DRIVER FOR MC-E"S(10)

[big beam]

I have a friend that want's to run 10 warm white MC-E's for fixed lighting.Using flashlights we came up with1.2-1.4A per emitter.I know that this is modest amp draw for an MC-E but I think heat will be EZ'er to control and the light it produces will be perfect at 1.2 amps or so.I've looked at the xitanium drivers but am a little confused when they say 12 or 24 volts.MC-E's run @ 3.7 don't they?There will be 4 MC-Es on one wall and 2 on the other 3 walls(10 all together)How will I wire this and what drivers will I need?I know I can wire 4 in series but that's 14 volts and the other set of 2 will need 7.2 volts if I wire them in series. Or 3.7 volts if a wire them parallel.Can I get a 3.7 V driver that will produce 12 amps(10 [email protected] anps a peice)
DON

 

[mudman cj]

Hi there Don! :wave: I would love to help you with this.

Check out the Lightech LED drivers as well. They also run straight off of line voltage, and will regulate to 700mA, which is perfect for getting the 1.4A you seek. The LEDs need to be wired 2s2p (hey, just like your KL4 ;)) so that each die gets 350mA (700mA / 2 = 350mA) The 1.4A case you reference is for all dies in parallel, in which case the die current is 1.4A / 4 = 350 mA as well.

You will need one of the 36W drivers to power 4 LEDs in series (7.2V per 2s2p LED) because the total voltage is over the maximum of 24V for the cheaper 18W version or the Xitanium 17W. Luckily, the other 3 walls can each use either the 18W version of the Lightech or the 17W Xitanium (even cheaper) to power 2 LEDs in series (again, 2s2p).

 

[big beam]

Hey CJ good to here from you again.That L4 is a great light.I have a one cell body for it also and with AW's IMR cells it's a pocket rocket.Now if I can find one of those 2 speed switches that some guys made for the L4 I would be :grin2:.

I didn't know that you can buy MC-E's 2S2P.I know the P7 are 3.7V only.
(But I should have known because the L4 is a 7 volt driver to run the Lux V and you said that you were using the stock driver in the L4 mod).Back OT.

For heat sinking I'm going to use soild 1" square aluminum the legnth of the wall.For the 4 led wall that will be 18'.I have some laying around so it's free.I was going to use thermal epoxy to mount the star to a peice of aluminim stock 1/8" x 3/4" x6" long and use thermal grease and 4 small screws to the 1" square stock so I could replace the the led's when better one's come out in the future.I would space the 1" stock 1/8" off the wall and the small peice of trim that would cover the bar stock and led's would also be spaced to allow for some air flow.

BTW it was your light that I used for a "light show" that persuaded them to go with warm white emitters and how bright they should be!

Thanks DON

 

[mudman cj]

You don't exactly buy MC-E's in a 2s2p configuration. They all come with 8 legs, one + and one - for each of the four dies. It's up to you how you choose to connect them. So any MC-E can be wired 4p, 4s, or 2s2p. See this post for a diagram on how to connect one 2s2p. Note that the jumper wire connects two of the - leads to two of the + leads on the two other dies. In other words, the jumper 'crosses over' from one 'corner' of the LED to the opposite 'corner'. Before powering the LED, I recommend checking to be sure you have electrical contact everywhere by using the LED checker function on a DMM. That will make the dies light up at a very low level, and it works for two dies in parallel just fine.

I am glad to hear you are liking the KL4. Good luck finding one of those two stage E series switches. They are very rarely sold. :mecry:

 

[big beam]

Well that looks EZ enough.I didn't know that the MC-E had 4 lugs on each side.The only mods I've done is Lux to SSC swaps and that was EZ.
DON

What do you think about my heat sinking? will it be enough?

Is that little metal pad on the bottom of the emitter isolated?

 

[big beam]

One more question CJ.
If the driver puts out 10-24V how do you get it down to 7V or so for the emitter?

Or does the emitter only"absorb" 7 or so volts @ 700ma.?
DON

 

[mudman cj]

Your heatsinking sounds adequate for driving them at a 5W level. Aluminum is an excellent heat sink after all, and you will have quite a bit of surface area for it to lose heat to the air as well.

The metal pad on the bottom (slug) is electrically isolated for the MC-E package, so no worries there. :twothumbs

For the three walls with 2 LEDs each, one of the 18W drivers will power them if they are wired in series. Each of the LEDs needs about 7V, but with two of them in series they need about 14V - well within the 10-24V requirement. You would not be able to run one MC-E wired 2s2p due to the 10V minimum. The LED would survive only a fraction of a second with 10V applied. :poof: But, it would be bright! :naughty:

 

[big beam]

Oh I get it now the emitter will be 2S2P and each emitter will be in series(the 2 on a wall ones)I may as well do the 4 leds on the front wall as a group of 2 and 2 so as to keep all the drivers the same.(5 sets of 2)Where did you get your 5A tints from if you don't mind me asking?

 

[mudman cj]

I got them from a group buy done by EpRoM, but he is no longer offering them. Since you need enough to be over the $100 order minimum, I recommend calling LED Lighting Supply to see what tint bins they have in stock right now. I have purchased 4B tint MC-E LEDs from them in the past, and they were quite nice. If you are lucky you might even be able to get them in K brightness bin (ask for order code MCE4WT-A2-0000-000KE4). If not, the J brightness bin should be available (order code MCE4WT-A2-0000-000JE4) They have about the best prices around.

 

[big beam]

Thanks cj I'll call Monday
DON

 

[mudman cj]

You're welcome Don. Please let me know what they have in stock in the neutral whites.

 

[mgc8]

Hi, I'm interested in doing a similar overhaul of my fixed lighting, and I would like to use 16xMC-E's in one room, same as above -- warm tint, at 1.4A, probably wired 2S2P each. Could I use 4x 36W Lightech drivers for this? I browsed the LedSupply page and figured out something among the lines of:
- 4x Driver (36W)
- 16x MC-E Warm-White
- 16x Heatsink ("half-brick" would be ok?)
- 16x holders + lens ("ripple wide"? what would be the best lens for maximum "flood" -- I am interesed in overall lighting, not spots).

Am I missing something? The only downside I guess is the price, at around 800$ it does get pretty expensive, but I should have the money available sometime this summer so I could do it all then...

Please advise if anything is wrong/out of place!

Regards,
Mihnea

 

[mudman cj]

Do you intend to group them with only 4 on each wall? Could you run wire from one group to the next? You could then possibly power as many as 8 LEDs in series wired 2s2p since 8*7.2=57.6V is just under the limit. And 7.2V per LED is conservatively high. Most MC-E dies have a Vf lower than 3.6V - more like 3.2V. This way you could get by with only 2 drivers. :huh:

The half brick heat sinks could work, but steer towards the higher power ratings. A 5W rated heat sink will run 60C over ambient, so you could easily be at 30C (inside the wall) + 60C (air to heat sink)+ 40C (heat sink to die) = 130C. At that point you are down to 70% of the rated output and only 20C away from damaging the LED. I would get the 11W version at least and orient the fins vertically to encourage natural convection.

I am not familiar with the optics, but it sounds like you are on the right track. You may find something in this thread helpful in choosing an optic.

 

[mgc8]

Do you intend to group them with only 4 on each wall? Could you run wire from one group to the next?

The led's would be 5,3,5,3 around the walls, since the room is rectangular, and fixed in a false ceiling with ample space to route cabling, so any series combination is possible.

You could then possibly power as many as 8 LEDs in series wired 2s2p since 8*7.2=57.6V is just under the limit.

Thanks for the tip, that is a great option! However wouldn't there be issues with the length of the wires from driver to leds? With 8 in series I think the furthest led would be about 4m (13ft) away, compared to half that for 4x.

The half brick heat sinks could work, but steer towards the higher power ratings.

Are you refering to this one? I am concerned about heat as well, since the lights will be recessed in the false ceiling, but I wasn't able to find a lot of affordable options -- and I don't have that many processor coolers to tare appart :) Is there a known good source for those? Otherwise I may just have to go loooking for Al scraps locally...

I am not familiar with the optics, but it sounds like you are on the right track. You may find something in this thread helpful in choosing an optic.

I've browsed that thread and a few over here, it seems it should work with the wide Carclo lens. Optics are cheap so can be easily replaced, the only problem is finding the right ones :)

Thank you for all the great information, it's much appreciated!

 

[mudman cj]

Thanks for the tip, that is a great option! However wouldn't there be issues with the length of the wires from driver to leds? With 8 in series I think the furthest led would be about 4m (13ft) away, compared to half that for 4x.

Extra wire will not pose a problem in this case. The circuit may need to provide a slightly higher voltage to overcome the extra resistance, but it will still regulate to the design current as long as the output voltage is within spec (and it will be). Since all of the LEDs are in series, they must all be driven at the same current. If 4 of them were wired in series with short lengths of wire, and then these 4 were in parallel with another 4 of them wired in series with long lengths of wire, then there could be a problem.

Are you refering to this one? I am concerned about heat as well, since the lights will be recessed in the false ceiling, but I wasn't able to find a lot of affordable options -- and I don't have that many processor coolers to tare appart :) Is there a known good source for those? Otherwise I may just have to go loooking for Al scraps locally...

Yes, that heatsink is exactly what I was referring to. I recommend using at least that much of a heatsink.

 

[blasterman]

Wire gauge resistance chart if anybody cares:

http://www.stealth316.com/2-wire-resistance.htm

I have a couple of runs about 100' with regular zip cord, and it doesn't amount to much, although those are series/series. Interesting the number of old timer electricians that bring up the same issue though; 'yuh can't run DC that far without thick wire'

 

[mudman cj]

The LEDs should amount to about 3.2*2*8=51.2V. That leaves 58-51.2=6.8V of overhead, but the NEC recommends staying under a 2% voltage drop, so let's use 0.02*58=1.16V to be safe. According to the 2% guide, with 13 feet of wire * 2 to get there and back, you could even get by with 28 gauge wire. But I suggest backing off a bit and going with 24 gauge instead based upon the results of this more conservative calculator.

Thanks to blasterman for pointing out the need to consider wire size - I was too hasty in only considering the effect on the LEDs and not considering the boundaries of NEC guidelines.

 

[mgc8]

Extra wire will not pose a problem in this case.

Got it :) and thanks for all the info regarding wire length/size and distance, it sure comes in handy! I guess I'll stick to 24awg since it's common and easy to work with...

Yes, that heatsink is exactly what I was referring to. I recommend using at least that much of a heatsink.

The reason I asked is that the specs from that page puzzle me some, I mean compared to the higher-price "5C/W" heatsink, the "10C/W" appears to be half the height but with more fins; would the added fins compensate and even double it's effectiveness? I'm not sure (tm) but it seems a little awkward. Then there's a third one that's even smaller, but with no "C/W" rating and even pricier, it has a lot of fins but would that make it the best? I know that lots of fins increase surface and dissipation, I just find it strange that there's so much difference in this case.

 

[purduephotog]

Got it :) and thanks for all the info regarding wire length/size and distance, it sure comes in handy! I guess I'll stick to 24awg since it's common and easy to work with...


The reason I asked is that the specs from that page puzzle me some, I mean compared to the higher-price "5C/W" heatsink, the "10C/W" appears to be half the height but with more fins; would the added fins compensate and even double it's effectiveness? I'm not sure (tm) but it seems a little awkward. Then there's a third one that's even smaller, but with no "C/W" rating and even pricier, it has a lot of fins but would that make it the best? I know that lots of fins increase surface and dissipation, I just find it strange that there's so much difference in this case.

Check out this page:
http://www.mmmetals.com/pages/extru...t_back_heatsink_aluminum_extrusion_1_to_2.htm

You can see 1-2" heat sinks with fins and different profiles. The measured C/W values are (more likely) for vertical orientation only- horizontal will be at least 40% less (higher number).

 

[mudman cj]

The reason I asked is that the specs from that page puzzle me some, I mean compared to the higher-price "5C/W" heatsink, the "10C/W" appears to be half the height but with more fins; would the added fins compensate and even double it's effectiveness?

You may be confused about this rating method. More effective heat sinks result in a lower temperature rise over ambient for a given power dissipation. So, the 5C/W heat sink is more effective than the 10C/W heat sink.

Earlier in this thread I was quoting heat sink ratings based upon the amount of power they can dissipate before reaching the same temperature (60C), and in that case a higher value was better.