Photography-equipment-friendly UV torch

[Akira]

I've been enjoying occasional shooting of UV induced visible fluorescence (UVIVF) but have found difficulty in holding my Tank007 TK-566 UV torch in the effective angle.  Also, I have wanted a bit more powerful torch.

 

Luckily a net shop based in Akihabara, Tokyo, has started to handle UV LEDs.  So, I got a 5W 365nm LED and the holder/reflector kit designed to fit the LED.  The forward voltage of the LED is 4V (typ.) and I wanted to drive it using a 5V AC/DC adapter, and a 1 ohm resistor connected with the LED in series was necessary to drive it at 4W.  Here are the parts I collected.

 

Front row, left to right: 5W 365nm LED, retainer, reflector, thermal grease, 1 ohm resistor, a set of screws to fix the LED and the retainer.

Second row, left to right: back plate for the case, aluminium die-cast case which is also expected to function as heat sink, SmallRig cold foot, DC adapter input jack.

 

I bought the 1 ohm resistor that was rated at 5W which was overkill.  A 3W one should be enough.  I was thinking that a 5W one was necessary because the LED was rated 5W (stupid me!).

 

[Akira]

And this is the completed torch and an example of how it can be held.  I used a 3/8" to 1/4" thread adapter to fix the cold foot (the 1/4" thread of the foot was cut short.)

 

 

 

 

 

 

 

 

[Akira]

And this is a quick test shot of ten quail eggs fluorescing reluctantly and their case fluorescing merrily.

I held a Baader U filter to cut the visible portion of the LED.  I need to find the way to fix the filter onto the reflector.

 

[colinbm]

Very interesting Akira, How long was the exposure please ?
If you can find a Zeiss T* uv filter 55mm, it will cut the visible light at 400nm & block it getting into the camera, no other UV cut filter is as good as this one.
The 365nm LED shouldn't have any IR so a cheap Chinese ZWB1 filter will be good.

[ulf]

Even a 2mm thick ZWB2 would be good for this if the LED's plastic structures do not have fluorescence.

I have seen some with the mechanical structure like in the pictures that do.

[ulf]

It is nice to create something and you already have done that.

Otherwise this might be an alternative that could be powered by both a suitable USB-C power bank or wall adapter.

The filter is already in place.

https://www.aliexpress.com/item/1005005236028768.html

Do not expect it to be a real 20W consuming unit though, but it is likely more powerful than your build. 

[Nate]

That's a nice build Akira.

I bought a couple cheap Alonefire 10w 365 UV torches that came with ZWB2 filters when I first started UV. The one's I bought are no longer available, but you might just find one so you don't need to use the Baader UV

[Akira]

11 hours ago, colinbm said:

Very interesting Akira, How long was the exposure please ?
If you can find a Zeiss T* uv filter 55mm, it will cut the visible light at 400nm & block it getting into the camera, no other UV cut filter is as good as this one.
The 365nm LED shouldn't have any IR so a cheap Chinese ZWB1 filter will be good.

 

The fluorescing eggs image was shot at f8.0, 1/6 sec., ISO 1250 with -0.7 stop compensation.  The rated power of TK566 torch is 3W, but I would doubt that based on this data with this image with the 5W LED operated at 4W.

 

The 1.9mm thick ZWB1 has about 5% transmission at 400mn which is well visible, so ZWB2 would be a better choice, as noted by ulf.   To cut the visible portion of the torch entering into the camera, I have a 440nm cut gel, but it gives a slight yellowish hue to the resulted image.  An L39 filter originally developed for B&W films would work as well.

[Akira]

11 hours ago, ulf said:

It is nice to create something and you already have done that.

Otherwise this might be an alternative that could be powered by both a suitable USB-C power bank or wall adapter.

The filter is already in place.

https://www.aliexpress.com/item/1005005236028768.html

Do not expect it to be a real 20W consuming unit though, but it is likely more powerful than your build. 

 

Thank you for the suggestion.  The unit looks nice.  The product image indicates 395/365nm, which could be an apology in advance in case the buyers claim the light is too purple?  Nevertheless, a UV-pass filter would extract the real UV portion from this powerful light source.

[Akira]

11 hours ago, Nate said:

That's a nice build Akira.

I bought a couple cheap Alonefire 10w 365 UV torches that came with ZWB2 filters when I first started UV. The one's I bought are no longer available, but you might just find one so you don't need to use the Baader UV

 

Thank you.  The Alonefire 10W UV torch seems to be still available:

 

https://www.amazon.co.jp/Alonefire-UV懐中電灯-ブラックライト-ペット用尿検出器-レジン硬化、アルミニウムケース、充電器、18650/dp/B07SWW5FHB

 

But one of the main reason for this self-made torch is that I'm reluctant to use torches driven by the 18650 battery which is said to be shaky in terms of quality and specification (the existence of the controller circuit).

[Akira]

FWIW, I ordered a cheap extension tube for the 1 1/4" eyepiece.  I will glue it around the LED and use the Baader U that I already have.  :)

[ulf]

On 5/3/2023 at 5:45 PM, Akira said:

The product image indicates 395/365nm, which could be an apology in advance in case the buyers claim the light is too purple?  Nevertheless, a UV-pass filter would extract the real UV portion from this powerful light source.

No, there are two variants to chose from when ordering.

 

Filtering a 395nm UV-LED with a ZWB2 is really not a good idea as the filter will absorb most of the LED output. 

The 395nm LEDs are not pure UV, but UV+deep Violet. (UVdV ?) Might be interesting for UVdVIF with a sharp cut light yellow filter on the lens.

[ulf]

5 hours ago, Akira said:

FWIW, I ordered a cheap extension tube for the 1 1/4" eyepiece.  I will glue it around the LED and use the Baader U that I already have.  :)

The Baader U is rather expensive.

For illumination it might be a good idea to get and use a cheeper alternative to lessen the risk-exposure to the Baader U and save it for photography.

Even if you are careful accidents happens.

 

Getting one of these of suitable size can be an alternative:

https://www.aliexpress.com/item/4001191819104.html

 

[ulf]

6 hours ago, Akira said:

 

Thank you.  The Alonefire 10W UV torch seems to be still available:

 

https://www.amazon.co.jp/Alonefire-UV懐中電灯-ブラックライト-ペット用尿検出器-レジン硬化、アルミニウムケース、充電器、18650/dp/B07SWW5FHB

 

But one of the main reason for this self-made torch is that I'm reluctant to use torches driven by the 18650 battery which is said to be shaky in terms of quality and specification (the existence of the controller circuit).

That is yet another impossible "10W" design. If powered with 10W it would overheat quite soon. 

It is likely to have a controller circuit just as the Convoy S2, (3W) 

 

It seams like the stated power values (MarketingPower, without any connection with real physics power) for UV torches is getting higher and higher.

I call them lies.

 

There has not been a big quantum leap in UV-LED technology for 365nm LEDs, lately, just a gradual improvement.

This means that of the supplied electrical power, the majority becomes heat, not light. 

Heat is an enemy of LEDs that makes them break down if too high.

To get rid of thermal energy you need heat sinks with enough surface and efficiency or fans.  

[ulf]

18 hours ago, colinbm said:

If you can find a Zeiss T* uv filter 55mm, it will cut the visible light at 400nm & block it getting into the camera, no other UV cut filter is as good as this one.

Be aware that there are cheapish, very well made copies of the Zeiss T* uv filters that look quite convincing.

They do not block UV as a proper Zeiss filter would, but the build and coatings look advanced.

Sometimes bad products can be sold via Amazon by third party sellers.

 

I got my copy-filter via Aliexpress to test. Too cheap to be true and it was.

 

Checking the blocking is easy with an UV-torch and some paper with any fluorescence.

[Akira]

My Baader U was given me in token of my friend's thanks for my effort to find equipment that was difficult to buy outside Japan.  I knew that it was an expensive filter and have only used in my place.

 

That said, a filter for the torch can only be concerned with the transmission of the appropriate wavelength and not with its optical quality (image degradation).  And, as Colin mentioned, the LED doesn't emit near IR / deep red portion, so the costly multi-layer coating of Baader U is not necessary.  I also have Hoya U360 filter which has been intact for nearly a decade.  So, a cheaper uncoated equivalent may work.

 

I don't trust the 10W rating of the torch either.  My TK566 is rated at 3W, but it is noticeably weaker than my new contraption rated at 5W and operated at 4W.  Apparently, their difference is much larger than the difference between 3W and 4W.

 

My contraption gets warm after about 30 second of continuous operation.  The heat sink is absolutely necessary.  Even my TK566 gets very warm after 1 minute or so.

[ulf]

If you have a multimeter you can estimate the current flowing through the resistor and LED.

The resistor have a tolerance of 5%. It might get quite hot and possibly melt the insulation of the cables it is touching.

If you measure the full supply voltage and the voltage over the LED you can roughly find the input power.

 

V_LED / 1(Ohm) = Current

Power = Current x V_"5V"

Assuming a perfect 1 Ohm resistor and 5.00V supply I would guess that you will have ca 3.8V over the LED and 1.2A flowing through the system.

The total power consumption will then be 6W. Of that 4.56W into the LED. A very optimistic conversion value of 35% would give 1.6W optical power.

That leaves 4.4W that is heating the case. 

[Akira]

1 hour ago, ulf said:

If you have a multimeter you can estimate the current flowing through the resistor and LED.

The resistor have a tolerance of 5%. It might get quite hot and possibly melt the insulation of the cables it is touching.

If you measure the full supply voltage and the voltage over the LED you can roughly find the input power.

 

V_LED / 1(Ohm) = Current

Power = Current x V_"5V"

Assuming a perfect 1 Ohm resistor and 5.00V supply I would guess that you will have ca 3.8V over the LED and 1.2A flowing through the system.

The total power consumption will then be 6W. Of that 4.56W into the LED. A very optimistic conversion value of 35% would give 1.6W optical power.

That leaves 4.4W that is heating the case. 

 

Based on the spec. sheet of the LED, I figured that 1 ohm resistor was just about right.  The designed current in the LED circuit is 1A, and the resistor is rated 5W, so there should be enough margin compared to the 1W of the designed power for the resistor (1V x 1A)... hopefully.

[ulf]

You will definitely have more than 1A through that circuit with an UV-LED as the typical forward voltage is closer to 3.7V than 4V.

If the Maximum rating of the LED is 1A you are overdriving it.

Can you please share the data sheet of the LED?

[Akira]

9 hours ago, ulf said:

You will definitely have more than 1A through that circuit with an UV-LED as the typical forward voltage is closer to 3.7V than 4V.

If the Maximum rating of the LED is 1A you are overdriving it.

Can you please share the data sheet of the LED?

 

Well, I don't think so.  Like I mentioned in my initial post, the data sheet says that the forward voltage of 4.0V typ. (3.5V min. and 4.5V max.) , unlike other ones like white or blue LEDs.  Please refer to the data sheet uploaded here.  If my sample is rated at the minimum Vf value, I would be overloading the LED.  I might change the resistor to the one with a higher value (1.5 ohms, or 2 ohms maybe?).

 

OSV1XDE5E1S.PDF

[ulf]

The condition for V_F 4.0V typ. (3.5V min. and 4.5V max.), (third column) is I_F = 1400mA

V_F is the result of a combination of the semiconductor-chip's diode curve and the internal resistance in the component.

That unusual rather high values is due to a less optimal series resistance design internally in the component. 

For other high power UV-LEDs I have seen that curve would not  change voltage that much for different current.

I'll see if I can find a suitable data sheet to show you.

 

1400mA is also the first item  in the table for Absolute Maximum Ratings at T_A = 25°C.

 

You risk running very close or above the ratings limit as the environment T_A will be above 25° but if the total on life time is low that might be worth the risk.

If you want to get more margin I think changing to 1.2Ohm would be enough

 

 

 

[ulf]

Here is a link to a quality UV-LED in the 3-4W region:

http://www.s-et.com/upload2/CUN6GB1A_181016_R5.4.pdf

It is just the SMD-component that should be mounted on a suitable metal-based PCB, coupled thermally to some cooling structure like a heatsink.

 

Compare the data sheets for how the intensity continues to rise with increased current and that the forward voltage is not that high.

This component also have a good thermal derating diagram fig.9 that helps designing the cooling and choose a reasonable working current. 

 

The 1400mA in the graphs is for designs for intermittent pulsing or super efficient cooling structures. 1400mA should never be exceeded.

[Akira]

Thank you for the detailed explanation and advice, ulf!

 

The voltage between anode and cathode of my particular sample shows 3.9V, so the current running through the circuit is 1,100mA.  It would be 1,157mA even if the resistor would be actually 0.95 ohm (lowest possible value within the 5% tolerance), which is about 82.7% of the absolute limit of 1,400mA.  I'm not sure if the margin is large enough, but at least it should be safe to assume that I'm not overpowering the LED.  I might add 0.1 ohm for the safety (and the peace of mind), though.

[ulf]

I think it is OK as it is, as long as you are not running it several minutes at the time and letting the box get very hot.

At first I was afraid that your unit would die rather soon. Now I am not.

 

However as you see from the intensity graph of your data sheet, a bit lower current will not give very less intensity.

Also for higher temperatures the intensity goes down so the light loss might even be marginally less.

 

I as a professional electronics designer must design with enough margins to handle the spread of tolerances, to cope with production spread of many devices.

 

Good luck with future experimenting.

[Akira]

4 hours ago, ulf said:

I think it is OK as it is, as long as you are not running it several minutes at the time and letting the box get very hot.

At first I was afraid that your unit would die rather soon. Now I am not.

 

However as you see from the intensity graph of your data sheet, a bit lower current will not give very less intensity.

Also for higher temperatures the intensity goes down so the light loss might even be marginally less.

 

I as a professional electronics designer must design with enough margins to handle the spread of tolerances, to cope with production spread of many devices.

 

Good luck with future experimenting.

 

I think I will run the LED at most for around a minute continuously.  Currently, I don't have any camera capable of the reflected UV image and inevitably must limit myself only for the UVIVF.  I can frame and focus in visible (room) light and use the UV LED torch only during the exposure.  In addition, I will only use it in a room temperature.  Under these use condition, there should be no concern about either overheating or overpowering.

 

Your solid advice helped nevertheless.  Thank you, again!