Why is my arduino uno's voltage regulator frying?


#1

I am making an arduino project for a high school class that I teach and I can’t figure out why the voltage regulator on my uno keeps blowing.

I uploaded my diagram so you can take a look and see if there is an obvious error I am missing. It can be found here: http://fritzing.org/projects/arduino-bluetoothcar-with-obstacle-avoidance-senso

My current thoughts are the culprit could be:

  • I need two seperate grounds, as opposed to a single common ground, though I don’t think this is causing the issue
  • I was using an 11.1 V lipo, switching to a 7.4 could solve the issue
  • I am drawing too much current through the regulator, not sure what to do to fix this
  • I accidentally shorted something or made a mistake assembling my circuit, but I did have it functioning before the regulator melted.

Any assistance would be greatly appreciated.


#2

EDIT - just noticed the black wire GND on the left bottom is connected to power.

EDIT EDIT - That L298N is going to need some fixing, because pins haven’t been assigned in SCH and PCB.

EDIT EDIT EDIT - Make sure you delete traces in SCH before you change stuff in BB, or it will muck up the connections, ie they will persist.


#3

The wires at the bottom right were simply a mistake when I was drawing the circuit, I have them in the correct rails on my actual circuit. What do you mean about the L298N needing fixing?


#4

Your motor driver part appears to be broken, I’ll fix it in a while (the red square in schematic indicates there are connections missing, likely the motor driver part. That said, voltage regulators are pretty tough, and they are short circuit and thermal protected (and the one in the arduino only has about 30ma to spare, but as noted that should not blow the regulator, it should just shut down and not put out enough voltage. Now back emf from the motors getting in to the regulator could blow a regulator, but the motor driver should be suppressing that. Let me fix up the motor driver and clean up schematic so I can look at how this is connected and see if I see a problem.

edit: Is the breadboard layout accurate? Is there a connection to the 5V connector (unconnected in the breadboard) of the motor driver to the Arduino +5V pin? If that is there and the internal regulator jumper on the motor driver is connected you will have two regulators trying to regulate 5V which is one of the few things that can damage a regulator.

Peter


#5

The 5v pin of the L298N is not connected to anything.


#6

When you mention back emf, it got me thinking. One of the times that the regulator blew was when the wheel got stuck. The motor was running but not able to spin. Could this cause a damaging back emf?


#7

I have uploaded a slightly cleaned up schematic. Not sure how to fix the L298N part, this is literally my first time using fritz.


#8

The part in core parts is utterly broken. I’ll fix it up and submit a correction. That said, I don’t think that a motor stall should kill the regulator. The most likely cause would be connecting the 5V pin on the Lm298 to the Arduino 5V pin with the 5V jumper still connected on the motor driver. There is an internal 7805 in the L298N board and you probably want to use. that as I expect the lm298 may want more power than the arduino can supply. As I said if you connect the2 regulators to each other that is one of the few things that can blow one. Do you have to use the L298N? If not I would suggest one of the modern driver modules using MOSFETs as they don’t need the heat sink and are much more efficient (the L298 is said to eat 1.5V of your motor drive voltage, thus the heat sink as it would be dissipating 1 to 1.5Watts at 1A of motor current. Modern MOSFET H bridges dissipate next to nothing. They are available for cheap (but long delivery time) on Ebay or for more money from Sparkfun or Adafruit and the like.

Peter


#9

I can 100% confirm that the 5v input on the L298N is not connected to anything, i am powering it directly off the battery.


#10

OK, that’s what breadboard showed, but that is also the most likely regulator killer. This is odd in that it is usually hard to kill regulators.

Peter


#11

I’ve overloaded 7805s and they just shutdown, so I would assume the AMS1117 would be the same. Is it a Chinese one, because maybe they don’t have thermal shutdown.

I’ve had a shield GND through a USB, but that blew the 16u2 USB chip and one set of MEGA2560 RX TX pins.


#12

So it seems that the consensus is that this schematic in theory has no issues? I swapped to a smaller battery (7.4 V down from 11.1) but I have been afraid to put a new uno in and fry it. I guess i can give it a shot now.


#13

I remember reading “10 Ways to Destroy an Arduino” a couple of years back, maybe check some posts.

https://www.google.com.au/search?client=opera&q=10+ways+to+kill+arduino&sourceid=opera&ie=UTF-8&oe=UTF-8


#14

I wouldn’t say that. I just finished a corrected H driver part and substituted it in to your sketch. There is a short from power to ground and that looks to have corrupted the routing database. I’ll correct the whole thing and then look it over.

Peter


#15

Ok then. I was gonna try it out but I’ll wait and see if you find any issues. Fortunately I still need to finish the sketch so I can work on that in the meantime.


#16

Just remember to delete any traces to parts in other views before you make a change to a part connections in the view you are working on. We’ve had to fix a lot people’s sketches when more than one view thinks it’s correct.


#17

If you are working on the sketch, you will want this part in place of the current H bridge. It isn’t quite correct yet as I haven’t figured out which motor pins are ground (as they need to be bused) but that won’t hurt for the moment. Schematic on this one actually works.

H-Bridge with L298N-fixed.fzpz (14.7 KB)

Just download the fzpz file and load it in to fritzing via open and it will appear in your mine parts bin ready for use.

edit:

Fritzing is giving you a poor first impression :slight_smile: , I’m so far no further ahead on why your regulator(s ?) failed. How do you know the regulator has failed? With Vin present 5V doesn’t appear on the 5V pin? In any case, there are two additional problems (both Fritzing). The hc-05 part is also defective, it will not connect properly to the breadboard because its pin spacing is incorrect. Below is yet another replacement part for that. The second problem is a new (at least to us) bug found by someone a couple of weeks ago. You can not connect parts via wires through the breadboard (which I’m fairly sure is a Fritzing bug). Part to part connections work, but part to breadboard via a wire and then back out via a wire to a part does not work. That is why your sketch is having routing problems. In addition, the voltage divider on the HC–05 receive pin is wired wrong. If the resistors were not there, that would possibly explain the regulator burn out, as connecting a 5V pin to a 3.3.V pin clamped by another voltage regulator (inside the HC-05 board) could cause burnout, but likely not with a 1K or 2K resistor in the way. I’ll continue poking at the sketch tomorrow now I have an HC-05 part that can be made to work on the breadboard (with it not being able to connect to breadboard it needed wire to breadboard to other part which won’t work).

Edit:

An updated version of this part (the first one’s schematic is wrong!). If you have downloaded the old one please replace iit with this one.

Bluetooth HC-05 Macho-fixed.fzpz (41.6 KB)

I will also submit this one as a fix to core parts.

Peter


#18

This must be a record of some kind, you have tripper across 4 bugs in a single sketch. I finally late last night figured out what was wrong. As well as the ones already listed the LIPO batteries are configured incorrectly and won’t connect to the breadboard without a wire (which of course due to the previous bug is fatal!) So here is a new copy of the battery as well as the others and a sketch which is complete and I think correct (I corrected the level shifting resistor voltage divider which is incorrect in your original). With that I don’t see any issues that should burn out voltage regulators, and motor back emf is unlikely as that should be suppressed by the L298D. So the cause of the burn outs is still a mystery.

A new LIPO batery that connects to breadboard without needing a wire:

LIPO-2000mAh-fixed.fzpz (6.6 KB)

and the fixed up sketch (other than pcb which I ignored), it has the new parts in its temp parts bin so you don’t need to download them. You can either export them to save them outside of Fritzing (which is what I do in case of database corruption) or just drag them to the mine parts bin to reuse them in other projects.

fritz_arducar-fixed.fzz (87.9 KB)

Peter


#19

Thank you Peter! I will incorporate the changes you made to the schematic and let you know what happens. I was considering a slight modification in which I powered the uno directly from the 5V output of the L298N. This would limit the voltage to the arduino and because the L298N has a large heatsink, I believe it will hold up just fine. Do you think this would work? I can’t think of anything that would prohibit it working, in fact I found a few sources online that suggest using this method to power the arduino, though they don’t explain why.