Arduino micro voltage regulation issue

[damienz]

Hi everybody,

Yesterday, I tried to monitor some voltages on a HD44780 display. I started by measuring the voltage of the onboard 3.3v regulator, but the value was too high. I checked with my voltmeter : 3.3v was good, but I got only 4.6v on the +5v pin, which explained my false reading (as I used the default reference for analog inputs). Then I decided to power the arduino from an external 12v regulated power supply (through the vin pin). This time I got a value too low for the 3.3v so I also checked the +5v pin, and this time I got about 5.5v. I disconnected the power supply, checked it, checked my voltmeter but both where good.

I decided to test a second time. I connected the 12v power. The HD44780 display became nearly black, I put a finger on the atmega and it was very hot (I now have a little square mark on my thumb...). I disconnected the power just after. I want to precise that I do not have modified any wiring on the breadboard between the tests.

The arduino is still working after that when powered by the USB port (and I have a good reading of the 3.3v now and also a real 5v on the +5v pin). I tried to power the arduino through vin againt but with an adjustable DC/DC converter this time. I slowly rised the voltage from 5v to 5v5 on the converter, and there is no more voltage regulation on the arduino. I get the same voltage on the +5v pin as the one on vin pin. I'm suspecting the 5v voltage selection circuitry to be dead, but I dont think I have done something bad with the board.

What do you think of my problem ?

Edit: I think I can explain the 5.5 volt during the second experiment. The arduino must have been still connected to the USB port of my computer. I tested again with my adjustable converter. Got 5.6 volts on +5v pin with USB connected and 6v without it. This must explain why I didn't burn everything on the second test, as the computer "limited" the voltage...

[damienz]

My problem looks very much like this one : http://forums.adafruit.com/viewtopic.php?f=25&t=40424

I buyed the arduino from Adafruit on 21 December 2012. Is it still under warranty ?

[adafruit_support_bill]

It does sound like a similar problem. Unfortunately it is beyond the warranty period.

Since you have an adjustable supply, there is one more thing to try: Set your supply to 5v, and try powering the Micro directly from the 5v pin.

[damienz]

The board is working correctly when powered by the +5v pin.

When powered this way, I measure 5v on the vin pin and also on the power pin of the USB connector. Is it normal ?

[adafruit_support_bill]

Sounds about right. That pretty much confirms that the problem is with the 5v regulator (the big component on the bottom of the board.) These are not too hard to replace. Although they are surface mount, the leads are pretty big and you don't need any specialized soldering equipment.

[damienz]

I'm sorry but that look strange to me. I'm not an expert so I may be wrong, but considering the MOSFET T1 on the schematic (http://arduino.cc/en/uploads/Main/ardui ... ematic.pdf), if vin and +5v are both at 5v (as I measured), T1 must be blocked and there should be no voltage on VUSB. VBUS pin on the atmega is an input and there is no other connection explaining the +5v potential on the USB connector.

If I'm not wrong, the same applies to T2, and there should be no voltage on vin pin.

I will try to unsolder the voltage regulator tomorrow in order to check it, and maybe the two MOSFET.

I would really like to understand what appened as I'm certain to not have done something wrong with the board.

[adafruit_support_bill]

I haven't analyzed the Micro, but I believe the power circuits are basically the same design as the Uno. On that board you do get 5v on VIN when powering via the 5v pin. You won't be able to pull much current from it, but the 5v is there.

As for what went wrong, you mentioned powering it from a 12v supply. These boards use linear voltage regulators. That means that when it is powered from 12v, 60% of the power has to be burned off as heat by the regulator. 12v is allowable under the board specifications, but for a development board, I like to leave a larger safety factor. 12v is fine for the board alone sitting on the bench. But when you add more circuitry or put it into an enclosure, you will start running into heat dissipation problems. And in the end, it is usually the heat that kills components.

[damienz]

I unsoldered the voltage regulator, and the two mosfet T1 and T2.

The voltage regulator is good. I have a steady 5.05v if I power it from anything higher than about 7.5v.

I have soldered the mosfets on a perfboard and added pins to put them easyly on a breadboard. I have connected the source to +3.3v, the drain to a led + resistor :
- T2 is working correctly : the led is off when the gate is at +5v and on when 0v
- T1 is dead. The led is always on, but more, if I put +5v on the gate, i measure 5v on the drain... I think this explain why the voltage is no more regulated on the arduino as vin seems to be connected to +5v by T1...

[damienz]

I haven't analyzed the Micro, but I believe the power circuits are basically the same design as the Uno.

Not really. There is an operational amplifier in the Uno and none in the Micro.

As for what went wrong, you mentioned powering it from a 12v supply. These boards use linear voltage regulators. That means that when it is powered from 12v, 60% of the power has to be burned off as heat by the regulator. 12v is allowable under the board specifications, but for a development board, I like to leave a larger safety factor. 12v is fine for the board alone sitting on the bench. But when you add more circuitry or put it into an enclosure, you will start running into heat dissipation problems. And in the end, it is usually the heat that kills components.

I agree but in my case there was not a big load on the regulator, just an HD44780 display. Also, the regulator used has a thermal shutdown protection. Is it not enough to limit the heat dissipation to an acceptable level ?

Also, if 12v is not safe, I think a lower limit should be written in the description of the product, as 12v is in the recommended range (maximum but still inside).

[adafruit_support_bill]

Thermal shutdown helps avoid instant catastrophic failures, but prolonged operation at high levels does accelerate aging of the device. Although in this case it seems that it was not the regulator that failed. It would be interesting to know what killed T1.

The voltage specs are set by Arduino, we don't manufacture these ourselves. My personal preference is to stay well away from maximum ratings of all kinds, but that is a conservative habit developed from years of work on transportation systems and medical devices.

Although your Micro is well past the warranty period, we do appreciate your efforts to diagnose the problem. If you contact [email protected] with a link to this thread we can offer you a discount on a new Micro.