32mm Aluminium vs 20mm Copper

Too late… I’ve blown mine already! I can see 2 little components with black dots in the middle; they both have ‘3400’ on them. After some Googling I understand they are Sot-23 MOSFET’s? Am I correct? Can I replace them?

Damn. I’m sorry if you blew your driver because of what I wrote. I forgot that some lights use alu reflectors, of course they can short the emitter connectors :confused:
SOT-23 is a form factor. You could replace the mosfets if you have a reflow soldering station, but they’re probably not the only destroyed component. Replacing the driver might be easier, if you find one that has the same size (38.5mm) or if you can make an adapter.

They’re probably the only destroyed component, IMO.

I still wouldn’t bother replacing them, I’d just solder a better driver on top of the stock one though. A pair of AO3400 in a light like this is not a great match I think. Led4power’s LD1 or anything with STAR_momentary firmware installed seem to jump out as options. (Such as something from RMM or another member.) Or get started building/flashing your own. It’s not that hard, I promise. Some new guides make it even easier IMO.

No it was my fault, after reading about everyone else shorting theirs out I completely forgot about it… 0:)
So if i wanted to replace the MOSFET’s, what ‘value’ or type do I need? I don’t have a reflow station anyway so I guess replacing them is out of the question.
I wouldn’t mind having a Nanjg driver in there but obviously this light has a momentary switch which makes it awkward

What do you mean “not a great match”? I’ll have a look at his website at some stage :slight_smile: I might get into flashing my own later, I just want this light going |(

well copper is always better as far a heat transfer goes, that is if we talking about direct transfer boards, like noctigon and sinkpad. another question, how hard will you be driving it? if you planing to overdrive the crap out of the led, definatly copper, if you be using no more than max current speced out by manufacturer, al board may be enough.
if we taking about none direct transfer boards, than it depends on the particular star, thicknes of insulation level, amount of “foil” under the led. sometimes correctly designed and made al board will be superiour to poorly designed copper board.

I mean that AO3400 FETs are weak (high resistance). This is a thrower light, so a strong driver makes sense to me. OTOH we don’t want to burn the LED and as I mentioned in tatasal’s Courui thread I’m afraid that a strong DD driver and 3x strong cells in parallel may destroy the LED. I really don’t know.

In any case I don’t like a situation where the losses inside a weak FET are the principle thing “regulating” (not really) the output of a light.

There’s another possibility. I had a similar problem with my Trustfire X6. I got some flat copper strip by pounding flat a copper pipe hangar from the hardware store. I cut a small strip from it and soldered it to the mcpcb so that it made a ‘bridge’ out over the notch in the side of the mcpcb and then carefully soldered the wire to the underside of that, taking care not to solder the wire to the mcpcb itself. This gave me a very solid, low-resistance electrical path to the emitter without being any higher profile than the emitter substrate itself. I just covered the copper strips with some kapton tape and it worked perfectly.

Yeah I just thought that the larger surface area of the 32mm alu might outweigh the shit thermal properties. I have a 20mm noctigon here I’ll use, I was thinking 4.5-5amps would be nice!

I see. Well I’m prepared to find out if the LED will handle it or not! What is this driver like? Would I be able to easily attach it to the existing driver in my light?

The surface area can be increased to equal the transfer rate across the bond line but the substrate contact area remains the same. That’s where the higher transfer rate of the copper is needed most. Think of aluminum as a better thermal insulator.

A lot of people seems to misunderstand that it is not really the aluminium material itself that hindering the LED performance, but the REAL CULPRIT is actually the dielectric layer that limiting the thermal transfer.

If we mount the LED on an aluminium MCPCB without the dielectric layer at all you can still obtain the comparable light output performance to your copper Sinkpad/Noctigon, even if the LED is driven very hard.

I know some people may argue that copper has much better thermal conductivity than aluminium - Yes it does, BUT in our LED flashlight application the heat produced by the overdriven LED is not even hitting the thermal conductivity bottleneck of the aluminium material.

You can read this comment from comfychair in the other thread. He was talking about the pill material but the idea is same.

Most importantly you should read djozz’s test HERE to get a good grasp of the idea. If you read the light output graph in this thread you will notice there is actually very little light output difference between the aluminium Sinkpad (w/o the stupid dielectric layer) and the copper Sinkpad, and that subtle light output difference can also be due to wear-&-tear of the LED from repetitive reflow process and high current operation.

As to answer OP, since the Courui D01 has very wide and flat reflector bottom I would suggest you to go for the 32mm Maxtoch copper MCPCB or the 32mm Noctigon MCPCB, as both of them have direct thermal path (a.k.a DTP). The stock 32mm board (having the annoying dielectric layer) is usable if you are not going to overdrive your LED (i.e. >3A for XM-L2) at all.

I thought di-electric layers were a thing of the past with custom aluminum stars. My statement was in general - copper is best used where cross-sectional area is most limited. Further out away from the heat source, the sheer mass is usually enough keep things cool. At least for a short period.

That said, I won’t argue against real world results. :beer:

Only SinkPAD has those (aluminum stars w/out dielectric). For a proper rundown on the performance of widely available star types (and the not-so-common aluminum SinkPADs) see djozz’s writeup here: 6x 20mm XML-ledboard comparison

Yes you are right, if the produced heat is reaching the “bottleneck” of the aluminium material then only it does matter.

Until one day when we have 100W LED in 3535 footprint (small footprint), then the difference in the thermal conductivity between the copper and aluminium might be relevant until then, and that also depends on all the involved cross sectional areas as you mentioned. :beer:

I don’t quite understand why all of these generic aluminium boards have to come with the dielectric layer. The only reason I can think about is that the solder will always find a hard time attaching itself directly onto the aluminium material, therefore the dielectric layer is there for easier soldering and reflowing. Then again is there nothing else better to replace the dielectric layer for this purpose? And how does Sinkpad deal with this problem with their aluminium MCPCB?

There are several challenges, I don’t know them all…

  • you need a thin dielectric layer for this to work
  • you need to prep the aluminum for solder
  • you need to more tightly align the parts of the sandwich (dielectric, copper, & mask all must be tightly aligned)

None of those things is very critical for manufacturing traditional MCPCBs. At low current aluminum non-DTP stars perform fine; the performance is more than good enough to justify their lower cost!

That make sense, but I still don’t know how aluminium Sinkpad is different from the generic aluminium board. If Sinkpad leaves their aluminium MCPCB completely DTP (“bare”) the end user will have a hard time reflowing their LED on it, maybe special soldering flux is needed, I don’t know.

What’s difficult to understand about it? All you have to do is go to their website and look at the promotional material: the Aluminum SinkPADs are certainly DTP. Aluminum SinkPADs are pre-tinned. They probably use a high-end HASL process. I don’t know what prep is used on the DTP center pad, but it’s also got a HASL finish. (eg it is tinned with solder already) EDIT: to clarify, generally speaking once something is successfully tinned you can just solder other stuff onto it without issue.

This explains it all, thank you Alex. :beer:

Yes, soldering/tinning to Aluminium even at home with a standard soldering iron is tricky but not impossible. I’ve done it myself, just Look on youtube for videos explaining how to get solder to stick to aluminium foil for example.

Critical is just that you don’t allow the aluminium to develop an oxidation layer (happens almost instantly in free air), which I imagine isn’t too hard to accomplish in a production environment under a protective atmosphere. Or however it’s done, but as wight says once it’s pretinned and the aluminium can no longer oxidize then you’re all set for an easy reflow.