I've noticed that a ds3232 rtc chip had a time difference of 0.06 ppm after 16 months (2.5s / 16 months). This seems ridiculous low (datasheet: 2ppm), so I would like to check this with several chips simultaneously.
I have not yet any good idea's on how to accomplish this. The I2C address is the same. Is there a simple method to multiplex I2C?
Nick
you will need separate busses. You devices cannot share the same address on the bus.
You may be able to find i2c transceivers which can be enabled/disabled via some control line. (LTC4300 may work)
Hmmm--apparently Dallas/Maxim didn't think of the case where more than one "extremely accurate" RTC would be needed on the same I2C bus. ;)
Burn an Xmega for your long-term test, with 4 I2C ports?
With "normal" RTCs I tested a batch for a short time, and the results were a bell curve. Yours could just be at the nominal. Also, the 2ppm is over the entire temperature range. Performance is likely to be better at constant temperature. It >>is<< rated from -2ppm min to +2ppm max, so you are really in the "TYP" range.
This seems ridiculous low
Yeah, it's a cool chip. Great RTC. Comes with an integrated temperature sensor.
Oh, I think I'll go with some ti chips (PCA9548A - 8-Channel I2C Switch With Reset or a PCA9545A - 4-Channel I2C And SMBus Multiplexer With Interrupt Logic And Reset Functions) to solve my little problem.
I have not yet any good idea's on how to accomplish this. The I2C address is the same. Is there a simple method to multiplex I2C?
The original project bitbsahed 8 independant ports.
No Way Am I Going To Do That!
I suggest using an 74HC4051 style multiplexor to rout the "Data" to the addressed device, since the clock alone doesnt enlicit a response.
Ron.
Thanks
If you need it for a test, checkout www.i2cchip.com. They have several cheap i2c pcbs, also one with a 4-channel i2c mux.
Or use several pins on your avr with software-i2c
Patrick
