3
\$\begingroup\$

I've a board with an esp8266ex and a flash IC. The main power rail is +5 and it goes into a MIC5504 for 3.3 voltage regulation. The enable pin of the MIC5504 is tied up with a 10k resistor to the +5 power rail.

The problem is that I'm going literally mad with this voltage regulator, I've already replaced 4 of them, while some simply burned on power up, others either gave me a 4.4 volts in output (and I burned out the ICs) or gives me a 2.4 volts without any visual difference (I'm using the same pcb)...

If I remove the MIC5504 and power the board with a 3.3v direct voltage, everything works wonderfull and the current is about 0.02-0.06 amp , so there's no short around the PCB or excessive current draw (that by the way should shutdown the MIC5504 from its datasheet).

During one of the tests, I magically got a 3.4v output from the MIC and did some test. After this , I just noticed that I forgot to put 2 decoupling capacitors around my flash IC and indeed I was able to flash it , but the flash failed to work on a regular startup (maybe due to the missing capacitor). After soldering the capacitors back, bhooom, MIC5504 down to 2.2v . I cannot see any sense to this...

I cannot sort out this thing in a comprehensive way, I simply have no clue.. Do you have any hint ? Any expert that worked with this voltage regulator that I'm starting to hate ? Or do you think the problem is somewhere else ?

MIC5504

enter image description here

enter image description here enter image description here

enter image description here

\$\endgroup\$
  • 2
    \$\begingroup\$ Please edit your question and add: (a) a full, accurate schematic diagram of the board, and (b) one or more clear, in-focus photos showing the whole board, especially the power input connector and regulator area, with enough resolution that we can identify the components to match them with the schematic diagram. Thanks. \$\endgroup\$ – SamGibson Oct 24 '17 at 11:29
  • 2
    \$\begingroup\$ What are you using as an output capacitor on the regulator? The datasheet says it MUST be a (minimum) 1uF ceramic with low ESR. If you don't have that right, it will oscillate. That would cause the DC output to be incorrect, and probably cause it to burn out as well. \$\endgroup\$ – JRE Oct 24 '17 at 11:42
5
\$\begingroup\$

What you describe is typical behaviour for those LDO regulators (like the MIC5504) which require specific input & output capacitors, and those capacitors are either:

  • not fitted; or
  • are fitted, but they are too far from the regulator; or
  • are fitted, but they are the wrong type of capacitor.

Here is the relevant section from the MIC5504 datasheet, with some important points highlighted:


Input and output capacitor requirements from MIC5504 datasheet


With missing / incorrect capacitors, especially the output capacitor, the regulator can oscillate causing incorrect operation of the connected device(s) and an abnormal reading on a multimeter (if you used an oscilloscope to look at the output voltage, then the output oscillation becomes visible).

You got different behaviour when you were missing the decoupling capacitors for your flash memory, as they are also effectively across the regulator's output and therefore altered the problem.

I've asked you to add a schematic and photos into the question, in order to see your specific implementation of those requirements. If those updates lead to any further analysis, I'll add it here.

|improve this answer|||||
\$\endgroup\$
  • \$\begingroup\$ I'm using for both input and output capacitors, 1uF 25V X7R SMD 0603 capacitors, VISHAY VJ0603Y105KXXCW2BC , by the way I suspect the problem is there, you're right that the missing capacitors on the flash changed the capacitance of the circuit and this points the capacitors as the problem...I'll post a photo of the assembled regulator (capacitors are just near it) as well as the schematic, but I would like also to "probe" it to see if it's swinging .... \$\endgroup\$ – Leonardo Bernardini Oct 24 '17 at 12:21
  • 1
    \$\begingroup\$ I've updated the post with the picture of the PCB, ok indeed I probed the voltage and it's swinging as hell, also it seems higher than 2.4 , my oscilloscope reports it swinging between 3.5 and 4.5 ... Now the question is, are the capacitors that I put in the comments valid for this ? \$\endgroup\$ – Leonardo Bernardini Oct 24 '17 at 12:40
  • 1
    \$\begingroup\$ @Leonardo - Thanks. Your description of the 'scope trace confirms the regulator output is oscillating. However the schematic is still missing and the photo isn't clear enough for me to see all the traces :-( Although the capacitor part number, if that is the one fitted, should be adequate, I can't see the full path of the Gnd trace to the output capacitor - it seems long and thin (not good). How did you confirm that those capacitors are the part number given? Where did you buy them? In short: If the Gnd connection is good enough, then the behaviour fits with lower value capacitors. \$\endgroup\$ – SamGibson Oct 24 '17 at 13:00
  • 1
    \$\begingroup\$ @Leonardo - (a) "the capacitors are on the ground plane" That does not match what I think I can see on the blurry PCB photo. I can't advise further without a better quality photo. (b) Notice that I asked for a "full, accurate schematic diagram of the board", which is not what you have supplied :-( Therefore I cannot see the various output loads, nor can I see whether any other concerns are visible elsewhere on the PCB. So with the limited info supplied (small, blurry photo & minimal schematic), I can't provide any further analysis. I hope my suggested test causes a change in the output. \$\endgroup\$ – SamGibson Oct 24 '17 at 13:21
  • 1
    \$\begingroup\$ Hi Sam, I've add a live screenshot at an higher resolution , also I've add you a render of the gerber file as well as the relevant schematic, keep in mind that all the other components you see into the schematic are still no soldered so they should not play any role in the game.....what's on the board is the complete SOC with its resistors and capacitors, the flash with its decoupling capacitor, and the LDO with the damn 1uf capacitors...from the photo you can see also that I've add an extra capacitor on top of the other .it no longer swings but stays at 1.45v stable... :/ \$\endgroup\$ – Leonardo Bernardini Oct 25 '17 at 7:43
1
\$\begingroup\$

Well to anyone that may be so unlucky like me to face an absurd similar problem, I've passed a week with Sam Gibson doing tons of tests. Basically we never found the reason of the problem, and in the end , I swapped the MIC5504 with On Semiconductors NCP114 , as soon as I used the new IC, everything worked perfectly... My opinion is that I got a faulty furniture of MIC5504, or the same is too sensitive to swinging power sources maybe because, as stated in the datasheet, it's mostly suited for battery powered items.... Who knows!

|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.