HELP...it went poof

[737mech]

Ok I am in need of some guidance. I was "trying" to make this light:

2D mag host with modified switch to be momentary
DSWOI P7
H22A heatsink
2 AW IMR18650 1600mah cells
1 D2Flex driver controlling HipCC driver

Got everything hooked up just like recomended by George on the TaskLed site
http://www.taskled.com/techhipcc.html

When I turned the light on and I had some smoke. Turned off the light and found Q1 on the D2Flex fried! Any thoughts? ANY and all help appereciated.

 

[russthetoolman]

did a metal reflector happen to short across the led leads?

 

[737mech]

no I never made it that far...lol

 

[Justin Case]

Did you install the jumper wire to put the d2Flex in PWM mode?

 

[Illum]

Did you install the jumper wire to put the d2Flex in PWM mode?

especially note the bottom of one of those PCBs....


switch- to MOSi

Easiest way for a FET to go poof...probably would be overvoltage from switching spikes
or it could be a short from solder/damage during the jumper installations..shorting between G/S, or G/D...

One thing to note, is that MOSFET usually fail short-circuit, as opposed to "burning" open-circuit.

 

[737mech]

yes I did the jumper wire, I double checked all my wires and soldering and I could not see any obvious shorts.

 

[Illum]

hmm...can you provide us a pic of the problematic driver?

I'm unfamiliar with the H22A heatsink, is it anodized?

While there are severe limitations by looking at the burnt FET and reverse engineer where the fault may lie...I'm out of ideas:shrug:

Unless its one of those "obvious" errors that will not be obvious until someone identifies it like inserting an battery or soldering an LED backwards...which I have repeated the mistakes so many times I question myself whether I should just put the iron down, head out and play ball instead:shrug:

 

[Justin Case]

Re: H22A DHS heat sink

Bare or anodized? Aren't some H22A heat sinks solid on the bottom, vice the anodized heat sinks typified by 4sevens and LITEmania with a driver cavity in them?

 

[737mech]

I could post a pic but I already disconnected all the wires.. The heat sink does not have a cavity and is solid.

 

[Mirage_Man]

I guess since George is on vacation there's no way to ask him about this right now huh?

 

[737mech]

He was kind enough to respond to an email while on vacation, talk about good customer service!

"To smoke Q1 means there was a short somewhere. Q1 provides power to the hipCC, so if there is a short somewhere in the hipCC path then Q1 would see a short/overload."

 

[Illum]

short on Hipcc...component testing time...

 

[Mirage_Man]

short on Hipcc...component testing time...

I wonder if George meant there's a short on the board from board assembly or during 737mech's assembly?

 

[Illum]

I wonder if George meant there's a short on the board from board assembly or during 737mech's assembly?

assuming nothing else is connected yes...but what if the short lies in the PC board that the wires leading from the hipcc board and shorting the hipcc out?

 

[Bimmerboy]

*note to self - avoid getting on any 737's.*

Just kidding! :nana:

Too bad about the board, man. I'm about to do a very similar build, except using a Britelumens sink, and have been planning a way to heatsink the hipCC, and protect from any possibility of the boards touching. There's not much space between the Mag switch and emitter heatsink.

Do you think your two boards possibly shorted out against each other?

Also too bad you can't tell us about how well this setup works, as I've been waiting for someone to comment on the 240Hz PWM frequency.

 

[Justin Case]

Why not cut off the Mag switch tower to provide more room? Also bend or cut the metal ground strip so that its height doesn't stick up past the top of the remaining switch housing.

 

[737mech]

I did cut the mag switch tower. As far as the boards shorting against each other no. I was bench testing before I was going to install the heatsink into the mag. I disconnected everything and will try this setup again, triple checking all connections and wires for shorts.

 

[greenLED]

I was bench testing before I was going to install the heatsink into the mag.

Right, but was the emitter already on the heatsink? In that case, is the heatsink anodized? Was the ground path on the emitter isolated?

 

[matrixshaman]

I'm not totally up to speed on this but I do have one of the D2Flex drivers running a P7. I see one possible problem that I don't believe has been mentioned. You list that you are using 2 of AW's IMR 18650's which are 1600mah. However a single one of these batteries can by their special chemistry probably put out at least 5000 ma to possibly 8000 ma. 5 to 8 amps is my guess based on what my little IMR 16340 puts out when I fire up the P7 (about 2.8 Amps). So if you don't have a current limiting resistor for the D2Flex as suggested it may be possible it was fried by exceeding it's rated load which is 3.4 Amps. I don't know much about the hipCC but I see it can drive a bunch of P7's. These are just mostly guesses on my part but since the IMR's are fairly new and very capable of some extreme current output it might be a factor. Other than that I can only think to double check all wiring - I always try to use color coded wiring to avoid problems. Of course it's also possible you just got a bad D2Flex board or maybe even the hipCC board has a problem that caused the D2Flex to exceed it's current limit.

 

[737mech]

Heatsink is anodized.