I want to run around & set the world on UV

[colinbm]

Knowing the LED's brand & specifications would be a good starting point.

[Stefano]

Does the LED chip have any markings? If not, should we ask the seller?

[ulf]

I have now measured the "Nemo" lamp-head without the rest of the torch.

Doing so takes the batteries and torch switch out of the equation.

I used my high current power supply I normally use for driving high power calibration lamps.

There the background voltage-noise from the switching supply is not that important.

 

The power supply have displays that show it's output voltage, current and power.

Those indications are reasonably accurate within 1% or so.

 

I measured the voltage directly at the lamp-head with my good DMM.

An Agilent U1252 with much better accuracy.

 

The test setup at 4.2v by the lamp head:

 

The result indicates that with a voltage at and above 3.8V the torch consumes essentially constant current.

Then the consumption is 7W:

 

This result show that it is very important to use fresh fully charged quality batteries with a low inner serial resistance to have full power light out from this torch.

The other measurement results presented here are very likely biased by flawed methods or bad batteries.

 

As the optical output for a constant LED temperature is proportional to the current, the blue curve above gives a hint about the light output for different battery voltages, when the battery is delivering current.

[colinbm]

The other measurement results presented here are very likely biased by flawed methods or bad batteries.

 

There was only one other tester....." flawed methods "...........?

[ulf]

I have had communications with others, possibly not on this thread.

Then it should be flawed metod instead, if it was only you, testing with one point with a "full" battery, at 3,7V.

Sorry

[colinbm]

Thank you Ulf, I was using the torch & LED as supplied.....

I wouldn't put the LED on an adjustable Bench Power Supply without knowing the LEDs specifications.

[dabateman]

Ulf your graph says current (amps) not voltage. So you held voltage constant at 4.2 V and varied the current?

I would think you would want to vary the voltage as our batteries may start at 4.2V, but they drop off quickly to 3.7, then trickle down to 3V.

What is the Watts output from 4.2V dropping down to 3V?

I don't know what the internal flashlight driver is so hard to guess its Amps to fix that. You test here show it needs a lot more amps than I would have thought. Or you mixed up the units.

I have one Convoy with 2x7135 so 700mA and now two with 3x7135, so max 1050mA.

Not sure if this is driven similarly or with a totally different class of driver.

[ulf]

Col, your way is the safest way to approach the problem when you do not know how things work.

No problem.

 

I have a better knowledge of such designs and thus more freedom to setup a better test circuit.

No offence intended.

I have just been working with these matters for many decades and can see potential measurement road-bumps more clearly.

 

____

No, David, that is not how I did.

 

I varied the voltage on the power-supply carefully to reach the indicated 0.1V steps at the lamp's head, starting with 3.0V.

I started at 3.0V and stepped up to 4.2V. The direction here is of no importance as no battery is involved.

It just felt safer and more logical to increase the voltage.

That is how you normally do when you power up a design for the first time.

 

I did not even look at the voltage by the power supply.

That must be higher to overcome all resistive losses in cables and connectors.

The goal with the test was to see how much current the lamp-head draws for different input voltages at the poles of the head.

 

Your description of the LiPo cell's behaviour isn't quite true.

They drop a bit from 4.2V but are then decaying rather slowly until the charge is almost depleted.

Here is a discharge curve of a rather extreme battery, but in the capacity range of 26650 used in the torch

https://www.google.c...QAAAAAdAAAAABA1

 

Observe that the lowest current for the top-line in the graph is 35A!

 

A good quality Lipo-cell would be able to deliver full intensity UV by the torch, at least the first 2/5 of the battery's total capacity, maybe more.

Then the current and intensity will decay for quite some time until the fast drop at the end.

 

The power- and current-input is the only thing the graph is telling.

The light output depends directly to the drawn current. Look at the diagram for the current.

I have absolutely not mixed up the units.

 

I would guess there is 5x7135 or even 6x7135-drivers inside this torch from the final current level of 1.88A, that I measured.

There are tolerances in the constant current levels from those drivers.

 

It could be driven by a different type of constant current IC, but the principle is identical to the 7135.

[dabateman]

I was just able to carefully unscrew the back brass ring. My light has five 7135 drivers back there. There is room for 6, but only 5 are in place.

So that should tell you exactly what this light is capable of doing.

5x 350mA is 1750mA maximum.

[Andy Perrin]

1750mA vs Ulf's measured 1880mA is good agreement.

[dabateman]

Sorry for the bad black and white image but this is what the Nemo driver looks like. You can carefully unscrew the back the brass ring with a lens tool in the slots.

 

Image 1

 

Image labelled:

[colinbm]

Good to see thanks Dave.

[ulf]

Thanks Dave.

I could not find a suitable tool to do that and had no time to make one.

 

So the Nemo "15W" is actually drawing around 7W max.

That difference between marketing "information" is rather typical for torches

If the LED is a 15W type that information would even be true in some sense in the marketing world of deception.

[ulf]

If someone have noticed, in my test picture, that there is a difference between the voltage value on the top display of the power supply and the value measured by the multimeter.

That difference is due to the losses in the cable and connections.

 

The cable is rather long and the test clamp is not ideally connecting to the lamps head.

For the measurement, that is not important, as I measure the voltage directly at the lamps head, just to avoid introducing these losses in the measurement.

 

Such a loss is similar to what you see from the internal resistance in a battery and in this case in the same size as from a good quality battery.

[dabateman]

Ulf,

Your curve has two slopes. The main up to 7W at 3.8 Volts. Then up to 8W at 4.2 V.

Can we push this to 9W at 4.5V? I have a AAA three battery holder that might fit the flashlight. Using 3x 1.5V batteries I wonder if it could be pushed a little farther.

 

Are we voltage limited? Or are we limited by only having five 7135 drivers?

[Andy Perrin]

That depends very much on whether this is truly a 15W LED being driven at 7W, or a 7W LED being driven at 7W. If the latter then pushing the voltage up too much will burn out the LED. They have a saturating I-V curve and if you change the voltage a little you can change the current a lot...

 

Also it might just all turn into heat.

[ulf]

Ulf,

Your curve has two slopes. The main up to 7W at 3.8 Volts. Then up to 8W at 4.2 V.

Can we push this to 9W at 4.5V? I have a AAA three battery holder that might fit the flashlight. Using 3x 1.5V batteries I wonder if it could be pushed a little farther.

 

Are we voltage limited? Or are we limited by only having five 7135 drivers?

You can push the input power a bit further by increasing the voltage, but there will not be more light, only more heat.

It would be completely meaningless and counterproductive.

This is what I ment before about that it is not really the input electrical power, that is important, but the optical output.

Also we do not know the specs of the LED and it.s maximum ratings.

 

These 7135 drivers are linear constant current devices and cannot benefit from a higher voltage.

To do that you'll need a switch converter and there is no space internally for that.

Ideally it should be a constant current step up/down converter with a low voltage lock out to protect the battery

I also highly doubt that the internal resistance in any AAA battery would make that exercise meaningful.

 

Andy, your answers are correct. I read them after writing the text above.

[ulf]

To push things a bit further I think the torch/rubber-band bundling with these torches is the best solution.

With three torches it is still manageable and you get a 21W input power to a reasonable cost.

 

Just now I really do not need anything more powerful than a few Nemos.

 

Solutions for higher power quickly becomes more complicated needing active cooling and more massive batteries or power sources.

 

It is always fun to have some design projects evolving slowly though.

A 150W UV-curing LED array might be used for a high power lamp head.

If I have time and inspiration enough this might be ready some time next year:

This is only the core mechanics of the device.

It is growing almost organically when I solve the different design problems.

I have sketches that change all the time and build slowly.

For safety reasons there will be a lot of electronics handling activation and cooling.

With these power levels the optical power will be directly dangerous and I want to think about the safety, a bit like for a smaller weapon.

 

The lamp head has to be supplied from a separate constant current source, to be designed later.

I will start using a lab power supply, when finalising the lamp head design, but the goal is to make the complete lamp fully portable and battery powered.

With some luck I can try it outdoors after next summer.

[colinbm]

150w Ulf......whooa......that must get very hot in front of that light, can the filter cope with that heat ?

[ulf]

TBD.

 

It will cope at least for shorter time periods until it needs to cool down.

I might have to have a thermal sensor against the glass too, but it will be removable if needed for using it longer periods of time.

 

Glass can normally tolerate elevated temperatures and I think it will survive at least 150°C.

Then the passive cooling by the air will be rather good.

 

I would never risk a proper Schott or Hoya glass here or my Baader U.

[colinbm]

TBD.....& I sweat about filters in front of 50w UV LED.............

[ulf]

The LEDs are comparatively very much more sensitive to heat than the filter glass.

As long as the glass do not crack due to stress I do not think there will be any problem with it being rather hot, except possibly accelerated ageing or touching it.

Then the UV energy itself might be a bigger factor for ageing.

[colinbm]

I have broken filters in front of 30w HID lamps....

[colinbm]

Fresh news on our mystery 'Nemo' 15w 365nm torch...

A new one arrived today, my second, & protruding from the multiple address labels is a contents label, including battery details...

 

3700mAh, 26650, then sum Chinese characters.

 

I wonder how spot-on these figures are....?

[ulf]

If I take values from my test the current drawn max is 1900mA and the torch head needs at least 3.8V to do so.

Assume that the battery fast drops from 4.2V to 4.1V.

Then to be able to get 1900mA through to the LED the total voltage drop with that current flowing through switch + battery + contact areas in the body cannot be more than 0.3V.

That allows the total serial resistance to be no more than 0.3V / 1.9A = 158mΩ as there must be some margin for the other components losses my guess is that you need a battery with less than 125mΩ

Quality batteries designed for high current have quite a bit less resistance while cheap no-name batteries or old worn ones can have more.

 

It is rather easy to test if you have a voltmeter and a power resistor with a resistance of around 2-4Ω. Just measure the voltage drop when the battery gets loaded.