I accidentally smoked the the 3v3 regulator on my uno and I am trying to find an adequate replacement part. The schematic points me to a part labeled (on my board): LM358-8AK-YD1118. a brief internet search reveals that the LM358 is a family of op-amps and my particular part is not listed for sale (that I can tell) on ebay or other parts distributors, though there are other model numbers available. I would imagine that this would be fairly common part to replace as I destroyed it in the first week or so of experimenting. Anybody have a link or know where to order a suitable replacement part?
You're looking at the wrong part.
On the R3 it's an LP2983-33DBVR
I think it's the same on the earlier models too.
They're pretty common. Most good suppliers will have them.
Actually, are you sure it's the regulator you smoked? It's more likely that the comparator is what you smoked by applying too much input power. It receives 50% of the input voltage and compares it against 3.3V to see if the input voltage is high enough to power the system. If that input goes too high it will smoke the op amp.
That component is a LMV358IDGKR - again a pretty standard component.
Are you identifying the component by the burn marks on the board, or by what you think is the component that has gone?
this is what i thought was bad. but i could most definitely be wrong.
All I know is that there is .003 volts coming out of the 3.3 volt pin. that and i definitely let the magic smoke out of something. no noticable scorching btw. How should I trouble shoot which part is bad?
That part is the comparator. The regulator is the little 5 pin device down and to the right.
Does the board function when run from USB, and do you get 3.3V OK from the 3.3V pin on the main power header?
Ok, that helps clear up my understanding of those two parts. Yes, the board functions on usb power as well as 12v external. i measured no voltage from the 3.3v pin on the power header. Measuring voltages at the regulator, I get 6.20v on the first and third pin on the side with three pins and 0.0v on the middle and opposite two pins.
having just looked at the schematic again it is for reals that the most likely place that let the magic smoke out was the LM2XXX and not that 8 pin thingie (LM358XXX) unless you could have hit a very small area with a lot of volts. The LM358's are pretty tolerant otherwise... kinda hard to kill but you might have blown the inputs... maybe but if the board turns on when plugged into into the USB connector W/o an external PSU then the reggie is shot, I think. Look in the upper left corner of the Arduino R.3 schematic pdf and find C5 it is the bypass for the LM358... notice that one end connects to 3 places. Pin 8, U5, C5 and +5V If the LM358 were shot you would more than not kill the 3V3 without killing the 5V source as it takes that through the USB port when disconnected from a PSU... most likely, I haven't tried it but I do know that "usually" those things just go away... the chip bond wires vaporize. Even that would be unlikely. When Vin goes high it causes the N Channel Mosfet to turn off through the bias established by RN1A & B (10K res. X 2) to switch the board from the USB port to the voltage regulator to prevent interaction between the two power supplies, the USB+5 or the +5V source. all is at this point conjecture. I just re read the schematic because I was curious about it and that logic makes sense to me. IMO
Doc
That does sound like the 3.3V regulator has been smoked. Pin 5 (top right) should have 3.3v coming out of it.
The datasheet for the part is available at http://www.ti.com/product/lp2983
Hmm looks like that will be a little trickier to solder than i would like :/. Thank you much. I greatly appreciate the information you share. this forum is quickly becoming my favorite site!
If the output of the regulator is open circuit (not shorted to ground), it might be easier to connect a different LDO regulator to the 3.3v pin on the header. Make it external to the board - at least for testing, to be sure that it's the regulator that's gone. Beats trying to get that puppy off without lifting tracks and ruining the board.
Putting a new one on in its place, if the pads are clean, is easy enough - I work with SO-2* chips all the time soldering by hand.
so would something like an lm117 work?
Overkill (it can supply 10x the current of the existing one), but yes - as long as you get the adjustment resistors right.
Looking again at the schematics, I don't think the 3.3v is used for anything other than the comparator that drives the mosfet that switches the power inputs. It's mainly there for the shields - the mosfet's comparator could be linked to a simple voltage divider on 5V.
With a soldering tip JUST wide enough to hit all three pins at once and not too much hotter than it takes to melt the solder you can heat the solder joints on one 'side' and flip that side up then heat the other side and 'flip' the 'chip' off the board. It's easy but a much practiced motion for me... So I don't know if you would want to try that. The other accepted method is to use a Very sharp hobby knife (Xacto knife) and very gently Cut the leads away from the body of that little bugger. You only need to do one side, an iron will work well to melt the solder and remove the other side next... "reasonably safely". in my experience.
Doc
Doing the flipping thing without practice you run the risk of lifting pads. I'd go with the cutting idea until you're more practiced 
one side of the comparator, Pin 2 the inverting input already does go to the 3V3 source, Pin 3 the non inverting input goes to a divider to Vin
Doc
Thank you for the advice. Think i will try to just snip the legs, because I desperately need a new tip for my soldering iron.
It takes a special type of flush cutter usually as the little buggers are really close to the body and close together too... but I am an old man and the eyesight is the third thing to go, comes right after memory. I won't go into the first one. That was why I suggested a knife.
Doc
one side of the comparator, Pin 2 the inverting input already does go to the 3V3 source, Pin 3 the non inverting input goes to a divider to Vin
I know. What I was musing over is whether the 3V3 is actually needed on the UNO at all, other than for shields, and they could include their own 3.3V regulator. I guess it keeps shield cost down, although for one shield I was instrumental in the design of we had to include our own 3V3 regulator, as it drew 68mA, and the dualanimallove takes its 3V3 from the FT232, and is consequently limited to 50mA.
Tip: although the UNO's regulator can provide 150mA, don't design for drawing more than 50mA if you want backwards compatibility.
The comparator needs a stable reference not tied to Vin (which will vary) and although even a string of silicon diodes (6 X .6 = 3.6V 'close' enough) might provide a ref, 'somewhat' stable the 3V3 source is also a supply for interfacing/powering external devices. It does have one major failing for interfacing and that is it cannot sink current... Tie a 4K7 resistor from 5V to 3V3 and measure 3V3... it will be about 4V7~. this is an issue where one is attempting to mix 5 and 3V3 devices as one cannot connect a 3V3 I2C device to the Arduino and use pull-ups to 3V3. The Arduino will pull the ports to 5V and 're-bias' the 3V3 source. (2 10K resistors tied to the 3V3 source are effectively in parallel thus the 4K7 (nearest value) ohm resistor).
Doc