I would like to drive 3 mce's at full rated output with dimming control. I have 3 MC-E's in prewired parallel configuration on stars. I would like to connect them all in series as Vf is about 10.8 ( 3.6 each) and shark set at 3A. I am using a 14.6 Volt Li-ion 10amp battery (external). How can I solder the dimmer to the Shark ? Where can I purchase a 20K alogarithmic pot? and what value ( 1amp, 5amp ?)Also will the Shark handle 3 amp continously for at least 3 hours? Sorry for many questions at once.
Driving 3-MC-E's with Shark + dimmer?
- Thread starter videoman
- Start date
Help Support Candle Power Flashlight Forum
HarryN
Flashlight Enthusiast
Hi - interesting project.
This question is a perfect one for the battery / electronics section, and it will have good appeal there. Perhaps you can ask a moderator to move it for you.
This question is a perfect one for the battery / electronics section, and it will have good appeal there. Perhaps you can ask a moderator to move it for you.
Gryloc
Enlightened
videoman,
It may be a bit tricky to solder an external potentiometer unless you have a steady hand, and maybe some good soldering equipment. I would request when ordering the Shark Buck drivers that you have the trimmer potentiometers removed (an option on the sales site).
Here is a large pic/diagram of how to wire a Shark Buck: here. Look for the white plastic trimmer potentiometer. Notice it has two pins on the left and one on the right. These pins are soldered on matching pads on the surface of the Shark. Here is a pic I combined together in PSP. Notice I oriented the little trimmer pot like the one in the above pic of the Shark:
(click to enlarge)
Notice that most pots are wired the same, so wire the pin 1 to contact 1, and the same with 2 and 3. Notice that you must use a 20K ohm (20,000 ohm) logarithmic pot. The logarithmic pot has an area of resistance that mimics a logarithmic curve (of resistance) relative to the turning of the knob. When this curve is combined with the way the Shark works, then the dimming of the LEDs will look about linear to our eyes. Here is a thread showing the conversion of a trimpot to full sized potentiometer on a Shark: here.
Any pot will work because power handling is of little concern since a tiny trimmer with a low power rating was used before. I would reccomend a larger pot like what you see in my above picture. The metal construction means that the pot will be much more durable and the sweeper life (turning the knob back and forth) will be much better compared to some tiny trimmers. In addition to naturally better power handling, they are great for attaching your own knobs to for aesthetics/ergonomics.
I would recommend checking out your nearby Radio Shack because of the wonderful convenience of being able to visualize every product in stock. The potentiometers usually provided there of the common shape and orientation you would be comfortable using, and they are of course always of high quality (name brands like Alpha on pots). In addition, you can find an assortment of knobs to fit the pot you choose.
You could find these pots on online electronics distributors like Mouser and Digikey, but I found it a bit of a pain to have to reference every datasheet PDF for each series and brand to make sure I search for and choose the exact model that has the features needed (the three metal contacts can be pins, pegs, loops, and can be straight or bent 90 degrees, the shaft can be a round shape, have a flat side (D-shape), or have splines, and so forth). If you cannot get to a Rat Shack, or find the exact potentiometer you need, I can find the exact parts you may want on Mouser and PM you with links.
When it comes to running the Shark at full power, you will need to heatsink the driver board or else the driver may fail or run in an unstable fashion. The Shark Buck has an efficiency of 78% (when running a single emitter, so I am not sure if it is better when more emitters are wired in series), so that means that the input power minus the output power is the amount of power in heat that is dissipated by the Shark Buck. So, ~3.6V * 3A = 10.8W per LED. Three LEDs: 32.4W output power. Now the power in would equal 32.4W / 0.78 = 41.5W. Now, 41.5W - 32.4W = 9.1W dissipated in the Shark. This is why the Shark must be heatsinked, as it will dissipate nearly as much power as one of the MC-E emitters.
So, what do you do? You need to thermally attach the driver board to a metal surface. If you plan on machining out your fixtures, then make sure there is a raised pedestal to mount your driver. You can also use a separate piece of metal and use thermal epoxy to fix it to a heavy part of the chassis. I believe that you can use an Alumina-based thermal epoxy (like Arctic Alumina thermal epoxy) to attach the back side of the driver (which has a smooth surface) to your metal. Just make sure that contacts on the driver does not come in contact with the metal to short them out. There should be a large metal foil area on the back of the Shark Buck meant to attach to metal, but you would have to ask around for pictures because I never handled a driver yet. Once attached thermally, the driver should operate for long periods of time.
Here are some links that I found where modders thermally attached their Shark Drivers to a metal part using AA (Arctic Alumina) thermal epoxy. Heatsinking the Shark driver: here and here. These little curved bits of metal are shaped to interface with the inside diameter of their Maglite host bodies (to transfer the heat). For you, you would be doing this with the side of the chassis, a separate chunk of metal, or maybe a part of the fixture where there are large finned heat spreaders located (like next to the LED emitters). Again, check around to see which areas can be glued using the slightly capacitive AA thermal epoxy. Also, you have to make sure the wires you insert into the little hole contact areas before soldering does not pass too deeply where it will contact the flat metal surface that you plan to heatsink the Shark to. This would result in a short, and possibly damage or overheating to the driver.
Finally, you should get around 3.5 hours of battery life when running this device at full power. The entire circuit will consume 41.5W at full power, and the battery has a capacity of about 146Wh (14.6V * 10Ah). So, 146Wh / 41.5W = 3.52 hours (theoretically). As you dim the lights, your runtime can only increase. Three MC-E emitters running at ~750mA per die will be stunningly bright. BTW, which tint of MC-E emitters did you choose at the end, cool or neutral white?
I hope this was helpful. The processes are pretty simple (but time consuming) once you understand them, but there are many simple processes in such project to get the desired results. It is like a giant puzzle consisting of big, easy to connect parts, but you have to assemble the puzzle in slow motion. Just be patient, work at your own pace, and make sure everything is connected properly, then triple check, before trying to power everything up. It will be worth it at the end!!!
-Tony
It may be a bit tricky to solder an external potentiometer unless you have a steady hand, and maybe some good soldering equipment. I would request when ordering the Shark Buck drivers that you have the trimmer potentiometers removed (an option on the sales site).
Here is a large pic/diagram of how to wire a Shark Buck: here. Look for the white plastic trimmer potentiometer. Notice it has two pins on the left and one on the right. These pins are soldered on matching pads on the surface of the Shark. Here is a pic I combined together in PSP. Notice I oriented the little trimmer pot like the one in the above pic of the Shark:
(click to enlarge)
Notice that most pots are wired the same, so wire the pin 1 to contact 1, and the same with 2 and 3. Notice that you must use a 20K ohm (20,000 ohm) logarithmic pot. The logarithmic pot has an area of resistance that mimics a logarithmic curve (of resistance) relative to the turning of the knob. When this curve is combined with the way the Shark works, then the dimming of the LEDs will look about linear to our eyes. Here is a thread showing the conversion of a trimpot to full sized potentiometer on a Shark: here.
Any pot will work because power handling is of little concern since a tiny trimmer with a low power rating was used before. I would reccomend a larger pot like what you see in my above picture. The metal construction means that the pot will be much more durable and the sweeper life (turning the knob back and forth) will be much better compared to some tiny trimmers. In addition to naturally better power handling, they are great for attaching your own knobs to for aesthetics/ergonomics.
I would recommend checking out your nearby Radio Shack because of the wonderful convenience of being able to visualize every product in stock. The potentiometers usually provided there of the common shape and orientation you would be comfortable using, and they are of course always of high quality (name brands like Alpha on pots). In addition, you can find an assortment of knobs to fit the pot you choose.
You could find these pots on online electronics distributors like Mouser and Digikey, but I found it a bit of a pain to have to reference every datasheet PDF for each series and brand to make sure I search for and choose the exact model that has the features needed (the three metal contacts can be pins, pegs, loops, and can be straight or bent 90 degrees, the shaft can be a round shape, have a flat side (D-shape), or have splines, and so forth). If you cannot get to a Rat Shack, or find the exact potentiometer you need, I can find the exact parts you may want on Mouser and PM you with links.
When it comes to running the Shark at full power, you will need to heatsink the driver board or else the driver may fail or run in an unstable fashion. The Shark Buck has an efficiency of 78% (when running a single emitter, so I am not sure if it is better when more emitters are wired in series), so that means that the input power minus the output power is the amount of power in heat that is dissipated by the Shark Buck. So, ~3.6V * 3A = 10.8W per LED. Three LEDs: 32.4W output power. Now the power in would equal 32.4W / 0.78 = 41.5W. Now, 41.5W - 32.4W = 9.1W dissipated in the Shark. This is why the Shark must be heatsinked, as it will dissipate nearly as much power as one of the MC-E emitters.
So, what do you do? You need to thermally attach the driver board to a metal surface. If you plan on machining out your fixtures, then make sure there is a raised pedestal to mount your driver. You can also use a separate piece of metal and use thermal epoxy to fix it to a heavy part of the chassis. I believe that you can use an Alumina-based thermal epoxy (like Arctic Alumina thermal epoxy) to attach the back side of the driver (which has a smooth surface) to your metal. Just make sure that contacts on the driver does not come in contact with the metal to short them out. There should be a large metal foil area on the back of the Shark Buck meant to attach to metal, but you would have to ask around for pictures because I never handled a driver yet. Once attached thermally, the driver should operate for long periods of time.
Here are some links that I found where modders thermally attached their Shark Drivers to a metal part using AA (Arctic Alumina) thermal epoxy. Heatsinking the Shark driver: here and here. These little curved bits of metal are shaped to interface with the inside diameter of their Maglite host bodies (to transfer the heat). For you, you would be doing this with the side of the chassis, a separate chunk of metal, or maybe a part of the fixture where there are large finned heat spreaders located (like next to the LED emitters). Again, check around to see which areas can be glued using the slightly capacitive AA thermal epoxy. Also, you have to make sure the wires you insert into the little hole contact areas before soldering does not pass too deeply where it will contact the flat metal surface that you plan to heatsink the Shark to. This would result in a short, and possibly damage or overheating to the driver.
Finally, you should get around 3.5 hours of battery life when running this device at full power. The entire circuit will consume 41.5W at full power, and the battery has a capacity of about 146Wh (14.6V * 10Ah). So, 146Wh / 41.5W = 3.52 hours (theoretically). As you dim the lights, your runtime can only increase. Three MC-E emitters running at ~750mA per die will be stunningly bright. BTW, which tint of MC-E emitters did you choose at the end, cool or neutral white?
I hope this was helpful. The processes are pretty simple (but time consuming) once you understand them, but there are many simple processes in such project to get the desired results. It is like a giant puzzle consisting of big, easy to connect parts, but you have to assemble the puzzle in slow motion. Just be patient, work at your own pace, and make sure everything is connected properly, then triple check, before trying to power everything up. It will be worth it at the end!!!
-Tony
Last edited:
Tony, I appreciate your interest in my project. I have decided to go with the MCE's and Shark . I also noticed interest in the upcoming Cree XP-G leds that will be hopefully available in the Fall. I hope that they can directly replace the MCE's, but that is another topic altogether.I am looking forward to your diagram as the project case is almost determined and the only hurdle is a diffusion panel to disperse the light evenly. I have ordered the shark 3A (4 of them ) from the Shoppe and am presently trying to locate the MCE's on a round or star board prewired in parallel die config. Some have stock but not neutral white, some have stock but bare emitter, some can get in a month. It is frustrating. Anyone out there has a star or round mount MCE's neutral white (highest lm) for immediate shipping that I can get within a week??
to power regulate the Shark, You "only" have to dissolder the small poti, then wire
a) a large external one to the contacts, or
b) built a few (three) "power levels" with resistors that equal three settings of the poti
imho the b) is better, because the levels are fixed. With the poti one would always run the circuit at much more power than thought. Runtimes can not be calculated good, then.
and 2nd:
STOP!!!
what do You want to do?
ordered SHARK BUCK?
four of them?
MC-E wired in parallel?
what about "wiring the (parallel wired) MC-Es in Series?
I am not sure, if the Shark Buck works similar to the Shark, and thus if it grants the current is controlled, no matter how much led are wired in series. My guess were, that the Buck can drive a single parallel wired quad led, dunno of adding 2 additional led. If that is possible, You wont need another three of the drivers.
Anyone has surefire infos?
PS: that poti can be a normal one. It is a sense resistor, there wont be a 3 A current on it.
(which is good, because the 3 or 5 W poti You asked for are expensive, difficult to get, HUGE and get very hot)
PPS: the XP-Gs can "replace" parallel wired quad-led, but the full 3A current will kill them immediately. Their max safe will be some 800-1000 mA, like the single die led they are.
a) a large external one to the contacts, or
b) built a few (three) "power levels" with resistors that equal three settings of the poti
imho the b) is better, because the levels are fixed. With the poti one would always run the circuit at much more power than thought. Runtimes can not be calculated good, then.
and 2nd:
STOP!!!
what do You want to do?
ordered SHARK BUCK?
four of them?
MC-E wired in parallel?
what about "wiring the (parallel wired) MC-Es in Series?
I am not sure, if the Shark Buck works similar to the Shark, and thus if it grants the current is controlled, no matter how much led are wired in series. My guess were, that the Buck can drive a single parallel wired quad led, dunno of adding 2 additional led. If that is possible, You wont need another three of the drivers.
Anyone has surefire infos?
PS: that poti can be a normal one. It is a sense resistor, there wont be a 3 A current on it.
(which is good, because the 3 or 5 W poti You asked for are expensive, difficult to get, HUGE and get very hot)
PPS: the XP-Gs can "replace" parallel wired quad-led, but the full 3A current will kill them immediately. Their max safe will be some 800-1000 mA, like the single die led they are.
Last edited:
Gryloc
Enlightened
I thought that the lack of info in the OP would worry some. Videoman was continuing from another thread about making a video light that uses three neutral MC-E emitters (each die wired in parallel, and all three emitters wired in series) powered from a single Shark Buck (3A model). The Shark Buck acts like any normal bucking driver where the output voltage can be any voltage below the supply voltage (minus any overhead for the driver itself), but this particular type of buck can deliver up to 3A of current with dimming capabilities.
So, each emitter has four wired dies (able to take 3A -meaning ~750mA per die), then because the input voltage is high enough, then all three separate MC-E emitters are then wired in series (meaning ~3.6V * 3 = 10.8V at 3A).
From what I have read, it appears that the Shark Buck can drive multiple LEDs in series at 3A, as long as the output voltage is lower than the input voltage. If I am mistaken, please someone chime in! I know that TaskLED's HipCC is meant to handle several quad die emitters in series, so what would make the Shark Buck any different? I do admit that if the driver is only 78% efficient, then that means a lot of heat created by the driver when the LED output power is ~32.4W. Thanks for any additional info!
-Tony
So, each emitter has four wired dies (able to take 3A -meaning ~750mA per die), then because the input voltage is high enough, then all three separate MC-E emitters are then wired in series (meaning ~3.6V * 3 = 10.8V at 3A).
From what I have read, it appears that the Shark Buck can drive multiple LEDs in series at 3A, as long as the output voltage is lower than the input voltage. If I am mistaken, please someone chime in! I know that TaskLED's HipCC is meant to handle several quad die emitters in series, so what would make the Shark Buck any different? I do admit that if the driver is only 78% efficient, then that means a lot of heat created by the driver when the LED output power is ~32.4W. Thanks for any additional info!
-Tony
Last edited:
Tony, that's exactly what I want to do. 3 MCE's configured in parallel so that I string all three in series and put a shark (one shark) to drive all three at 3 amps. Can the shark handle that or is it too much to ask a shark for? If not how can I get FULL light output from the 3 MCE's with dimming control? Reliability is paramount. We charge 2 to 5 thousand dollars to videotape a wedding and we hate to have to explain to a customer why a light went down during an important speech in a dark reception. If there is a more expensive driver that is reliable, so be it. Thanks
Why not try the Taskled Hipflex? It's a larger board and may not fit your host, but I did manage to sand it down just enough to fit into a mag D driving 3 x MC-E in series (each die wired in parallel). Works great and the UI gives many options. However, if you like external pots better, this option would not work.
Just a thought....
Hill
Just a thought....
Hill
Gryloc
Enlightened
Ok, I am not 100% sure if the Shark Buck can handle (waste heat wise) three or more MC-E and P7 emitters that are wired in series. I was starting to doubt this because I was growing worried about the huge amount of waste heat that would be produced.
When I asked a while back whether the Shark Buck could drive 2 or 3 P7 or MC-E emitters from 13.5V (alternator charging voltage) in this Shark Buck post, dat2zip said:
So, it is definitely safe to run two P7 or MC-E emitters from that input voltage range with the Shark Buck, but what about three? Theoretically, the Shark Buck should drive any amount of LEDs in series given Vout<Vin, as it acts like any other buck driver.
Changing gears here, can anyone inform us how the efficiency changes as the number of LEDs in series increases? This may help determine how much output power can be handled by the driver.
With the HipCC and HipFlex, the efficiency actually increases as the number of LEDs (P7 or MC-E with dies wired in parallel) used increases. See the efficiency charts on the hipCC and HipFlex technical data sites.
The only figure I got for the Shark Buck was ~78% for driving a single P7 (or MC-E) from nearly any input voltage. See Dat2zip's efficiency table or graph in this post in his Shark Buck thread. What about multiple emitters in series? With 78% efficiency, then the Shark Buck would have to dissipate 9.1W when the output power is 32.4W. That is rather alarming to me!
Thanks for any additional clarification ahead of time!
-Tony
When I asked a while back whether the Shark Buck could drive 2 or 3 P7 or MC-E emitters from 13.5V (alternator charging voltage) in this Shark Buck post, dat2zip said:
So, it is definitely safe to run two P7 or MC-E emitters from that input voltage range with the Shark Buck, but what about three? Theoretically, the Shark Buck should drive any amount of LEDs in series given Vout<Vin, as it acts like any other buck driver.
Changing gears here, can anyone inform us how the efficiency changes as the number of LEDs in series increases? This may help determine how much output power can be handled by the driver.
With the HipCC and HipFlex, the efficiency actually increases as the number of LEDs (P7 or MC-E with dies wired in parallel) used increases. See the efficiency charts on the hipCC and HipFlex technical data sites.
The only figure I got for the Shark Buck was ~78% for driving a single P7 (or MC-E) from nearly any input voltage. See Dat2zip's efficiency table or graph in this post in his Shark Buck thread. What about multiple emitters in series? With 78% efficiency, then the Shark Buck would have to dissipate 9.1W when the output power is 32.4W. That is rather alarming to me!
Thanks for any additional clarification ahead of time!
-Tony
Thanks Hill for a suggestion, but it's not for a flashlight but a Video Light. Size of the drivers is not really a problem as the enclosure (4X3X2.75 in.) can accomodate 3 emitters as well as 3 drivers and a pot or two as need be.Ideally I would like 3 emitters one driver and a dimmer. That 78% efficiency is sucking quite a bit of power. That's why I am interested in those upcoming XP-G's if they ever materialize, in which case that 78% will not be too bad. I've talked to my video guys and they really need that dimmer in there. So for now I'll use those 4 Sharks for 4 video units and wait till something comes up. These units are only prototypes for evaluation purposes, and if the configurations and components work as they should, all the better. Thanks again
Thanks Tony, perhaps those efficiency figures for the Shark are for 1 emitter used, as I can understand that if Vout approaches Vin (using more leds in series) that the efficiency increases as the driver will have less work to do. Perhaps I'm wrong but I am beginning to see the picture. Are there constant current drivers out there even a bit bigger that can handle say 5 or 6 A but capping them at 3A (reduced mode) for less heat and reliability? Hmm just a thought.
Must have dimming capability. I heard of putting 2 shark bucks in parallel for a 6 amp max, therefore underdriving the drivers should run them cooler and more reliable when connected to 3 MCE's. My question is can I use one pot on one of them drivers to control both drivers? Any wiring schematics ? Also would like to see an efficiency graph with the 2 driver setup as opposed to 1. Will there be any gain or loss? Goal is driving 3 MCE's at full pop with utmost reliability.
get a stereo-poti (= a poti that has 2 "levels", both activated by the same turn)
dissolder the smd poti on the drivers
connect the stereo poti to both drivers (the same way the smd was. But only the center contact is important, the other two only do the direction up/down is made)
dunno about the 2 drivers, I would use one (cost+size)
dissolder the smd poti on the drivers
connect the stereo poti to both drivers (the same way the smd was. But only the center contact is important, the other two only do the direction up/down is made)
dunno about the 2 drivers, I would use one (cost+size)
As soon as I get the 3 MCE leds I will hook them up an use one driver(shark 3A) for now with a single 20K log pot. I will drive them at full output for at least 24 hours with a regulated 14.4 DC supply. They will be heat-sinked in the die cast aluminum enclosure along with the driver and pot. This gives me a chance to verify heat, light output, beam angle, reliability. If the test passes, then one driver is enough for the task. My only concern after that point is the wide angle even spread of at least 60 degrees horiz. X 40 degrees vert. I hear Calclo has a dome like lens that disperses at 110 degrees evenly (saw that at Futureelectronics) gonna check that out.
Capo_au
Newly Enlightened
I dont have the time right this second to read though some of the quite in depth post in this thread but i have done the exact same as the opening post asked with a $4 AX2002 driver, a 10k Pot and 3 XR-E's running at 1A dimmable.
Just requires a few modifications to the driver.
Hope thats of some interest.
Sorry if its already been covered! :duh2:
Just requires a few modifications to the driver.
Hope thats of some interest.
Sorry if its already been covered! :duh2:
