Yezl Y3 - a picture breakdown

But we're talking about, what, a max of 35 watts for one XML2 before the bond wires fail? And usually at that level it'll be on a 20mm board? That's a huge area and not very much heat, compared to a test designed to replicate a CPU heatsink scenario. The differences between a good paste and a not-good paste will be small enough at our power/heat levels that they will not show up in the amount of light output. Is that not what matters?

In the CPU world, back when CPUs had a tiny exposed die, quality of the paste and the clamp loads present were ultra-critical. That's when most of the thermal paste shootout comparisons were done. Currently, with nearly everything having a bonded-on heatspreader on top though, the performance of the paste is much less critical, the critical joint has moved inside the CPU between the die and the heatspreader where (most times) you can't get at it, and Intel/AMD use very good quality pastes inside there. Some heatspreaders are even soldered to the die.

And in the other thread where I asked about taking a star-pill contact print when using screws, the reason is that there are only two very small areas around the screws where the clamp load is high enough to transfer dye from one piece to the other. It's pretty horrible, really. And thermal pastes only work when the clamp loads are high enough to transfer the dye, anywhere else the layer is too thick to transfer any meaningful amount of heat, and the paste just acts as an insulator instead. And yet... it still works. There's very little if any difference between using screws where there is only a tiny contact area due to MCPCB distortion, and soldered where there is 100% contact. At the levels possible in a flashlight these minutiae just do not make the slightest bit of difference. If they did, if the difference between a good paste and a shitty paste actually showed up at flashlight power/heat levels, I'd be all for it. I'm not for stuff that doesn't make any difference.

Well, sounds like the screws is better than no screws on an aluminum pill? Again, I wish I had all the machining tools, jigs and skills, but I don't now. I've done some soldered MCPCB's, and I'll probably do more, but it's gotta be the right setup for me - I don't have the options you do.

I'm thinking I've seen improvements with screws vs. no screws, specially on zoomies, but I'll try to leverage soldering more, specially on high amp lights.

I’ve wondered why we use paste at all. Especially if someone goes to the trouble to lap both components, how can it really help to put an insulation layer between em? Fujik, especially, as it seems to me to just be a sheet of rubber between 2 pieces of metal.

I use thermal paste or thermal adhesive, because it’s what I’ve been taught to do as a trained chimp. Don’t understand it, just do it.

I too solder or reflow the star to the brass or, more rare copper, pill when it’s there.

First time I'm hearing it's an insulation layer. So, a star sitting on it's edges over a hollow pill is better than a star sitting on a solid pill top with AS5 in-between? I'd say prove it - make a believer out of me, because I'm not sure how this makes any sense. I don't have time or the instruments to do the proper measurements. I thought I understood the theory of why to use thermal grease, and I thought I saw measurments that show thermal grease is better than air - I'll believe it until someone proves otherwise.

Whats the purpose of off thermal paste? There is no such thing as two surfaces that have 100% surface contact. It does not matter how highly lapped together two surfaces are there will always be microscopic pits etc. The purpose of the thermal paste is to fill these voids. Nothing else. The same goes for gasket goo.

And if that’s the case, then a microscopic layer is all that’s needed…and of course the better that material carries heat the more effective it will be, right? So Fujik, white silicone, rubber, as it were. A 1/16” thick layer of the stuff like comes on so many lights from the factory, seems exactly like the silicone pad we have on the table to PROTECT the table from the heat of a pot fresh off the stove. It doesn’t allow heat through, but insulates against it…right? Same goes for our Silicone oven mitts. Grab a 425º aluminum cookie sheet out of the oven and not get burned…because the Silicone does not flow heat through, but insulates! Exactly like the silicone insulation on our top end wires.

Thermal paste, yeah, I can see if it’s applied properly that it would fill those microscopic voids. But it’s not to be used like glue or mortar. I guess then that our radiant stove tops are horribly inefficient, glass to metal pan with no paste between them. Point being, for the relatively low wattage of our lights, is it really a big deal?

Place a copper disc 1/2 on 1/2 off a burner, holding the part that isn’t on the burner. Think you’re gonna hold it for very long? And that, with no paste and nothing holding it tight against the hot burner. Heat is going to transfer, microscopic pores or not. Filling microscopic pores might indeed make a difference in an extreme instance, but is it really necessary for us?

That’s my question. If a bare copper star is pressed firmly against an aluminum pill, no paste whatsoever, will it kill anything? Will the difference be measurable? I’m not arguing for or against, just trying to learn something. Sure seems to me the difference is very thin.

From a practical standpoint, if the reflector and lens fit the light properly then the star is going to be pressed firmly against the heat sink of the pill. The only way I can see it making a huge difference to have no thermal paste is if it’s loose, allowing the star to not make contact. Screw it down and it seems to me it’d make little difference. I say that again not in argument, but in lack of understanding.

The efficiency of the led is directly related to how well it gets rid of heat. Theres plenty of graphs around showing this. The holes or microscopic holes are filled with thermal paste to assist in the heat transfer how ever good or bad the surfaces are. If there is an air gap heat is not transferred efficiently. The past takes away this air gap. Ideally the led will have star to pill contact as much as practical with the thermal paste providing a path for anywhere there will be air. When applying thermal paste it pays to put a small dollop in the centre of the star or pill and the two items pressed together so the paste is squeezed from the centre outwards to eliminate any air pockets that may form.

I keep all thermal paste away from my table and stove as I believe it may well be toxic.

I guess what puzzles me is the over use of these fillers. I would think pure metal to metal contact would have optimum results. If microscopically the surfaces are made up of ridges and valleys, then the paste should theoretically fill the valleys while allowing the ridges to make direct contact. With the paste having less ability to transfer heat than the actual metal, these valleys would be the weakness in the thermal pathway.

So, if too much paste is used, there is NO direct metal contact and a considerably weakened thermal path would be the result. Would this seem logical? I guess I’m trying to figure out the best way to use the least amount of paste possible.

Correct. Simply put, as much metal to metal contact as possible with no air at all.

The coollab stuff is 100% metal btw - 100% liquid metal alloy.

Metal to metal with very thin grease to fill the microscopic voids, but that's all in the ideal world with perfectly flat surfaces. I can't seem to find a flashlight with a flat pill top, so, for the bigger gaps, what do you do? Is the grease better than nothing? I would think so...

I'm thinking my sanding is only optimizing the pts of contact, but I still think that's an improvement. I've given up hope that the full surface of a MCPCB will contact. Soldering down the MCPCB of course eliminates this issue, if you have that option.

The coollab manual says the same things: http://www.coollaboratory.com/pdf/manual_liquid_pro_englisch.pdf. Couple of interesting things - one about fingerprints, one about solids or liquids compared to air, etc.

Boy did this Y3 thread get side tracked... Smile

Sorry, thought it went along with the build to some degree.

Thanks for the info on that specialty stuff.

I’ve used a liquid solder from Radio Shack that seems to be a form of graphite. Very thin coming out of the tube, black liquid, that sets up pretty quickly and makes an electrical connection. I wonder how it would perform as a thermal paste? It’s really thin, but acts almost like a glue when set up. This would maintain continuity to the light, if nothing else, if that’s desired.

…and now we return you to the previously scheduled program already in progress.

I dunno if I'd buy this coollab stuff - dunno the cost, but it may be perfect for some applications: knowing you have a flat surface on aluminum, or don't want to reflow the MCPCB. If there were some unevenness like this Y3, hate to waste a lot of it because it sure sounds pricey...

So, the list of poor pill tops for me lately are:

  • this Y3 (uneven)
  • XinTD X3 (serious dimple impression in the center)
  • MaxToch M12 (two levels of indent, center dimple)
  • All Warsun's (major machine 14mm center indent)
  • think there's more...

How are they getting that crucial point so wrong? I’d think the machining process itself would ensure it’s a flat surface, even if not finished with machining lines in it.

To me, that conical dimple from drilling it out first is unacceptable. It removes critical mass from directly underneath the die. Now, I’ve thought about this (is this why my head hurts?) and wonder if that is actually a Pro instead of a Con? The heat is going to travel, with the sides thicker it might actually serve to wick the heat into the pill/body quicker. Dunno, that might be reaching. I’ve seen a spectral image on heat that showed a pattern from the heat under the emitter almost identical to the light output from the front. So if you invert that cone of light, that’s the natural path the heat wants to take. It’s almost like an X with the emitter in the center, light out the top, heat out the bottom. So I prefer a large copper sink under the emitter to allow that ^ of heat to follow it’s natural course. I have no idea what is actually the most efficient there. But perhaps it’s a placebo that makes me feel good about it. Much better than thin shelves for the emitter, for sure.

Have you found the K50 satisfactory? Just got mine yesterday, debating with myself whether or not to open it up at all.

Wish that Yezl Y3 would ship!!!

It just occurred to me that the spectral picture I saw followed the available path. It would only have relevance if different options for the heat path were exploited, charted, measured and timed to show what method kept the emitter coolest longest.

This is why R&D is so expensive, right?

Some reason my ears were burning and I had to check this thread :slight_smile: Yeah lapping is good for the best contact on concave/convex surfaces (especially intel sinks). Taking the heat spreader off and mounting directly to die was pretty good too and worked well. But after coolabs came out with the stuff, I found lapping and applying it was pretty much guaranteed to last. I can’t say that about other tim I’ve tried. They all needed to be re-applied as they broke down over time :frowning:

btw you can get a almost perfectly flat surface by taping wet sandpaper down to a glass table. Lap back and forth on the wet/dry paper with a bit of soap and water. I liked to do 600 to 800 wet/dry grit and finish with 1200 if I wanted a nice surface. I didn’t see any difference in temperatures when getting a mirror finish going any higher. Even 800 was good enough.

Anyways coolabs lasts over many installs, you use so little on a lapped surface. You have to rub it into both sides to be bonded before the bead, it’s kind of weird that way or else it won’t stick and shoot out like a ping pong ball. I loved it cause it allowed me to stabilize an overclock and keep it- like for instance at one time I liked to build 2.8 intel G6950 to clocked to 4.1 (non-binned)0 on cheap gigabyte boards, no complaints down the line of any failure :slight_smile: The stuff doesn’t degrade. It’s the only tim I’ve tried that could do that. The bus and video cards it worked great on too. Water cooling or custom aluminum sinks, it works.

To answer dbc question. The moose guy answered it pretty well (sorry moose guy- I’m tired and too lazy to look past avatars). There is no such thing as a perfectly flat surface. If you look under a microscope, it’s ridges and valleys even with apparent perfectly flat mirror finishes. The tim fills these in and allows heat to transfer better than air.

Oh yeah the internet is full of fail, except forums like this and maybe Ownage pranks every sunday, the ggn

That's all good info, but we are talking about flashlights, specifically the Y3. Flat surfaces, like a glass table, doesn't pertain to a LED shelf (has an outer ridge - can't be sanded openly, you have to use something to go inside). In general, budget and maybe not so budget lights have uneven surfaces, unlike CPU heat sinks and CPU's. They don't care about machining this critical surface smoothly - they don't get it, or they don't think it's necessary under 3 amps.

You all probably know this, but when you look at the instructions and HOWTOs, videos, etc. for applying thermal grease, like Arctic Silver, to CPU heat sinks, they always emphasize using a very thin layer, almost nothing, like just a drop, of the thermal grease, and applying using something like the edge of a credit card or razor blade to get that thin layer. The grease is only suppose to fill in the microscopic grooves, etc.:

Tom, how do you think this Y3 would handle an MT-G2 application? (have I already asked that? hmmm)

I’ve ordered some of the CoolLab liquid metal to see just how that performs. Looking at your Yezl Y3 breakdown, you drilled and tapped holes to mount the 20mm star. How did you ensure the star was perfectly located first? And can you tell me where to find matching drill bits, taps and screws to be able to do this?

I think for myself and many others the thermal paste is used as a glue to sort of hold the star in place while mounting the reflector assembly, and I know I’ve used adhesive thermal compounds to do just this. I’ve put on the adhesive (much thicker than necessary I see now) and assembled the reflector, and bezel, no lens, so I could slide the star around on the adhesive before it set up and properly locate the emitter in the center of the reflector opening.

So, how do you ensure the emitter is in just the right spot before drilling and tapping those holes? I want to mount the star with screws when using the Liquid Pro from CoolLabs so I’ll need to have the right tools and understand how to apply them. I have a spring loaded center punch, not too worried about the bit walking, but locating the holes is a big step. And of course, I need the proper screws and tap.

I’ve drilled a close estimation hole and forced a screw in so it cut threads in the aluminum. Screw taken from old hard drives or whatever. But I’ve also had the heads snap off. So I guess it’s time I go about this properly…

No clue whether you are serious, but...

For the screws, some may debate any advantage or disadvantage for thermal purposes, but it's nice to know the star can't twist around when tigthening up the reflector - I've torn thru the silicone on wires from twisting effects and fried LED's that way.

For placement of the screw holes, I measured as precisely as I could to center the SinkPAD, then marked positions to screw the holes with a Sharpie. I did not want the screws too tight in close to center because I did not want the flat head screws to apply too much force inwards, bowing up the SinkPAD in the center when they are tightened down.

I used a drill press to drill the holes, and an old tap/die set that my father had in one of his toolbox's. This set has to be 40+ years old - I have to be real gentle because I broke one already. Dunno sizes - I experimented first in scrap aluminum to get the right drill bit and tap size. I do know the screw sizes however - they are:

#6809 - Metric machine screws, Phillips flat head, Stainless steel 18-8, 2mm x 0.4mm x 4mm

or

#6844 - Metric machine screws, Phillips flat head, Stainless steel 18-8, 2mm x 0.4mm x 5mm

from BoltDepot.com