rsdio presents: 1-Wire network via an SPI-compatible display controller.
To produce the 3-wire SPI™ interface required by a MAX7221 display controller (active-low CS, DIN, and CLK), this 1-Wire network serially addresses three 1-Wire switches (DS2413). The first switch creates Chip Select directly (active-low CS), the second creates the serial-data line directly (DIN), and the third switch—with the help of three exclusive-OR gates—creates the serial clock (CLK).
I think it would be easier just to find a uC that has SPI/1-Wire and use it as a bridge. Should be about 1 dollars verse all those extra chips.
all those *expensive* chips… yeah.
Can you implement a bit-banged 1-wire receiver (transceiver?) implementation on the average microcontroller? Are their micros with HW support for 1-wire?
Can you imagine the overhead? A 64bit ID (I’m guessing) to updated each of two IO chips a minimum of 18 times (assuming 0x00 or 0xff) per byte? Yikes. My original post included a comment to that effect, but I couldn’t make it sound not lame.
Atmel apnote AVR318 shows a 1-wire bitbang, so I’d guess that you could do
this with any avr that has more than 17 I/O lines (8 + 8 + 1). I don’t get their
ref to IC3 either. Each ds2413 has two switches, so they used 3 out of 4.
I think this apnote is more of a wow-look-at-this than a practical thing.
And the max7221 is $12 in quantity 1! An atmega48 that can do the whole
1-wire and display driver is $2.50.
I had in mind making a simple 1-wire to SPI bridge. In this app-note, the max7221 is just a “sample” SPI device (expensive, but the “additional parts” needed to do an 8-digit 7-segment display more than just a cpu are a bit annoying.)
Ap note avr318 says that the code will run on the attiny series, so my guess
is that something like an attiny85 (8 pin dip or soic) could be programmed
to bridge spi to 1-wire. I looked at the code and it’s targeted to atmel’s
compiler, not avr-gcc so I couldn’t try it out.
Maybe the advantage of this circuit is that it could be a remote PCB that doesn’t require any firmware, just 1-wire and a battery. An SoC could handle this, but you’d have to load firmware onto every board before it would work.
I’m not trying to defend Maxim; I just like these circuit notes because they sometimes show interesting ideas that might be useful in an entirely different context.