At Samstag, 19. Juli 2003 09:50 you wrote: > > I'll check that again, if this is the case all could be reduced to > > one tri-state sysctl (0=display all four digits, 1=display 1+3, > > 2=display 2+4). What do you think? > > I'm not sure. I read the datasheet again, and I find a total of 7 > possibilities, not 3 as you seem to think. I even took a sheet of paper > and a pencil to draw a nice table. Yes, really. I can scan it and send > it to you if you want ;) In the meantime, I'll try to recap my findings > here. > > [...] I've checked this, and - well your assumptions were not quite correct (for static mode) - neither were mine :-). I don't know why this might have been implemented, here are the results: C2C1C0 1 2 3 4 static 0 0 0 8 8 0 1 0 8 8 1 0 0 8 8 1 1 0 8 8 dynamic 0 0 1 0 1 1 8 8 1 0 1 8 8 1 1 1 8 8 8 8 The only difference is, that in static mode the digits are slighty brighter than in dynamic mode (even when only displaying two digits). I think just leaving them out and using dynamic mode with all digits lit might be just fine, unless anyone tells me he actually needs this. > Oh of course, it's all clear to me know. That's rather logical. Anyway I > guess I'll end up reading the I2C specs (at least this part) to know > more about the different modes, and what happens if one uses a mode the > chip does not particularly like. I really would like to know that, > because that's what I do for the SAA1064 in sensors-detect. I use the > READ_DATA_BYTE mode while the SAA1064 want READ_BYTE mode. I wonder what > will happen... Well, you tell me when you try ;) I guess this chip just ignores any read-address, since all reads (with read-address and without) result in 0x80 if the device has just been powered up, 0x00 otherwise. > The PCF8574 has a readable register, really? Could not find in in the > datasheet. That could help me detect it in sensors-detect... (Some > reading after) Hm, it ain't a register but the I/O value, so I can't > assume anything about it. It can have any value and even change if it is > in input mode (I'm not sure I really understand how input vs. output > works though). I think it is implemented by setting an output bit to "1" , then re-reading it and it will change to "0" when pulled to GND. Just an assumption from what I remember about that subject :-) > Of course. You know it could increase your disk usage by up to 180%? :/ > My perl package uses 18MB. You don't need all the modules though, and I > guess you'll remove the whole think as soon as the tests are over. I think I'll have two installations, then I can zip the large one whenever I don't need it. Who knows what changes after I delete it :-) Greetings, Sascha
