> After almost a whole day of trying to look at code and fixing mine, I
think I
> finally got somewhere. The procedure for obtaining the temperature for a
TC74
> seems to be this (correct me if I'm wrong:
>
http://www.microchip.com/download/lit/pline/analog/thermal/tempsens/serial
/2146
> 2c.pdf). Anyway, the way I'm doing it is:
>
> Send 7-bits of address (I have the part # with address 1001 000) and
then one
> bit of R/W. The first phase of transfer always seems to be write, so I
send a
> 0.
> First byte out then becomes: 1001 0000.
> I then get a successful ACK (I set SDA high intentionally, and the slave
> succesfully pulls it low).
>
> I then send the command to read the temperature,

You should do a bus STOP here, followed by a bus START before the next
command.

> which is 00h. I also get a
> successful ACK on this. According to the datasheet, it says to re-send
the
> slave address accept with the R/W bit set this time, so I send: 1001
0001.
> And again, I get a successful ACK.
>
> Next, I should just send 8 clock pulses and be able to read the data on
SDA,
> but instead I get all 0's.

Good.  That's better than all 1s because somebody is at least pulling the
bus low.  Did you make sure the master released SDA before sending the 8
clocks?  If it did, then the slave thinks it's addressed but probably
confused.  That could be explained by the missing STOP/START.  Or, is it
really cold in your office <g>?


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.