18F4550 Power Draw

[rackley]

I ported my code from an 18F452 to the 18F4550 to enable USB connectivity. Some registers had to be tweaked, but overall it wasn't too horrible. Everything is working fine, but I noticed that my heat sink is getting warm. It never got warm with the 452 in there (it's relatively big,) so I did some power draw measurements. My contraption draws 300ma or so with the 18F452 in the circuit (almost all of that is the LCD), but when I pop the 18F4550 in the circuit it draws 700ma. Why the heck would the 4550 be drawing 400ma by itself? That's way too much power! The 4550 gets to about neutral temp when touching it (not warm but not cold), and the 452 always seems cold.

What do you guys think, am I somehow sinking a bunch of current into a ground? It's the same code on both, the only thing I can think of is maybe the chip itself draws more power because of the USB support, but 400ma seems an AWFUL lot to do that. Maybe a misconfigured register setting a pin wrong? It would have to be one I'm not using since everything is working fine I/O wise..

Thanks,
Ray

 

[Nigel Goodwin]

I ported my code from an 18F452 to the 18F4550 to enable USB connectivity. Some registers had to be tweaked, but overall it wasn't too horrible. Everything is working fine, but I noticed that my heat sink is getting warm. It never got warm with the 452 in there (it's relatively big,) so I did some power draw measurements. My contraption draws 300ma or so with the 18F452 in the circuit (almost all of that is the LCD), but when I pop the 18F4550 in the circuit it draws 700ma. Why the heck would the 4550 be drawing 400ma by itself? That's way too much power! The 4550 gets to about neutral temp when touching it (not warm but not cold), and the 452 always seems cold.

What do you guys think, am I somehow sinking a bunch of current into a ground? It's the same code on both, the only thing I can think of is maybe the chip itself draws more power because of the USB support, but 400ma seems an AWFUL lot to do that. Maybe a misconfigured register setting a pin wrong? It would have to be one I'm not using since everything is working fine I/O wise..

It sounds like you already think, you've got something wrong somewhere!, PIC's don't run warm unless you're abusing them - I agree that USB support shouldn't take 400mA extra, although you could always check the datasheet, the specs are near the end.

Assuming it's in a socket?, try leaving pins unplugged, one by one and see what happens.

 

[rackley]

Hmm, that's a good idea. I"ll just bend a pin out each time and see which one fixes it

On the datasheet issue, I already tried it and it doesn't give any power consumption figures at all, only a spec for power dissipation max of 1.0W.

Thanks,
Ray

 

[justDIY]

perhaps you have a different ground level on the gizmo supply compared to your USB host... which is throwing off your regulator, causing it to draw extra current (making up for the difference in ground levels)

 

[motion]

Several ports configured as outputs might be shorted to ground. Pay attention to pins that have a different use on the PIC18F4550: Vusb and the formerly unused pins ICPORTS, etc.(on the 44-pin package).

 

[rackley]

Well the USB connector is disconnected when it's sinking that much power, and the USB GND is tied into the main circuit ground.

VUSB goes to + on a .22uf tantalum cap which is then tied to ground for stability.

I've tried bending almost all the pins out and it still draws the power. Next step is to stick it in a breadboard and just connect Vdd and Vpp and see what she draws.

Thanks,
Ray

 

[Oznog]

On the datasheet issue, I already tried it and it doesn't give any power consumption figures at all, only a spec for power dissipation max of 1.0W.

18F4550 datasheet, section 28.0, shows definite power draw figures for most all scenarios.

PRI_RUN, Fosc=40MHz, Vdd=5v, no pins loaded, is 21mA min, 40mA max. I believe that's with no peripherals enabled, but peripherals are not going to increase power draw astronomically.

You also mention an LCD and 452 drawing 300mA?? No little microcontroller LCD module I know draws anything like that and the 452 doesn't draw that much. You're like 10x what it should be. Are you sure you're taking current measurements correctly? At 1.5W (300mA 5V) an LCD and 452 would get quite noticibly hot in a few minutes!

So for starters I'd recommend checking your meter for accuracy. Still, a 4550 generally would not be described as operating "warm". The dsPIC chips can get quite hot when maxed out at 5v but generally not the 18F series. What Vdd, Fosc, and oscillator mode are you using?

Is this thing actually functioning ok?