Desperately need help! ESP8266 causing transient voltage dip

[shinew]

Hi I'm building an air quality monitor using shinyei ppd42 particulate sensor, DHT22 & ESP8266 to upload to thingspeak. I've had it setup using the ethernet shield running for few months without any issue, so I just want to swap out the ethernet to the ESP8266 wifi module.

My setup is as following for power supply:
9V DC(500mA) -> LD33V voltage regular -> ESP 8266
onboard VCC(have tried Uno & Pro Mini 5V) -> PPD42, DHT22 & low power 0.96" OLED display

The problem is that whenever the ESP8266 is connected, it creates ~80mV transient dip. See attachment for both DSO screenshot measured at the VCC input for the Shinyei sensor, one with and one without the ESP8266 connected. Everything is operational including OLED, DHT22 & ESP8266 are functioning fine, but when ESP8266 is connected, PPD42's reading is way off. According to the spec sheet, the PPD42 can handle 30mV of ripple. So my goal has been trying to get to that but due to my limited experience, I haven't been able to figure out how

I've multiples of everything, so I've tried swapping out sensors, esp8266 modules, changed to different power supplies(9v, 5v & variable bench PSU), swapped voltage regulator to LM317, adding tons of caps up to few thousands uF without success. Can someone please point me to the right direction to fix the issue? thanks alot!!

[Ice-Tea]

What kind off caps? Electrolytes won't kill HF. You'll need a(n assortment of) ceramic(s) for that. Might also be worth to look into beads. Also: be sure to measure right. With long leads you might be chasing ghosts...

[zapta]

The esp8266 requires high current for its size. Make sure your LDO can handle it.

[shinew]

The esp8266 requires high current for its size. Make sure your LDO can handle it.

LM317 should be able to handle it, but even if it's not, the problem is not esp8266, which is functioning normally as well as the rest of the circuit. It's the Shinyei PDD42, which is not powered by a regulator. I've even tried to power the whole thing with my bench power supply at 5V and have it connected directly to the PDD42, it behaves the same way.

[shinew]

What kind off caps? Electrolytes won't kill HF. You'll need a(n assortment of) ceramic(s) for that. Might also be worth to look into beads. Also: be sure to measure right. With long leads you might be chasing ghosts...

Thanks I'll try some film caps. I've tried few 0.1uF ceramic before and it didn't seem to help.

[sleemanj]

Where are you adding the caps?

[Ice-Tea]

Place them with short leads over a power and GND pin directly at the module. And you may need a few different values (per example 22uF, 4u7, 1u, 220n, 10n) to do a decent cleanup.

[shinew]

Where are you adding the caps?

5V -> caps -> Shinyei PPD42.

[diyaudio]

Hi I'm building an air quality monitor using shinyei ppd42 particulate sensor, DHT22 & ESP8266 to upload to thingspeak. I've had it setup using the ethernet shield running for few months without any issue, so I just want to swap out the ethernet to the ESP8266 wifi module.

My setup is as following for power supply:
9V DC(500mA) -> LD33V voltage regular -> ESP 8266
onboard VCC(have tried Uno & Pro Mini 5V) -> PPD42, DHT22 & low power 0.96" OLED display

The problem is that whenever the ESP8266 is connected, it creates ~80mV transient dip. See attachment for both DSO screenshot measured at the VCC input for the Shinyei sensor, one with and one without the ESP8266 connected. Everything is operational including OLED, DHT22 & ESP8266 are functioning fine, but when ESP8266 is connected, PPD42's reading is way off. According to the spec sheet, the PPD42 can handle 30mV of ripple. So my goal has been trying to get to that but due to my limited experience, I haven't been able to figure out how

I've multiples of everything, so I've tried swapping out sensors, esp8266 modules, changed to different power supplies(9v, 5v & variable bench PSU), swapped voltage regulator to LM317, adding tons of caps up to few thousands uF without success. Can someone please point me to the right direction to fix the issue? thanks alot!!

I just recovered last week from some serious issues, I ended up chasing ghosts  (for a week) with an ADC "ground problem" and discovering a "new" breadboard having a 0.002 volt drop as I moved my the meter probe every 60mm from the  bread board ground traces relative to the ground reference point. (This was partially fixed using double traces)

Another issue, if you using breadboard wires, make sure its good quality (low resistance)  another blow to my project was using new but crap ebay breadboard wires.. a trap for young players as dave would say.

I would always recommend doing wireless prototypes on a PCB (two layer and more) with a continuous ground plane. if you really cant do a PCB, use good interconnections especially for the power supply.
 


 
   

[shinew]

diyaudio, thanks for the advice. I also suspect it's some interference, but I just can't figure out the real issue, most likely because I'm not an expert at this...

So I've tried various values of caps from 100pF to 1000uF, none of them seem to do anything at all. Then I noticed a strange "fix", if I disconnect the power from regulator to esp8266 then reconnect it, or boot up the arduino, wait few seconds then connect it, the problem goes away without any filtering cap.

Basically I'm clueless of what the issue could be...

UPDATE: NVM, it wasn't a "fix", after 30seconds, once my code has detected the esp8266 not connected to the internet, it resets the wifi module and then the problem comes back...

[sleemanj]

You say you are putting the caps at the PPD42.

Try adding both a ceramic (100nF) and a larger bulk capacitance onto the ESP, across it's VCC and GND or as close to as practical.

You could also maybe try an inductor on the VCC to the ESP.

In short, try to arrest the problem at it's source, the ESP, rather than it's destination, the PPD.

PS: you're not running any of your wires to close to the ESP antenna maybe?

PPS: I know nothing of dealing with interference, just following my own logical course.

[shinew]

I've tried filter the ESP originally without success, that's why I though it might simpler to just filter the PPD which is the only component it's affecting. But I haven't tried placing caps directly on the module yet,  so I'll give that a try both on the PPD and ESP.

Btw I captured few zoomed in view of the drop, it's actually not that fast, only about 500us. Should I use larger cap to filter that? Thanks!

[JoeO]

Pictures.  We need pictures.

[macboy]

Your problem will be fixed by using separate regulators for the Wifi, and for the rest of the circuit. Make sure to return the ground for the Wifi back to the common star ground. That should be the negative leg of the bulk capacitor at the input of the regulator(s). You could try to isolate the Wifi module's power using series inductance or resistance plus a bunch of capacitance, but that will only reduce noise. In the end, no regulator has perfect load regulation, so when the Wifi module loads down the power rail, it will sag down by some amount. If you can't make the rest of the circuit work properly when that happens (which is something you should do I think), then you need separate power for it.

[Ice-Tea]

@matboy his scopeplots do not indicate the  voltage regulation would be an issur.  Sags 15mV or so where he can tolerate 30mV.

[mikerj]

Are you certain this is a power rail problem?  Could it not be radiated RF from the WiFi module causing the problems?