Adam, Adam, Adam!! Are you Blue because you always keep spouting the same stuff without taking time to catch your breath and pausing to think through just what you are saying? You really should take heed of what some of the oldsters in this MCU business are trying to tell you. Fail to do that and you'll find your self sitting right in the middle of the frying pan with hot oil and sausage all around.

The "power on problem" you are having is only a problem because you did not take proper steps in the design of your circuitry. Do it properly and there will be no "power on problem".

Virtually all MCUs in the market place have their pins power up in "input mode". Some like the 8051 family of legacy compatible parts have internal weak pullups on the pins to keep them in a defined state during the power up sequence. That is actually a very good thing because it means that I/O pins can be left open if they are not in use. Other MCUs have pins that are simply just inputs at power up and it behooves the designer to add the necessary other components to the design to make sure that the inputs do not float around and cause problems.

There are a wave of newer large devices that the chip makers have decided to also add weak pullups or pulldowns to their GPIOs which default as inputs. The choice of what you may find on a particular ARM controller, system I/O chip or SOC may vary. Some of the Intel system chips have some GPIOs default with some pins with pullups and others with pulldowns. Some of the fuse or flash cell programmed FPGAs have programmable selection of adding an onchip pullup or pulldown to help with input pin state at power on.

It is rather rare to find a general purpose MCU that has fuse or flash config bits for enabling power on enabled pullup or pulldown resistors on board the chip. I have encountered a few that had such feature but it was normally a single global config bit that applied an optional pullup to ALL the I/O pins. And many of those were not even non-volatile and wouldn't take effect until the software started to run and configure the register.

In the past I sometimes used things like relay or lamp drivers that had a potential for a "power on problem" that was circumvented via a single enable pin on each package. A properly designed circuit from a port pin could hold off the drivers until software could come and enable all the drives after the normal port pins that go to the drivers are setup and active in the way they have to be. In the case of the 8051 if the enable pin to the driver was active low (most common) one can hook this directly to a port pin along with maybe a pullup resistor to hold off the chips. If the driver has a high true enable you can get by with a single PNP type circuit to the enable instead of having to put one at each individual driver.

You keep crowing about how some PIC does not have the problem that you have experienced with your 8051 design. You should really take the time to evaluate and then explain why this is the case. At least provide a PIC MCU part number in the discussion to add credence to your claims.

Save your self soon. Otherwise you'll be getting pretty crispy.

Michael Karas