1
\$\begingroup\$

I have some trouble with my LM2576-5.0WT. I have an input source of 12 V (8 A max), and I want to convert it to 5 V (3 A maximum.)

Hence I built my circuit as the one on the first page of this datasheet..

On my PCB (with other things) I have trouble with the 5 V. It changes with the load, and it's 10 V with no load (at the beginning I got 5 V, but not anymore).) My PCB is more complex but the schematics of that part is like this:

Main PCB schematic (only DC-DC part) Main PCB layout (The green is a ground plane)

Hence, I made a simple circuit (same as in datasheet, so without the delay (Cd1 and Rd1)) with breadboard. I got 5 V without load. But when I put a motor as a load, to test it, the voltage decreased to 4.1 V. Why is the voltage not stable? How can I have always 5 V? I need to supply a RPi, so I need 5V - 3 A. You can find below my test circuit (same as in the datasheet).

Schematics Test circuit

NB: list of component (the codes are from RS, sorry I cannot put more than 8 links)

  • LM2576-5.0WT: 9101966
  • Cin 100 µF: 7111731
  • D1 - 1N5822: 6870877
  • L1 - 100 µH: 1048443
  • Cout - 1000 µF: 7111312
\$\endgroup\$
11
  • 2
    \$\begingroup\$ There is not enough information to answer the question. Which parts you used to build the circuit, and how you built it? Do you have a schematic, PCB layout, or list of used components? \$\endgroup\$
    – Justme
    Commented Oct 15, 2020 at 7:24
  • \$\begingroup\$ Did you measure your input voltage too when observing that behavior (voltage decreased to 4.1V)? Like Justme said we require more info to help you. \$\endgroup\$ Commented Oct 15, 2020 at 7:37
  • 3
    \$\begingroup\$ Wireless breadboards do not work well for power circuits. The resistance of the connections can be 1/2 ohm. \$\endgroup\$
    – Mattman944
    Commented Oct 15, 2020 at 8:12
  • 3
    \$\begingroup\$ Pro tip: get a module with LM2576 (or any other suitable chip). The PCB layout of DCDC converters is quite critical as current loops have to be small. Also, as ready-made modules (from China) are cheap and also just work it really is a waste of time to make your own DCDC circuit. I could do it (have done so in the past) but don't as the modules are cheap and just work. \$\endgroup\$ Commented Oct 15, 2020 at 9:17
  • 2
    \$\begingroup\$ 2nd Pro Tip: for any circuit where more than about 100 mA is flowing: a solderless breadboard is completely useless. The resistance of the contacts and wires is too high and that usually prevents the circuit from working properly (if working at all). If a small part is using high currents, you could use soldered connections only for that part. \$\endgroup\$ Commented Oct 15, 2020 at 9:20

2 Answers 2

2
\$\begingroup\$

Wire lengths + breadboard!

If you look at the datasheet, chapter 8.2.1, they show that the input and output caps and the diode should be connected at a single point close to the GND pin of the IC.

Looking at your picture, you are using a GND rail, as well as "long" (non negligible inductance) GND wires everywhere.

I strongly suggest that you try to rewire your design on the breadboard trying to minimize as much as possible the usage of wire.

Best would be to solder everything together instead :

  • dead bug soldering (google it! it's artistic...)
  • use a perf board (this is my personal preference)
\$\endgroup\$
1
  • \$\begingroup\$ Oh yeah, I forgot that point, I'll correct it! On my PCB, I'm using a ground plane for the connection. Is it ok, or could it be a problem? \$\endgroup\$ Commented Oct 15, 2020 at 8:38
1
\$\begingroup\$

While breadboarding is not ideal, I'm not so sure that's the problem.

Look at page 5 of that datasheet : notice the conditions for the output voltage specification:
0.5 A ≤ ILOAD ≤ 3 A

It's fairly common for switching supplies to require some minimum load before they can regulate properly. Otherwise the output voltage drifts up out of specification.

I'm not suggesting 0.5A is the required minimum load : 10 or 50mA is probably quite sufficient if you can tolerate a few mV extra.

Page 7 : "Output leakage current" 2mA min, 7.5 mA to 30mA max may be hinting that the minimum load current should be in this range.

But test your circuit to see how it performs with a few minimum loads : then tidy up the wiring as the other answer suggests, and test again to see if there is any improvement.

\$\endgroup\$

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.