LED light . . . im stumped

zachtheterrible said:

R1 would likely cause problems the way you have it mounted on the board. 33mOhm is pretty low and the contact resistance of the board is probably higher than that, inconsistent too.

Click to expand...

I hope that this change won't screw anything up when i put it on a PCB.

Click to expand...

No your PCB will be fine, though don't return the main current through ground through the same trace leading onto the ground pin of the chip. Of course this is already a pretty inaccurate circuit so it's not really anything to worry about. Soldered connections will not have resistances on the order sometimes possible with the protoboard. And the protoboard is hard to predict, you could have a great connection on one pin and a lousy one on another. Or it may heat up and change temp. No prob for most jobs but a 33mOhm current sense is a special issue.

The pcb layout in the Silicon Chip magazine project looks like it has the high current paths separate from the sensitive ones. 8)

The silicon chip magazine's PCB is not the correct size at all. In my PCB the track sizes are also much bigger, mostly .032". The different loops are also kept away from each other.

audioguru said:

The pcb layout in the Silicon Chip magazine project looks like it has the high current paths separate from the sensitive ones. 8)

Click to expand...

I have just come into this thread.

I saw a data sheet for a switching regulator a few months ago and it recommended that the ground tracks should all be connected at one point, ie. a "star" connection.

This prevents voltage drops across the tracks carrying the high currents affecting the sensitive areas of the circuit.

The same could also be done on the positive supply side also.

Len

I found the data sheet I referred to above. See page 24

Len

So, im having absolutely NO luck with the PCB. It only seems to work on the breadboard. The two PCBs that I have created for this project have not worked at all.

I want to use the one from the chip-torch article, but i cannot figure out how to size it down to the right size. Whenever I try, the quality becomes terrible. Can anyone get it to the right size? I have tried changing the file formats to .gif and .bmp, but the same thing happens.

Is that any good? You'd have to tell me the exact size you wanted i though. Alternatively, download "irfan view", it's well worth having anyway.

What is irfan view? Is that what you used to do it?

Oh yeah, i just realized it would be impossible for you to give me the right size without actually printing it out and stuff. ill have to do it on my computer.

Irfan view is a free program that nobody should be without in my opinion, it solves most image problems there are. Yep, its what I used.

https://www.irfanview.com/

I ALMOST had this thing finally mounted and working in a flashlight (i got the PCB to work), when i blew out another $11 LED!! :evil: :evil: :evil:

I cant believe it! I think what may have happened is that a decoupling capacitor accidentally got disconnected while I was working on the cirucit and it let large voltage spikes through the LED. Do you think that's the problem?

EDIT: I figured out why it burnt out. It was not the decoupling capacitor. When there is no load on the circuit, the circuit produces around 30v. When I connected the LED I saw a small spark. It was obviously enough to zap the LED. The project is almost done though. When I get my next LED, it will finally be complete and ill post pics :lol:

zachtheterrible said:

I just realized that my circuit is oscillating @ 370khz! Almost twice the nominal frequency that the datasheet recommends: 200khz.

Anyone know how to slow it down?

Click to expand...

Hi,

I SUSPECT THE INDUCTOR MAY BE PARTIALLT SHORT. PLEASE CHECK.

2. ple check the electrolytic 220uF for leakage.