3-bulb-simple-and-inexpensive.html

Many thanks. with a 24VAC plugpack and a 10W 33 Ohm resistor in
series, what remains of both globes seems to work much better
than on straight 12VAC.

I found the specifications sheet mentioned here:
http://www.ushio.com/files/specifications/germicidal-low-press
ure-mercury-arc.pdf

The GTL3 globes discussed there seem to be missing the little
plate positioned between the filaments on mine, which seems to
be coated with something (perhaps to make it emit more visible
light, the spectral range in the specs for the GTL3 is
surprisingly narrow). Anyway, they're clearly both mercury
discharge lights and that makes a lot more sense of the
behaviour.

I'll have another go at erasing EPROMs with the damaged bulbs in
this circuit.

[Computer Nerd Kev]

On 06 May 2017, FMurtz wrote:

> Computer Nerd Kev wrote:
>>>> A while ago I bought a pair of these 12V 3W UV-C light
>>>> bulbs to use to build an EPROM eraser:
>>>> https://www.aliexpress.com/item/3W-E17-UV-Germicidal-Ligh
>>>> t-Bulb-Intermediate-3-Watt-UV-C-Replacement-Bulb-for-GG10
>>>> 00-1100/32665735988.html - http://tinyurl.com/h45emmj
>>>>
>>>> There are more details in this listing:
>>>> https://www.aliexpress.com/item/3-Watt-UV-E17-Replacement
>>>> -bulb-UV-C-Replacement-Bulb-for-GG1000-1100-Air-Sanitizer
>>>> s/32665910312.html - http://tinyurl.com/m5rqmg9
>

> could someone get one working and see if it will cure
> loca glue?

No LOCA glue here I'm afraid. From a breif search it looks
like that glue uses longer wavelengths than the narrow UVC
sort that EPROMs require. There might be easier sources of UV
light that you could use (ebay listings seem to indicate that
UV LEDs work).

[Computer Nerd Kev]

On 06 May 2017, Computer Nerd Kev wrote:

> On 06 May 2017, Andy Wood wrote:
>
>>
>> The bloke in the link below seems to be talking about the
>> same sort of lamps. He suggests an AC voltage of at least
>> 24V with a series resistor to keep the current around
>> 300mA.
>>
>> http://russellsrandomthoughts.blogspot.com.au/2013/05/the-g
>> t l3-bulb-simple-and-inexpensive.html

> [snip]

> I'll have another go at erasing EPROMs with the damaged
> bulbs in this circuit.
>

After forty minutes, I corrupted the data on a 27C256 and
completely erased a 27256, so that's a sure improvement.
Hopefully with a new, intact, bulb it will work even better.

Thanks again.

[Clocky]

On 6/05/2017 9:31 AM, Computer Nerd Kev wrote:
> A while ago I bought a pair of these 12V 3W UV-C light bulbs to use to
> build an EPROM eraser:
> https://www.aliexpress.com/item/3W-E17-UV-Germicidal-Light-Bulb-Intermediate-3-Watt-UV-C-Replacement-Bulb-for-GG1000-1100/32665735988.html
> - http://tinyurl.com/h45emmj
>
> There are more details in this listing:
> https://www.aliexpress.com/item/3-Watt-UV-E17-Replacement-bulb-UV-C-Replacement-Bulb-for-GG1000-1100-Air-Sanitizers/32665910312.html
> - http://tinyurl.com/m5rqmg9
>
> SHORT STORY:

<snip long short story>

long for the small device I had in mind).
>
> P.S. I know about the $20 EPROM erasers sold from China, but I have a
> strict policy of not buying mains powered devices direct from
> China. Especially devices that I'll be tempted to leave unattended
> during use.
>

I've never considered how they work, I just assumed it was similar to a
fluoro.

I've had one of those Chinese ones for years and it has worked
flawlessly. I believe it was ex-university as it wasn't brand new but it
was in excellent condition. Less than $20 at the time, so no regrets.

I bought a spare tube a few years ago but I'm still on the original one.

[Computer Nerd Kev]

Clocky <not...@happen.com> wrote:
> On 6/05/2017 9:31 AM, Computer Nerd Kev wrote:
>> A while ago I bought a pair of these 12V 3W UV-C light bulbs to use to
>> build an EPROM eraser:
>> https://www.aliexpress.com/item/3W-E17-UV-Germicidal-Light-Bulb-Intermediate-3-Watt-UV-C-Replacement-Bulb-for-GG1000-1100/32665735988.html
>> - http://tinyurl.com/h45emmj
>>
>> There are more details in this listing:
>> https://www.aliexpress.com/item/3-Watt-UV-E17-Replacement-bulb-UV-C-Replacement-Bulb-for-GG1000-1100-Air-Sanitizers/32665910312.html
>> - http://tinyurl.com/m5rqmg9
>>
>> SHORT STORY:
>
> <snip long short story>
>
> long for the small device I had in mind).
>>
>> P.S. I know about the $20 EPROM erasers sold from China, but I have a
>> strict policy of not buying mains powered devices direct from
>> China. Especially devices that I'll be tempted to leave unattended
>> during use.
>>
>
>
> I've never considered how they work, I just assumed it was similar to a
> fluoro.

Most are, I understand that all fluorescent tubes produce UV light, but
normally the white phosphorous coating on the inside of the glass blocks
this (while also producing more generally useful wavelengths of light).

Halogen bulbs also produce UV, so I guessed that the bulbs in question
were halogens that had been optimised for the purpose. Wrong entirely
as it turns out.

I preferred them because they're smaller and looked to be simpler to
rig up in a small device (and they still are, now that I know how).

> I've had one of those Chinese ones for years and it has worked
> flawlessly. I believe it was ex-university as it wasn't brand new but it
> was in excellent condition. Less than $20 at the time, so no regrets.
>
> I bought a spare tube a few years ago but I'm still on the original one.

Yeah, I don't doubt it (although I'm a little surprised that a
university wouldn't have better standards). I just figure it's safer
to have a blanket rule than try to work out exceptions.

Actually you'd have a hard time convincing me at the moment, because
after getting the DIY EPROM eraser sort-of working, now my cheap
Asian EPROM programmer is giving me trouble. You can't win...

[Clocky]

Dunno, they work well enough and it hasn't failed, that's a pretty high
standard for Chinesium stuff.

I just figure it's safer
> to have a blanket rule than try to work out exceptions.
>
> Actually you'd have a hard time convincing me at the moment, because
> after getting the DIY EPROM eraser sort-of working, now my cheap
> Asian EPROM programmer is giving me trouble. You can't win...
>

Heh ;-)

Which EPROM programmer do you have? My old faithful is an older Willem
model 3.something or other. The only thing to fail was a transistor but
that was probably my own fault and easily fixed.

Can't say it's great feature or quality wise, but it does the job quite
reliably.

[Computer Nerd Kev]

Well it's actually an old 80s programmer named "MEPX4" which connects to
a PC ISA port. It reads fine, and I programmed a chip with it not long
ago, but now it seems to just program the first byte (correctly) and then
give up while claiming a verification error. The hardware looks pretty
good, but the MSDOS software is pathetic (spelling errors and all). It
might be that I've changed something about how I run the software (booted
to real DOS from a floppy (a different floppy to before though, it just
occoured to me... they're both MSDOS 5 though...)). I got it second-hand
without documentation unfortunately.

I've been planning to build my own Willem programmer anyway because the
software supports a lot of newer devices, in fact I've already made the
PCB. The old one has the boast that it can program four chips at once
though.