• You must be a Supporting Member to participate in the Candle Power Forums Marketplace.

    You can become a Supporting Member.

BB750 shutting down after 10-12secs?

hotbeam

Flashlight Enthusiast
Joined
Jan 2, 2003
Messages
1,737
Location
Melb, AU
I just finished wiring up 2 BB750s ready for installation into the McModule. The BB750's were tested fine. Both lit up my 5 watters when they were powered from my 6V battery holder. I then fitted a BB750 into a McModule, powered it up for a few seconds to make sure contacts were not lost or loosened during the fitting. No problems there. The next step was to pot the circuit in the McModule with Artic Alumina to aid the transfer of heat from the components to the McModule.

Note the added copper heatsinking on the LTNH and the diode /ubbthreads/images/graemlins/grin.gif
bb750_heatsink.jpg


After potting and letting it cure for 30 minutes, I proceded to test it inside a PM6 host running 2 x 123s. What I found was frustrating! After about 10-12secs on, the light dropped from max brightness to moon-mode. By moon-mode, I mean I can look directly at the 4 dice in the 5 watter lit.

I also tested the BB750 which I haven't installed into the McModule and it does the same thing. That is, it would go to 'moon-mode' after about 10-12 seconds of power on. So what's going on?
 

shiftd

Flashlight Enthusiast
Joined
Nov 17, 2002
Messages
2,261
Location
CA
Holly molly, what a nice heatsink /ubbthreads/images/graemlins/grin.gif

Actually my BB sammich behaves in exactly the same way. However, mine took longer time to drop into moon mode (about 30 seconds or so).
I believe it was because thermal shutdown, but after seeing your heatsink, i don't think that is the problem anymore. /ubbthreads/images/graemlins/icon3.gif
 

hotbeam

Flashlight Enthusiast
Joined
Jan 2, 2003
Messages
1,737
Location
Melb, AU
So what the @#$%? I thought the added heatsinking on the sensitive components would help move heat from the components! That board was also potted in the McModule with Artic Alumina... right up to the level of the anode cap.

You don't want to ask which LED was epoxied to the McModule /ubbthreads/images/graemlins/mad.gif /ubbthreads/images/graemlins/mad.gif

Why would this configuration overheat? There is plenty of thermal transfer path for the circuit. It is only a BB750. Obviously the circuit has gone into thermal shutdown mode. But why?

What to do with this now?
 

kakster

Flashlight Enthusiast
Joined
Feb 6, 2003
Messages
1,903
Location
London, UK
I've had a BB750 do that too. It was in a regular McLux bezel driving a 5 watter, im 99% sure it's overheating for some reason. Mine would run about 2-3 minutes then then just dim right down. Potting it in AA solved the problem for me, Im not sure why your's would be made worse with a heatsink attatched. Does the sink actually touch the McModule or flashlight bezel/body? It might be transferring heat from the Luxeon to the BB.
 

shiftd

Flashlight Enthusiast
Joined
Nov 17, 2002
Messages
2,261
Location
CA
ah ya, in view of kakster reply, i just remembered that my module only thermal shutting down when it is attached to the led. I think it was the heat of the led that actually caused it to thermally shutdown. when i tested the board, separate from the lux, it ran undisturbed at all (i tested for about 2 mins or so).
Maybe your heatsink is the one that actually made it worse? /ubbthreads/images/graemlins/confused.gif
 

hotbeam

Flashlight Enthusiast
Joined
Jan 2, 2003
Messages
1,737
Location
Melb, AU
Kaks, the copper heatsink does not touch anything else but the top of the chip and the top of the diode where it was AS'ed (Artic Silver epoxy). The heatsinks do not touch each other. The Artic Silver does not touch any legs on the chip. Further the whole thing was AA'ed to the McModule.
 

hotbeam

Flashlight Enthusiast
Joined
Jan 2, 2003
Messages
1,737
Location
Melb, AU
[ QUOTE ]
shiftd said:
...i just remembered that my module only thermal shutting down when it is attached to the led. /ubbthreads/images/graemlins/confused.gif

[/ QUOTE ]
So what you are saying is the BB750 board shuts down when it is powered with the LED attached? /ubbthreads/images/graemlins/grin.gif So what good is the curcuit in isolation then? /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif
 

shiftd

Flashlight Enthusiast
Joined
Nov 17, 2002
Messages
2,261
Location
CA
Actually howie, what i meant to say was to use the board not connected straight thermally to the lux. I mean, like in sammich, the board lies below the lux, and the lux' emitter board help transfer the heat from the lux to the flashlight body, while at the same time transfer the heat to the epoxy, which in turns transfer the heat to the converter. I believe the same thing is true for Mcluxes mod, which makes me wonder, if those Mclux mods can works without thermally shutting down, why could not yours? /ubbthreads/images/graemlins/confused.gif

Ok, maybe i am just babbbling (it is late here now), but i believe thermal shutting down is the cause of the dimming.
 

jtice

Flashaholic
Joined
May 21, 2003
Messages
6,331
Location
West Virginia
Hmmm this is strange.
That deffinately shouldnt be happening.
If you have it potted in AA real nice, with no or very little air pockets, it should be fine.
15 seconds shouldnt get it too hot.

And the fact that you have TWO BB boards doing this?? thats very odd.

Maybe Wayne will have some suggestions, but, I am at a loss. /ubbthreads/images/graemlins/frown.gif
 

Doug S

Flashlight Enthusiast
Joined
Jun 20, 2002
Messages
2,712
Location
Chickamauga Georgia
[ QUOTE ]
hotbeam said:
I also tested the BB750 which I haven't installed into the McModule and it does the same thing. That is, it would go to 'moon-mode' after about 10-12 seconds of power on. So what's going on?

[/ QUOTE ]
I am not surprised at all by your results. You are simply asking the LT1618 to dissapate more power than it possibly can. Even with your heroic efforts to enhance the heat transfer, it just isn't possible. Using 2X123 driving a 5W luxeon, the IC is likely operating at, or close to, its current limited condition. I have personally been surprised that so many people have had success with the BB500 using 2X123. At 750mA output, the LT1618 is trying to dissapate close to twice the power. It is simply not viable.
A related thread that may be of interest is this one:
Heavily loaded 123 plus BB
 

jtice

Flashaholic
Joined
May 21, 2003
Messages
6,331
Location
West Virginia
Doug,

I have heard ALOT of ppl using a BB750 with a 5W and 2x123.
I never really heard this being a problem, unless the BB wasnt potted in well.

Could it be the 5W led he is using? Has too high a Vf?
 

dat2zip

Flashlight Enthusiast
Joined
Jan 5, 2002
Messages
3,420
Location
Bay Area
The experiments I've done have shown the exact same thing.

The heat problem is a global one. Just heat sinking the top of the IC won't help. The diode too dissipates as much heat as the IC does and this is spread to the whole board. Feel the board, it's toasting.

I found the only way to tame these conditions is to fully pot the top of the board with Arctic Alumina and then bond that to a heatsink. Again, contact surface area is important, and just extruding a rod into the thermal bog on the board to a massive heat sink won't cut it. I have found that thermal bogging the Ecan is sufficient heat sinking to run most any conditions and keep the BB from thermally shutting down.

Remember, that the IC is trying to dissipate close to 1/2 W and so is the diode.

Also, I think the top of the IC is thermally isolated and makes a poor thermal conductor for the heatsink. I have suspicion that the IC legs are moving more heat, which is effectively the board. Maximum constack to the IC legs and the board is where I think the heat must be pulled from.

I've been successful in just pouring a thin layer of Arctic ALumina in the Ecan after it is in place. This coats the board and makes perimeter contact with the Ecan. This has about the same effect as if I had potted the whole board and saves on Arctic ALumina. That stuff is expensive.



Wayne
 

McGizmo

Flashaholic
Joined
May 1, 2002
Messages
17,290
Location
Maui
I respect Doug S.'s take on this but I am certainly confused as I have made quite a few BB750/5W/2x123 and have not experienced shut down once? As I recall, Charlie had a few problems with a McModule run that he did and it was determined that proper thermal relief of the IC needed to be more agressively addressed. To my understanding, these issues were resolved and there are quite a few of these units out there??!?!

What Hotbeam is experiencing here doesn't make sense to me and is contrary to the experience (limited as it may be) that I have had???

Initially, I too wondered if the great heat sink that he has applied hasn't provided an express path for heat from the LED to be directed at the IC. Now that he states the IC sink is isolated from the LED sink, this premise is weaker. Something certainly is wrong and as far as I can tell, atypical??!?! Certainly the IC needs to dissipate heat and not receive additional thermal loads from external sources. The 10 to 12 second duration prior to shutdown makes me wonder if there isn't possibly some type of short common to both drivers or something?

Hotbeam, can you measure the circuits current demand from the batteries?

- Don

EDIT: I see Wayne has answered while I was slowly typing my post. Nevermind
 

shiftd

Flashlight Enthusiast
Joined
Nov 17, 2002
Messages
2,261
Location
CA
Also, i believe there was even a BB917. Wait, is that BB or DB? /ubbthreads/images/graemlins/confused.gif
 

McGizmo

Flashaholic
Joined
May 1, 2002
Messages
17,290
Location
Maui
shiftd,

That's a down boy 917 and the good news here is the buck circuit is running less current but a higher voltage through the IC. It is my understanding that the current is partly the culprit in shutting down the IC. The 917 is driving a Luxeon III so the total power required and being generated is less than the 5W driven at 750 mA. At least this is my understanding.
 

dat2zip

Flashlight Enthusiast
Joined
Jan 5, 2002
Messages
3,420
Location
Bay Area
I have run the BB on the bench at 1A output and the IC pumping max power (min Vin). I can tame it with air alone so I know that if you heat sink the whole board (IC and components) that it should be OK.

I also know that I tested a rod to the top of the IC and no matter what I did with this rod, stick in a massive water container, clamped it to large chunk of metal that the board temperature was hardly effected. Granted I used aluminum, but, in almost all cases where there is shutdown is because there is air or air pockets touching the board.

10-12 seconds is atypical. From a cold start, it takes more than that before the board thermally heats up.

Also, in the picture provided, I see some of the extra copper holes that the CM kindly added. Those could be connected to GND and in this case is OK if they short to the wall of the module.

You say that it ran fine on the bench? What was the condition on the bench setup? What was the voltage and did you measure the current the converter board drew under these conditions.

BTW: All 5W configurations are running max power when it reaches low Vin and dropout occurs. This occurs on all configurations and on some 1W/Lux3 configurations.

So, there are quite a few 5W modules out there that run just fine. On a 5W, A BB300 would still fall out if regulation assuming the 2X123 can huff out the current with low Vin. I know BB500, 700, and 750s will do it all the time and I am unaware of any problems if HS is done properly.


Wayne
 

Doug S

Flashlight Enthusiast
Joined
Jun 20, 2002
Messages
2,712
Location
Chickamauga Georgia
Wayne's description on how to make it work makes sense, that is, to minimize the path length through the thermal bog to the real heatsink. Hotbeams added heatsinks have the effect of lengthening this path. Remember, that thermal bog is a very poor thermal conductor relative to copper or aluminum. I agree with Wayne that the primary thermal path out to the IC is likely via the pins and I suspect mostly via the ground pin.
Now about those 2X123 setups driving 5W luxeons that work, I propose a worthwhile investigation. When the BB is in regulation, the sense resistor(s) will have 50mV across them. If less than 50mV, it has fallen out of regulation. It would be easy enough to bring a pair of fine wires out of the potted module to monitor this sense voltage to see if the modules are operating at the currents that folks are assuming.
 

shiftd

Flashlight Enthusiast
Joined
Nov 17, 2002
Messages
2,261
Location
CA
Ah ya, thanks gizmy. /ubbthreads/images/graemlins/blush.gif
 

dat2zip

Flashlight Enthusiast
Joined
Jan 5, 2002
Messages
3,420
Location
Bay Area
It's actually very easy to see the effect with a power supply. As you drop the voltage say starting at 6V, you will see the current increase. At some point it could start wavering or peak close to 2A. Continuing to drop the supply will not increase the current. This is the MAX power (heat wise) at this corner.

If this corner continues to move to higher Vin, then the board is heating up. A well heat sinked board will hold the corner dropout location, If you bring the voltage up and then ramp it down and note that the min Vin is now higher, the IC temperature has risen and probably quite a lot.

That's actually a good test to run before and after potting is determine if the converter board will hold regulation for the total range of the battery setup you plan on using. If you are using 2X123 you could estimate or guestimate that nominal is around 6V and that pooped batteries are 4V for ezample.

You can now test your configuration on the bench with a power supply and verify the converter board ISW and all the other tolerances indepedant of the what the computed dropout point is and measure it confirming where it really is in this setup.

If, by chance you got the IC with skewed ISW set at 1.5A instead of 2A (typ), you might see the dropout occur at 4.5V for example. All the numbers I'm providing are example only and I am not stating that the min V on a 2X123 is 4V, but, for apply an example of how you would use this information only.

This is what I do with my converter boards before I pot them. (I use a stiff breeze from a fan over the converter board and the LED).

A good bench supply is needed to perform these experiments.

This is also good to perform after potting to verify that the completed module or assembly is thermally heat sinked and will not thermally shut down. By using the bench supply should simplify and take out any of the guesswork before you button up the flashlight.

I hope this helps. I've run enough supply runs to visualized them with my eyes closed. I've also noted power supply fluctuations that I can tell if the IC is screaming, whining or blissfully ignorant.

Wayne
 

Doug S

Flashlight Enthusiast
Joined
Jun 20, 2002
Messages
2,712
Location
Chickamauga Georgia
[ QUOTE ]
jtice said:
Doug,

I have heard ALOT of ppl using a BB750 with a 5W and 2x123.
I never really heard this being a problem, unless the BB wasnt potted in well.

Could it be the 5W led he is using? Has too high a Vf?


[/ QUOTE ]
High Vf does aggravate the problem since higher Vf means that the IC must provide more power to deliver the same current.
BTW, I "hear" a lot of things on CPF that I take with a grain of salt /ubbthreads/images/graemlins/grin.gif
 
Top