0
\$\begingroup\$

I have a test circuit connecting an ESP-8266 WiFi module to a MAX7219 LED driver. The setup is powered by a modified ATX power supply for both 5V (the MAX7219) and for 3.3V (the ESP-8266).

The setup worked fine for days with the first ESP-8266 unit I purchased. Photos of the setup are below.

I later purchased two additional ESP-8266, flashed them with the identical code, and substituted each of them into the circuit for testing. They each work fine for about 15 - 30 seconds, and then the whole ATX power supply shuts off.

The only time this power supply shuts down like this is when I have short circuited something. But I am not sure how different ESP-8266 modules could have a time-delay short?

I am really puzzled because I would think a short would happen right away, not after a few seconds. This behaviour happens with both of the new ESP-8266 modules I purchased, which makes me think manufacturing defect is unlikely.

While I did buy all ESP-8266 modules from the same e-bay vendor, I noticed that the last two (non-working; B & C in photo) are not identical to the first (working; A in photo). The circuit traces seem to be slightly different.

But all the spec sheets I found indicate that all of these ESP-8266 should be wired the same way. The only device differences I read about are with different amounts of on-board memory, but I am not sure how / why that could possibly lead to a hardware short circuit.

I am not really sure what is going on and would love any insight.

setup

front

back

As requested, I monitored current / voltage for each of the various ESP8266 devices. I do not have an oscilloscope to look for spikes - I used two multimeters and watched manually. Here is what I observed:

GOOD 3.35Vdc 20mA-60mA

BAD (#1) - repeated 3 times, same behaviour 3.32Vdc 70mA cut out ~20s

BAD (#2) - repeated 3 times, similar behaviour 3.32Vdc 74mA cut out ~60s, 30s, 10s

GOOD (repeated) 3.35Vdc 20mA-60mA (fluctuating; seems a lot "noisier" than the BAD chips)

\$\endgroup\$
4
  • \$\begingroup\$ can you monitor current through and voltage across the ESP device after powering on? and post it back \$\endgroup\$
    – User323693
    Feb 9, 2017 at 3:35
  • \$\begingroup\$ ATX supplies typically require a minimum load; what have you done to provide one? Also, these are grossly oversized for running an ESP8266, unless you are trying to power something else as well, you really should choose something more fitting. \$\endgroup\$
    – Chris Stratton
    Feb 9, 2017 at 3:36
  • \$\begingroup\$ Why do people insist on using computer power supplies as general-purpose power supplies? Computer power supplies are designed specifically to power computers, and require haywire work-arounds to be used otherwise. \$\endgroup\$
    – Peter Bennett
    Feb 9, 2017 at 4:23
  • \$\begingroup\$ @Umar - added measurements @ Chris Stratton @ Peter Bennett I understand the ATX power supply is overkill. I just use it here as a quick-and-dirty temporary test rig because it conveniently gives both 5V and 3.3V without having to setup a voltage regulator. I've used this power supply for years, so I don't recall what minimum load I provided (if any) - I seem to recall putting a power resistor inside. If it means anything, the power supply will happily run all day with nothing connected to my power outs. \$\endgroup\$
    – Roberto
    Feb 9, 2017 at 5:20

1 Answer 1

Reset to default
1
\$\begingroup\$

It certainly sounds like you have two fried ESP8266's. To add some more to Chris's and Peters comments above:

  1. If you don't have minimum loads on the ATX power supply your rails may be out of spec, and typically they will be high. By how much they are out of spec is an unknown, but it could easily be 10% or more. On the 3.3 V supply this would take you just outside the ESP8266 absolute maximum of 3.6 V.
    Most ATX power supplies have what is called cross regulation, so the +3.3, 5 and 12 V rails don't have independent regulation, it's ratiometric.

Here's a graph of the rail association for a typical 220 W ATX power supply:

enter image description here

Notice that there are minimum currents specified for the +3.3, 5 and 12 V supply. If you don't have these minimums in place then the supplies output voltage becomes questionable. If you have a more expensive supply it might gracefully shutdown, but most cheaper units do not.

  1. I don't see any problem using an ATX power supply for hobby/bench use, after all the other choices might be to have an adjustable 30 V, 5 A bench power supply which is just as ridiculously over powered for an ESP8266 as an ATX supply.
    If you are going to use an ATX supply it might be prudent to fit output current limiters so you can adjust current limits down to suit the devices you are powering.
\$\endgroup\$
1
  • \$\begingroup\$ thanks for the background - this makes sense. I guess I'll try yet another ESP8266 in case these two are fried. I'm a little confused about one thing: if they were fried, why would they function perfectly for several seconds (including correct display on the 7 segments)? Since you all seem to be suggesting the ATX might be causing issues here, for further testing I will work from my bench power supply and use a voltage regulator to split off 3.3V from a 5V output. Personally, I like the ATXs because they are (literally) free on the street and give useful output power! \$\endgroup\$
    – Roberto
    Feb 9, 2017 at 5:29

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.