Re: TI-H: Complaining about memory...
[Prev][Next][Index][Thread]
Re: TI-H: Complaining about memory...
>I'm not flaming, but please tell us why the EuP isn't good?
>Grant
>Much of the EuP design was borrowed in the EII.
Hehe, is THAT what bryan told you? About 4 months ago I'm the one who
first told bryan about the Atmel memory chips and that I was going to
use them in the expander II (I should have kept my mouth shut, he
would have never seen them :). At that time I was unsure about which
processor I was going to use, but I eventually settled for the AVR.
Development was pretty natural after that, with bryan using the PIC
and I using the AVR. I was advising bryan to switch from the PIC at
the time, because I knew that the PIC sucked. But aside from me
giving bryan some design pointers, there was really no "borrowing"
after that. I will say this with certainty: I have never taken any
of bryan's ideas about memory expansion!
In any case, the EuP design definitely has shortcomings. As bryan
will tell you, the PIC is one of the worst microcontrollers on the
market today. Slow, expensive, non flash-programmable, and poor
instruction set. (How much is bryan charging you guys for
preprogrammed PICS ? :) The EuP design is becoming a bit more popular
now, but what happens if a bug is found in the firmware? Everyone is
stuck because they can't reprogram! So the efficiency of the EuP
design is now fixed, with no more improvements ever possible without
alienating all the existing EuP owners.
In addition to this, the core design of the PIC was created back in
the early 80's. If bryan didn't already have a lot of time invested
into using the PIC he would have switched to the AVR a long time ago
(he should have listened to me! :). All of his new (non-calculator
related) projects are using the AVR.
Using the PIC means that the EuP is inherently slow-as-snot, and not
to mention expensive and non-reprogrammable. Even the expander SF
gets over 5k/sec transfers while the EuP barely gets 1/3 of that on
the 92. The situation will be the same if/when the EuP is ported to
the 85. But it's the same difference as going from a 9600 baud modem
to a new 56k modem. I have some new theories on the linkport
operation on the 85/92, and if they're correct I believe that I can
get nearly 15k/sec transfer rates on the 85, and if another one of my
theories are correct (I haven't told anyone about this theory yet, I
must contact david ellsworth to further explore this idea) I can have
a single-chip, easy-to-install, low power internal expander for the
92.
There are a couple other (major) comparative differences between the
EuP and the upcoming EII, of which I won't elaborate on right now.
So in the end, basically you should all save your bickerings until the
expander II/III comes out... then make your decisions :).
-Mel
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-The TI-Memory Expansion Homepage
-http://www.egr.msu.edu/~tsaimelv/expander.htm
References: