Strange I2C problem (and solution)

I build a lot of data logging devices, and I frequently use a real time clock and a EEPROM memory IC with various PICS using an I2C interface. While testing a new prototype I noticed that voltage would frequently shift from ~5.6V (normal) to ~3.3V. After much desoldering and breadboarding, the problem appeared to be related to the the 10K pull-ups I was using for the I2C interface. Without the pull-ups the voltage stayed at 5.6. But the pull-ups are necessary. So after trying several hardware solutions, I simply set the SCL and SDA pins high early in the program, and keep them that way. This seems to fix the problem, but I am still a bit baffled by it. Why would the pull-ups cause a voltage drop?? Just thought I would share this in case anyone else sees something similar.

If the "other end" of the I2C bus is running off 3.3v, then the SDA and SCK will be pulled down to 3.3 + 0.6 ish volts. through the devices protection diodes to Vcc.

Just a thought!

Timmers.

Thanks Timmers, but everything is running off the same 5.6V. I'm still a bit mystified, but my best guess is that the regulator cannot put out enough current (I'm using an onboard regulator from another component, which is "supposed" to be able to support a microprocessor).