AAAAA!!!! I hate driver boards

[monkeyboy]

Look what I just built:

It's a fivemega M*g 2D body with finned head and machined body. It runs quad Cree XR-E Q4 WH bins with Mc-R 19XR reflectors. The heatsink is a copper modamag PXR and the LEDs are driven at ~830mA with the Shark driver.

Now here's the flipside; when I run it on 2 Li-ion C cells (7.4V), the shark driver overheats and cuts out after about 15 mins continuous use. I tried using 8AA (9.6V) so that the shark driver runs cooler with less boost. It runs for about 45 mins before cutting out. Better, but no cigar. And in case you were wondering- yes, I heatsinked the crap out of the shark board not to mention the copper PXR and the finned head and the conservative drive current of 830mA. Not a lot I can do now seeing as I AA'd the shark into the PXR.

Man, this really grinds my beans. I spent so long waiting for parts to come, so much effort assembling the thing and so much money . All I have to show for it is a light that is only good for intermittent use.

I wonder if I would have been more sucessful with a Luxdrive driver? These claim not to need additional heatsinking.

 

[wintermute]

Voltage of Cree at 830mA should be around 3.6V.

4 x 3.6V = 14.4V in series

P (in watts) = I (current) V (voltage)
P = .830A x 14.4V = ~12 Watts

Assume 90% efficiency for driver. ~12 W / 0.9 = 13.3 W from the batteries.

2 x LiIon C cells (7.4V)
13.3 W = 7.4V x I ===> I = approx 1.8 amps.

8 x AA (9.6V)
13.3 W = 9.6V x I ===> I = approx 1.4 amps.

1.4-1.8A should not make the shark overheat. (Shark = Maximum input current: 4A) The problem might be you AA'd the shark to the PXR...those 4xCrees are getting 12 watts of power, for perspective 12W is more then some small soldering irons. Copper will obviously pull heat from the emitters very effectively - so you are taking heat from the emitters and feeding it to the shark, causing it to overheat.

In the future - when you are using AA adhesive, mix in 25% AA compound...then you can remove without killing whatever you have attached - but it still won't just fall off.

It's not a problem of the driver, it is a problem of the driver placement. Try to get that driver board off the PXR. Once you do - you'll be at less then half the max current of the Shark and you can find some place else to mount it (like the body of the flashlight itself).

 

[monkeyboy]

Thanks wintermute, I'll try that. First time I've heard about heatsinking being too good.

 

[JamisonM]

Thanks wintermute, I'll try that. First time I've heard about heatsinking being too good.

I have thought about doing the same thing you have done here. My idea for heatsinking the driver was to use the C version of the hotlips heasink. Firstly you would sand it smooth. It would mount on the solid plastic bulb holder with AA epoxy with both wire coming through it. If it was too tall and the PXR heatsink would hit it, just cut the plastic bulb holder back a little until you get the fit you want. I haven't done anything or tried it out yet, but would love to see the idea work. Beautiful light BTW.

 

[monkeyboy]

I have thought about doing the same thing you have done here. My idea for heatsinking the driver was to use the C version of the hotlips heasink. Firstly you would sand it smooth. It would mount on the solid plastic bulb holder with AA epoxy with both wire coming through it. If it was too tall and the PXR heatsink would hit it, just cut the plastic bulb holder back a little until you get the fit you want. I haven't done anything or tried it out yet, but would love to see the idea work. Beautiful light BTW.

That might just work. Only problem I can see is that the PXR heatsink extends down a bit so you wouldn't be able to screw the head down very far resulting in a longer length. You could probably shorten things by cutting down the PXR slightly and cutting off some of the threads on the M*g body. (I've actually cut down the threads on this light to make it a bit shorter).

UPDATE: I've repostioned the shark onto the M*g body and I'm charging batteries for another continuous run test. In the process, I broke off some of the ceramic material on the inductor but it still works.

UPDATE 2: AAA NUTS! it still cuts out. Managed 11 mins on 7.4V this time. I'm going to try 2 x Buckpuck 700s. If they'll fit.

 

[RCatR]

To remove epoxy:
Put on a gas mask(seriously)
Go outside(with a breeze)
Put a little rubbing alcohol onto the epoxy in question
Light

Once the alcohol has burned off the top layer of epoxy should be gooey, scape and repeat.

 

[zelda]

Some inspirations:

http://www.candlepowerforums.com/vb/showpost.php?p=2063063&postcount=29

Zelda

 

[monkeyboy]

Some inspirations:

http://www.candlepowerforums.com/vb/showpost.php?p=2063063&postcount=29

Zelda

Nice build. So you've had no issues with overheating even at 980mA? Have you tried running it continuously at that level? I'm going with buckpucks since it claims to run fine without additional heatsinking and the power will be split between 2.

 

[zelda]

I run it for max. 15 min on high, but I will check it now..
I hope that I have not the same problem

 

[monkeyboy]

I haven't had much luck with shark boards. I've bought 3 shark boards in total. 2 of them have died while adjusting the pot and now this one overheats. I've heard lots of good things about the shark so maybe it's just me?

 

[zelda]

Mine shut off after 15 min on high

 

[LukeA]

I haven't had much luck with shark boards. I've bought 3 shark boards in total. 2 of them have died while adjusting the pot and now this one overheats. I've heard lots of good things about the shark so maybe it's just me?

I've never gotten a Sandwich Shoppe board to work for more than maybe an hour total before it quit. One board didn't work from the start, another worked for a while, then refused to work when resoldered (correctly). The other one I have I haven't tried yet, but I'll probably be dissappointed. Again.

 

[monkeyboy]

Mine shut off after 15 min on high

Oh no, sorry to hear that but at least you have the option of turning it down. I got 15min with 7.4V and 45min with 9.6V. I'm guessing if you run it near the limit ( 3 x Li-ion C, 11.1V (12.6V max)), it might just be OK unless Vin exceeds Vout and everything goes up in smoke!

 

[monkeyboy]

I've never gotten a Sandwich Shoppe board to work for more than maybe an hour total before it quit. One board didn't work from the start, another worked for a while, then refused to work when resoldered (correctly). The other one I have I haven't tried yet, but I'll probably be dissappointed. Again.

This is what I've found too but I've never had any problem with GD boards.

 

[zelda]

15 min for shocking peoples, its enough :devil:

zelda

 

[wintermute]

To remove epoxy:
Put on a gas mask(seriously)
Go outside(with a breeze)
Put a little rubbing alcohol onto the epoxy in question
Light

Once the alcohol has burned off the top layer of epoxy should be gooey, scape and repeat.

This is seriously why you should mix in 20-25% AA compound into the AA adhesive. It will still hold, right up to the point where you remove the part. But you can remove the part without breaking anything.

I haven't had much luck with shark boards. I've bought 3 shark boards in total. 2 of them have died while adjusting the pot and now this one overheats. I've heard lots of good things about the shark so maybe it's just me?

Why not try the Fatman from TaskLED?? Up to 2A is fine...technical info on the board is here.

Oh no, sorry to hear that but at least you have the option of turning it down. I got 15min with 7.4V and 45min with 9.6V. I'm guessing if you run it near the limit ( 3 x Li-ion C, 11.1V (12.6V max)), it might just be OK unless Vin exceeds Vout and everything goes up in smoke!

Yeah - 3 x fully charged LiIon batteries might release the magic smoke.

 

[wintermute]

dup

 

[wintermute]

dup x 2

 

[AvroArrow]

You could also try George80's Fatman driver. It should fit your voltage setup and does pretty much the same thing as the Shark. Plus it's got connections for an external potentiometer for manual dimming capability. I must confess though, I don't have personal experience with either board, so YMMV. You could just e-mail George and ask him how well the Fatman can handle your setup and if the driver will need additional heatsinking or not.

EDIT: Doh! Looks like wintermute beat me to it regarding the fatman. He managed 3 posts in the time it took me to write 1. heh.

 

[wintermute]

You could just e-mail George and ask him how well the Fatman can handle your setup and if the driver will need additional heatsinking or not.

Now, Input_current should be around 2A or less for optimal performance of Fatman. Fatman can run up to around 2.2A max. When running at high input power it is recommended to thermally epoxy the back of the Fatman PCB to a heatsink or the body of the flashlight.

*4 x Crees @ 1000mA and 14.8v = 14.8 W / 0.9 (approx driver efficiency) = ~16.5 W from batteries
*4 x Crees @ 875mA and 14.5V = 12.7 W / 0.9 (approx driver efficiency) = ~ 14.1 W from the batteries
*4 x Crees @ 750mA and 14.2v = 10.7 W / 0.9 (approx driver efficiency) = ~11.8 W from batteries

Using 2 x LiIon cells would be 8.4V fully charged, 7.4V under a load, but only 5.6V near dead.
Using 6 x AA NiMH (Eneloop) cells would be ~8V fully charged, 7.2V under a load, but only 6.0V near dead.

2 x LiIon near dead
16.5 W / 5.6V = 2.95A <==too high
14.1 W / 5.6V = 2.52A <==too high
11.8 W / 5.6V = 2.10A <==max of the driver board, must heatsink

6 x NiMH AA near dead
16.5 W / 6.0V = 2.75A <==too high
14.1 W / 6.0V = 2.35A <==too high
11.8 W/ 6.0V = 1.96A <==OK, but top of operating specs. should heatsink

***Now, if you went to a 3D mag body and used 9xAA NiMH Eneloop batteries:
Fully charged = 12.15V --- Under load = 10.8V --- Near dead = 9.0V

9 x NiMH AA near dead
16.5 W / 9.0V = 1.83A <== ok
14.1 W / 9.0V = 1.56A <== ok
11.8 W / 9.0V = 1.31A <== ok

So, with a Quad Cree, the most you can do with 6xAA is 750mA to each LED --- but with 9xAA you can do all the way up to 1A.

This is why I have planned on a 3xCree build - works better with 2xLiIon rechargeables or 6xAA battery setup.

*3 x Crees @ 1000mA and 11.1V = 11.1 W / 0.9 (approx driver efficiency) = ~12.5 W from the batteries
*3 x Crees @ 925mA and 10.95V = 10.1 W / 0.9 (approx driver efficiency) = ~11.25 W from the batteries
*3 x Crees @ 875mA and 10.8V = 9.45W / 0.9 (approx driver efficiency) = ~10.5 W from the batteries
*3 x Crees @ 750mA and 10.5V= 7.88 W / 0.9 (approx driver efficiency) = ~8.75 W from the batteries

2 x LiIon near dead
12.5 W / 5.6V = 2.23A <==too high
11.25 W / 5.6V = 2.00A <==ok, but top of operating specs. should heatsink
10.5 W / 5.6V = 1.88A <==ok
8.75 W / 5.6V = 1.56A <==ok

6 x NiMH AA near dead
12.5 W / 6.0V = 2.08A <==ok, but top of operating specs. should heatsink
11.25 W / 6.0V = 1.88A <==ok
10.5 W / 6.0V = 1.75A <==ok
8.75 W/ 6.0V = 1.45A <==ok