PIC24FJ over heating

[raphaeldavidf]

Does it still heat up with /MCLR tied low?

Yep

[raphaeldavidf]

Quite apart from the heating issues, 50R is ridiculously low for a LED resistor for what appear to be status LEDs. You rarely need anywhere near 20mA for modern LEDs.

And you may get more help if you uploaded the PDF here instead a site with a lame PDF viewer and login required to download to a decent viewer.

How much current is it pulling from the supply ? That info would be far more useful than "getting too hot"

Sorry about the file, my bad, I'm quite new to the forum. Didn't see that I could upload the files here. I'm going to edit and upload the files to the first post.

Thanks a lot mikeselectricstuff

[raphaeldavidf]

Quite apart from the heating issues, 50R is ridiculously low for a LED resistor for what appear to be status LEDs. You rarely need anywhere near 20mA for modern LEDs.

And you may get more help if you uploaded the PDF here instead a site with a lame PDF viewer and login required to download to a decent viewer.

How much current is it pulling from the supply ? That info would be far more useful than "getting too hot"

You say you haven't really connected anything up apart from the few capacitors and pull up resistors, a PIC will run with a 10k or so pull up to VCC, decoupling caps near the power in and a 3.3 / 5V power source depending on the variant.

In that case the 50R resistors should not be fitted, thus not causing issue, I wouldn't connect those 50R resistors by the way, I would use more 270 - 330 depending on the Vf of the LED.

Things tend to heat up when shorted, even if the PIC was in permanent reset it should still stay cool to the touch, if you have no external oscillator and no code running it won't even be really doing anything apart from running it's internal regulators etc, certainly not enough to start heating up.

You may have to post your board files.

Did you mention you were powering from a pick kit or programmer?  Try programming the device then powering it from a different bench PSU or something to see what current is drawn, put the current limit on low!! You will soon know if it goes short circuit, the current limiter will kick in.

Even the low spec regulators when shorted can draw 100mA, if it is higher it can theoretically draw more, a USB port is capable of drawing over 1 Amp (out of spec but it can still draw over this!), and that will do unknown damage to your chips.

You might have a short in one of your capacitors. You can try running without them.

Can you measure the voltage on the Vddcore pin and measure the total current being drawn by the MCU?

Murphy got me and it is really stupid.
The problem is that the VReg tantalum cap was backwards. In the case A smd cap the ground is on the side without the mark, not with the mark like an electrolytic cap 

It's working like a charm now!

Many thanks to all for all the help and patience 

[Howardlong]

Been there, done that, and almost certainly others on here have too.

Not really sure WTF they were thinking to mark tantalums the opposite way to non-tantalums but what I will say is that I will never forget the experience, and I'm pretty sure you won't either.

[mikeselectricstuff]

Why are you using a tant on Vreg ? a 10uf ceramic works fine.

[raphaeldavidf]

Been there, done that, and almost certainly others on here have too.

Not really sure WTF they were thinking to mark tantalums the opposite way to non-tantalums but what I will say is that I will never forget the experience, and I'm pretty sure you won't either.

Hahahahahaha I'll never forget, the fire of shame burned it into my mind 

[raphaeldavidf]

Why are you using a tant on Vreg ? a 10uf ceramic works fine.

This was my first try with this pic, so I just copy the datasheet. The next board I'll try with a ceramic, thanks for the tip 

[Howardlong]

I haven't designed in tantalums for some years now, favouring ceramics, and as Mike says ceramics are fine in this application. On a product I've had in the field for over four years which uses a PIC24FJ64GB002 I use an 0603 10uF Farnell part number 1760126, they are a couple of pennies each by the reel. It's worked fine in many, many thousands of units, and DC bias effects often cited about using MLCCs haven't caused any problems. You don't need to voltage derate MLCCs like you do electrolytics either.