Yes, it's SMD. This is actually funny, but I burned the last chip... I didn't connect it correctly. Before I was putting the chip correctly. But last time I didn't. I lost one led because of this. I also have no more PT chips. Tomorrow I'm gonna buy a few more. Last chance;]
I'm just hoping, the problem was a DC supply. So I won't destroy more circuits.

 

I tried again. 10V 5A supply. I changed the coil to 68uH. I connected multimeter instead of led. It showed about 500-600mA. But the chip was burning, so I switched off. I really dont know what to do...

SW pin is grounded...

Correct?

 

I made another pcb. I think it was working - current on the output was like I wanted. But, only for a few sec, then it's starting to burn... I have to turn it off and oon after a few secs. Why is that chip heating like that? Is it because I use lab power supply? Wrong schottky diode? Pad isnt soldered well? I tried with another power supply with 15V max 5A, and it started to burn very fast...

 

Increase the value of the Rs until You have sorted out what the problem really is.
Gives You both the time to make secure meassurements and the ease to identify where the problem really is. Try 1 Ohm first. At least.

 

SW pin is grounded..

It shouldn't be. Is that intentional or do you mean there's a permanent short circuit from SW to GND?

 

#1, use a breadboard when testing circuitry, not solder!

According to documentation here:

Translate: http://www.cheapled.ru/info/schems-pt4115.html

Rs = 0,1 / Iout

So with a .13 ohm you are over driving a 700 ma LED according to my calculator.

A .15 resistor would drive .666 ma

 

When I turn it off its making a short. I tried with higher value resistor, but it's still burning. (but now its shorted so it can be burning) Before, I used 0.33ohm and the output current was more less ok. I use breadboard, but I must solder the chip cuz its smd.

 

Can you take a picture of the circuit for us? Just a thought...

 

Can you take a picture of the circuit for us? Just a thought...

Yeah that might help. I would assume burger meant the actual circuit board. Might be a little hard to follow in a pic but what the heck.

 

Ill take a pic tomorrow. Now I just attach the pcb file.
R3 is a coil. Not a resistor. PT's pad is in front of R1.

 

I use breadboard, but I must solder the chip cuz its smd.

That might be your issue. You might want to just go for it and create the whole board. That is want I did.

 

That might be your issue. You might want to just go for it and create the whole board. That is want I did.

I built and rebuilt the circuit a few times. Now I'm helping myself with a breadboard.
I have a broken soldering station, Im using the other without temp. control. Maybe I'm heating the chip too much? But then, every single one? Normally i just touch for a sec the chip...

 

I built and rebuilt the circuit a few times. Now I'm helping myself with a breadboard.
I have a broken soldering station, Im using the other without temp. control. Maybe I'm heating the chip too much? But then, every single one? Normally i just touch for a sec the chip...

Where are you located?

 

Hmm...China? Could be bad chips. Here's as simple a diagram as I can make:

The dotted line is for connection to ground or v + dimmer control. It may not require full voltage to trigger. A 555 oscillator might be able to control that.

I'm not impressed with this chip as everything goes through it to ground. That means that all current flows through it at its expense. A typical idea to save parts and money. Is the LED heatsinked? Maybe that chip tab needs one too?

 

Is the LED heatsinked? Maybe that chip tab needs one too?

That tab is for a heatsink but the datasheet calls for a large pad. I don't remember the size it calls for. I am pretty sure a traditional heatsink is not needed.

Can you recommend a easy to use buck current regulator? I would be interested as well as the OP.

 

Ebay has a ton of them, but I never tried them yet. I'm still working with smaller LED's in Joule thief circuits as once you go over 120 ma you are using as much as some light bulbs. Plus the heat and current problems.

I've had 100 ma through a garden lamp 5252F 4 pin chip, but that was by accident testing an LED in circuit. I'm about ready to test them with more current and LED's. The chip is rated up to 5 volts in battery power which it can double with charging input and discharge limiting.

Is there any kind of chart with this PT4115 for coil and resistor ratings? It would seem that the coil value might change the current output too.

 

Ill take a pic tomorrow. Now I just attach the pcb file.
R3 is a coil. Not a resistor. PT's pad is in front of R1.

Where is Your current sensing resistor here?
It should be connected between Vin and Csn ON the board.
An external resistor is at no use.

And verify the pinplacing on You chip.

 

I made some pics. The pcb was made very fast, just to test. I was resoldering/replacing parts many times. So its full of mess. On the pics there are 2 different boards.
The chip should be ok, because I bought it in two different shops.

burger2227 - you can find a chart in the datasheet

 

You want those leads on your sense resistor much, much short. Looks like lots of cold solder joints there. I will check your routing once I get home.

Do you have access to a laser printer? Are you familiar with the toner transfer method?


Are these shops local? I wish I had local parts shops.

 

One local shop, and one via net. Yes I know transfer method.