> On May 31, 2018, at 6:14 PM, Brad Spencer <[email protected]> wrote: > > Unfortunate behavior. Looking back over the sensor driver I worked on, > it appears that I always read something to determine if the device was > actually there.
I spent some time reviewing NXP’s i2c spec this evening (well, during timeouts, etc. — GO DUBS), and I’m becoming convinced that there is a subtle error in our i2c_bitbang code… the spec seems pretty clear that a START-address-ACK should occur if a device is really there, and thus the “quick read” method for device detection should work 100% of the time. I suspect we’re not properly setting the SDA line as an input at the critical time… we should be ensuring that any internal pull-up is activated so that any responding slave really does have to drive the line low, so that we can get a clear reading on the ACK. I might experiment with this using an external pull-up to see if the ghosting behavior goes away with the “quick read”. (Not sure I'll get to it this weekend, though… despite my kid’s spring soccer season being finished, I seem to have a full calendar nonetheless…) -- thorpej
