I want to run around & set the world on UV

[colinbm]

My new baby Ulf....

50w dual LEDs 10 x 365nm + 10 x 375nm

 

 

 

 

[Stefano]

Nice baby! And nice spectrum too! Yes, there's no "green", but still, 360-400 nm is a nice range, it will yield more colors than an (almost) monochromatic 365 nm LED.

[colinbm]

Thanks Stefano

I need to find some lower wave length LEDs, but they will need to match these closely in volts & current, but I or the maker haven't found any yet.

It would be nice to have a spread from 320 to 400nm like this.....one day....

[ulf]

Looks very nice.

If you put a filter on this the filter will be hot as much of the light's energy is within lower transmission areas of both UG1 and UG11!

All energy that is not transmitted will be converted to heat.

[colinbm]

Thanks Ulf

I have a nice copper heat sink of similar size to your aluminium one, with silver paste on the PCB, it is as cool as a cucumber....

Plus the collimating lens is big, about 65mm dia x 25mm thick of UVA passing glass & is just warm, I am running an 80mm fan, half above the fins & half above the LED & collimating lens.

 

[Andy Perrin]

Colin, that looks like you are building the Eye of Mordor please be careful.

[ulf]

Nice build Colin.

 

Each generation is a step forward and the last step was long.

It is a really good improvement in cooling and design.

 

Until I switched mainly to flash illumination for UV photography, I considered building a multi-type LED arrangement like yours.

I though maybe using four LED types, including 385nm and 405nm.

I was speculating about having separate current paths for each type, with a custom PCB, with all mixed types mounted as closely as possible to avoid coloured shadows.

If each type can be intensity controlled separately there is more freedom to balance the false colours.

 

I have dropped that idea completely now and the 150W design, in progress is for fluorescence only.

[dabateman]

So the evolution of a UV photographer:

1. Buy a filter.

2. Get a full spectrum converted camera.

3. Buy better lenses. Maybe even quartz.

4. Buy a spectrometer to optimize filters, light and lenses.

5. Start building odd UV LEDs, capable of taking down small aircraft if focused from a distance.

 

I think that about summs it up.

 

Stefano may have jumped from step 1 to 5, maybe saving some money.

[Stefano]

Well, with my initial USB cameras I had to use a portable PC to take photos outside, and I couldn't even WB...

[colinbm]

Thanks Guys

I could of had four different wavelength LED's combined on this PCB, 365, 375, 385, 395nm, but I didn't want the light spread going too much into the visible above 400nm, so I just asked for 365 & 375 nm.

I do want to find compatible, similar powered LEDs to cover the UVA wavelengths, on one PCB, that our converted cameras can see.....one day....

[colinbm]

First, I would like to thank all the regulars (you know who you are) for all the help & encouragement in building these LED's, without your collective help it wouldn't have been able to happen.

 

I have re-built the nine times quad star LED's that I built some time ago. One of its problems was excessive heat from the light front, but we have learnt how to tame that.

Each star has one of each LED, 365, 375, 385, 395 nm. I can get these made on a single PCB, with like 20, 30 or 40 LEDs with this mix of wavelengths.

But I went with the dual LEDs as shown earlier in this post, because I was concerned that the spill-over into 400 plus nm would spoil its use for UV photography & particularly UVIVF.

So here is a knock-up of a 90w quad UV LED. I would like your opinions of its usefulness or not, please.

 

 

 

 

[Stefano]

It does have some "bleed" past 400 nm, we discussed filtering them before and for that I suggest you to put small filters on each star (9 in total). A huge filter would not make a lot of sense.

 

It seems that the deeper 365 nm LEDs are less powerful than the others. This usually happens as 365 nm LEDs are not as efficient as 385+ nm ones. That's the main advantage in Ulf's idea of separate channels, to tune the "color".

[Andy Perrin]

Yeah, I think that the 385 LEDs are drowning out the 365s and the camera sensitivity will make it seem monochrome-ish. The last iteration was less powerful but it looked better balanced.

[dabateman]

Depends what the purpose is.

I think this will nearly perfectly reproduce the UV level from the sun that most cameras would see.

Slap a Baader venus filter on your lens wnd I think the color would be the same outside or inside with this.

 

[colinbm]

I haven't gone out & taken any photos yet, but I have been thinking about the 'Nemo' torch......

The driver is capable of 1750mA constant current, so I decided to try a higher voltage & tried the 6V battery.

It works a treat, the Nemo now consumes 1650mA & 7.5 watts, which is better then the 4w before.

[ulf]

I haven't gone out & taken any photos yet, but I have been thinking about the 'Nemo' torch......

The driver is capable of 1750mA constant current, so I decided to try a higher voltage & tried the 6V battery.

It works a treat, the Nemo now consumes 1650mA & 7.5 watts, which is better then the 4w before.

Col,

 

I'm sorry, but I must unfortunately tell you some unwanted facts.

 

Yes it is really better in making the torch hotter, but not brighter.

You only increased the supplied power. The added power will only be manifested in more heat from the drivers.

That definitely leads to a lower efficiency of the LEDs.

Depending on how much the increased heat from the constant current drivers is coupled to the LED the light intensity loss might vary.

That is only useful if it is cold and you want to use the torch to warm cold hands.

 

The constant current drivers is not taking any advantage of increased input voltages, on the contrary.

They create constant current making a voltage over the LEDs.

The rest of the unused voltage, the difference between input voltage and diode voltage is burned away as heat in the drivers.

 

To benefit from higher voltages you need a dcdc-converter that is transforming high voltage and a lower current to lower voltage and high current.

With such a device I would set the DCDC-converters output just above where the current starts to decrease. From the other measurements that is around 3.7V

 

This do not increase the optical output much but at least keeps the LED temperature as low as possible.

 

There are two ways to increase the optical output.

Cooling the LED to a really low temperature and/or increasing the current through the LED.

The first is mechanically and thermally difficult.

The second demands the you replace the driver board with some other constant current source.

 

If the current is increased I suspect that there might be problems with the cooling that might be close to the margin just now with current design.