gwoho@nntp-server.caltech.edu (g liu) (05/14/91)
On the isa buss (the one in an ibm at), connector b8 is called 0ws, which suposedly tells the motherboard not to insert wait states when it is pulled down. since it is no connect on the xt buss, it must insert the wait states when it is not connected to work with xt cards. none of my cards connect this pin to anything, thus i assume that they are using wait states, and actually working like 4.77 mhz xt cards. i tried just shorting this pin to ground, resulting in a system that would not boot. no video, nothing. i dont know if it beeped, because i got rid of my speaker long ago. is it true that these cards are running much slower than the buss allows, and is there anything i can do about it? i thank you in advance for any advice. gwoho liu.
torbakke@loke.idt.unit.no (Torbjoern Bakke) (05/14/91)
In article <1991May13.175633.25495@nntp-server.caltech.edu> gwoho@nntp-server.caltech.edu (g liu) writes: > On the isa buss (the one in an ibm at), connector b8 is called 0ws, which > suposedly tells the motherboard not to insert wait states when it is pulled > down. since it is no connect on the xt buss, it must insert the wait states > when it is not connected to work with xt cards. This is correct. > none of my cards connect this pin to anything, thus i assume that they are > using wait states, and actually working like 4.77 mhz xt cards. This is also correct, but the ISA-bus is usually running at 6 or 8 MHz, so (at least some of) your cards operate faster in an AT than in an XT. > i tried just shorting this pin to ground, resulting in a system that would > not boot. no video, nothing. > i dont know if it beeped, because i got rid of my speaker long ago. This sounds very reasonable. This signal is usually generated on a cycle-by-cycle basis. The expansion cards that are capable of supporting zero wait-state transfers do this by decoding the address, and then driving 0WS low (active). If you just short this pin to ground, the microprocessor assumes that all bus-cycles should use zero wait-states. The consequence of this is that there will be a lot of unfinished memory-cycles and I/O-cycles. The results are unpredictable. > is it true that these cards are running much slower than the buss > allows, and is there anything i can do about it? Very few expansion cards use the 0WS-signal. In fact, most cards even use more than one wait-state (using the IOCHRDY-signal). So the first answer is yes, most cards run slower than the bus allows. There is not much you can do about this. You could of course buy other cards that run faster, or even build your own expansion cards. It *is* possible to use the 0WS-signal. > i thank you in advance for any advice. You're welcome. ----- Torbjorn Bakke (Research Assistant) Internet: torbakke@idt.unit.no Div. of Computer Systems & Telematics torbakke@solan.unit.no Norwegian Institute of Technology -- ----- Torbjorn Bakke Internet: torbakke@idt.unit.no Div. of Computer Systems & Telematics torbakke@solan.unit.no Norwegian Institute of Technology Bitnet : TBAKKE@NORUNIT