On Tuesday 04 October 2005 07:22 am, Jan-Erik Soderholm scribbled:
> PicDude wrote [about Ramtron FRAM devices...] :
> > IIRC the datasheet said something like "instant"
> > writes,... but in some other doc I had seen
> > something about 72us write time, which is noticeable.
>
> As far as I remember, they say that you can
> write to the device at the full speed of the serial
> interface. And if so, the actual write-time doesn't realy
> matter. This might have changed in more reasent devices...

I found the doc that mentioned 72us...
  http://www.ramtron.com/lib/literature/appnotes/24C16adv.pdf
(see the table at the bottom).


> And at (e.g.) 400kb/s, you're not *that* far from 72us/byte.
> And 72 us is far better then the normal 5-10 ms in regular
> EEPROMs, not ?

Not really in my app -- I have enough time between successive sets of data 
being received (and needing to be saved).  But to write multiple pages, I 
need to setup and kick-off one write, then let the PIC run off and handle 
other things, then later come back and write the second page, etc.  I 
implemented a sort of queuing mechanism to handle this -- create the 
datasets/addresses, then save them in various variables with the addresses 
they need to be saved to, and set some flags indicating that the datasets are 
ready.  Each time the program counter comes back to this section of code, it 
checks if this mechanism is free (last write complete), and if anything else 
in the queue is ready to be saved, etc.  That all works fine.

But... if I could write a page, then immediately setup and write the next 
page, it would let me eliminate the queuing mechanism code, which would be a 
nice saving of processing cycles, and RAM/code space.  72us = 360 
instructions (at 20Mhz) between successive writes.

Cheers,
-Neil.



>
> Jan-Erik.


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist