I didn’t understand it, and forgot to give it a look.
I asked about the cavity style of thermal pad and got a misleading response. Something about a 1:1 ratio of thermal paste?
I still don’t see how my choice of solder will perform better than copper. I know there would be solder between the emitter and thermal pad anyway, but this is typically kept to a minimum, not utilizing a cavity of solder.
I didn’t realize they were ver II until I’d placed the order.
Their charts and graphs show it working as good as or better than the original version. Reckon we’ll see.
SinkPAD has been using cavities for a long time now, I didn’t realize anything was still using the punched center pad? The only numbers I’ve seen indicate that SinkPAD-II is almost as good as SinkPAD-I… but it’s not better. are you sure you’ve seen something indicating that it’s as good or better than the original version?
They mention 1:1 because Cree’s normal guidelines are for around a 50% coverage on the XHP70 and even less on the XHP50 (I think). Cree’s recommendations do not apply to MCPCBs with a cavity like SinkPAD-II and Noctigon V2.
On the stencil, what Deannae is telling you is the same thing I’m telling you. Cree’s recommended stencil is for flat MCPCBs, no cavity. Approximately 2:1 on footprint to paste. SinkPAD’s recommended stencil is for their cavity-having MCPCB, so it requires more, approximately 1:1 footprint to paste. Simple.
As far as that graph… I’ve got to tell you that you are misreading the graph. The GREEN line is a SinkPAD-II, note that it is always above the BLUE line which is a SinkPAD-I. Lower temps are lower on the graph. The RED line way above the other two is a conventional non-DTP MCPCB.
The text written in purple is misleading at best. I would call it untrue, or a fib.
And I can read far better than most. I’m not color blind either. Not being facetious but wondering why SinkPAD would put such obvious discrepancies in their PDF brochures.
They show me a spectral photo of the SinkPAD II running cooler. They show me a graph with it just holding it’s own, albeit slightly warmer. Then they SAY it’s better or as good as. Where’s the middle ground?
This is why I am suspect on the II. But, all that being said, it’s still far better than an aluminum board with masking between the pad and metal. Or strips of copper laying on top of an XM class board.
I even have an XHP70 mounted on an XP-G footprint…
Something tells me that the majority of purchases for these mcpcb’s are for flashlight mfg’s who won’t push more than 3 amps through them, I suspect that the traces are just fine for that current. Its wackos like us who are going to push it hard. Lets not forget that its a packaging issue and these LED’s are fundamentally being misused.
Heh, my mistake! I misread your post until you followed up on my comment. Then it finally sank in that you were referring to the labels on that image. I’ve always been drawn to the graph in that PDF, I may never have even read those labels in the past! Maybe the graph is an error. Or maybe the rest is an error, or a fib as I suggested. Who knows. Sorry for misunderstanding you.
I figure it’s reasonable to think that the II should be ‘close enough’ to the I, but like you I can’t imagine why it would actually be better… a thicker layer of solder certainly shouldn’t be helping anything.
I wonder if the thin traces can carry serious amperage? 12A or more? I was thinking of using 18ga wire, not a lot of point in that if it’s stepped down to threads going in on the board…
I agree, more or less, but this is a strange product. In a production environment you don’t want or need the extra steps required to handle solder jumpers or the extra pick-n-place step & cost of the 1206 jumper. AFAIK for 6v operation the MK-R PCB and XM-L PCB are perfect fits to the XHP70 and XHP50 respectively, so making the PCB configurable isn’t really necessary. Putting 12v on a DTP PCB does require an airwire though because you’ll be unable to implement Cree’s recommended 12v footprint. That’s why they added the 1206, that’s their ‘airwire’ (a trace runs in between the pads!). I assume that they added the solder jumpers as an afterthought aimed at enthusiasts and engineers building prototypes. Cree is targeting higher operating temperatures with this line of LEDs… industrial designers will have to do a cost/benefit analysis to see it it’s worthwhile to use SinkPADs at all rather than conventional MCPCBs.
I’m curious about what the Noctigons are going to look like.
If the traces in the MCPCB are too thin to carry the desired amps in the 6V config, wouldn’t there be huge advantages to running in the 12V config for the same desired total watts? ie: assuming the same power, doesnt higher voltage/lower amps travel much better through the same copper than lower voltage/higher amps? Its just for the sake of discussion, and I understand the sinkpad might have additional choke points.
Also, could a standard MCPCB (MK-R or XM-L) be modified by properly cutting the traces to put the emitter into serial config for 12V? Im not sure if the thermal pad not being electrically isolated while in the 12V config would prevent that.
That jumper looks silly on the sinkpad. I think we need a dedicated 12V MCPCB to prevent reflector clearance and other shorting hazards from botching up our builds.
It’s not silly, but looking at the datasheet would answer your question. You can modify a standard MCPCB but not to make it look like Cree’s recommended 12v footprint. Instead you’d have to use an airwire from one corner of the emitter to the other corner of the emitter. It may not be obvious how big this hurdle is until you draw out the 4 LED chips on a piece of paper, complete with the two sequential connections, and then attempt to draw the wiring in for 12v. Isolating the thermal pad requires either an airwire or a big jumper (like 1206 size).
Yes, 12v would result in less resistive losses.
EDIT: to clarify, the thermal pad is always electrically isolated unless you connect it to something. Cree recommends that you connect the thermal pad in between the two strings of 2 chips for the 12v config. It’s a convenient way to route electricity from one corner of the LED to the other. Please take a look at Page 26 of the datasheet to see what I mean.
Thanks Wight. Ive actually been staring at that darn schematic for over a week now and wondered if the std mount could be modified. Gah! Lets hope Noctigon comes up with separate dedicated versions for 6V and 12V.
You’re welcome. The most practical way to use the 12v configuration commercially is to use a non-DTP PCB like what Cree shows. I just can’t see anyone pulling a rabbit out of the hat to produce a great 12v DTP board for enthusiasts. Modifying an XM-L Noctigon with two lengthwise cuts and an airwire is the most practical thing for enthusiasts who want DTP 12v XHP50’s I think. For XHP70 there may not be an ideal DTP 12v high current solution, SinkPAD’s product will be as good as it gets.
It’s not that bad for 12v! Enthusiasts have been using those thin traces to run XM-L and XM-L2 emitters for years, this is only a bit worse due to the trace length. For 6v I think sticking with the MK-R boards is probably in order though.
@RMM - sorry to fill your post with nonsense. PM me and I can delete with no hard feelings…
@wight - You’re right, those thin traces have held up well for the short lengths they must travel while surviving many 4 volt high amp ranges.
Alright, so here’s some FP math for you:
My thoughts about the XHP70 were in regards to the efficiencies realized at the higher voltages. I had conjured an extension tube for a BTU Shocker without a cell carrier. Basically cutting the existing BTU battery tube in half and inserting another longer tube inside the ID of the BTU battery tube to lengthen it. After head and tail mods, you could stack as many 18650’s as you wished (only limited to the length of the tube mod and a suitable buck driver). I was considering a 4S3P tube config and a 12A 7135 based driver. So, 3 x 12V XHP70 in parallel driven at 4A each. Extrapolating from Crees data sheet (amps for 6V/2 = 12V version) djozz graphs net 4750 lumens @8A, 7.5ish volts (that should net 14,250 lumens total for 3 emitters. Pushing through the CN11770_IRIS with 90% optical efficiency nets 12,825 lumens OTF.
A 50 cal ammo can would actually make a far more suitable host since the shape would more easily accommodate cell fitment and active cooling. Oh hell, now we might as well just consider a 400W HID and go for some serious lumens and throw and not have to worry about heat sinking and complex circuitry. I think I’ll just go boating , drink beer and steer… before insanity sets in.
If you glue the mcpcb with Arctic Alumina Thermal Adhesive, wouldn’t that break electrical continuity between the ground and positive, allowing the 12V circuit to be used mounted like we normally mount and given that the pad is now “hot”?
I have an X6 with the XHP70 under a Minnie-M and it pulls 7.45A for 4426 lumens out the front, so that correlates pretty well with djozz’s figures if that helps any in figuring the output. 6V, of course.
But you also have to consider that it’s not a throwy beam profile by any stretch of the imagination. What it DOES do is spread light evenly across a wide area with no rings or artifacts. Having used 3 of em now, I’ve seen an odd hue to each one. Not really greenish, but headed that way. That shows up in the beamshots pretty well but you really don’t notice it by eye until you turn on another light.
Cree has 6V variants that are designed for the MK-R. No fancy business with airwires or resistors. Then they have the dedicated XHP boards. How do I know this? I’m looking at em! :bigsmile:
Can I preorder this beast of a light!?
Damn, and I just received my 8500 lumen tn36 vn last night, to get dethroned by another coke can light in less than a week!
