EEVblog #831 - Power A Micro With No Power Pin!

[Zeph]

One of my first Arduino tests was connecting two Uno's via serial & ground, each powered by a different laptop via USB.  The first Uno sent simple commands to the second, which then turned the onboard LED on or off accordingly.  All worked as expected.

But when I happened to unplug the receiver first, I was rather surprised to see the LED continue responding to Serial commands, absent power.  That's when I learned about parasitic power.

In this case, the power was coming from one ATmega328's TX output, to the other ATMega328's RX input.  Serial commands very briefly interrupted the parasitic power (not long, a couple of 38400 chars every second), but apparently the bypass caps on the second Uno kept it running.  Not just the uC, but the D13 controlled LED, and even the "power" LED on the receiving Uno.  That uC can supply up to 40 mA on an output pin, and the uC can run on very little current and much less than the 5V normal supply, so it's not very hard for one uC to power another via connected I/O lines.

Voila - one pin power + signal, using standard Uno's and two jumper cables!

After learning why this worked, I've been surprised people don't run into this all the time, when connecting two uC's via I/O pins, and then removing power from one of them.

[apis]

I just love those PowerBricks from Diligent. Don't get me wrong, the video was great but that thing caught my eye!

I can use one of each, even the 5V ones, too bad they are pricey at $17 + Shipping.  Also, no 1.8V? 

Any quality cheap alternatives out there? (not to build but to purchase)

http://www.ebay.com/sch/i.html?_nkw=breadboard+power+supply

Those things supply two separate voltages to top/bottom rails of a breadboard. Voltages are selectable with jumpers. They have a switch and take either USB or barrel jack power.

The voltages are set by linear regulators on the PCB. Easy to change if you want other values.

They have no protection built in though. If you short one out for a second they are gone, not ideal for experimenting. If they were made properly they might have been useful.

[fubar.gr]

Most linear regs have built in short circuit protection

[apis]

I had one and managed to kill one of the regulators when probing it by accidentally shorting out two of the jumper pins, maybe I was just unlucky but mine was very fragile.

[westfw]

Most linear regs have built in short circuit protection

I don't know if they're counterfeit chips, or just especially cheap and inferior regulator design, but there have been a "significant" number of reports recently where the regulators on Arduinos and derivatives have failed in ways that I thought that the common 3-terminal regulators were just not supposed to.   "I had by Nano connected to a 12v wall wart (~16V, perhaps) drawing 100mA (1.1W dissipation in the poorly-heatsinked SO223 package, not really within spec) and it failed permanently."  or "It let Vin onto Vout and fried everything."
(other possibilities: substituting a regulator that has a different pinout without changing the PCB, or using one with different electrical requirements for the "tab".)
(The Arduino derivatives are so frequently operated via USB power that a vendor could ship a version with totally defective regulation for a long time, without anyone noticing.  Sigh.)

[Fungus]

Most linear regs have built in short circuit protection

I don't know if they're counterfeit chips, or just especially cheap and inferior regulator design

Is that even possible? There's no reason even the cheapest of clones of a voltage regulator can't use a proper design.

The "genuine" regulators haven't changed since the 1970s.

Well... I guess they could use really low-grade silicon or something.

[Ian.M]

There are a few nasty situations that can kill a linear reg that the cheap clone board designers don't bother protecting against.  e.g. if there is significant output capacitance, or an alternative source of power to the output, a reverse current protection diode is recommended, from Vout to Vin to shunt the reverse current round the regulator.  Otherwise if the input is shorted to ground the charge in the output cap can blow the pass transistor.   A series diode in the input may be needed as well (or may be sufficient on its own) if a very large output capacitor is used or an alternative supply can feed the output.

The LM317 is one of the few regulators that has an internal shunt reverse current protection diode. Ignoring this for regulators that don't have such protection is probably the commonest way in a bench environment of blowing a regulator so its shorts in to out.

Other ways for a regulator to let through excessive voltage is if its common pin goes O/C either due to external causes or just due to it being cooked to death with excessive dissipation and insufficient heatsinking.   

My experience is that persistently overheated 78xx series regulators normally fail with the pass transistor O/C, or high resistance. I haven't seen enough of any other individual type on my repair bench to notice their commonest failure mode.

[SeanB]

I have had a number fail with them oscillating slowly going into and out of thermal overload, probably from the die separating from the internal heat spreader. Some do it cleanly, others are very noisy in going into and out of thermal shutdown. Others just become very noisy on the output, not oscillating, but just a lot of noise, as even new capacitors does not fix it, but a new regulator does.

Fake ST parts with the thick tab, while the genuine part has the skinny copper saving tab. Even seen a few with tinned steel as the tab and leads.