Posted on April 23, 2005 at 02:341. LS245 buffer used in IDE path should not be 'VERY HOT', it only can be compromised to say 'it is WARM'. if it is HOT, check the PCB status between its output pins and its GND pin. or VCC pin. check the soldering and wash out the flux that could remain. or, replace the chip with new one.
2.yes. ISAOE# is issued low when it accesses ISA bus. but remeber. it multiplexes ISA bus signal out of signal from CPU so seen from IDE connector. it reflects directly ISA bus activity to be easily watched. RMRTCCS# is issued when it accesses boot ROM. but strange. if CPU tries to read KBD or SER port data from SMSC SIO, the ISAOE# should be togled.
3. running DOS, if you do the codes to poll the SIO you can get the message from outer world through SIO (which is totally absurd in real aspect of application..) if SIO misses interrupt. there seems to be no accommodation for real software to run on it.
4.if DDx shows nothin, it means ISAbus in that address space shows no action. but if chip multiplexes IDE and lower ISA address bits. it is probable that ISA part died. check PCI siganls and ISA signals. if they shows no action. (ISA activity is reflected on PCI bus.) the bus core died. check all the basic clocks. if it doesnt work, it means the whole hardware is not in stable level.
if all clocks are working as start- strapped but no PCI and ISA clocks running, then the bus signal integrity is bad. the chips power plane seems to be designed without significant error.
5. probing IRQ on board with osciloscope may make glitch on that line. and make erronous notification to BIOS. though, you can't easily conclude that BIOS died. because the solid code will mask interrupt as soon as it enters critical section. if the programmer of the BIOS code are sane enough. (oh i admit that critical section is abuse of the notion. i never heard a BIOS supporing multi threading or multi tasking..) but in this case. you have to make glitch synchronous to 14.4Mhz clock. and you noticed IRQ1 action. so lets forget about the probing problem.
6. regarding your CPLD, check it works properly. short IRQ inputs at randome pattern to gnd and probe ISACLKx2 and ISAclk and each muxouts. and see if it multiplexes lines corectly. after all that has no misfunctioning. then you have to go back to SIO side.
Felix pointed right thing. check the device list that BIOS makes.
7.General BIOS by default supports headless operation. (if things is the same with VEGA..) have the 9pin serial cable , Female to female or cross cable (pin #2 and pin#3 should be connected at each end.) and test pulg out the VGA and see serial port on host side. use hyper terminal with 115200 bps 8 bits data 1 stop bit and no flow control. if it doeesnt show message on hyper terminal, the SIO - MUX/DMUX - CPU signal flow path has problem.
8. update BIOS to brand new on
http://www.gensw.com/pages/prod/bios/boards/stpccon2.htm
for fear that older BIOS may set the chipset initialization value not properly.
9. DOC200 is slower. i dont know about in DOS cases DOC2000 runs a block device emulator sw so booting from it is lazy. compared to magnetic HDD.
BTW. are you working on the ref eval board or your own board? i am quite confused so far..
