BuckPuck 3021-dn-700 open circuited? blew up my sample MC-E

I was testing out my Cree MC-E tonight by using 4 LedDynamics Buckpucks 3023-dn-700's running in parallel with each other to power the 4 dies all in parallel.

All of a sudden, after 5-10 minutes of use, i heard a faint pop, the LED light went do a mean deep blue, and instantly shut off the power supply.

Looked at the LED and all 4 die's were completely black around the edge!!!!!!! so F'ing pissed

Anyone else have problems with the LEDDynamics BuckPucks ??

I hear your frustration in losing your led so quickly... I have to ask some questions to give you an answer.
Did you mount to a heatsink? If not then heat killed the led, they can't be ran unmounted without sinking.
I believe it is a heat issue, I use the buckpucks and have been sucessful in many different experiments using them.

overheating is the cause of a die turning blue
Russ

With 4 die and 4 buckpucks shouldn't they all be wired independently? 1 die going out took out the other 3?

something has me thinking the slug isn't electrically isolated:candle:

Interesting. Got any pictures?

Did you measure the Vf of each die to see if they were well matched so you could run them in parallel?

Maybe there was a problem with the heatsinking, or possibly the drivers aren't meant to be wired in parallel with each other, but you wouldn't think they would fail in a way that would cause excess current to the LED.

Have you tested the drivers after the incident? Do they all work fine independently?

Illum_the_nation said:

something has me thinking the slug isn't electrically isolated:candle:

Click to expand...

What, the fact that you can run the dies in series?

The slug has to be isolated for the dies to be run in series.

well, from what I've read, unless all four pucks failed at once which is very unlikely, how can you explain how 1 failing die killed all 3 under the dome?

I'm baffled as to how this is all wired together, pics would be helpful

i had the 4 dies of the led wired up in parallel with each other. I also had 4 of the puckpucks wired in parallel, each giving off 700ma so a total of 2800ma for current.

Heatsink - that wasnt the issue, i had a massive CPU heatink and it never even got warm.

it was working perfect, and all of a sudden it just turned angry blue and stoped working. I checked afterwards and the - and + leads of LED 1, and with the continuity checker found that there was no resistance, and there was perfect continuity

I think what happened was that one of the buckpucks more or less open circuited, pushing a full 14v down to the led, pushing unlimited current to the led's burning them.

i'll take a pic tomorrow even though i ripped the lens off already

Check the drivers to see if they're all still working properly. I'd like to know whether it was the LED or the driver. Too bad you never tested the dies to see the Vf of each one.

If the heatsink never even got warm, was the MCPCB making good contact with it?

VanIsleDSM said:

Check the drivers to see if they're all still working properly. I'd like to know whether it was the LED or the driver. Too bad you never tested the dies to see the Vf of each one.

If the heatsink never even got warm, was the MCPCB making good contact with it?

Click to expand...

yes i'm sure it was propery heatsinked, the whole MCPCB was glued down with thermal epoxy and the board / led was cool to the touch afterwards.

The LED driver's however were extremely hot and dont work anymore - so ya it was a bad driver

What was your input voltage to the drivers? 14 volts? The forward voltage on the MC-E is about 3.5 per channel. Your drivers (each of them) were working really hard and probably over-heat and were destroyed. In some situations a driver can release the full voltage down the line and blow up the LEDs when it fails.

Also, you spoke about all these being run in parallel. You used the words "with each other" You didn't tie all those outputs together by chance? They should have all been independant circuits....

What you should have really done on this to match everything up was wire all the LED circuits in series and applied a single driver channel to the whole LED. That 14 volt input then would have been matched to about a 13 volt voltage drop over the LEDs and the single driver channel would have ran very cool. You should try that on the go-round.

KeithInAsia said:

What was your input voltage to the drivers? 14 volts? The forward voltage on the MC-E is about 3.5 per channel. Your drivers (each of them) were working really hard and probably over-heat and were destroyed. In some situations a driver can release the full voltage down the line and blow up the LEDs when it fails.

Also, you spoke about all these being run in parallel. You used the words "with each other" You didn't tie all those outputs together by chance? They should have all been independant circuits....

What you should have really done on this to match everything up was wire all the LED circuits in series and applied a single driver channel to the whole LED. That 14 volt input then would have been matched to about a 13 volt voltage drop over the LEDs and the single driver channel would have ran very cool. You should try that on the go-round.

Click to expand...

I agree. You should only need one 700mA buckpuck if the dies are run in series from a 14V or greater source.

Another option would have been to reduce the input voltage to something just above forward voltage -- like 4 or 5 volts (I don't know much about your driver but most drivers will accept a wider range of input voltages so you would need to read you specs on that).

Well technically I think what you mean is "short circuited".

I have run buck pucks in parallel configurations quite a few times without any trouble. In general, you should be able to parallel most constant current drivers without any troubles. It is not a guarantee, but for a standard buck regulator, it would normally not cause any issues.

I highly doubt as many have said that the mismatch between forward voltages in the dies was an issue. If that was the case, one of the dies would have gone bad first, and the others would have gotten brighter till one of them died and so on and so forth.

The buck-pucks are in general quite efficient. While there is an efficiency improvement by having closer matching of the input and output voltages, even with 14V and 3.5V out, the dissipation in the Buckpuck will be minimal and while it may be warm, it should be no where near hot to touch at 700mA.

Now that said, the Buckpuck is only rated for 15 seconds short circuit. LEDs do fail short circuit unfortunately. I would expect that only to kill one die though and if it was shorted, in theory there is no current and hence no blackening of the other dies.

Going out on a limb (or taking some experience and knowledge into account), I would take a look at what was happening on the input to the buckpucks. You said 14V ....what were you using for that 14V? You imply all the buckpucks are dead? That is a bit unusual.

One last thing to consider is static electricity. Buckpucks are robust, but you hit it with a good static discharge and you could blow one. I do not know what the internal circuitry is, but if the internal switch fails on, bye bye LEDs.

Semiman