Buckpuck 3023-dn-1000 Blew up??? burned up??

[mds82]

So i sell some LED lighting units for BMW 3 series. they are a replacement light that goes in the headlight. I have 1 3023-db-1000 powering 2 Cree XR-E's run in series. The power supply is highly regulated at 12volts. Had a customer complain that the light stopped working and his headlight was getting an error message. I get an e-mail with an attachment.....

WTF!!!! look at this. I have had 3-4 of these buckpucks break, 2 of them had smoke coming out of them. WIth this picture the customer had the unit installed and working since 4/3/08, 5 months ago...

 

[Oznog]

Did customer install it with BOTH + and - terminals isolated from ground?

 

[mds82]

In the car there are only 2 connections, + and - . there are no other connections available to install this.

 

[LED Boatguy]

Ringing

ASSuming everything was wired correctly, I'm thinking harmonics did that. The power systems of new cars are chalk full of noise and all it takes is a switcher like an unfiltered Buckpuck to wreak havoc. In this case the Buckpuck itself was the victim and not one of the big-bucks computers in the car.

It's been a while since I read the Buckpuck's datasheet, but there was something in there about adding a capacitor or two if the lead length was over a certain distance. Those caps could have saved the day here.

$.02

Found it. They say to add a 50V 220uF cap across the input terminals if the leads are over 18" long. I'd add a 1uF ceramic cap to that too if it were me.

http://www.leddynamics.com/LuxDrive/datasheets/3021-BuckPuck.pdf

 

[chadne]

I've blown a number of these in straight battery applications, so I'm not surprised it would happen in a car. In all of my configurations, I was well within the capacity of the driver, so I don't really understand why it would happen.

 

[Oznog]

In the car there are only 2 connections, + and - . there are no other connections available to install this.

I mean output + and output -. Some people see that and think they can tie the output - to ground on both the driver and LED. That won't work with most of these converters.

Or they see 2 lights over there and 2 drivers over here and think they can save wiring by tying the two - together as long as the two + are separate. 3 wires total. Again, probably won't work.

Or insulation failure in the wiring could have caused that problem too.

 

[Paul Baldwin]

Re: Ringing

Found it. They say to add a 50V 220uF cap across the input terminals if the leads are over 18" long. I'd add a 1uF ceramic cap to that too if it were me.

http://www.leddynamics.com/LuxDrive/datasheets/3021-BuckPuck.pdf

Hi, I've just setup a buckpuck using a 12v battery as the power source for my leds. I've wired a 50V 220uF capacitor as it says in the pdf as the battery is some distance from the bp. I'd really not like my one to go the same way as the one in the photo above! Can you please explain why a 1uF cap would be a good idea please? Or doesn't this apply to my application as I'm not connecting it to a car?

Cheers

Paul.

 

[snarfer]

Can you please explain why a 1uF cap would be a good idea please?

I think you meant to write 220uF cap. The capacitor compensates for the resistance of the long wires by storing power when the buckpuck is not drawing any, and providing it when the buckpuck is drawing it. Average current over the wires will stay the same, but peak current is reduced. As instantaneous power dissipation in the wires is proportional to current squared this will reduce average power dissipation in the wires feeding the buckpuck and thus tend to minimize temperature rise in the input wires.

I don't see any reason why this would explain the failing buckpuck though. I would check into whether the failing devices are consistently burning up in the same corner, and what is in there producing the heat. Perhaps it is the cheap inductor manufacturers usually use in these types of mass-produced devices.

 

[WeLight]

Motor vehicles can be difficult with respect to loads during starting and current surges, I would recommend the use of varistors or other transient suppressor device when powering leds of an auto rail

 

[LED Boatguy]

Re: Ringing

Can you please explain why a 1uF cap would be a good idea please? Or doesn't this apply to my application, as I'm not connecting it to a car?

Among a few other things, I'll put a 1uF ceramic or tantalum cap right next to the 220uF one. The reason is those big electrolytics may be good for absorbing the occasional electrical upset, but they suck at attenuating high frequency noise.

I still won't put a switcher in a newer car or boat. Three whites in series run just fine off of 12-14V and a $1 CC linear regulator.

$.02

 

[Paul Baldwin]

Re: Ringing

Cheers for clarifying that

Paul.

 

[Oznog]

Motor vehicles can be difficult with respect to loads during starting and current surges, I would recommend the use of varistors or other transient suppressor device when powering leds of an auto rail

Well, what you can do is a polyfuse and a 20V or whatever TVS (transorb) zener on a main power rail. Use the unidirectional TVS that clamps if the rail goes more than 0.7v negative when it's off (inductance may cause it to be more negative). So you've got a battery, alt, starter, then all the sensitive stuff fed by resettable fuse + surge protected line. The self-resetting polyfuse is nice but you do have to pay close attention to part selection. They have an internal resistance (as does any fuse) and if you have a normal fuse in series before it then obviously we want the polyfuse to be rated to open before the conventional fuse does.

 

[Illum]

EDIT: forgot the 3023 and 3021 difference



Also, does it need to be heatsinked or is the heatsink for the LED? the low thermal conductivity of aluminum will allow it to maintain itself to whatever its ambient temperature takes it...if the buckpuck is glued to the heatsink somehow it sounds like its receiving bad harmonics in addition to unnecessary heating

 

[WeLight]

I'm confused regarding the setup, it looks like the buckpuck had the wires directly soldered to it...considering the design of the Buckpuck uses DIP pins soldering heavy gauge wire directly to it would pose some sort of shorting risk wouldn't you think?

Why not using the wiring harness they have for the buckpuck?

Also, does it need to be heatsinked or is the heatsink for the LED? the low thermal conductivity of aluminum will allow it to maintain itself to whatever its ambient temperature takes it...if the buckpuck is glued to the heatsink somehow it sounds like its receiving bad harmonics in addition to unnecessary heating

The 3023 buckpuck have wires coming directly out of the potting, 3021 is the pinned version

 

[znomit]

The 3023 buckpuck have wires coming directly out of the potting, 3021 is the pinned version

Welight whats the difference between the old buckpuck and the new buckplus?

 

[Paul Baldwin]

Welight whats the difference between the old buckpuck and the new buckplus?

I'd like to know that too as I wasn't aware of a buckplus.
As an update as this thread has been resurrected I can state that the buckpuck I used is working fine It powers my Cree Q5 lounge light install for 50+ hours (since late 2008) a week powered by a roughly 30 year old scalextrix transformer that quickly kills DX style cheap drivers!
Thanks for selling a quality product Welight :and to CPF for the guidance :thumbsup:

 

[SemiMan]

It's called "Load Dump" and this is what it does to electronics never designed to operate in a motor vehicle without some sort of protection.

I have used quite a few buck-pucks and boost pucks with never a problem...but I do treat them right.

Semiman