A little off topic but not totaly! PIC <-> UV lamp

[Jody]

Hello,

I have made a control for a UV lamp.
It is a 400W UV lamp to kill some cell's.. it is a research they want to do..
All is mounted inside a cabinet. The lamp is mounted on a steppermotor. The PIC controls the door lock (UV is harmfull for the eyes), the stepper and the door controls. Reads back the IR switch..
So there are wires going to the PIC's pcb and the cabinet.

It is working if we using a "normal" light bulb.
If we are using the UV lamp the PIC goes beserk. opens/close the door, reset's etc.

I already have put copper tape all around the wires wich are going to and back to the PIC.. It seems to get a little better but is is not working like I want.

Is there somebody who has an idea where to look further??

Best regards,
Jody

[temtronic]

It sounds like the 400W UV lamp is creating a lot of EMI. More info about the UV lamp power source is needed. When you say 'normal' lamp, is it just a lamp 'swap',so using the UV PSU or totally different power setup?

All of the I/O of the PIC should have EMI protection(chokes, caps, NE2) whatever it takes to prevent radiation getting to the PIC.Also it should be in a completely metal box and grounded. EMI will easily go through plastic.

[Jody]

Hello Temtronic,
I will try to clearify the thing:
Starting the project the lightsource was claimed to be a standard lightbulb (like we use to have in our home).
That's how we tested it and that works fine. I switch the lamp on/off using a SSR and that worked fine...

But in real life they are using a 400 watt medium pressure mercury lamp (Using it as a UV light source) and now I have problems. The lamp is producing a lot off EMI. The lamp itself is also running on 130VAC, so I don't control the lamp itself but the power source of the lamp (we are running on 230VAC over here, The Netherlands)

Mine wires for the stepper motor and the IR-light switch are really close to the lamp. Have put everything in a grounded metal tube but it is not working like I want. Putting the wires in a grounded metal tube help a bit but not enough.

The PCB I can put in a metal box... but what to do about the wires.... How can I prevent the EMI getting on the wires??

Best regards,
Jody

[temtronic]

For EMI suppression where the wires enter the metal box the PIC is in, you'll need cholkes (coils) and caps. The values need to be selected to allow the PIC signals in and out but 'filter' out the EMI (usually high frequency).
You should also 'suppress' the EMI at the source, again some form of L-C network, before it can travel down the wires to the PIC.

The wires should be in a shielded cable, grounded only at the source ( the lamp unit), left floating at the PIC.

Eliminating EMI is 1/2 proper design, 1/2 correct components and 1/2 luck.
Much is 'trial and error'.

In a past employ, I could get 15 bits from a 16 bit, 32 channel ADC data acquisition system mounted inside an optical emission spectrometer, far,far more EMI than your UV lamp ,so I know it can be done.

[Jody]

Monday the total cabinet is brought to mine lab..
I will start to change and measure.. I will keep you informed..

Thanks for your assistance!!
keep you informed!!

[Ttelmah]

A mercury UV lamp, is an _arc_ lamp. Just like a fluorescent lamp, it'll have
a ballast (may be built into the base), and this produces a huge amount
of EMI. Have you considered whether you might be able to switch to a
UV LED source instead?. These are more efficient, and now offer output
powers competing with many mercury lamps. Also some mercury designs
will now use electronic ballasts that with careful design can reduce the EMI.

You need to start by working out whether you are dealing with radiated or
conducted EMI. Probably both. Add an EMI powerline filter as close to the
lamp as possible. Then look at Faraday shielding round the control electronics.

[temtronic]

I had a very quick look for 400W UV LED lamps....
About $650 Canadian compared to $100 for the EMI generating ones !
I'd seriously consider LED version,providing it's the same UV (635nm ?) ,life span and power rating.
While it is a lot more money, you could easily spend 1-3 weeks trying to eliminate the EMI. R&D time is a real cost though many never consider it.
Sad part is, even though you think you've found the EMI, once the lamp ages, ZZAP...it generates a new type of EMI.......arrrrghhhhhhh.

Friend's garage system has done that. Rock stable for 5+ years,reporting data every 5 seconds....last week some 'glitch' is getting to the PIC causing 'weird stuff to happen'...

[Jody]

I don't think that the price for the lamp was a big problem for them...
But if you can give me a part number for a UV lamp which has the same amount of UV radiation then I will suggest that to them.

I also ask the researcher if it is possible to change to LED's

[Ttelmah]

There is more to this than may at first appear.
The Mercury lamp produces 254nm. The peak frequency for effective
destruction of bacteria is up nearer 270nm. You can get LED's giving
266nm, which for the same light level is about twice as effective as
the mercury lamp.
Then for a given power, the LED produces more light.
So I'd actually expect something like a 100W LED array to be as
effective as the 400W mercury lamp.
There are off the shelf germicidal 60W LED lamps, for only about $70.
Two of these would probably be as effective as the existing mercury
unit.

Needs a bit of experimentation/testing to see how effective this is.

[temtronic]

I'd need the actual UV lamp make/model info to get a comparable LED version.
Using Google, I found several here in Canada, many for AV use ( sound stages ) !

I have no idea the 'Lumens per watt' ratio ,just think it might be a good(better ?) use of time to 'search the Web' than fight the EMI.

Jay

[Jody]

The lamp they are using right no is the :
400W medium pressure mercury lamp 3040/PX0686
Site:
[url]
http://www.photochemicalreactors.co.uk/html/immersion-well-reactors.html
[/url]
They are afraid that, if we change to LED, not have the same wavelength.
And that irradiance of the LED would be too weak that they have to extend the reaction time to several weeks, which is what they am not happy with…

maybe moving the LED's closer to the vials...."

what they say:
"Major emission peaks of the current lamp at UV region are 365, 312, 302, 297, 280, 265, 254 nm, but to be honest I have no understanding of which wavelength is the most important.

[temtronic]

Hmm 7 major peaks.... I can understand why they want a 'broad spectrum' UV source ( lamp ).... LEDs are very 'narrow' in comparison...

sounds like you're back to suppressing the EMI.......

[Ttelmah]

254 is the one you are worried about for germicidal applications. As I said
these have a spectrum to which they respond centered around 270nm. This
is bell shaped, and spreads out for several nm each side. The LED sources
get closer to the most sensitive point than the mercury lamp does.

There are many people doing LED germicidal lamps. They are being used in
hospitals, and laboratories.

We use pulsed LED's in water treatment applications. Using a high pulse
intensity gives good efficiency, while massively extending the life of the
LED's.

For viruses for example, you don't 'kill' these (since they are not alive), but
you 'deactivate' them. We went from a mercury solution to an LED solution,
but did extend the tube area where the light source impacted. Testing
though shows the efficacy is better than the older solution....

Talk to a specialist lamp supplier. They will know the products, and power
available. They will also be able to give you recommendations on EMI
suppression on the mercury lamps.

[temtronic]

What you need, is something from 'them' explaining in detail, EXACTLY what 'UV source' you need to control. Do they need 'broadband UV' or 'just '254' ?
Those lamps emit several peaks, maybe '254' is the critical one ?, I don't know.
Now I'm wondering if the energy given off by 2,3, 10 bulbs is IDENTICAL ? Say bulb #4 emits more '254' than the others...is that critical to the experiment ?
It's really important to get the 'specs' nailed down now.

If it's just one specific, that can be done with LED, everyone is better off.

[Ttelmah]

Germicidal UV lamps, don't normally have phosphors, so produce
only 5 main peaks. The strongest is at 254nm, and this is what matters
for the germicidal applications. The others are closer to visible, which
is why you see the blue light. Normally 313, 404, 437 & 546nm. These
have basically no effect for the germicidal application, and are effectively
wasted energy.
Hve a look at the Wiki article on this:
[url]
https://en.wikipedia.org/wiki/Ultraviolet_germicidal_irradiation
[/url]

If you want to see real LED 'power', look at devices like:
[url]
https://www.inui.co.uk/collections/led-uvc-lights/products/uvc-fld400
[/url]

These were being used near to me, by the labs working on the Covid
vaccines, and replacef kW mercury units they had previously fitted.