Shiningbeam 2.8A driver gets super hot???

[The_bad_Frag]

Hi there! :wave:

My shiningbeam 2.8A driver gets hot. REALLY fricking hot!

I soldered a setup for testing together with a cree XM-L on a cpu cooler, a shiningbeam 2.8A driver hanging in the air connected to 8xAA Ni-MH(4x2). The current from the batteries is exactly 2.80A and the voltage at the driver-in is 3.78V under 2.8A load.

The datasheet of the XM-L says 3.3V at 2.8A. This would mean 0.48V need to be droped at the driver which makes 1.34 watts of heat.

After 1 minute runtime I can easily burn my finger on it. I mean its a linear driver but is this normal? :duh2: Or do I make anything wrong? I never had any driver getting that hot.

 

[Mattaus]

Really interested to see what people say about this as I have a few on the way as well.

Just a random thought, and my understanding of your post could be waaaaay off...but those drivers are rated to 6V max. How are your batteries connected? By '4x2' do you mean 2 in series (3V) then 4 of these in parallel?

Could be too many packs maybe? I intend on running mine off 4 D-Cells (6V) so I hope this is not an issue.

I'm probably wrong. Could always just heatsink the driver!

 

[yazovyet]

can you link the driver?
4 AAs at 1.2 volts each would give around 4.8 volts, more on the initial high right after a charge. pulling 2.8amps at 4.8 volts would be 13.44 watts with the LED useing 10.2 watts (LED at 3.8 volts 2.8 amps). so that's about 3 watts of heat. that's around 76% efficency. if the driver you're using is one of those 8x7135s that sounds about right. and if your batteries at putting out more than 1.2 volts each then you would experience a lower efficency (and thus more heat for the same light)

so i guess it gets fairly hot off 3 or 4 watts of heat without anything to transfer it to. (i cant really tell you how much 3 or 4 watts is and if it should get something burning in a short time or not but i would think 3 watts will toast a small PCB if given time and not heatsinking)

this is a post about a driver that uses 4x7135s, i link it just as a refference as to their efficency at various voltages.
http://www.candlepowerforums.com/vb/showthread.php?192925-AMC7135-Specs-Inside-**UPDATE**

edit:
could you measure the voltage over ALL the batteries (what the driver would see) and describe or even better take pictures of your battery set up?

 

[hellokitty[hk]]

Hot off the charger NiMH's are 1.4v+.

 

[Justin Case]

I soldered a setup for testing together with a cree XM-L on a cpu cooler, a shiningbeam 2.8A driver hanging in the air connected to 8xAA Ni-MH(4x2). The current from the batteries is exactly 2.80A and the voltage at the driver-in is 3.78V under 2.8A load.

How did you measure the 3.78V under load for your NiMH pack? When you say "4x2", do you mean two sets of four series AA NiMH cells, with the two sets in parallel? If so, something is wrong since your cells are averaging about 0.95V each, which is basically the cutoff voltage.

 

[archer6817j]

I use these drivers in my custom lights. Mac uses the same ones. I just fired one up for 2 mins (1x18650) and I can barely say that it's warm. I've never run one in air however. The board is pressed into a hollow copper tube but there is no significant thermal path other than where the (-) rim on the board makes contact with the tube. Can you try it with different batteries or maybe just three in series...for troubleshooting purposes?

 

[Aepoc]

Just got one in the mail today. I wasn't planning on having to provide heat sinking... Keep us posted.

 

[The_bad_Frag]

How did you measure the 3.78V under load for your NiMH pack? When you say "4x2", do you mean two sets of four series AA NiMH cells, with the two sets in parallel? If so, something is wrong since your cells are averaging about 0.95V each, which is basically the cutoff voltage.

Yes I have 4 cells in series and 2 sets in parallel. I hold my multimeter on the + and - on the driver while it was under load. This is a normal voltage drop for these cells when having 1.4A load. They are a bit older also...

When I continue to discharge them at this rate (I tested it in other flashlights before)the voltage does not drop any further. When they are complety discharged the voltage drops with within 1 minute really hard. Then I stop discharging them. Now they have a idle voltage of about 0.95-0.93 which is good. If I run them on less load the voltage stays up at 1.0-1.2 depending on the discharge rate.

I will try a few things in 8 houres when Im back from work.

 

[The_bad_Frag]

Ok I did some research im the name of flashlight science. There were big explosions, stone melting leds, XM-Ls driven up to 200A and more! ...ok this was maybe a exaggerated description but I found out why my driver was getting so hot and some other things.

First to the batteries: My AAs are about 4-5 years old. They had a lot of charge cycles so they are not the freshest anymore. Those are 2900mAh Accu Power AP2900-1. Also they didnt get up to 2900mAh anymore. When I take them out of the charger they have 2000-2400mAh.
OK.
THIS RESULTS MAY BE DIFFERENT WITH OTHER, NEWER CELLS!
3 AA NiMH in series does not work. XM-L is only driven at 1.5A or less.
3 C NiMH in series does not work. XM-L is only driven at 1.9-2.2A.
4 AA NiMH in series does not work. XM-L is only driven at 2.2-2.3A.
4 C NiMH in series works fine with this driver. Also works fine in my Philips XM-L mod with a DX driver.


Now back to the super hot driver and my 4 AA in series 2 sets in parallel setup.
Reason for the super hot driver: The combination of this hot headed driver plus the light comming from the XM-L. :devil:

As you can see on the picture there is around 25mm space between the XM-L and the driver. So I fired up the XM-L at 2.8A and held my finger at the same distance. I noticed a lot of heat.


Then I took my soldering holder and told him to hold that copper sheet right there.:nana: Now the driver is still warm but not burning hot. It heats up to its max temperature within 2-3 minutes but then stays there. My "finger thermometer" says 50-55°C. (my finger thermometer is pretty accurate...I trained it at work measureing the temperature of car engines just by touching it with my fingers.)

The copper sheet refelcts a lot of light but still getting super hot at this distance! Watch out for the XM-L-Death-Ray! :devil:


But now the next problem. The driver runs good in 20°C air but I want to use it in a special mod where its not 20°C fresh air. So I need to heatsink it somehow.

How do I heatsink a driver without short circuting the complete driver?
My first thought was I make it the same way like I did in my test mod. I take a thermal glue pad put 1 side on the driver and the other on the heatsink. But the driver in my test mod had 1 completly flat side. This driver would only have contact with those 5 plastic chips on the upper side.
Is this enough heat conduction?

 

[archer6817j]

This is how I do it. I don't think you really need to heat sink it that much. I've got it to where I can just press the driver into the tube. No need to solder or peen the edges over. Just watch out for shorting on the back side if the driver.

 

[The_bad_Frag]

Hmmmm your driver looks pretty good heatsinked. The outside ring is the hottest part of the driver.

My driver without heatsinking would be in a plastic head hanging around with 50°C air temperature. I dont think it will survive that.

 

[Justin Case]

The large mounting tab on the 7135 chip is connected to the junction, so that's what you want to connect to a heat sink.

An XM-L driven at 2.8A is pulling close to 10W of power, most of which goes out as waste heat. It's no surprise that you feel a lot of heat coming from the LED. You have roughly a 7W-8W heat source.

Your battery configuration tests highlight the trickiness in working with 7135 drivers. In theory, the voltage overhead is only 0.12V. But in practice, it seems to be more than that. The reason is parasitic resistance, such that your cells have to deliver additional voltage to stay in full regulation.

I've been testing an XM-L driven at about 1.6A-1.7A. Vf is about 3V. Vin to reach regulation when using a 7135 driver is about 3.3V. Voltage overhead thus looks to be about 0.3V in this case. When you run at 2.8A drive, the XM-L Vf is that much more -- probably about 3.3V. Thus, even if the voltage overhead stays at 0.3V (which I doubt is actually the case), then you need to deliver at least 3.6V to stay in full regulation. My seat of the pants guess is that you will need at least 3.8V, which kind of defeats the purpose of using a 7135 driver if you are planning on using 1xLi-ion. With NiMH, it also suggests that you need to run with 4 cells (assuming those cells are healthy enough to hold 1.2V at 3A draw) to reach full regulation.

My takeaway is that if you want to power an XM-L with 1xLi-ion or 3xNiMH, you have to cut back on the drive current from your 7135 driver. 2.8A is too much.

 

[The_bad_Frag]

I will definitely stay with my 4s2p setup. ...or have no other choice.:thinking: Yesterday I noticed that those old cells get warm after around 20 minutes. I decided to use the old cells in my airsoft which doesnt need highpower super cells.

Now I ordered 8 black eneloops. Anyone tested them already? But I think they can easily hold the voltage up to 1.2V when discharged at 1.4A.
They cost me 29,44€ at amazon with 0€ shipping. On a local store I will get around 20 alkalines for the same money. I mean how stupid are the most people who still buy alkalines? :fail: I need to charge them 3 times and already got nearly free cells which are a lot better than stupid alkalines. Over the last 15 years I maybe saved about 10000€ worth of batteries.

Now I need to fire up my mini lathe to make a decent heatsink for the driver.

 

[Changchung]

Hi, nice thread, I am not sure if how I sink my mod can help you...

http://www.candlepowerforums.com/vb/showthread.php?314351-A-small-CR123-XP-G-mod

 

[vestureofblood]

Hi Frag,

I know the driver spec on these says up to 6V but that is very hard on the driver. If you cut the Vin back from 4.8V to 3.6 ( ie use 6 cells in 3S2P rather than 8) you will get a HUGE increase in efficiency. If you run the light the way it is now the current will drop off sharply after a minute or two anyway.

If you really really just want to use the full pack, not that it will increase your runtime at all because 7135 chipsets just burn off all extra voltage. You can remove the spring from the bottom of the driver (put some flux in it and remove it with your ironr and lay all your wires flat. Then take some thick bars or plates of copper and attach them directly to both sides of the board using arctic silver epoxy. If you get enough sinking on the board this solves both the heat and current issues.

Hope this helps

EDIT:

I dont have a picture of one of these boards with the sinking added, but you can see in this pic the srping is removed and the wires lay flat below the height of the chips.



In order to get the current stable in this mod I had to use 3/8" thick copper bar on each side of the board.

 

[The_bad_Frag]

The problem was with the old cells that the Vin with 4S2P was not 4.8V but 3.78V. With 3 cells the Vin was a lot lower and the XM-L was not even at half power.

I will test it again with the black eneloops and look if I can get it to work with 3P2S then.

But I already have an idea how I can heatsink it. :thinking: Sadly I cant get the material for it on the weekend.

 

[ma_sha1]

I run into the same problem when I did this light:
http://www.candlepowerforums.com/vb...-Yard-Sweeper-Mag-2C-3-mode-with-3x-IMR-22430

I solved it by wiring a inline power resistor. No difference in efficiency, as the same amount of excess heat will be burned off ,
just a matter of where, either on the driver or in my case, on the resistor.

 

[Justin Case]

I run into the same problem when I did this light:
http://www.candlepowerforums.com/vb...-Yard-Sweeper-Mag-2C-3-mode-with-3x-IMR-22430

I solved it by wiring a inline power resistor. No difference in efficiency, as the same amount of excess heat will be burned off ,
just a matter of where, either on the driver or in my case, on the resistor.

I don't think that is correct. a 7135 acts like a variable resistor. You have a fixed resistor.

 

[Justin Case]

The problem was with the old cells that the Vin with 4S2P was not 4.8V but 3.78V. With 3 cells the Vin was a lot lower and the XM-L was not even at half power.

I will test it again with the black eneloops and look if I can get it to work with 3P2S then.

But I already have an idea how I can heatsink it. :thinking: Sadly I cant get the material for it on the weekend.

I doubt 3xEneloops will be any better. At 2.8A draw, the cells probably will quickly settle down to around 1.2V each, giving you only 3.6V (even less than the 3.78V you are getting now).

Like I wrote above, if you want to run a 7135 driver at 2.8A to an XM-L, IMO you are going to have to use 4xNiMH and take the hit on driver efficiency.

I know that there are tutorials on CPF suggesting that 3xNiMH can hold their voltage at 2.8A draw to give you enough Vin to run in full regulation with a 7135 driver, especially if you use a low Vf LED. And that's correct in theory, based on the 7135 datasheet value of 120mV voltage overhead. But IMO the reality is that the voltage overhead is going to be a lot more than that for most flashlight mods. Now if the NiMH are D sized and can hold 1.3V each under load for an extended time, then you may be ok when using an XM-L emitter. But any lesser cell that sags down to 1.2V under load IMO is not going to deliver enough Vin to reach full regulation. Also, if you can eliminate the major sources of parasitic resistance in your flashlight, that may also allow you to use 3xNiMH.

 

[Al Combs]

Now I ordered 8 black eneloops. Anyone tested them already? But I think they can easily hold the voltage up to 1.2V when discharged at 1.4A.

Here is the Silverfox NiMh Battery Shoot Out thread. Do a browser Ctrl-F search for Eneloop. These are the white tops but it should give you some idea. It looks like 1.24 volts for 1.5 amp load (guesstimated midway between the 1 and 2 amp curves). That takes you out to about 40% of capacity. So if 3.78 volts on 4 over the hill batteries didn't do it, 3.72 on 3s2p Eneloops might not either. These are excellent batteries though.

I don't think that is correct. a 7135 acts like a variable resistor. You have a fixed resistor.

Actually ma_sha1 was using both an AMC7135 and a fixed resistor in a Download Poor man's Multi-Lux Setup. That seems like a good idea. Something like 1/3 ohm would bring 4.96 volts @ 2.8 amps down to 4.0 volts. The Shining beam could easily handle that without creating a heat problem.