right. If parallel ports have integrated pullups, hopefully they aren't too strong for the external i2c chip. Two other possibilities: 1) the i2c chip doesn't have a good ground 2) the parallel port is driving high on that pin rather than changing to an input, so you're getting contention. This could be caused by a driver bug, or your connecting to the wrong pin. Fred Labrosse wrote: > Mark Studebaker writes: > > What value of pullup resistor are you using? > > 5K is probably about right. Definitely not less than 1K. > > I note that doc/i2c-pport doesn't make a recommendation. > > > > We just did some measurements using a digital scope on the clock and data > lines on the compass. > > Just to remind people, the compass sits on address 0x60 (1100000 in binary) > and I'm trying to use i2c-pport. > > When I send the address on the i2c bus (trying to read a register of the > device), I do read, synchronised with the clock, the address (1100000), > followed by a 0 (write) followed by something at about 1.2V that should be > the 0 for the acknowledgement. So clearly, the device is not pulling down > enough. > > Anybody got any idea? > > TIA, > > Fred > >
