dlear@blake.acs.washington.edu (Dale Lear) (09/12/89)
8 mhz "turbo" mode problem I'm having trouble with an XT clone when using it in 8mhz "turbo" mode. While working away at 8mhz the keyboard shift status byte (Ctrl/Shift/Num lock/Scroll lock) seems to get randomly set. When this happens I might type "abc" but the program I'm using will think I've typed "^A^B^C", "ABC", or "alt-A alt-B alt-C". I have experienced this problem with a wide variety of programs but using a Borland product seems to exacerbate the trouble. This problem *never* manifests itself when I run at 4.77mhz. If I'm lucky enough to be able to get to a DOS prompt and the caps bit is the one that's set, I can fix the problem by running the following batch file which clears the keyboard status byte at 0:417. FIXIT.BAT mode mono mode com1: 1200,N,8,1 c:\dos\debug < c:\bat\fixit.dbg The file FIXIT.DBG contains: e 0:417 00 q On *rare* occasions, just waiting a few minutes will correct the problem. If I can't run FIXIT (i.e., DOS thinks I'm trying to run a program called ^F^I^X^I^T) and waiting doesn't work then I'm forced to reboot. Pressing the Caps Lock/ Scroll Lock/ Num Lock/ Shift keys has no effect. I've experienced the problem while using different keyboards so I don't think it's the keyboard's fault. I use the following programs to switch CPU speeds. 8MHZ.COM (all numbers are in hex) IN AL,61 OR AL,04 OUT 61,AL XOR AX,AX INT 21 477MHZ.COM (all numbers are in hex) IN AL,61 AND AL,FB OUT 61,AL XOR AX,AX INT 21 Just for the record; the mother board is a no-brand-name Taiwan made job called a "TURBO-640", the CPU is an 8mhz rated NEC V20, the ROM BIOS is a true blue IBM BIOS dated Monday, August 16, 1982, and I'm using MSDOS 3.30. Finally, CONFIG.SYS contains: shell=c:\command.com /p /e:512 device=c:\dos\ansi.sys files=24 buffers=24 fcbs=32,32 If you have any suggestions or ideas please let me know via e-mail as I don't usually read this group. If somebody is tells me how to fix the problem I'll post the solution. Thanks. -- Dale Lear e-mail address: lear@kaleetan.ms.washington.edu
jim@jose.uucp (Jim) (09/13/89)
Dale Lear writes: > 8 mhz "turbo" mode problem > > I'm having trouble with an XT clone when using it in 8mhz "turbo" > mode. While working away at 8mhz the keyboard shift status byte > (Ctrl/Shift/Num lock/Scroll lock) seems to get randomly set. When > this happens I might type "abc" but the program I'm using will > think I've typed "^A^B^C", "ABC", or "alt-A alt-B alt-C". I have > experienced this problem with a wide variety of programs but using a > Borland product seems to exacerbate the trouble. This problem *never* > manifests itself when I run at 4.77mhz. > > <stuff deleted> > > Just for the record; the mother board is a no-brand-name Taiwan > made job called a "TURBO-640", the CPU is an 8mhz rated NEC V20, the > ROM BIOS is a true blue IBM BIOS dated Monday, August 16, 1982, and > I'm using MSDOS 3.30. Finally, CONFIG.SYS contains: I had a 8 mhz turbo xt motherboard. It came with a standard 8 mhz 8088, and I bought the 8 mhz v20 for it. It generally worked fine except for ocasional problems (similar to the ones you mention) and then the problems were only in turbo mode. I removed the v20 and put the 8088 back in, and the problems disappeared.
ching@pepsi.amd.com (Mike Ching) (09/15/89)
In article <1989Sep13.150422.6926@jose.uucp> jim@jose.uucp (Jim) writes: >Dale Lear writes: >> 8 mhz "turbo" mode problem >> >> <stuff deleted> >> >> Just for the record; the mother board is a no-brand-name Taiwan >> made job called a "TURBO-640", the CPU is an 8mhz rated NEC V20, the > >I had a 8 mhz turbo xt motherboard. It came with a standard 8 mhz 8088, >and I bought the 8 mhz v20 for it. It generally worked fine except for >ocasional problems (similar to the ones you mention) and then the >problems were only in turbo mode. I removed the v20 and put the 8088 back >in, and the problems disappeared. The problem is that the 8088 uses a 33% duty cycle clock (clock source is divided by three) and the V20 uses a 50% duty cycle clock. The 8088 clock in the PC doesn't meet the worst case timing requirement for a V20 at the same speed. You could try another V20 in hopes of getting a faster one or use one spec'ed for a higher speed. mike ching