Serial number rom ic, with spi or i2c interface?

If you can't find one... a cheapo 8-pin microcontroller from Microchip, Atmel, etc.

There are also some eeproms with write-protect block locking facilities.

Note that most of these type of devices are fairly easily circumventable so should not be relied upon for secure applications.

should not be relied

With a micro you can implement a reasonably secure challenge-response system. The weak point then probably becomes the chip's code protection mechanism.

should not be relied

Agreed. This would be more secure than most of the EEprom type devices which use password type access control, or write-once/password-write devices that can simply be replaced or emulated.

Thanks guys :-) Anyway, my problem is not security, but a simple matter of lazyness :o) Most important for me is that it's preprogrammed, so I dont have to do anything other than mount the device and read the id number.

they ponyprog eeprom writer software has an auto-increment serial number function

martin

"Adrian Hjelmslund" schreef in bericht news:43d7565f$0$12400$ snipped-for-privacy@news.cirque.dk...

FAIK Dallas is unique. But what's the problem with that 1-wire interface? It's as easy as :) or as difficult as :( SPI or I2C. Only if you have absolutely no pin left for it and you need to connect it using an existing SPI- or I2C-interface you can put a micromicro in between. So an eight or six pins Microchip or AVR. Of course it makes it little easier for the bad guys to fiddle with it but you can easily find this out if you need to.

petrus bitbyter

It's an absolute pig if you have multitasking or interrupts, because unlike SPI and I2C, you can't just hold levels in mid- process. So he'll have to read it before that bit of the program starts.

Paul Burke

I've used the 2401 in a number of projects. What I usually do is to read it as part of the startup sequence and store the contents into RAM before interrupts are enabled.

If the serial number is not "special" then maybe you can implement it in on-chip EEPROM. When you power the unit up for the first time at the factory, you can either have the tester enter a number on the local interface, or use an external or internal [pseudo]random source to set the serial#.

he'll

Well.......that statement requires a bit more qualification. There are places you can't pause due to timing or power constraints, but you can pause between the bit sequences for days if you like.

Thanks guys, you've been very helpfull.

I've been thinking bit more and I think I have to stick to the 1-wire solution eventhough I'm not too happy about it.

Problem is that it is timing dependent and that I don't have other 1-wire devices, so it will feel bit messy to implement it just because of a serial number, but the alternatives aren't much better I must say.

I came to the conclusion that Bobs solution about reading during the initialisation process, before enabling other interrupts or timer, is the solution that suits me best.

Thanks for advices and inspiration :-) Adrian