1
\$\begingroup\$

I am building a DC-DC buck converting circuit from first principle with no on chip solution. Output Voltage must be 5vdc with a max current @500mA. input is 20+-4vDc. The problem that I am having is that my output voltage is 7.5vdc once bread boarded, but a 5vdc is required. I can solve this by increasing pull up resistance to the TIP122G but there is still a problem of the 5vdc output not being constant and dropping when a load is added. I think this my be due to my feedback circuit but I'm not sure as to what the problem may be and what I should change in it. Saw tooth generator has a frequency of 50kHz. This is a outline of my circuit on Multisim

\$\endgroup\$
  • 2
    \$\begingroup\$ Could you show a photo of the construction? Breadboarding a switcher is not easy. I need more time to study this but at first glance: your ICs don't seem to be decoupled, and 1N4007 is too slow for this application. \$\endgroup\$ – user133493 Apr 29 '17 at 0:23
  • 2
    \$\begingroup\$ Building a discrete differential amplifier, with an ancient 741 at that, is going to be anything but accurate, you are at the grim mercy of component tolerances. Note that the LM393 has an open collector output. \$\endgroup\$ – user133493 Apr 29 '17 at 0:30
  • 1
    \$\begingroup\$ Try evaluating at 100mA and at 500mA Iload. If ZERO LOAD, the switcher needs to be on for ZERO time. \$\endgroup\$ – analogsystemsrf Apr 29 '17 at 3:52
  • 1
    \$\begingroup\$ A few quick observations... you have a maximum of 12.5mA to power three ICs, your 741 is going to have trouble with inputs that close to its negative supply and if you're going to drop about 12V across your "switching" transistor you might as well drop the other 3V as well and make it a linear regulator! \$\endgroup\$ – Finbarr Apr 29 '17 at 8:27
2
\$\begingroup\$

The main problem is the 741 op-amp - you appear to be powering it from a 7.5 volt zener diode yet, if you read the 741 data sheet the implied minimum supply voltage is 10 volts. See table 7.2 in the uA741 data sheet. Try using a modern device and not this crappy dinosaur. Think "rail-to-rail".

Also, I don't see any decouplers on the op-amp power rail.

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$
  • Q5's base is driven from 9.1V through high valued resistors, therefore it cannot saturate properly.
  • Switching will also be extremely slow due to the 10k base drive. There is no way this schematic will work.
  • 1N4007 is way too slow
  • 741 should be replaced with something more modern
  • There is no overcurrent protection
  • C5 is way too low
  • etc

You should scrap the entire schematic and use a proper, modern, MOSFET-based DC-DC solution (or an integrated, ready-made one, which you can purchase for a few bucks these days).

| improve this answer | |
\$\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.