2
\$\begingroup\$

I recently acquired some 'ESP8266 SMD Adapter Boards' from Electrodragon. Up till now, they've been working fine. Except... I think I just broke one by attaching it to a 12v power supply. The power supply says it's a 12v 2A DC supply, which would seem to play nice with this (from that linked product page):

On board LDO for 3V3 power, 0.5-1A output (NOT for serial port and other IO pins), input voltage range 5-12V

I have a water sensor and a buzzer soldered to the ESP board, and the requisite GPIO16-RST connection so that sleeping the chip works. I also got some simple DC connectors to take a standard 5.1mm plug and use it as a power supply. Here's a picture of the overall setup.

This has been working perfectly when powered through my FTDI cable, but it's tedious holding the power/ground pins manually (this build doesn't have headers on it, the plan was for this to be a stepping stone to 'mass production' (which is seven units, but still).

I tested the 12v power supply, my multimeter said it was outputting 12.18v. I plugged it into the system, and... poof, magic smoke and a red-hot pin. It came from the little black chip at the bottom of the board:

enter image description here

The middle pin on the near side of the chip is where the smoke was apparently coming from. I unplugged the connection, but the board wouldn't power on again.

The chip that was smoking is imprinted, in barely legible type, "352". It might be "L352". Assuming this is a voltage regulator, is it rated for 12v? Was I sold a falsely-marketed product? I can't find a spec sheet for it, and before I ruin another one I'd like to know if these are 12v tolerant.

Or, is there something wrong with my circuit/soldering? I'm brand new at this, so it's messy, but here are some additional pictures: bottom of the ESP, side view showing some connections better, water sensor module. Don't hesitate to ask if there's any additional information I can provide!


Edit: I've just hooked a fresh board up to 12v power, and watched it a little closer. I seemed to confirm that the voltage regulator couldn't dissipate heat:

enter image description here

It desoldered itself. Impressive.

I'm going to get some 5v power supplies.

\$\endgroup\$
  • \$\begingroup\$ I looked up a few devices, such as "ML62352". They have an absolute maximum rating of 12V, but they dont seem to come in the package of the photo so, not sure. \$\endgroup\$ – Wesley Lee Apr 13 '16 at 4:24
  • 1
    \$\begingroup\$ Who knows what the chip actually is, if the supplier won't tell you? Quite a few (popular in China) devices in that package are rated at 6V maximum, and some are rated at 12V absolute max. I would not put either on 12V. Put a 78M05 ahead of it and be safe, IMHO. This has nothing do with the ESP8266, and everything to do with whoever is supplying 'Electrodragon' this adapter board thing. \$\endgroup\$ – Spehro Pefhany Apr 13 '16 at 4:59
  • \$\begingroup\$ At 12V, and a 400mA draw for the ESP peak, that's 3.6 Watts wasted by the linear regulator. That has no heatsink and a plainly undersized landing pad. You may have just cooked it. IF it's not a fake IC. \$\endgroup\$ – Passerby Apr 13 '16 at 5:29
  • 2
    \$\begingroup\$ The 7805 in a TO-220 through hole package, can only reasonably take 1~1.5 Watts without a heatsink added. At 12 - 5 = 7 * 0.4A = 2.8W your still looking at thermal issues. If you keep the 3.3v linear regulator, add a 12v to 5V switching regulator in front. Like any common car usb adapter. Dollar store/ebay/walgreens. Then the 3.3v linear regulator is only wasting 1.7V * 0.4 = 0.68 Watts. \$\endgroup\$ – Passerby Apr 13 '16 at 5:47
  • 1
    \$\begingroup\$ Or just start with a 5V supply. Home usb chargers are everywhere. \$\endgroup\$ – Passerby Apr 13 '16 at 5:56
5
\$\begingroup\$

There are three reasons this can happen.

  1. Fake IC. Either not actually a regulator, or it's specs do not match what you were told.

  2. Damaged IC. Static electricity, Reverse Voltage, or some other reason. Not always visible damage.

  3. Heat. Heat kills. Under-designed applications may not allow full use of a part's specs.

In this case, it seems to be reason 3. A Sot-23 IC with little to no copper pour or external heatsink to keep it within it's allowable junction temperature. If you don't take this into account, then it will fry itself. Some ICs may have over-temperature lockout circuitry, but don't count on it.

Here, the issue is highly susceptible to Voltage and Current concerns, re: Power/Wattage. As a Linear Regulator, the input voltage * input current is wasted in heat. So while it could do higher voltage at a low current, or a high current at a lower voltage, it can't do both.

The application, a ESP8266, will draw up to 400 mA at 3.3V. With a 12V input, that's (12V - 3.3V) * 0.4A = 3.48 Watts. Too much for no heatsinking.

This application really only allows for 5V input. (5V - 3.3V) * 0.4A = 0.68 Watts. Much more manageable.

As OP has shown, 3.48 Watts was enough to UNSOLDER the part from the board. Reverse reflow soldering ha. The first board had physical damage instead.

The Regulator itself would most likely actually be 12V compatible, with careful selection of input voltage and current requirements. If you hold the ESP8266 in reset, and power something small, like a 20mA LED, you will see it working just fine from 12V.

\$\endgroup\$
3
\$\begingroup\$

This appears to be designed to be used with Li-Ion battery + charging circuit

Dissipation with fixed 3.3V output: Pd = (Vin-Vout)*Load

| Load  | 3.9 Vin | 5 Vin  | 12 Vin |
|-------|---------|--------|--------|
| 0.1mA | 0.06 mW | 0.17mW | 0.87mW |
|  50mA |   30 mW |  85 mW | 435 mW |
| 150mA |   90 mW | 255 mW | 1305mW |
| 500mA |  300 mW | 850 mW | 4350mW |

The SOT23-5 with a minimum footprint has typically a Junction to Ambient Thermal resistance ranging between 220ºC/W to 260ºC/W

Calculating the maximum Power the package can dissipate:

Pd(max) = (Tj_max - Tamb_max)/ ThermalResistance_ja

Tj_max: 125ºC

Tamb_max: 25º --> Pd(max) = 450 mW to 380 mW

Tamb_max: 50º --> Pd(max) = 340 mW to 290 mW

Running at 12V should be possible, but your average consumption should be kept below 40mA.

Even with 5V input you are pushing the regulator to its boundaries if indeed your application consumes more than 200mA on average.

The MIC5219-3.3YM5-TR is the closest regulator I could find with the package, SOT23-5, and the claimed specifications.

\$\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.