1
\$\begingroup\$

I'm having an issue with this SMPS design where the voltage is correct at approx. 19.45V but I'm getting nowhere near the target of 3000mA

The zener voltage on D3 is 18V. R3 is a 1K resistor.

When I hook up a 650mA load the voltage drops to around ~ 3V. When I hook up a 300mA load the voltage sits at about ~12.5V. Without a load it is at the target voltage of 19.45V.

I would really like to figure out where I've gone wrong with this one :(

SMPS Schematic

NCP1200D60 Sheet: https://www.onsemi.com/pub/Collateral/NCP1200-D.PDF

FQP6N60 Sheet: https://www.mouser.com/datasheet/2/149/FQP6N60-189115.pdf

Wurth 750343068 Sheet: https://www.mouser.ca/datasheet/2/445/750343068-1483750.pdf

Update: I replaced the full bridge rectifier with a MUR460 diode as per The Photon's suggestion and it has improved the current issue considerably. I can reliably power a 700mA load, but I still cannot power a 2200mA load. It seems to spike every 3-5 seconds, powering the load and then dropping immediately. The MOSFET's heatsink get's pretty warm.

Update: I've designed a new PCB and done some more testing. Thank-you for everyone who responded, you all had helpful feedback.

I went on to run with this:

updated schematic

I then made a few more modifications, including upgrading the 47uF caps to 100uF caps.

At the end I'm still stuck with ~600mA but it is working perfectly, no MOFSET heating up, no issues at all.

Andy made a point that I disagreed with at first

Finally, the driver chip is not “man” enough to control a MOSFET that is capable of delivering anything like 60 watts to the output. You need to rethink this in my opinion.

But I now know he's right. The chip isn't, not without some help. I managed to dig up examples where the chip is used for 40-80W power supplies and seen that they almost always make use of the AUX winding on the transformer (I have one on mine that I'm not using)

See this random example (No idea where it came from):

random schematic 70W supply using

I have some more work to do at the drawing board to get my current levels, but I'm marking this question as answered, as all the major design issues have been addressed.

\$\endgroup\$
9
  • 1
    \$\begingroup\$ What gave you the idea to use a full bridge rectifier at D2? With the full bridge you no longer have a flyback design. And I don't know how well the current-mode control of the NCP1200 will deal with that. \$\endgroup\$
    – The Photon
    Commented Nov 9, 2019 at 3:02
  • \$\begingroup\$ I'm used to using the traditional (and gigantic/heavy) transformers. SMPS/flyback is a new concept for me and I'm not in my element with it. I did notice that other designs use a single diode and attention is paid to the polarity of the transformer. Could this be that simple? \$\endgroup\$
    – jonny5556
    Commented Nov 9, 2019 at 3:09
  • 1
    \$\begingroup\$ Like I said, I'm having trouble working out what effect the full bridge will have. I kinda think that it will prevent the PWM control working at all, leaving you with just the skip-mode control to adjust output voltage. But I would expect this to result in run away high voltages at the output rather than sagging output. But if you don't know exactly why you chose the full bridge, then I'd say try using the part as designed first and only try variations after you have it working in the way On expected. \$\endgroup\$
    – The Photon
    Commented Nov 9, 2019 at 3:13
  • \$\begingroup\$ It may be that with the full bridge, the overload protection circuit in the NCP1200 is being triggered, see pg 9 in the datasheet. \$\endgroup\$
    – The Photon
    Commented Nov 9, 2019 at 3:15
  • \$\begingroup\$ I'm excited to try this suggestion. It will be super easy modification to make. I have no clue why I went with a full bridge other than habit, and having several hundred of them. \$\endgroup\$
    – jonny5556
    Commented Nov 9, 2019 at 3:28

1 Answer 1

Reset to default
2
\$\begingroup\$

Using a bridge rectifier In the output instead of a fast diode is missing the point of what a flyback converter does and partially ruins performance. Then, most bridge rectifiers have awful reverse recovery time and will kill any decent performance when switching at high frequencies.

These are two massive issues that caused your problem. The next is lack of snubber network on your primary winding. Circuits not using a snubber can get away without one because they are low power designs less than a few watts but, you are wanting maybe 60 watts. No flyback snubber will likely harm the MOSFET or degrade it significantly.

Finally, the driver chip is not “man” enough to control a MOSFET that is capable of delivering anything like 60 watts to the output. You need to rethink this in my opinion.

\$\endgroup\$
3
  • \$\begingroup\$ I think you're probably right about everything except about how 'manly' the chip is. I seen the chip driving an 80W supply that I opened so I know it's capable. Since scrapping the bridge the circuit works beautifully at low current levels (700mA and lower) and I'm inclined to believe the problem is centered around the MOSFET & primary winding just as you say. I will add a snubber network and a few tweaks based on the datasheets. \$\endgroup\$
    – jonny5556
    Commented Nov 9, 2019 at 17:48
  • \$\begingroup\$ Thanks Andy, turns out you were right on all accounts :) \$\endgroup\$
    – jonny5556
    Commented Nov 25, 2019 at 19:31
  • \$\begingroup\$ A pleasure to help. \$\endgroup\$
    – Andy aka
    Commented Nov 25, 2019 at 21:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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