Hello,
 I'm trying to build a project which should interface with a 93LC46
with a possible "Memory upgrade" to 93c86...both with a 16 bit
organization and a 3-wire interface ( DI and DO tied together...)
This should be obtained simply changing the chip in the socket and
resetting the micro (16C56 or similar..)
I think I can use the lowest address (000) as a scratch-area.

The problem is to find the simplest mode to detect that the EEprom has
a greater size.. and change the addressing number of bits (from A0..A5
to A0...A9) once detected the RIGHT size.

This is my analysis: if I should try anyway to read the address 00 0000
0000 considering it a 93c86 ( I use binary values ... it's a matter of
bit-count.. Startbit and Opcode are 3 bits )
If it is a 92c86, I will get an output after 13 clocks ...and with 16
clock pulse I will get the data at 00 0000 0000.
If it isn't , the EEPROM will try to output the content of address 00
0000 after 9 clock pulses....which will be wasted during the first 4
clock (D15 D14 D13 D12) and then recorded....so that the original D11
will be recorded as D15 and so on.. until the last four bits which will
be padded to .....I think the same value as D4 ...
Anyway I should get a "left shifted version" of the data at 00 0000.

The same should happen trying to write .....
If I try to write a  0008h to the address 00 0000 0000 and I will
operate always (in the "measuring routine" ) as it were a 93C86,
according to my analysis in the case of a 93C46 I should write 0000 at
that address as the four added "0" clocked in as "address" will be used
as "data", so right-shifting the data..and the single "1" will be not
be recorded as it will fall "outside".

So writing 0001h at 000h and reading the same address could give me a
different response : 0 if it is NOT a 93c86, 1 if it is a 93c86..

Is this analysis correct? Can you suggest a simpler way ?

I have not yet a 93C86 on the shelf to make a test, can anyone help ?

Best regards

==============================
Bruno Rodeghiero i3RGH
Bassano del Grappa (VI) ITALY
mailto:brurodeg@tin.it
==============================