1
\$\begingroup\$

I have an audio processor board that has a 5V -> 3.3V LDO. For some very strange reason, when I measure the voltage of the output of the LDO, I read 5 instead of 3.3V -- which by the way is a lethal voltage to the chip that's on the board, but for some reason, it doesn't seem to mind it.

So here's some background information, this is the board: enter image description here

And this is the full schematic:

https://drive.google.com/file/d/1OV1CHaihA_eIsqDjlkCkUZPVncHjFkQO/view?usp=sharing

It's a very simple board, just an audio DSP chip and an eeprom, designed for quick prototypes and DIY solutions.

The DSP chip is ADAU1452. For it's "IOVDD" rail, it asks for 3.3V, and absolutely no more than 3.6V. The way this board achieves this voltage is using a typical 1A linear regulator.

The regulator that came shipped with this board was AMS1117, apparently, a very popular choice for Chinese made designs.

When I power the board using a 5V supply, it seems to work just fine. Here's the problem: when I measure the voltage between output pin of the LDO and the ground, I read 5V. Input is 5.1V, output is 5V. How can this be?

The multimeter I am using is Fluke 115, a reputable name brand multimeter, so I know it cannot be lying to me. Why would I possibly be measuring 5V instead of 3.3V like I should?

As a step to debug, I took the original LDO out, soldered jumpers to its pins, and supplied it a 5V input out of the circuit, without any caps. And lo and behold, I read on my multimeter that it was outputting the correct 3.3V! As soon as I put it back into circuit, I read 5 V again...

I thought maybe there is some problem with stability related to input or output caps, so I swapped the part with a brand new pin campatible LDO ADP3338. And again, when I measure, I read 5V at the output.


What am I not understanding here? When I plug my multimeter in this board by connecting the ground probe to the ground lead of the LDO and positive probe to the output lead, why shouldn't I read 3.3V? Is this a measurement fault, or is there some other components in the circuit that can change the voltage of the LDO output? When I look at the schematic, I can't pinpoint a possible cause for the output voltage of the LDO to climb back to 5V. I simply don't understand what's going on.


By the way, ignore that weird, backwards looking Q1. The board came shipped like that. Somebody must have made a blunder when designing the board and later realized oops, gotta put the part backwards. That transistor doesn't seem to be the culprit though.

\$\endgroup\$
  • \$\begingroup\$ What voltage do you measure on the regulator's ground pin, vs a known ground on the board? \$\endgroup\$ – Chris Stratton Oct 10 '18 at 22:23
  • \$\begingroup\$ I read 0 volts. \$\endgroup\$ – darksky Oct 10 '18 at 22:25
  • 2
    \$\begingroup\$ Is there a path that bypasses the regulator? That might have been easier to check while it was out of the circuit. \$\endgroup\$ – Chris Stratton Oct 10 '18 at 22:27
  • 2
    \$\begingroup\$ I suspect a path around the chip. With the part out, try plugging the board in and measuring that output voltage -- see if it's 0 or 5V. \$\endgroup\$ – TimWescott Oct 10 '18 at 22:41
  • 1
    \$\begingroup\$ @TimWescott you were right! With the LDO out, the board turned on as if nothing happened. Measured the output pad of LDO and read 5V! \$\endgroup\$ – darksky Oct 10 '18 at 22:46
0
\$\begingroup\$

The linked schematics does not seem to be exactly right one for the board.

There are no R25, R26 on it. Since both are just jumpers, and R26 is located right next to LDO and R25 is on connector pin 13 marked "VDD5V" on schematics, I suspect they (or at least one of them) are used to bypass LDO.

\$\endgroup\$
  • \$\begingroup\$ Removed R25 and R26. The board still turns on without the LDO :/ \$\endgroup\$ – darksky Oct 11 '18 at 2:49
  • \$\begingroup\$ Found the problem jumper! R24! Removed and there is no longer a short between vin and VDDIO. This answer put me on the right track. Thank you. \$\endgroup\$ – darksky Oct 11 '18 at 3:32
  • \$\begingroup\$ It turns out R24 and R25 were settings. R24 bypasses the LDO in case Vin is already 3.3, while R25 connects Vin to the LDO. sigh It always sucks when datasheets are outdated. \$\endgroup\$ – darksky Oct 11 '18 at 4:12

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.