0
\$\begingroup\$

I'm building an SMPS in flyback topology. In order to do it, I use UC3845CB IC. In first step, I wanted to see if it even works at all with pulse transformer that I bought, made especially for flyback pulse power supply. Below is the schematic for circuit that I've built. It is supplied from bench power supply with two outputs (+10V from output A and +20V from output B) connected via ground.

I connected all pins responsible for current and voltage control to ground in order to get simplest possible configuration.

enter image description here

When I turn only the +10V output, there is quite little current flow (0.01A).

But as soon as I turn on +20V, current on output A (not B!) rises dramatically (up to 3A), while current on output B is much lower (something like 0.3A). Strange thing is that IC does not heat up at all! And it is not burned either - when I turn off B output, I get 0.01A again on output A, and waveform on MOSFET gate is correct (50% PWM with 17kHz).

When I measure voltage on C2, it shows around 300V.

Also, when I turn ON only output B, there is no current flowing in the circuit.

Ratio of winding on the transformer is 56:14.

I'm totally confused, but again, I'm a total newbie regarding pulse power supplies. Anyone got a slightest idea what is going on?!

Edit: I changed the picture to better illustrate what is going on

\$\endgroup\$
  • \$\begingroup\$ Have you calculated the peak current in the primary of your coupled inductor to be sure you're not saturating it, given your switching frequency? With your feedback pin grounded you will be operating at max duty cycle meaning your secondary voltage can get very high maybe even breaking down your capacitor. \$\endgroup\$ – John D May 15 '18 at 15:38
  • \$\begingroup\$ your oscillator is not free running, so you need to complete the loop by sending back the signal from a secondary winding to the feedback pin. That is why the ecessive current. when there is no oscillator, the driver transistor (next stage) goes into saturation. I remembered back in the vacuum tube television days without the oscillator, the driver tube would get so hot, that it melted its glass envelope. \$\endgroup\$ – drtechno May 15 '18 at 16:25
  • \$\begingroup\$ @JohnD Ok, so what you are suggesting is that I'm saturating primary side of the transformer? By the way, max duty cycle is only 50%. Can you suggest me any article on calculating peak current on primary side and saturation of the primary inductor? I would be very grateful. \$\endgroup\$ – Em Ka May 15 '18 at 17:06
  • 2
    \$\begingroup\$ you have to have oscillator action first. Otherwise, the output will be "off" this circuit runs in a matter that it switches its current 100% on and off with the oscillator signal.. alsoI'd put a reverse bias diode either across the mosfet or the transformer because I'm thinking the datasheet drew the zener across the device just to illustrate its construction. So that probably doesn't have a real dampening diode. If you don't have the oscillator running the predriver inside the chip goes into 100% conduction and sometimes can short completely out. \$\endgroup\$ – drtechno May 15 '18 at 18:00
  • 1
    \$\begingroup\$ "when +20V output is off and +10V output is on, circuit should draw about 3A, right?" No, that is way too much current. It should draw much less, like maybe 0.01A. If the gate is damaged, when you add the 20V supply you could get conduction through the gate causing the extra current draw. Bottom line there's no way you should be seeing 3A on the supply to the IC. Post some scope shots of the gate, the drain of the FET, the clock oscillator and the voltage across the FET sense resistor and we should be able to tell you what's wrong. \$\endgroup\$ – John D May 15 '18 at 19:59
1
\$\begingroup\$

Here's a slightly shrunken version of your circuit with letters A, B and C in red on it: -

enter image description here

  • A - you HAVE (corrected, sorry) followed the correct dot notation for a flyback transformer. Good for you!
  • B - you have no load connected and you'll need a minimum load most likely
  • C - you have no mechanism to provide feedback control of output voltage to the chip such as this: -

enter image description here

I'm not ruling out that there are other problems but B and C are the glaring ones. I'm also unclear about what you are referring to when you mention letters A and B.

\$\endgroup\$
  • \$\begingroup\$ @Andyaka What I mean by letters A nad B are outputs of my bench power supply. What I found very strange is that On the power supply's output that provides supply for the IC, there is very high current (like 3A) while current going through the primary side of transformer is low (max 0.5A). How that can be explained? \$\endgroup\$ – Em Ka May 15 '18 at 17:08
  • \$\begingroup\$ @Andyaka Please also keep in mind that in my circuit, there is 56:14 ratio of the transformer (in your example, ratio is 4:50) \$\endgroup\$ – Em Ka May 15 '18 at 17:12
  • 1
    \$\begingroup\$ @EmKa that's irrelevant - you don't have output voltage feedback and the circuit I added does have it. That's the relevant bit, not the turns ratio. I have no idea about your power supply - please check it with a meter. \$\endgroup\$ – Andy aka May 15 '18 at 17:14
  • \$\begingroup\$ @Andyaka Power supply is OK, but what is strange is that it sends more current through the IC that through transformer as soon as I turn on both +10V and + 20V! \$\endgroup\$ – Em Ka May 15 '18 at 17:20
  • 2
    \$\begingroup\$ @EmKa without a flyback suppressor on the primary you may have unwiitingly destroyed the MOSFET with the knock-on effect that the UC3845 may be damaged. \$\endgroup\$ – Andy aka May 15 '18 at 17:23
0
\$\begingroup\$

Alright, I have found the mistake. As easy as it is... it was the lack of filtering at the input that was making indicators on lab benh powe supply going crazy. As soon as I add large 470uF capacitor on input, everything stabilized itself and now it is working fine:)

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.