Folks...
Took a look late last nite at the various schematics uploaded to the
files section - specifically, the controller. [It's in the file
FRONTschematic.pdf, IIRC.]
The CPU here is a high-integration "all-in-one" controller that has
quite a nice little CPU inside - besides ROM, RAM and peripherals.
It's part of Hitachi's H8S (actually, H8S/2000) series. The part # is
HD64F2345FA, meaning it's a -2345 variant architecture, with flash
memory ("F") in a 100-pin ("A") SMD package. The Hitachi website is:
Drill down to find the .PDF file for this controller. I did a quick
perusal (I've seen H8 stuff before; Yaesu seems to like it; my VX5
and VX150 have smaller 8bit H8s on 'em) but this is a quick,
approximate rundown...
This very capable 16/32 bit CPU can be clocked up to 20MHz. It has
128KB flash ROM, 4KB RAM, 6 x 16bit timers, 2 x 8 bit timers, 8
channels of 10bit ADC, 2ch 8bit DAC, something akin to a DMA
controller, and 2 multimode serial channels. Various memory ranges
can be configured to be of differing bus widths. There's some
smartcard support features too. Not all of these features are used
on the FT817 - for example, no external memory-mapped peripherals are
used. Other chips (PLL, DDS, etc.) are controlled via fast serial
data transfers - even "user memory" for settings, freq memories, etc.
is stored in an external serial EEPROM (24C64).
A quick read indicates this CPU apparently has a small fixed factory
boot ROM space so that its (empty) flash can be programmed via a
serial port. It can also be programmed/verified as a freestanding
chip on a chip writer (it's a 27F101). I note *no* ROM code
protection features [other than some anti-go-wild stuff, there's no
code-lock or encryption]. Thus it appears the code can be somehow
dumped and analyzed & tweaked, worst case requiring de/re-soldering
of the chip. I did NOT immediately see a graceful way for code to be
dumped out of the flash ROM to the outside world while still soldered
to the controller board. There *may* be some way the code can be
dumped from this chip without desoldering it and reading it on a
burner; I will have to look into it. [It is likely this code is
mostly written using compiled C. Speaking from quite a bit of
reverse-engineering experience, reading disassembled C is sometimes
much harder than reading human-written assembly code; we humans think
somewhat alike, and not much like a compiler ;-).]
There's probably quite a few "hidden" features, modes, etc. here we
should find out about and that the more brave amongst us may wanna
tweak. After all, I am indeed a charter member of the International
Association of Warranty Violators :-P
73s Bill N6AOT
wiese@...
-------------------------------------------------
William M. Wiese Jr.
Application Engineer
Philips Semiconductors
SIIG/Storage & Information Infrastructure Group
Sunnyvale, CA
408-991-2667
william.wiese@...
"The opinions expressed herein are mine
and are not necessarily representative
of those of my employer or colleagues."
--------------------------------------------------
