Relatively cheap DMM, sorry.
But I think the proof is there, heat sink the snot out of it and it will run.
But how to heat sink it in tight spaces? I will do my utmost to improve the thermal path to the ground ring. I have not had time to revisit the design since yesterday. Tonight is a write-off because it's footy night and after my game I will be sticking around for a team BBQ which inevitably does not end until the wee hours of Saturday morning. It's a tough life.
wight wrote:
I can’t speak for the rest, but for me your work laying out PCBs is more valuable than refunds on parts that still operate. Everybody knew that this was an unfinished design when they ordered parts.
+1
Drink a pint for us! 
+eleventy hogillion!!!111oneoneone
I'm in for the refund. I hope you have lots of money.
Hey Dale, I may be able to get you that power supply after all.
We’ll just have to design a copper panel to lay in under the regulator with water cooled pipes running through it, pumped through a radiator and backup battery pack carried in a backpack. No problemo.
have fun playin footsy with the guys…. (cringe)
Footy. Football. Also known in this country as one of Rugby League, Rugby Union, Australian Rules (AFL), or soccer. My shoulder reconstruction 2 years ago forced me to stop Rugby League, so now it's the surprisingly-notmuch-more-gentle* soccer for me.
*seriously I've had more medium to major injuries playing soccer than I did with rugby. And no, I don't prance around like a girl when I get tackled. This is amateur soccer; we punch on to settle our differences :D
No worries Matt, we’re all real patient around here and won’t do a thing without your say so. 
I’ve seen what you can do with a pile of copper, surely you can create a cradle, a heat sink that sit’s under the regulator and covers it’s top with finned goodness that allows it to run cool!
guy’s a genius creating parts
Matt, no way man! I owe you, if anything! I know what I signed up for, I haven't had time to build my own yet but I've been trying to follow along although there is about 100 posts I need to read to get caught back up to speed.
I'm just glad we've got someone who can design these boards and give us some new stuff to try out. I think we'll get it working eventually and if not, it was a fun ride!
A few thoughts:
1. I think that we are pushing the chip too hard and more heatsinking is not the answer. It may be a small bandaid but it's not going to be enough.
2. How about a 15mm or 16mm slave board that only has the buck circuit, then we can try and run the ICs at around 2.5A-3A and see if they'll live. I guess first we should try and up the resistor a bit and see if the current board will last at around 2.5A-3A and go from there.
Yeah, I was going to replace the R025 with an R056 but already had the 25 covered up with heat sink. Figured I’d just run it and see what happened and try the 56 on a new board when they get here.
I ran the dog crap out of that one yesterday though, took it from 3.84A down below 3A on a this new set of Efest 35A cells and it worked fine. Modes kept working and all that. But the chunk of copper glued to it was itself stuck to 27 cubic inches of aluminum. Shows that it can work though. I don’t think I’ve ever ran an MT-G2 for 8 solid minutes before at 3A and more. The big chunk of copper the emitter is on was HOT, even sitting on the aluminum cube and in open air. Which makes me think that this driver and this emitter can’t really be used together. The emitter will heat up the host, the regulator won’t have heat sinking because the heat sink will be saturated by the emitter. Vicious circles I know.
So I’m thinking about that lower output, to an XR-E or XP-E2 or something. Long run time, intense beam profile, maybe a little pocket thrower? Don’t know, but we’ll figure out where it’s niche is I’m sure.
How much amperage can an XP-E2 on copper take? 2A? A bit more?
DBCstm, RMM, you two are starting to put your finger on the problem here. If we don't want these drivers to act like the KD beast drivers (20 - 60 minute life span) we need a fair amount of wiggle room in terms of performance and durability. Figuring out wiggle room by just turning down current arbitrarily and checking to see if the chips still pop is not really ideal. That's why I'm recommending that we base our next tests of these drivers on the datasheet's absolute maximum dissipation rating. Use a graph such as djozz's crash test data to figure out what current a given wattage will allow you to run with a particular LED.
| efficiency (Percent) | output current (Amps) | losses (Watts) |
| 75 | 6 | 2 |
| 76 | 6.333333333 | 2 |
| 77 | 6.695652174 | 2 |
| 78 | 7.090909091 | 2 |
| 79 | 7.523809524 | 2 |
| 80 | 8 | 2 |
| 81 | 8.526315789 | 2 |
| 82 | 9.111111111 | 2 |
| 83 | 9.764705882 | 2 |
| 84 | 10.5 | 2 |
| 85 | 11.33333333 | 2 |
| 86 | 12.28571429 | 2 |
| 87 | 13.38461538 | 2 |
| 88 | 14.66666667 | 2 |
| 89 | 16.18181818 | 2 |
| 90 | 18 | 2 |
| 91 | 20.22222222 | 2 |
| 92 | 23 | 2 |
| 93 | 26.57142857 | 2 |
| 94 | 31.33333333 | 2 |
| 95 | 38 | 2 |
EDIT: that's output power (watts) not output current (amps). Adjust accordingly.
Mattaus wrote:
DBCstm wrote:…the proof is there, heat sink the snot out of it and it will run.But how to heat sink it in tight spaces? I will do my utmost to improve the thermal path to the ground ring. I have not had time to revisit the design since yesterday. Tonight is a write-off because it’s footy night and after my game I will be sticking around for a team BBQ which inevitably does not end until the wee hours of Saturday morning. It’s a tough life.
The first recipe listed here seems to work fine and is fast, cheap, and reversible. Just need q-tip (cotton swab) dipped in alcohol to get off any residue or potting in difficult to reach areas.
Ok, somebody help me out with proof of concept here…when we talk about thermal conductivity I think Silver or Gold. Copper then naturally comes to mind as an economical alternative. When I think of ceramic or silicone, I think insulator. Stop the heat. Silicone rubber oven mitts let me grab a 425º pan out of the oven without getting burned because it DOES NOT conduct that heat to my hand. If I had a copper oven mitt I’d be burned in a millisecond.
So how can it be that in our flashlight world ceramic and silicone compounds are being described as thermally conductive? Seems to me a layer of silicone between 2 pieces of copper is going to shield the heat from going between those 2. What am I missing?
First of all, realize that when we say “best” in relation to thermal interface materials we are often talking about the best of the worst. What I mean is that the vast majority of greases, putties, and pads suck compared to copper/gold/etc.
That said, think about the difference between Copper and Brass. Brass is mostly copper, but it’s worse than aluminum.
2 nice thick gold discs, pure gold, a small fortune worth, with a thin layer of silicone separating the discs. Doesn’t matter that the gold is incredibly efficient and outweighs the silicone 100:1, the silicone is a wall that stops heat from transferring from one disc to the other.
Brass is like that. You’re taking nice copper and adding crap to it that kills it’s effectiveness. Doesn’t take much crap to ruin your supper.
And that is what puzzles me, we use copper as a heat sink but put silicone between it and the heat. A wall, as it were. Wouldn’t a direct contact with the hot surface and copper be more efficient? Affixing the copper with mechanical means if necessary?
I’m going to have to disagree here. Proper heat sinking is sop when dealing with high power components, sometimes using both a thermal pad and a top mounted sink. Much as I like the creativity of it, Dales coil on top represents more of a bandaid as it was done as an afterthought. The chip is designed with a thermal pad but in its current rendition, we can’t access it to attach a proper sink. Personally, I think Matt’s hit one out of the park here just by getting a 3A buck driver into a 17mm package and if he does nothing more it would already be a success but I won’t throw in the towel until I’ve seen what it can do when given the same attention we give to LEDs, namely, a direct thermal path.
I never would have thought of that! You’re a Genius Scott, pure Duck Genius! A copper woven towel over the heat source, throwing in the towel can be a win after all! Woohoo!
Sorry, happy camper here. Just got word I have a Ti light about to head my way. 
Matt, you gave away the possibility of financial gain by making this an open source project. No way should you be financially responsible for the open source testing as well. I’m still going to purchase kits when available (hopefully with flashed mcu’s) because I think it does work but is being restricted by heat. There may well be other ways to improve design so let’s not stop just yet. Puhleeese?
I think that adding more heatsinking from the bottom of the chip (the thermal pad) would be much more beneficial than trying to add it from the top...but really at the end of the day I think we're just pushing it too hard. It is probably much more efficient at lower driver current and it is probably more efficient (although more expensive!) to parallel two of these circuits together.
I've been looking a bit at the normal buck controllers that don't have an internal FET but that switch an external FET...that would give us the ability to have a lot more handling capacity, although again it gets harder to fit it all in a 17mm package.