Mcp23017 getting very hot within seconds

Hey there!

So I let someone create a PCB for me and it arrived this week.

I want to control up to 16 motors with SDA/SCL

So the PCB receives GND/3.3V/SDA/SCL and 2 Pins to control the direction of the motors.
I have one MCP23017 to expand to 16 IO-Ports. I use 8 TB6612FNG motor drivers to control 2 motors each.

I attached an image that shows part of the PCB.

But when I connect power (3.3V) and ground to the PCB only to power the MCP23017, it gets extremely hot within a few seconds. Nothing else is connected to the PCB yet.

Is the uc getting too much current? If yes, how can I lower that current?

I checked all the pins like 20 times and only the 4x gnd (A0,A1,A2, VSS) and the 3.3V (Vdd, RESET) are connected. No shorts anywhere. No SDA/SCL connected.

If I got it correct the PCB "top layer" is used for 3.3V.
When I connected the PCB to the 3.3v output of my D1, the D1 was not detected by my computer anymore (and after a few tries it seems like it killed the D1).

Any ideas what I can do here?

Thanks a lot

I think you should double check the IC's pins then check everything with a multimeter. Check you power source as well for it may be giving too much power. After that maybe you can add a resistor or something. Also do you have a circuit diagram, if so can you send that as well?

Yes I checked everything with a multimeter and the connections are correct.
It is a 2 amp powersupply. I thought the components only draw the power they need?
3.3V is correct as well.

Here is a picture of the circuit diagram (only for 2 of the total of 8 driver motors)

Your PCB has been laid out wrong. This is the pinout of the part.

Pin 1 is the top left hand corner as you look at it from the top side. That dosn't match your PCB layout.

No, that is very bad advice.

It is just turned 180 degrees.
The nose is "down" on my PCB, on your picture it is "up".
If you turn it around, it again matches

No it doesn't.

Yes, give me a second.
I will turn the image around for you

Pins are matching. The labels are rotated as well.
But it should work this way


  1. Wiring
  2. Layout
  3. Bad IC

Why should it?
And obviously it doesn't or it would not get hot.

This is what happens when you lay out a PCB and forget that from the underside pin 1 is the upper right hand corner, not the upper left.

Can you post the PCB track side for that part.

I don't really understand what you mean. I just ment that the pins should be connected properly.

If I look from the top down to the PCB, ignoring what is printend on my PCB and just comparing to the schematic you showed:

I get:
GND on A0, A1, A2, VSS
5V on VDD and RESET

And this is exactly what I would expect here.

As (in my opinion) I got the wiring correct and got GND and 3.3V on the correct pins of the IC and nothing else is should be the IC.
But am I correct, that it should not matter how much amps I give to the PCB, the MCP23017 should be working standalone without turning hot?

You could prove that out by plugging one into a breadboard simply enough.

Okay thank you. I will try this. Maybe my IC is just dead, although I tried it before with a breadboard.

If it was OK in a breadboard with just power applied but goes/went critical when placed in your circuit board then that's a telling thing.
I suggest the addition of an IC socket.
I will also note the absence of bypass capacitors in your schematic.

1 Like

Thank you for your replies.
I will try again and just ordered IC sockets.

Yes that at least is correct.

What about doing what I suggested and posting the track layout then we could see what sort of a mess the PCB has been made of the PCB.

The schematic is irrelevant, it depends on how that schematic is converted into copper tracks. Did the "friend" who made the PCB layout use a schematic capture program? If so what was it. Also have you told your "friend" what happens when you apply power?

Yes you have probably fried that IC as well.

Have you tried seeing if there is a short between the 3V3 and ground with your meter.

Sorry, not answering to that.
Your name totally suits you.
Pointing out false things and insisting on them, although I tried to show you it's the same thing.
Maybe the PCB is a mess, but interesting that you can see that only because one IC is rotated 180 degrees.

Saying "friend" (which I did not say a single time) like I am trying to hide it was me who did this.
And as I already stated, there are no shorts and I checked all pins multiple times.

So thank you for your help so far, but please do not spend any more time on this.

Yes and do you know how frustrating it is with some one who refuses to believe what you are telling them.

I won't and I will also add you to my ignore list.

Glad to know that all is well