Looking for some input about RTCs
Kye
Posts: 2,200
Hey guys.
I'm working on making a driver to allow the use of real time clocks on the I2C line the propeller uses to communicate with its eeprom and also to communicate with the eeprom. I also plan to make the driver somewhat expandable capable of running multiple I2C devices at the same time. As in the driver will handle all low level and high level commands.
I plan to make the driver run on a cog. That being said I want my I2C engine to handle as many things as possible.
So, for right now I need help choosing the right real time clock to add to the I2C bus. I'm not sure which one to use.
Some consideration are that I would like to have a RTC capable of using a backup battery and also capable of holding some values that will presit over power cycles. I don't really care if its nvram or sram.
I also need to protype with the chip. That said I would like to use one that has a DIP variant and SMD, QFN, ect. variants for production. I do however have a pretty nice facility to work on this project however, Carnegie Mellon's Robotics club has afew precision soldering irons I can use. If anyone can point me to a site details how to connect surface mount chips to something like the·propeller protoboard then I will be able to use any chip package variants, as long as·then can be modfied with a sodering iron.
Thanks for your help.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nyamekye,
I'm working on making a driver to allow the use of real time clocks on the I2C line the propeller uses to communicate with its eeprom and also to communicate with the eeprom. I also plan to make the driver somewhat expandable capable of running multiple I2C devices at the same time. As in the driver will handle all low level and high level commands.
I plan to make the driver run on a cog. That being said I want my I2C engine to handle as many things as possible.
So, for right now I need help choosing the right real time clock to add to the I2C bus. I'm not sure which one to use.
Some consideration are that I would like to have a RTC capable of using a backup battery and also capable of holding some values that will presit over power cycles. I don't really care if its nvram or sram.
I also need to protype with the chip. That said I would like to use one that has a DIP variant and SMD, QFN, ect. variants for production. I do however have a pretty nice facility to work on this project however, Carnegie Mellon's Robotics club has afew precision soldering irons I can use. If anyone can point me to a site details how to connect surface mount chips to something like the·propeller protoboard then I will be able to use any chip package variants, as long as·then can be modfied with a sodering iron.
Thanks for your help.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nyamekye,
Comments
and it's in common use thanks to the PPDB.
OBC
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
New to the Propeller?
Check out: Protoboard Introduction , Propeller Cookbook 1.4 & Software Index
Updates to the Cookbook are now posted to: Propeller.warrantyvoid.us
Got an SD card connected? - PropDOS
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nyamekye,
I'm using Mike Greens simple I2C driver and it's a doddle to talk to.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cardinal Fang! Fetch the comfy chair.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nyamekye,
Chips with RAM don't really interest me for a couple of reasons at least:
1) They consume more power
2) They are much more expensive
I prefer to power the chips with a 0.1F supercap such as the EDLSD104H5R5C etc as they do not require any holders or maintenance. Any application that is off power for more than a week can just as easily have it's clock parameters updated anyway. The DS1302 does a neat job of charging the cap as well.
*Peter*
consider ramtron FM31L278 (http://www.ramtron.com/products/integrated/processor-companion-product.aspx?id=29) solution: it will replace the standard eeprom with unlimited write capability, add a RTC, Counters, Serial ID ...
... and of course it is battery backed
[noparse][[/noparse]Post Edit]
It is double I2C one address for the 256K fram and other for peripherals.
Post Edited (dMajo) : 1/13/2009 9:59:08 AM GMT
Thanks
Donald
*Peter*
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nyamekye,
I thought the same until I tried. You can work with SMD but it takes more patience and a good magnifing glass. For prototyping, I usually look for an adapter. Usually, its a small PC board with the SMD layout to a dip type. They have them in normal sizes. I have just glued SMD chips onto perfboard and very carefully soldered small wire to the pins. I have some old 28 gauge wire wrapping wire left from my younger days. Try to stay away from the 0.5mm stuff though. Thats really hard. Keep the braid wick handy.
Peter,
Since it also has additional memory, it will be accomplishing multiple tasks with a single chip. The cost is reasonable for proto work anyway. The shipping to Shanghai is a bear though.
I've had a 10+ year old lithium coin battery on my DS1307 for a few months now and its still keeping time. I hope the FM31 is as frugal.
Donald
you have forgot to add the cost of the eeprom that is no more needed
shanghai_fool,
I think that should be better since the battery on DSxxx keeps alive also the ram locations where the time is stored while on FM the locations are fram and the battery is used just to count the time
Post Edited (dMajo) : 1/13/2009 2:50:43 PM GMT
- Most calendar RTC parts can handle leap years, but they can't handle leap seconds, and few handle daylight savings time automatically.
- it is easier to compute elapsed time.
- Don't have to worry about time zone, except on output for display.
Computing the date from a 32 bit count isn't that hard. Here it is in C (sorry, I don't think in SPIN):
Post Edited (Andrew E Mileski) : 1/13/2009 8:55:09 PM GMT
Intersil has an extensive line of I2C RTCs, some with advanced features and up to 512 bytes of SRAM.
For my own purposes, I consider the NV memory essential. I also like to have versatility in the timing/alarm output. Accuracy is determined primarily by the 32kHz crystal. A few RTCs have temperature sensing and compensation circuitry that allows at least a first order correction of the parabolic error curve of the typical watch crystal, and some at a premium $$ even have a built in crystal that is better matched to the oscillator.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com
Donald