One for any electronics experts

[Demonperformer]

DIY is unnatural (possibly dangerously so) territory for me, so please be gentle.

I am looking to purchase one of these as the basis for a flat-field light source for my Ne3 filter (386.9nm), having discovered that the ELT is way out of range for that wave length.

The specifications table puts the "peak wave length" at 385nm, which is pretty close. However, the product details above it state that it has a wave length of 365nm, which is way outside the band width of the filter. So what am I missing here?

On page 20 of the pdf, there are "directions for use", which don't actually mean a lot to my non-electronic mind. in my naivity I'm hoping I can just attach a positive lead to one pin and a negative lead to the other (I'm guessing that's what it means by "2-pin"). Am I being stupidly optimistic?

On page 2 of the pdf (initial electrical/optical characteristics) it says that the U385 should have a current of 500mA and a voltage of 3.6V (which strikes me as a rather weird figure). Any thoughts on the best way of achieving that?

Thanks.

[almcl]

I am definitely NOT an electronics expert although I can solder components to circuit boards and sometimes get things to work  but that component looks way beyond anything I would attempt to work with. 

The soldering conditions 'reflow soldering on to a PCB', the individual resistor calculations and the relative humidity requirements (>70% once the foil package is opened ) all look a bit challenging.

Having said that, one off cost of under a tenner it might well be worth a punt?

Oh, one final thought (haven't checked) but UV is hazardous, possibly why its 'only suitable for use in industrial applications' ?

[robin_astro]

The manufacturer's datasheet on that page says peak at 385nm (actually there is a tolerance so it could be anywhere from 380-390)  There is also a spectrum plot there showing how wide the emission line is.

https://docs-emea.rs-online.com/webdocs/1427/0900766b81427cb7.pdf

The LED will drop 3.6V and will pass 0.5A current so if you plan to use a 5V supply for example you would need a series resistor which would drop the voltage by 1.4v at 0.5A.  Ohms law says R=V/I so R = 2.8 ohm. (Go for the nearest next highest standard value)  The resistor would need a rating of  at least 1.4 x 0.5 = 0.7 Watts.  Make sure to connect the LED up the right way round

Cheers

Robin

[robin_astro]

The datasheet also has advice on eye safety.  The eye is  insensitive to this wavelength so you will not see it.

Robin

[Gina]

But it could damage your sight.  Surface mount packages need very specialist use.  I have used them but with great difficulty and I am well used to soldering small items.

[robin_astro]

Indeed, hence the pointer to the eye safety in the datasheet which says to not look directly into the beam and wear eye protection if there is a risk of reflections

[lenscap]

Have a look at this data sheet. (Scroll down to the table for the NCSU275)

http://www.nichia.co.jp/en/product/uvled.html

The peak wavelength  of the spectrum varies with slight changes to the voltage across the led.

So if you wish you can fine tune the peak wavelength. Wire a fixed and a variable resistor in series with the led.

Choose the values of the fixed and variable resistors so you can get the lowest & highest led voltages that you want. Then adjust the variable resistor to get the wavelength you need.

As Gina says surface-mount components are a pain. Through-hole uv leds are available.

[Stub Mandrel]

1 hour ago, Demonperformer said:

On page 2 of the pdf (initial electrical/optical characteristics) it says that the U385 should have a current of 500mA and a voltage of 3.6V (which strikes me as a rather weird figure). Any thoughts on the best way of achieving that?

It's quite complicated to explain, but as you pump more current through an LED the voltage only rises slowly, conversely a small increase in voltage can make a big current increase, destroying it.

By limiting the voltage to 3.6 volts or using a 500mA constant current driver ( a useful approach if you have lots of LEDs daisy-chained).

If you have a USB-type phone charger giving a 5V, 500mA to 1A supply you can use that. You will need to drop 5 - 3.6 volts  = 1.4 volts. 1.4 volts/500ma  = 2.8 Ohms. A 3.3 ohm resistor should work fine and avoid over-cooking the LED.

I also guess that 265 is a typo as the product code contains 385.

 

[Dr_Ju_ju]

What are you intending to mount this(these?) on ?? as being only 3.5mm square they will be very fiddly to handle, and mount.....

[symmetal]

If you want to make a led panel for flat fields then using these type of UV led strips would be easier to work with. They are available at 385nm, you get 60 leds per metre at £22.60 a metre and they work off 12V.

Alan

[Demonperformer]

Thanks for all the input. Needs a lot of thought & I will get back then.

[pete_l]

First off, about the wavelength. We can assume that 365nm is a typo and that the clue is in the part number viz, NCSU275T-U385 as the device is one of a family of similar devices - tuned for different wavelengths.

Now, about care and feeding. You will almost certainly be supplying this device through a fixed voltage supply (e.g. 5 Volts). The device spec states a forward voltage drop of 3.6 V (page 2). So the remaining 1.4 Volts must be "lost" in the current limiting resistor you will put in series with the device. The datasheet tells us the maximum current the device can handle is 700mA. So to get that current flow, you should invoke Ohm's law V = I * R, rearranged to calculate resistance, So R = V / I. Inserting values, V is the 1.4 Volts we wish to drop. I is the 700 mA working current so R = 1.4 divided by 0.7, which is 2 Ohms.

Be aware, this is the maximum current the device can handle. There is no requirement to run it at that level. If you wish to work with a lower light intensity (maybe one that isn't quite so eyeball-frying) you can reduce the current through the LED by using a larger value resistor.

In all cases you should pay attention to power dissipation: when the LED is running at full belt: 700mA dropping 3.6V it will dissipate 3.6 * 0.7 = 2.5 Watts - a significant amount of power. On its own, this would probably burn-out the LED. So it would need to be thermally connected to a largeish heatsink, to keep the device within its working temperature range.

Reducing the current through the device also reduces its heat output, meaning a smaller heatsink is needed. But then the resistor has to dissipate more power.

 

P.S. Considering that the light output of this LED is beyond visible range, you might find it helpful to put another LED (e.g. a  red, yellow or green one) in parallel with it and mounted some distance away. That will give you a visual indication of when your sun bed flat-panel is operating. This second LED can be driven at much lower current levels., say 5 mA. So place a 1k resistor in series with it and feed it from the same 5 Volt supply as your U-V LED.

[robin_astro]

5 hours ago, lenscap said:

Have a look at this data sheet. (Scroll down to the table for the NCSU275)

http://www.nichia.co.jp/en/product/uvled.html

The peak wavelength  of the spectrum varies with slight changes to the voltage across the led.

So if you wish you can fine tune the peak wavelength. Wire a fixed and a variable resistor in series with the led.

Choose the values of the fixed and variable resistors so you can get the lowest & highest led voltages that you want. Then adjust the variable resistor to get the wavelength you need.

 

The wavelength does not change with voltage. The table is for different versions of the LED with different wavelengths  (Actually strictly speaking the voltage across an LED is always constant provided the voltage source is above the specified value)

Robin

[robin_astro]

3 hours ago, pete_l said:

Reducing the current through the device also reduces its heat output, meaning a smaller heatsink is needed. But then the resistor has to dissipate more power.

Actually the power dissipated in the series resistor decreases as you increase it.  ie for a 5v supply, the power dissipated in the resistor is always 1.4V * the current flowing through the circuit because the voltage drop across the LED remains constant at 3.6V

[The Admiral]

Just a thought, what material are you going to use to diffuse the light source at UV wavelengths?

Ian

[Demonperformer]

I was thinking a piece of perspex.

Most acrylic plastics will allow light of wavelength greater than 375 nm to pass through the material.

[Demonperformer]

Thanks to everyone for their comments, particularly those pointing out that this is not the best option and suggesting alternatives. Also to those providing pointers as to how I could set up this device - even though I will probably go with an alternative, I am sure the comments will be transferrable. One suggestion I particularly like (without wishing to denegrate any others) is pete_l's suggestion of adding a led of a different colour in the system (but not as part of the light source) to indicate when it is running - will definitely be incorporating that one. Thanks.

What I did like about this device was its 120° field. This meant that the distance from the LED to the perspex diffusion sheet could be less than 3cm for my smaller scope and less than 6cm for the bigger scope, resulting in a compact set up.
What I didn't like about it is the cost and the fiddly nature of putting it together.

Now, considering the two alternatives that have been suggested.

Through hole uv led:
What I like about this is firstly that it is dirt-cheap and secondly it would be much easier to wire into a setup.
What I don't like is that it only has a 30° field. This would increase the distance to the diffusion sheet to 19cm for the small scope and 37cm to the big scope, making for a much more bulky set up.

Alan's suggestion of an LED strip:
What I like about this it would be reasonably cheap, getting just the length I require, and we are back to the 120° field. I also like that it runs on 12V which is the same as my current ELPs, so no need to remember to change the voltage depending on which source I am using.
What I don't like is that you have to use it in multiples of 50mm, which equates to 3 leds. Now I could always use three 50mm strips, producing a 3x3 grid in a 50mm square in the centre, but, as I see it, the problem is that by introducing multiple point sources I will be generating interference patterns on the diffusion sheet which will then, by definition, not be a flat field.

My gut-feeling is to go with the strip if I can find a way around the interference pattern problem. What do others think?

Thanks.

[The Admiral]

Just now, Demonperformer said:

I was thinking a piece of perspex.

Even the opalescent type?

2 minutes ago, Demonperformer said:

the problem is that by introducing multiple point sources I will be generating interference patterns on the diffusion sheet

Do you think that will happen? It's outside my experience certainly, but with an opalescent sheet and (relatively) broad band radiation?

Just a further option, and perhaps retrograde in this day and age, but black light fluorescent tubes can be had much more cheaply than LEDs. See for example, https://www.lightbulbs-direct.com/product-type/specialist-bulbs/ultraviolet/, and you might find one suitable for your needs.

Ian

[Stub Mandrel]

1 hour ago, Demonperformer said:

What I did like about this device was its 120° field.

Even if it emits light evenly in all directions, a short distance will give you very uneven illumination. At 60 degrees off axis the illumination at a plane ahead of the source will only get 75% of the illumination of on-axis. The LED strip suggestion is a way to get more even results.

Diffraction isn't an issue as they aren't point sources, the LED chips are significantly bigger than the light wavelength.

If you get clear perspex and  gently attack it with fine wet and dry paper you shoudl be able to make a suitability fine diffuser.

 

My preference would be for multiple light sources at least as long as it is wide, lined with slightly uneven aluminium foil. That will reflect the off-target light randomly and make a much more even field.

I uses this arrangement with four UV tubes  in a box for my PCB/etch resist exposing setup, with a very thin sheet (2mm) of glass (not frosted) and it works very well.

 

[alan4908]

17 hours ago, Demonperformer said:

DIY is unnatural (possibly dangerously so) territory for me, so please be gentle.

I am looking to purchase one of these as the basis for a flat-field light source for my Ne3 filter (386.9nm), having discovered that the ELT is way out of range for that wave length.

The specifications table puts the "peak wave length" at 385nm, which is pretty close. However, the product details above it state that it has a wave length of 365nm, which is way outside the band width of the filter. So what am I missing here?

On page 20 of the pdf, there are "directions for use", which don't actually mean a lot to my non-electronic mind. in my naivity I'm hoping I can just attach a positive lead to one pin and a negative lead to the other (I'm guessing that's what it means by "2-pin"). Am I being stupidly optimistic?

On page 2 of the pdf (initial electrical/optical characteristics) it says that the U385 should have a current of 500mA and a voltage of 3.6V (which strikes me as a rather weird figure). Any thoughts on the best way of achieving that?

Thanks.

Hi DP

On more fundamental issue - I don't understand why you are contemplating using a UV light source for flat fielding. 

An EL panel will allow you to correct for dust donuts and vignetting.  The fact that an EL panel has a poor response at your filters peak response is irrelevant since it is the dust and system vignetting that you wish to correct, not the wavelength response of the filters. 

Alan

[Demonperformer]

8 minutes ago, alan4908 said:

Hi DP

On more fundamental issue - I don't understand why you are contemplating using a UV light source for flat fielding. 

An EL panel will allow you to correct for dust donuts and vignetting.  The fact that an EL panel has a poor response at your filters peak response is irrelevant since it is the dust and system vignetting that you wish to correct, not the wavelength response of the filters. 

Alan

I guess the thing that has started all this was a test I did about a week ago taking flats with each of my filters because another member was getting weird effect with his 1600. The HOS filters all used exposures of a fraction of a second. The Ne3 used exposures of about 80s. The light from a standard ELP seemed to be deficient at this wavelength. This assumption was confirmed by discussions with the ELP manufacturer.

[alan4908]

2 minutes ago, Demonperformer said:

I guess the thing that has started all this was a test I did about a week ago taking flats with each of my filters because another member was getting weird effect with his 1600. The HOS filters all used exposures of a fraction of a second. The Ne3 used exposures of about 80s. The light from a standard ELP seemed to be deficient at this wavelength. This assumption was confirmed by discussions with the ELP manufacturer.

 OK I understand. However, why don't you simply take a sufficiently long exposure each filter ? For instance, for me this is a fraction of a second for my LRGB filters and it rises to 60s with my S NB filter.  

Alan

[Gina]

That's what I did too.

[Stub Mandrel]

32 minutes ago, alan4908 said:

 OK I understand. However, why don't you simply take a sufficiently long exposure each filter ? For instance, for me this is a fraction of a second for my LRGB filters and it rises to 60s with my S NB filter.  

Alan

At 60s you are probably going to want to use dark flats.

[Demonperformer]

4 hours ago, The Admiral said:

Just a further option, and perhaps retrograde in this day and age, but black light fluorescent tubes can be had much more cheaply than LEDs. See for example, https://www.lightbulbs-direct.com/product-type/specialist-bulbs/ultraviolet/, and you might find one suitable for your needs.

Ian

This was actually my initial thought. There are some really cheap black light torches on Amazon (3 weeks from China). I wondered about removing the LED unit from one and using that. But when I contacted them to check the wavelength they couldn't provide any details so I thought I would explore this option.