maxflex5

[georges80]

Finally received the first 60 maxflex5 drivers yesterday and expecting a whole lot more in a few days. So, around the 20th Oct I'll have plenty to ship. I have a bunch of emails from folk that have patiently been waiting. Anyhow, here's a top & bottom view of the maxflex5.

It's basically the same as the maxflex4 (same improved thermal handling due to 2oz copper and thicker plating of vias) with the major addition being a dedicated hole/pad for the ground connection to the momentary switch (SWG in the picture below).

The boards are also ENIG (gold plated) like maxFlex2, 3 and 4. The ENIG provides excellent 'flatness' of the pads/traces (and thermal interface area) and also better solderability. The PCBs cost more, but it improves the quality of the maxflex driver board.

cheers,
george.

 

[il2mb]

George,

Can I drill a through hole in the heatsick area? I want to secure a block of aluminum with a small machine screw.

Thanks

 

[georges80]

George,

Can I drill a through hole in the heatsick area? I want to secure a block of aluminum with a small machine screw.

Thanks

Sure - assuming you want to destroy the driver...

You do realize that the other size of the gold thermal pad area matches up with the switcher IC? That grid of 4x4 vias is a direct heat path to the bottom thermal pad of the IC.

cheers,
george.

 

[il2mb]

Damn that was a stupid question.

Thanks George. I'll go pull my head out of my a$$.

 

[yellow]

thermal paste between alum and heatsink area,
work the block in (press and move it around a bit), so that the paste is as thin as possible
epoxy glue around block (+ weight on the block), let cure
same on other side of alum block + light housing (or back side of led thermal/mounting plate),
done
--> perfect thermal path for driver

 

[marcopolo]

Yellow, a while back myself and a friend argued endlessly about this point, delving into datasheets for artic alumina epoxy vs paste and on paper it all looks good - the paste should conduct heat away faster.

Well, we eventually tried both and it made absolutely no practical difference to the temperature of the driver which method we used.

In short it's an awful lot of pointless mucking about using paste in the middle then epoxying th edges.

I would say thermal epoxy is fine, not the silver stuff though as is may get stuck on components nearby and it is capacitive. The artic alumina is better. In the end if the day the temperature equilibrium point will not be very different and certainly not enough to effect the performance.

If however you are dropping 100W of processor power - then you need all the heat sucking power you can get, but if you look up the thermal management for the maxflex - even under extreme conditions of low i/p voltage and high o/p voltage there's not that many watts to take care of - it also mentions heatsinking of the inductor in extreme conditions.

Marco.

 

[yellow]

when You see it that way,
both paste AND thermal glue make no sense,
a thin layer of normal epoxy glue works just the same
AND IS CHEAPER considerably,
(that is my main point against thermal glue. Imho that is totally useless and expensive)

 

[marcopolo]

AND IS CHEAPER considerably,

Dude how much are you using !!!!!!??

I bought a 3.5ml 2 part tube of arctic alumina 2 years ago for £8 and it has served countless projects!!