I have built a small LED panel from straw hat LEDs (blue), 4 in series each, and 12 in parallel. I have figured out forward voltage is about 3 volts.

Brightness distribution is 100% even.

I am using a MC34063 buck regulator. It even has a 1m piece of wire as a current sense resistor. Maybe it is too high a resistance (thin wire wrap wire).

Because the chip turns too hot even at only 280mA output current. It turns so hot the current is decreasing down to 230mA.

Maybe the inductor is not right? I use a 220uH toroid with 1 Amps spec.

Input voltage is 19.5 VDC from a laptop adapter. Output voltage is about 12V.

The diode is a good fast recovery diode, as well 2x 100uF caps.

I tried 1nF and 470pF but both frequencies result in too much heat dev.

Should I add an external IRF MOSFET? As they have a larger metal tab, and can easily be cooled. The type I want to use is a PNP MOSFET. I also have various BJTs around here. If I have PNP is it better to use than NPN for a buck regulator? But why not use a MOSFET.

True I have never done that with MC34063.

The light output I get from the 48 LEDs panel is very satisfying. But getting so much heat from one small panel already! There must be something wrong in my circuit, or is it really the MC 34063 is only good for a few 100mA?

Also I have a few different coils here to change if need be. I am not too bright to know the values for a buck so maybe if the value is not right, explain.

 

I think you need a current limiting resistor in series with each 4-LED string. This is a voltage regulator with current limiting, but is not a constant current source. Based on the current draw indicated you could actually raise the regulated voltage a few volts, then calculate a resistor to allow about 20 ma through each string of four LED's.

 

My problem is the MC 34063 chip. It turns hot.

 

Yes.
I'm thinking its turning hot because it's going into current limiting.
The LEDs will reach a certain forward voltage to start lighting up. At that point the current draw will increase rapidly as the voltage is increased slightly if no current limiting resistor is in the line.

 

Another thought.
Try putting a low value resistor in series with each LED string, say 10-22 ohms to drop the voltage a few tenths at 20 ma - see what happens.

 

Well I have an Amp meter inlined.

They are always a bit off for SMPS, but I know how accurate these are from measuring the input current for a TL494 board, having 12V bulbs, and calculating the efficiency.

The displayed current seems to be 10% off.

Since the frequency for the MC34063 is lower, it can't be that much.
Also from the heat which the LED board is developing the current must be about 20mA each.

I will try a 680uH coil now, unfortunately they are very large, so I would have to order 470uH coils to use with this chip.

The LED power is about 3,8 Watts or something, if you consider 80% efficiency, that makes about 1W for a small PDIP IC with no copper groundplane. Maybe it is indeed the wattage.

Not too sure, I an thinking about this myself. Current through the LEDs is nominal I think.

Or maybe the wire resistor I made is wrong (too high). I will shorten it and see if the heat goes away.

 

I believe that what I posted in post #5 is a good starting point.
Why not give it a try?

 

The LED current is not the issue.

Does it satisfy you if I replace it with a power resistor for the time being?

I have installed switch to short the current sense resistor, and to change to 680uH coil- no result. Maybe I should try 100uH as well if I have one around here.

Then the only resort will be to install a P-channel MOSFET. This would also get rid of the dropped voltage from the internal NPN.

 

Hi takao21203,
I use this ic to drive 8 LEDs in parallel from 12v without problems.
First, I measured the forward voltage of the leds at the desired current using my bench supply and then adjusted the feedback resistors of the MC34063 to obtain that voltage.

Did you include Rsc between pin 6 and pin 7 (0.22 or 0.33 ohms) ? Also, the wire you coiled may be causing instability. If you have a scope, look at the waveform at pin 2.

220uH is the recommended value.

Timescope

 

Hi takao21203,
I use this ic to drive 8 LEDs in parallel from 12v without problems.
First, I measured the forward voltage of the leds at the desired current using my bench supply and then adjusted the feedback resistors of the MC34063 to obtain that voltage.

Did you include Rsc between pin 6 and pin 7 (0.22 or 0.33 ohms) ? Also, the wire you coiled may be causing instability. If you have a scope, look at the waveform at pin 2.

220uH is the recommended value.

Timescope

Yes the wire coil is used for the Rsc. I have installed a switch to short it with no result at all.

I use 48 LEDs.

I have also roughly calculated the Ptot distributed through the PDIP package as 1 Watt. Maybe that is the reason and I should not bother too much.

I know also LM2596 heats up considerable driving a few cooling fans, and these are DPAK properly mounted on a copper PCB with many vias.

Unfortunately I do not have a small PDIP cooler here.

 

I bought these for use with the MC34063, as well I bought some MC34063 PDIP ICs.

All 220uH. The large coil is not the right thing for this chip

What I want to try for a long time is to use 3x BC327 in parallel for SMPS. Maybe a big effort to connect 9 wires anyway.

I also have various BJT here- most of them far overdimensionated. And of course the IRF MOSFETs.

I don't build so many analogue circuits means I don't have so much concrete experience with dimensioning.

All your answers may be useful for me.

 

The value of Rsc is important because it sets the peak inductor current.

From the datasheet, Ipk = 2 x Iout = 480mA in your case.

Rsc = 330mV / Ipk = 0.68 ohms

If the peak current is too high, the power dissipation will also be high due to losses in the Darlington output transistors.

Timescope

 

The value of Rsc is important because it sets the peak inductor current.

From the datasheet, Ipk = 2 x Iout = 480mA in your case.

Rsc = 330mV / Ipk = 0.68 ohms

If the peak current is too high, the power dissipation will also be high due to losses in the Darlington output transistors.

Timescope

Is it really employed to limit peak currents artificially? I don't use it on my TL494 board and it runs through approx. 200W continously.

As far as I understood it is a function of load current. Limiting the peak will directly limit the load current.

Anyway, I have bypassed the Rsc with no result at all.

How the MC34063 would somehow clip off the peaks but still output the same current is not clear to me.

The LM2576 does not have such a clipping only at 5 Amps, and efficiency is still regular all through the range.

 

I did actually another test with a 47 Ohms resistor (6x 270 Ohms).

Heat ramp is the same, it does not even take a bit longer.

So the LEDs are definitively not the origin of the problems.

I have tested with static load and it behaves the same so it also can be believed the LEDs don't draw excess current, and the DMM also roughly reads the right value.

 

Some things to consider:
You agree that the current through each string is around 20 ma's.
But the voltage drop across each led is about 3 volts.
Why not run 20 ma using an appropriate resistor through 1 LED (and try a number of them to get an average)
and measure the exact voltage drop. Then move on from there.

 

Anyway, I have bypassed the Rsc with no result at all.

That is the problem. Connect two 1ohm resistors in parallel between pin 6 and pin 7.

The TL494 does not sense the current and uses voltage feedback. Modern pwm controllers use current mode control that senses the current through the transistor and switches off the drive when the current reaches a certain value on a cycle by cycle basis.

Timescope

 

Some things to consider:
You agree that the current through each string is around 20 ma's.
But the voltage drop across each led is about 3 volts.
Why not run 20 ma using an appropriate resistor through 1 LED (and try a number of them to get an average)
and measure the exact voltage drop. Then move on from there.

I did measure the voltage drop.

All LEDs light up totally even.

Moreover, I have installed a 47R resistor which is a static equivalent.

You talk about a problem which does not exist as such.

 

That is the problem. Connect two 1ohm resistors in parallel between pin 6 and pin 7.

The TL494 does not sense the current and uses voltage feedback. Modern pwm controllers use current mode control that senses the current through the transistor and switches off the drive when the current reaches a certain value on a cycle by cycle basis.

Timescope

It does have two feedback comperators.

Actually I have installed a switch so I can use the wire-made resistor, or no Rsc whenever I fancy.

I did not mention it but on a previous circuit I made yesterday, I shortened the output, and the MC chip exploded instantly. With the piece of wire coiled up, that is not so likely to happen.

I have calculated the resistance for much thicker bell wire, 1m, as about 0.1 Ohms. So this wire is much thinner, very thin wire wrap flywire, and it is more than 1m. It is somehow likely 0.5 Ohms a little more or less. As I said, bypassing it with the LEDs running does not produce any effects.

 

Please show the schematic you are using!

The 34063 running buck 20v->12v at 280mA will run nice and cool, using the datasheet circuit for the buck regulator with no external pass transistor. I think you have a parasitic oscillation type problem with that weird coil of wire in circuit!

You can even ignore the voltage regulation and set the current sense resistor value to make it act as a constant current buck driver, perfect for LEDs. So set the current sense resistor to give 0.3v at 280mA (R = V/I = 1.07 ohms).

 

PDIP Package dissipation is 1.25 Watts.
Temp increases 100ºC/W after that without heatsink, since wiring is point to point, not much heat is removed by the suggested copper layer size on the board.

Even running 2W through it, let alone sourcing 4W, will cause the DIP to get hot very quickly.

The coil of wire in the sense is also an issue. Try laying the wire out in one big loop, or a bunch of lines run next to each other, going a few inches, going over a 1/4", back a few inches, and over a 1/4", then back down a few inches... glue down the ends. That will reduce size and inductance, though adding a touch of capacitance, until you can get a non-wirewound sense resistor.