Temperature of LED junction for forward voltage measurement?

[Flyback]

Hello,
Page 5 of the attached datasheet gives forward voltage measurements for the UV LEDs at 350mA , and says it's done at an ambient temperature of 25 degC...however, do you know what was the junction temperature for these measurements?.
I mean the measurment might have been done by simply pulsing the leds with 350mA for 1millisecond every 5 seconds, and thus the led junction would be at about 25 degrees C, although the measurement might have been done with 350mA flowing continuously in the LED, which would have meant the junction temperature being much higher. This has an obvious effect on the actual forward voltage since as you know its variable with temperature.

 

[audioguru]

Why is the temperature of the forward voltage measurement important? LEDs are fed a certain current, not a voltage. The forward voltage drops a little as it warms up so simply make a current regulator that delivers enough voltage to a cold LED.

 

[Flyback]

yes but our led load is fixed by the semiconductor company......and our vin is fixed at 48v (due to cable drop) as you just cant find off-the-shelf psu's that are sub $150 and deliver 600w at 90vac...so if our leds have too much forward voltage then our buck led drivers drop out.....we cant fit any other toplogy driver on the small driver pcb size.

We need to know if those readings of vf are at 25 degc..if they are then we can breath with relief that the 3.8v one will never occur as the junction temperature will always be well above 25degC...however,if that 3.8v one is at Tj=85degc then we are in trouble.

 

[alec_t]

Have you contacted the LED manufacturer?

 

[jjw]

It should have been announced if it is not measured with continuous 350mA current.
It is strange that there's no mentioning about thermal resistances, junction to case and case to ambient.

 

[audioguru]

It is a chip LED that will quickly vaporize if it is not mounted and cooled correctly.
The Korean manufacturer says they have no control over somebody's mounting methods.

 

[ronv]

They are pulse measurements. That way they can quickly bin them. The spec doesn't say that for sure so if still in doubt, call them up.

 

[Flyback]

I've not yet received a reply, as you know we need the info (max likely vf) to check out the led driver design. The statistical chance of receiving 3.8v pieces in any given batch would be good data to have also......I mean, we don't want to get the design specially for the highest vf if its hardly ever likely to occur.

They are pulse measurements.

...so you mean the junction temperature was 25 degc for these measurements?

 

[MrAl]

Hello,
Page 5 of the attached datasheet gives forward voltage measurements for the UV LEDs at 350mA , and says it's done at an ambient temperature of 25 degC...however, do you know what was the junction temperature for these measurements?.
I mean the measurment might have been done by simply pulsing the leds with 350mA for 1millisecond every 5 seconds, and thus the led junction would be at about 25 degrees C, although the measurement might have been done with 350mA flowing continuously in the LED, which would have meant the junction temperature being much higher. This has an obvious effect on the actual forward voltage since as you know its variable with temperature.

Hi,

Who cares what the junction temperature is? If they state 350ma at 25 degrees C ambient and they dont say pulsed, then that's a continuous current and the max voltage is then 3.8v at 25 deg C ambient. From the other chart we see that the extrapolated low temperature operation could be up as high as +0.2 volts putting the grand total at 4.0 volts, but that's only if the LED will ever be powered at an ambient temperature LESS than 25 degrees C. For operation at or above 25 deg C ambient it will always be 3.8v or lower.

-0.2 volts over 60 degrees C says it's got a temperature coefficient of -0.2/60=-0.1/30=-1/300 volts per degree C ambient at the max current level of 350ma.

So junction temperature never comes into play and it's a good thing because they dont seem to spec the thermal resistance junction to case. They are always working in terms of ambient temperature anyway so it is a wonder why they even specify the max junction temperature, but because they do work in terms of ambient temperature that means we can too.

What else is strange though is that they dont specify what they have the LED mounted to for the tests.

 

[alec_t]

I mean, we don't want to get the design specially for the highest vf if its hardly ever likely to occur.

If that means you are designing for less than the highest vf then you are going to have some unhappy customers when Murphy's Law bites.

 

[kubeek]

The statistical chance of receiving 3.8v pieces in any given batch would be good data to have also

Did you actually read that datasheet? It clearly says that they sell them graded by Vf, you can see five different bins from 3.2 to 3.8V. The proper question would be how exactly do they sell it, last time I have ordered some binned leds you could order for example groups 1-3 or 3-5, so check the availability of the bins with the manufacturer and/or distributor.

 

[ronv]

I based it on the range of binned values matching the basic spec.. I can't imagine waiting for each one to warm up to measure it.

 

[Flyback]

Did you actually read that datasheet? It clearly says that they sell them graded by Vf, you can see five different bins from 3.2 to 3.8V

It is not possible to order any bin of any led......it is possible to order a reel of leds, and you can be sure that all the leds on "that reel" will be from the same bin, but you have no say on what that bin is.....if it was possible to pick your bin, then everybody would only ever pick the lowest bin voltage.
Yes it sounds a little strange to me too......but I was convinced of this about a year ago on a forum when many people pointed it out to me......I told them it was like me going into a restaurant and ordering a meal and hoping it would be steak and chips....but ending up with fish and chips instead and just having to put up with it.

 

[alec_t]

you have no say on what that bin is

So you have to design for the highest voltage .

 

[tvtech]

You know what Guys..

Most/all LED applications are Current Driven. Like put a nice driver in there and all is good to go....I expect a firing squad here...unless of course I make a bit of sense.

Just say for instance you do some sweet calculations. Your Maximum Voltage across all can never go above the Maximum for any LEDs ever...
Therefore, a string cannot fail. No more Current ever that another one must handle. Voltage derived. And therefore safe.

Am I making sense ??

Yes, LED's are current driven devices....but think about it when many are connected together ...

Just to re-iterate on my basic point:

1oo LED's of 2 in a string @ 20Ma per string. Driven with a Supply Voltage of 9 to 10 Volts. With a little 150R dropper resistor per string.
Stable supply voltage is 9 to 10 Volts....

If strings decide to fail, it wont bother the other working ones. Voltage is the thing.....
No other string has to work harder to handle excess current from the broken string...

And then they all fail getting hotter and hotter..

Classic stability with my stuff that has been reliable for Years. I use Voltage to control strings..NOT Current. One string failure makes not another die.

Just a wee bit happier...

Yet to have a string die. So I cannot prove my theory to you all..but it seems to work.
Last I checked my little light had clocked 12240 Hours. Running 24/7.. on Mains.

Sometimes thrown around, dropped or abused. But always works.

@ 3 Hours a Day...(like NORMAL People) that is around 11 years of testing already.

Right is right.

Now I believe that is one cool friend. When you need it, it's there. Period...Especially in a strange room or a hotel..and need a pee ASAP....the more you think about a pee...the more urgent it gets...I peed on my (ex) Wives shoes once....could not find the damn light switch...

So I opened the Cupboard door....and peed there. On her shoes. That was that.

I thought it was the Loo. I kid you not.

My 2c here.

Regards,
tvtech

 

[MrAl]

Hi,

Yeah sometimes you can get certain bins from a company and sometimes you cant. They make 'wider' bins sometimes too and that at least gives some control over what you are getting although still not perfect.
The thing is, the better bins are higher priced too though and could be twice as much as a lower bin part. So there is a cost trade off at the very least if they do sell individual 'wider' bin parts.
Then some sites do their own secondary binning too and it's almost always wider so it becomes even more questionable what you are actually going to get. I suppose some might bin tighter too but i dont think i've ever seen that yet.

If you really wanted to you could work up your own bin system by buying a bunch and binning them yourself based on the probability of getting a certain bin part, and that would help determine the cost of resale when you sell the ones you dont want, if you can sell them.

 

[tvtech]

My Post modified above Mr Al's one. My nature

I built my little light to be indispensable to me.

Sleep well Guys.

 

[ronv]

The odds of getting 11 or 12, in the same series string, at the tippy top of the spec. will be very very low.
But they do start out at 25C. Right?
And of course they will still heat up, just not as bright, even at lower voltage. Or will your driver not power up???????