UC3842 basics

[rogfanther]

Re: UC3842 basics

One idea, if someone would please explain.

Suppose one has some transformers, like, one taken from a 5vsb ATX PSU, and from some small smps, say a 5V 2A one.

Knowing the inductance of the coils, and maybe the number of turns in them, could they be used with a 3842 in a circuit to create a different voltage, like 3.3 or 12V ? Just adjusting the values of duty cycle , frequency and the control circuit ?

Or is it that the transformer, in a 3842 circuit, or in others also, always need to be custom dimensioned to that circuit´s parameters ?

Thanks.

[Th3_uN1Qu3]

Re: UC3842 basics

Okay, no hand drawn version but i'm attaching this document. Take a look at the schematic in Figure 6. It shows a forward converter not a flyback but this is not relevant - the wiring of the opto is the main difference.

What they did here is wire the voltage feedback of the 3842 to ground, thus disabling the voltage feedback loop entirely, and using the TL431 as error amplifier. As you see, the collector of the opto drives the Comp pin (the output of the 3842's internal error amp), so it goes straight into the PWM comparator.

[Th3_uN1Qu3]

Re: UC3842 basics

I promise a crudely hand drawn version tomorrow.

[ben7]

Re: UC3842 basics

I have to get off my lazy ass and finish drawing the schematic i have gotten to, because it's pretty different from the one in the first post. I'll do a PCB too.

pft lol

[Th3_uN1Qu3]

Re: UC3842 basics

I have to get off my lazy ass and finish drawing the schematic i have gotten to, because it's pretty different from the one in the first post. I'll do a PCB too.

[budm]

Re: UC3842 basics

OK, since the original daigram as shown in post 1 is 150 Ohm, the only 100 Ohm resistor is in the DC output side. Or may be we are not talking about the circuit in post 1, see PDF. I am just confused.
I can see why there are misunderstanding.

[ben7]

Re: UC3842 basics

If you are getting 5V output for this 5V switching power supply, and if you are using 47 Ohm resistor to limit the LED current and it is burnig up with 5V supply, something does not add up, since 431 will maintain 2.5v, that means the voltage drop across the RLED and LED will be 2.5V. I do not see how you will burn it up.
Or you are designing 5v POWER SUPPLY?

There seems to be lots of misunderstanding, the smoking resistor is in series with a diode on the feedback of the smps transformer, it powers the UC3844 after the supply is started up via a 100k resistor to the +HV rail.

Yes, I am designing a 5v smps just for fun/learning

And I tried 100ohm as in the original schematic, but the smps just starts ticking... 47ohm makes it start up right away

[Th3_uN1Qu3]

Re: UC3842 basics

I think by "feedback winding" he refers to the resistor from the AUX, which powers the 3842. 47 ohms is too much current. It's 150 ohm 0.25W on mine, it does get hot but no smoke.

[budm]

Re: UC3842 basics

If you are getting 5V output for this 5V switching power supply, and if you are using 47 Ohm resistor to limit the LED current and it is burnig up with 5V supply, something does not add up, since 431 will maintain 2.5v, that means the voltage drop across the RLED and LED will be 2.5V. I do not see how you will burn it up.
Or you are designing 5v POWER SUPPLY?

[ben7]

Re: UC3842 basics

V^2/R =
25/47 = 0.53W

Not surprising it smokes.

lol

I measured it and the resistor only has 3.4v across it.
I changed the resistor for a 47 ohm 3w power resistor but it gets too hot to touch.

[tom66]

Re: UC3842 basics

V^2/R =
25/47 = 0.53W

Not surprising it smokes.

[ben7]

Re: UC3842 basics

Hi Ben, you need to bias the 431 properly and also the FB for your voltage output requirement.
See attachments and also here:

The example circuit in post1 has so many missing components and circuits to make it work properly, that 431 should have had 1mA bias for it.

Many thanks!

I got the power supply circuit to work, biased the 431 properly, and I get a steady 4.97v output

It does drive a 1A load, the output drops to 4.96v (not much drop). But when I connect the load (a lightbulb) there is a buzzing sound from the transformer, then as the lightbulb lights up, the buzzing/beeping goes away, so the circuit still must be a little unstable.

And the 100ohm resistor (I used 47 ohm) on the feedback keeps smoking, its a 1/2w.

[budm]

Re: UC3842 basics

Hi Ben, you need to bias the 431 properly and also the FB for your voltage output requirement.
See attachments and also here:

The example circuit in post1 has so many missing components and circuits to make it work properly, that 431 should have had 1mA bias for it.

[ben7]

Re: UC3842 basics

Mm. Opto and 431 are good? I've had bad 431s make all kinds of odd behavior. A flyback won't work with an output inductor in the feedback loop.

I changed the 100 ohm resistor on the feedback winding to a 47 ohm, the opto and 431, and now it does work, but it seems I cant set the output voltage by changing the voltage divider connected the 431's adj pin; it always stays at 4.27v +- 10mV. Adding a load doesn't really do much, and the transformer makes a slight hissing sound.

[Th3_uN1Qu3]

Re: UC3842 basics

Mm. Opto and 431 are good? I've had bad 431s make all kinds of odd behavior. A flyback won't work with an output inductor in the feedback loop.

[ben7]

Re: UC3842 basics

Little pulses on the output means the UC3842 isn't getting the right startup voltage, or it's in current limiting mode when it shouldn't be, due to high parasitics at the mosfet source triggering it. If Vcc of the 3842 is over 12 volts, try increasing the capacitor on pin 3 to 1nF.

The power for the UC3844 measures 12v, but when I tried a higher voltage external supply (20vdc) and it just shutup and didn't do anything. Removing the external power supply, it goes back to ticking...

And yes the 1nF cap didn't help either.

I tried using an inductor to filter the output but it still didn't work.

tried switching the connections to the feedback winding, but no change whatsoever

[Th3_uN1Qu3]

Re: UC3842 basics

Little pulses on the output means the UC3842 isn't getting the right startup voltage, or it's in current limiting mode when it shouldn't be, due to high parasitics at the mosfet source triggering it. If Vcc of the 3842 is over 12 volts, try increasing the capacitor on pin 3 to 1nF.

[ben7]

Re: UC3842 basics

I tried increasing the output voltage setting but it still doesn't work. Also adding a minimum load resistor changes the clicking sound to a beeping sound.

[ben7]

Re: UC3842 basics

Do you actually get anything on the output? Try setting the output progressively higher, until the supply fires up properly. The control method used by this circuit is different than the usual variable-frequency self-oscillating circuit used in a lot of 5vsb designs, so it may not actually go down to 5v stable at zero load.

Ok I will try that, I do get little pulses on the output (led on the output).

[Th3_uN1Qu3]

Re: UC3842 basics

Do you actually get anything on the output? Try setting the output progressively higher, until the supply fires up properly. The control method used by this circuit is different than the usual variable-frequency self-oscillating circuit used in a lot of 5vsb designs, so it may not actually go down to 5v stable at zero load.