1 MC-E vs multiple XR-E which needs more heatsink?

[yellow]

I have to admit my english might no be good enough to understand the difference?
:thinking:

 

[lightime]

Isn't it the same question from a different perspective? I want to know which needs less heatsink because I am limited by space. Whichever one needs less leaves the other one needing more or "better" as you say

I am being rather pedantic I know but the OP's question was about which needs more. Everyone has given answers based on what would be a much better question. "Which one needs better heatsinking?" That is what should have been asked.



I am OMG Lumens. See my sig line.

 

[saabluster]

Isn't it the same question from a different perspective? I want to know which needs less heatsink because I am limited by space. Whichever one needs less leaves the other one needing more or "better" as you say

As I see it.

More=quantity
Better=quality

More "can" be better but not necessarily. The outside of your light can only shed so much heat for a given set of conditions. You could add more metal on the inside to extend the amount of time the light could be on before overheating the LED or you could improve the quality of the materials used. Improving the quality can allow you to run the exterior at higher temps while still keeping the die temps low enough. Just adding more of a lower quality material will not appreciably add increased steady state performance but can increase short term performance.

You say you are limited for space. I am not familiar with your application but if your limit is one of of diameter I would go for the MC-E. If it is of depth go for the XR-E.

 

[Frozen]

I have 3 bike lights, each made by two DX housing sku.13741. The first one is driven by quad XR-E R2 and two are MC-E from Cutter. Design, driver Maxflex, battery, and other materials are almost same.
Quad R2 run 1 minute longer (5'30'') than MC-E before Maxflex get 60°C and switch to low mode.

 

[SemiMan]

The difference in forward voltage is 0.1 volts approximately at typical operating currents, so call the typical forward voltage at 700mA 3.4 volts. The MCE has an advantage of 0.1/3.4 or roughly 3%.

In a typical flashlight, with each die running at 2.5watts, I likely want my thermal resistance to be on the order of 40C/watt or less. We could even call it 50C/watt or less.

The XRE has a 4C/watt improvement in thermal resistance. That gives it an advantage of 4/40 or 4/50 or somewhere in the 8-10% range or far better than any small reduction in forward voltage.

The increase in die temperature, 10C max on the MCE is only going to reduce the die voltage by 20-30mV, so maybe that 3% goes up to 4%, but since it is a closed loop, the effect is less. 25% of the energy going in is converted to light, but since the efficiency difference is at best slightly better for the MCE due to the low forward voltage... so I would be maybe at 3.5%*(1+3.5%).... want to call that 3.6% ...heck call it 4.1%.

Anyway you look at it, for the same overall heat sink, the 4XRE die will be running at a lower overall temperature. As one designs thermally to either minimize die temp to maximize output and life, and other designs for minimum heat sink for a given die temp, the 4XRE wins just based on bulk heat sink.

A practical implementation means the 4XRE will likely have a better thermal path to the bulk heat sink which means that the 4C/watt advantage may actually be 5-6C/watt or possibly 15% which will be vastly better than the MCE implementation. Let's not forget the 4XRE have a larger IC to board bonding areas so in most practical implementations that may be another 1C/watt or more... so now we are at 6-8C/watt.... getting up there on 20%.

So my analysis stands, the MCE will require more heat sink, and/or the XRE will require less heat sink to meet a maximum die temperature specification at a given current drive.

Semiman

 

[saabluster]

So my analysis stands,

And so does mine. I purposely left open that side of the argument just because I wanted to see your comeback. And I'm lazy. As a matter of fact there is quite more to be argued in favor of the XR-Es. As I mentioned before my point was to answer the OP's question and since everyone was jumping on the XR-E only bandwagon I thought I'd side with the poor helpless MC-E . I agree with most everything you said but it is always fun to be the devils advocate. We are both right depending on the conditions and variables, the end goals from a design perspective, and depends as well on the definition given to some of the terms. I do concede that the XR-Es win in the vast majority of circumstances. But not all.

 

[SemiMan]

Someone doubles as a politician.......

 

[saabluster]

Someone doubles as a politician.......

I did not have sex with that woman!

 

[csshih]

I did not have sex with that woman!

That's what they all say.

 

[WeLight]

My 2cents worth, caus I found two and it dont buy anything else

The MCE, M bin at best is a Q4 emitter x 4 so its already behind the eight ball, it uses a leadframe as part of its heatsinking model and does not sink the same way as XRE so it becomes difficult to compare. As a package it has some marked improvement over XRE in terms of density and if you could get 4 XRE die as close together you would still have Q4 vs R2 as a comparison so certaintly MCE will require more heatsinking just on comparable efficiency

 

[saabluster]

My 2cents worth, caus I found two and it dont buy anything else

The MCE, M bin at best is a Q4 emitter x 4 so its already behind the eight ball, it uses a leadframe as part of its heatsinking model and does not sink the same way as XRE so it becomes difficult to compare. As a package it has some marked improvement over XRE in terms of density and if you could get 4 XRE die as close together you would still have Q4 vs R2 as a comparison so certaintly MCE will require more heatsinking just on comparable efficiency

The M bin MC-E is 430 lumens at 350mA per die. Divide that by four and you get 107.5 lumens per die at 350mA. That means the MC-E dies at minimum are averaging ever so slightly above Q5 minimum spec. That and the OP's question was comparing Q5 binned XR-Es. Still thats a good point. You could just as easily use R2s instead of the Q5s.

 

[lightime]

Thanks SemiMan for the great explanation. I wanted to thank everyone for all their input. Tons of great info here. Very much appreciated folks. :bow:

Now it's time for me to start applying it. I have a bunch of Q5 XRE's and even a couple MCE's coming for me to mess around with. MANY more questions will follow