1
\$\begingroup\$

I'm trying to provide a 3.3V source for my micro controller from a 5V power supply, however seem to be having issues with all the 3.3V regulators I've tried.

I'm currently trying to use an MT1117-3.3, datasheet here.

When I turn on my 5V supply, the output I get from the regulator is 4.1V and 4.5V if I remove the load.

I tried an L78L, which gives a similar result althoguh apparently this requires 5.3V to regulate.

I've also tried an LM2931A which also gives a similar result although it's input voltage requirement is less than 5v.

Would appreciate any suggestions on what might be the problem. Diagram below.

Test circuit diagram

\$\endgroup\$
8
  • 1
    \$\begingroup\$ If that is your load, it might be too low. Can you add at least 5mA of load (e.g. 470 ohms resistor) to check the output voltage? \$\endgroup\$
    – Justme
    Mar 14, 2021 at 13:18
  • \$\begingroup\$ I've increased the load by reducing to first 500 ohms, and then 333ohms, in both cases the resulting voltage was still around 4.1V. Current through the resistor is around 4.2mA and 6.3mA respectively. \$\endgroup\$
    – Geoff
    Mar 14, 2021 at 13:29
  • 1
    \$\begingroup\$ Your numbers don't add up. For example, 4.1/500 would result into 8.2mA of current, not 4.2mA. And 500*4.2 would be 2.1V, not 4.1V. Please describe your circuit in detail, and describe what are you measuring from where and how, because the measurements make no sense. \$\endgroup\$
    – Justme
    Mar 14, 2021 at 13:55
  • \$\begingroup\$ What does your 5V supply measure with the same voltmeter? \$\endgroup\$
    – Spehro Pefhany
    Mar 14, 2021 at 14:34
  • 1
    \$\begingroup\$ Output of LDO in some condition start oscillating. It may cause the high voltage you measured. Oscilloscope will help. Choose proper capacitor is important. Check datasheet. \$\endgroup\$
    – user263983
    Mar 14, 2021 at 14:43

3 Answers 3

Reset to default
1
\$\begingroup\$
  • The L78LXX requires 2 volts difference between input and output. 5V in to a 3.3V regulator is outside of the specifications. It's anybody's guess what it will do.
  • The MT1117A datasheet specifies a tantalum capacitor on the output. If you use anything else, there's a good chance it will oscillate. The output voltage may appear to be too high or too low when it oscillates - it depends on what your voltmeter does when hit with something it doesn't expect.
  • The LM2931 requires a 100µF capacitor to be stable - your 10µF won't be enough. On top of that, the LM2931 datasheet says that it requires that the capacitor have a specific equivalent series resistance (ESR) to be stable. If your capacitor is too small or if it doesn't meet the ESR requirements given in figure 34 on page 10, then it will oscillate - that causes the same wonky voltages readings that you would see if the MT1117 were oscillating.
  • Check that you have the regulator connected the right way around. It was really easy to connect regulators in TO-92 or TO-220 housings backwards. It's harder with SMD components, but it still happens.
  • Outside of those causes, you should check to see if your voltmeter is working correctly. Many voltmeters show incorrect voltages when the battery is weak.

Check things in this order:

  1. Check the battery in the voltmeter - or check if it reads a known voltage correctly.
  2. Make sure the regulator is wired correctly.
  3. Double check the output capacitor type and value.
  4. Hook the output to an oscilloscope and see if it is oscillating.
\$\endgroup\$
1
  • \$\begingroup\$ I've cross checked the voltages against multiple sources so fairly sure it's not the multimeter. Also checked the wiring. I used my arduino as a makeshift scope and got no oscillation, although it could still be at very high frequency. Sounds like the most likely cause is the caps, so I'll go over those. Thanks. \$\endgroup\$
    – Geoff
    Mar 15, 2021 at 7:47
0
\$\begingroup\$

According to the specification of datasheet you attached, the load current in your circuit seems to be too low. Recommended is a minimum load current of 5 mA. The load in your circuit is about 3mA or less. Try to decrease the resistor value of about 470ohms.

\$\endgroup\$
3
  • \$\begingroup\$ I've increased the load to above 5mA and still getting a voltage of around 4.1V out. \$\endgroup\$
    – Geoff
    Mar 14, 2021 at 13:41
  • 1
    \$\begingroup\$ Have you already doublechecked the correct pinning of the regulator in your circuit ? \$\endgroup\$
    – calenberger
    Mar 14, 2021 at 13:45
  • \$\begingroup\$ I've checked and rechecked the wiring - yes it can be confusing. \$\endgroup\$
    – Geoff
    Mar 15, 2021 at 7:44
0
\$\begingroup\$

If you want to provide a 3.3V you can just use a zener diode :

schematic

simulate this circuit – Schematic created using CircuitLab

with Vzbk < 5*RL/(R1+RL)

\$\endgroup\$
12
  • 2
    \$\begingroup\$ That'll give you 3.3V, but is it really a good solution? 3.3V Zener diodes aren't really precise. R1 will constantly waste a lot of power and it limits the current available to the load. A regulator is the correct solution. Geoff is trying to use a regulator, but there's something wrong with the circuit involved. Suggesting a worse solution isn't much of an improvement. \$\endgroup\$
    – JRE
    Mar 14, 2021 at 16:13
  • \$\begingroup\$ If the Rs is sufficiently high it is independent of input voltage. \$\endgroup\$
    – Miss Mulan
    Mar 14, 2021 at 16:30
  • 2
    \$\begingroup\$ 1. It "works" if you don't care if 3.3V is 3.3V or higher. 2. It "works" if you don't mind wasting energy all the time. 3. It "works" if you don't need high current. Rs of 500 ohms means you can get only a few milliamperes to the microcontroller. To get more current, you need a small Rs - which also wastes more power. \$\endgroup\$
    – JRE
    Mar 14, 2021 at 16:43
  • 2
    \$\begingroup\$ The 1N4733A is a 5.1V Zener diode. Maybe you meant the 1N4728A instead? \$\endgroup\$
    – JRE
    Mar 14, 2021 at 16:52
  • 2
    \$\begingroup\$ Crash course in the problems of Zener regulators. \$\endgroup\$
    – JRE
    Mar 14, 2021 at 17:00

Your Answer

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

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