UV EPROM eraser - Do it yourself with UV induction lamp

[Tantratron]

Hello all, I have no knowledge on how exactly UV eprom eraser works regarding the power radiation level, UV wavelength and time duration to completely erase uv-eprom chip without over stress.

I do have some spare uv-eprom (model HN27C101AG-20) from Advantest R3261/R3361 spectrum analyzer, these do store the firmware revision which I'd like to erase then reprogram but I've no standard UV eraser equipment.

Attached a set of 150W induction (electrodeless) UV bulb and 300 W induction UV bulb which coame from a 2022 contract (Covid19) to purify air, these are like visible light induction lighting or fluoresecnt visible light with electrodes except no tri-phosphor so you really get a power UV radiation output. I still have their respective 150W and 300W ballast so I was wondering to make a home or DIY project, inside I'd install few uv-eprom.

Now what is the theory or limitatio, what to do or not do if I would flash these uv-eprom in front of either 150 W or 300W induction flash: how many seconds, minutes, how far away should be the uv-eprom from say the surface of these induction lamp ?

Is there a risk if too much power to destroy the uv-eprom internal semiconductor, how about the case and pins of the HN27C101 chip ?

Thank you for any teaching, recommendation on the art of UV erase memory chips.

Albert

[tooki]

Well, look at a commercial EPROM eraser, for example https://www.mouser.com/datasheet/2/446/ERASER-1212238.pdf

It uses an 8W tube, producing intensity of 7mW/cm², producing an erase time of 10-35 minutes.

So what you have is massively more powerful. (Am I understanding correctly that it’s 2x150W+2x300W, i.e. 900W total, so literally on the order of 100x as powerful?)

I don’t have any experience with EPROMs myself, but if the erase times scale linearly with UV intensity, then you’d be looking at erase times of just a few seconds.

I don’t know whether the silicon die itself can be damaged from high levels of UV, but for sure the case and pins won’t; UV-erasable EPROMs use ceramic cases (with quartz windows), and I don’t know of any metal or ceramic that is affected by UV.

[rtv]

Couple minutes inside old 5W toothbrush uv sanitizer worked out for me

[Tantratron]

So what you have is massively more powerful. (Am I understanding correctly that it’s 2x150W+2x300W, i.e. 900W total, so literally on the order of 100x as powerful?)

In the picture which I did attach earlier, it shows a metal test bench which was used to EMC test these UV induction bulb, the ballast would be outside for a german customer. After all EMC pre-compliance engineering test in 2022, he never asked for returning the 4 bulbs so I've kept the whole test bench inclduing ballast. I can run each bulb separate so only 150 W at minimum, the reason of the 4 bulbs screwed up was to only store and multiple test then verify the performance. The metal case was simulating say a faraday case plus protecting from skin cancer the strong UV radiation as well as destroying my test lab furniture.

Note the 300W bulb has the same diameter as the 150W bulb, the only difference being the total lenth of the bulb racetrack or miniature Tokamak then of course, choosing the appropriate balalst power (150 W or 300 W) to ignite the stable run the plasma. The total UV radiation power is proportional to the doughnut or racetrack cylindrical surface.

[Renate]

How many seconds, minutes, how far away should be the uv-eprom from say the surface of these induction lamp ?

I don't know. You tell us.
No, I'm not being a smart-ass.

Program an EPROM, read it while it's being erased.
For more accuracy, wiggle Vcc+0.25V, Vcc-0.25V just the way (some) programmers do.
Figure out the time, then extend it 50% for safety.

[Tantratron]

How many seconds, minutes, how far away should be the uv-eprom from say the surface of these induction lamp ?

I don't know. You tell us.
No, I'm not being a smart-ass.

Program an EPROM, read it while it's being erased.
For more accuracy, wiggle Vcc+0.25V, Vcc-0.25V just the way (some) programmers do.
Figure out the time, then extend it 50% for safety.

You seem to not realize how powerful these UV induction lamp can be at 150W or 300W.
Just let one run say few minutes sitting on your wooden floor, its varnish and maybe more will be ruined in no time.
Same let some furniture, very quick to be surface destroyed.
So no way to install a ZIF eprom programmer while the UV flash is erasing, it will ruin about everything.
This is why for safety, these UV induction lamp will be always contained in a metal enclosure to protect furniture, test equipment, human eyes and skin

[Renate]

So no way to install a ZIF eprom programmer while the UV flash is erasing, it will ruin about everything.

Who said anything about a programmer?
You probably only need a half dozen wires.

[JimKnopf]

I use a plain old Siemens 6ES5985-0AA11 UV Eraser.
It works with a OSRAM HNS 10W/U OFR bulb with E27 socket https://www.lamptech.co.uk/Documents/Brochures/Osram%20-%20Liste%20Entkiemung%20-%201969-IV%20DE.pdf.
The bulb is placed 20 mm below the EPROM surface.

There is a timer for 0 to 30 minutes. A sticker says, 30 minutes are needed to fully erase eproms.

Edit: I added the characteristic wavelengths of this UV bulb, measured with Spectryx Blue-Box and Octave (script edited by Claude-Code).
        The important wavelength 253,7 nm is missing in my measurement due to the limited wavelength range of the Spectryx device (350-1000nm).

[Ian.M]

Photocurrent will be a problem.   Applying bias to the EPROM while it is under strong UV will cause unpredictable behaviour, maybe even damage.

I doubt your bulbs and ballast will like being rapidly switched off and back on, so if you want to do Renate's proposed erase test you will need some sort of mechanical shutter to cover the EPROM window before the programmer applies Vcc and other signals to the EPROM.

[madires]

Erasing EPROMs with a 4 W UV-C fluorescent lamp takes about 15 minutes. A 150 or 300 W UV-C lamp might be total overkill.

[Tantratron]

I use a plain old Siemens 6ES5985-0AA11 UV Eraser.
It works with a OSRAM HNS 10W/U OFR bulb with E27 socket https://www.lamptech.co.uk/Documents/Brochures/Osram%20-%20Liste%20Entkiemung%20-%201969-IV%20DE.pdf.
The bulb is placed 20 mm below the EPROM surface.

There is a timer for 0 to 30 minutes. A sticker says, 30 minutes are needed to fully erase eproms.

Edit: I added the characteristic wavelengths of this UV bulb, measured with Spectryx Blue-Box and Octave (script edited by Claude-Code).
        The important wavelength 253,7 nm is missing in my measurement due to the limited wavelength range of the Spectryx device (350-1000nm).

Many thanks @JimKnopf for these technical details about the electroded UV lamp, its length, power, distance between EPROM surface and bulb. This will help me do some computations where I'll renormalize the surface radiation because induction (electrodeless) UV lamp have much bigger diameter tube than electroded UV tube.

As for the UV spectrum, my UV induction bulb do have lot of 254 nm because they're ozone free and meant to purifiy air. There are models which generate ozone with 185 nm, in that case they're used for water purification and odor removal. If you are interested, attached datasheet from the manufacturers showing different wattage, dimension of these bulbs inclduing my 300W but never provided the 150W.

Erasing EPROMs with a 4 W UV-C fluorescent lamp takes about 15 minutes. A 150 or 300 W UV-C lamp might be total overkill.

Yes I agree hence my post #1 to see capture recommendation or get an idea of power level, duration of UV exposure and distance.

Photocurrent will be a problem.   Applying bias to the EPROM while it is under strong UV will cause unpredictable behaviour, maybe even damage.

I doubt your bulbs and ballast will like being rapidly switched off and back on, so if you want to do Renate's proposed erase test you will need some sort of mechanical shutter to cover the EPROM window before the programmer applies Vcc and other signals to the EPROM.

I still do not understand the 2 post from Renate so my plan is to proceed from the UV eraser data provided by members in the thread then renormalize or intuit a UV flashing procedure.

As for the possible damage of bulb and ballast, I'm not worried because these are induction (electrodeless) plasma technology. I've worked a lot on this forgotten technology wether visible lighting or UV lighting, you can switch on or off, actually you can dim with pulse skip technology in the ballast. There are no electrodes in the lamp, all the HF energy (typically 250 kHz) is transferred by induction to the plasma vessel thanks to power ferrite. This is why this technology can last 50 khrs and in practice 100 khrs without any failure compared to electroded plasma lighting.

[JimKnopf]

For educational purpose i will test my Commarker Omni 1 10W UV Laser next week to erase some spare Eproms.

[madires]

Erasing EPROMs with a 4 W UV-C fluorescent lamp takes about 15 minutes. A 150 or 300 W UV-C lamp might be total overkill.

Yes I agree hence my post #1 to see capture recommendation or get an idea of power level, duration of UV exposure and distance.

I guess the distance for my EPROM eraser is about 1 or 1.5 cm. The lamp is in a small aluminium case with a slit above the lamp to place the EPROMs. And there's a sliding cover with a safety switch.

[TERRA Operative]

Page 11 of the attached datasheet has some info on the light intensity required for a certain erase time, you can calculate a ballpark erase time of your lamps from this which should hold true for most chips (I'd go for a 200% fudge factor to be sure all bits are fully erased).

[LaserSteve]

Experience long ago as teen having to erase several hundred chips:

1. UV induced Keratosis of the Cornea is 2 days of severe pain using a 15 watt linear lamp around a meter away.  You'll be fine, the boss said, shielding, who needs shielding.... I soon found a new boss.   

2.  A detailed study after shielding the lamp and wearing polycarbonate UV adsorbing safety glasses showed that when the programmer said erased, the chip needed 100% more exposure.   10 minutes became 20 minutes to be sure.

3. Inverse square law is your friend when your lamp can probably easily erase chips at one meter distance. 10X over exposure in my youthful experiments exposed/revealed that some older chips did not like 10x exposure at 2 cm from the 15 watt lamp, resulting in a few blown bits.  It was brand dependent, some brands showed no issue, some cheaper chips did.   This was the 1990s.

From what I remember, I used a 18 Watt 9 inch / 22 cM  T4  diameter linear germicidal lamp

Steve

[Gyro]

My EPROM eraser. It has a 6" 4W, (maybe a little overdriven) tube and  will erase a 27xx EPROMs and PICs in 20 minutes. It has an internal timer and magnetic interlock for Sunburn (EPROM and human) protection.

Distance between tube and DIP package surface is around 10mm.