eric@necntc.NEC.COM (Eric Hanselman) (11/10/87)
To the net.at.large, this may be of interest to a large number of people, but is really a posting of something that I am unable to e-mail to the originator, Bruce Eckel at Washington.edu Bruce, This is a resend of my original message. The first to eckel@namu.ocean. washington.edu bounced pretty seriously. I'm sending this to eckel@ocean.washington.edu in hopes that it will find you. Is your name server broken? I'm also sending this a few other ways. Let me know which gets to you. To answer your question, I know very little about V20 PC problems. I'm the applications engineer for communications parts at NEC and try to stay as far from V series as possible. Actually, I should say that I knew very little. I have talked with the guy who does know about V series issues, as he gave me the whole story. The bottom line is that you and/or the dealers are right. A 10 MHz V20 (uPD70108-10) will not run properly in most PC's. The reason is a simple one; clock timing. The V20 doesn't like to run on the 33% duty cycle clock that is fed to most 8088's. The 10 MHz part requires a minimum high pulse width of 41 ns and a minimum low pulse width of 49 ns. The low pulse width is no problem, but a 33% 10 MHz clock will give you a minimum high of 33 ns, a clear timing violation. The trick that was available to 8 MHz PC users, namely using a 10 Mhz V20 to beat the timing problem, is not available to 10 Mhz users as we have no higher speed part. The V20 will run properly if you feed it a 50% duty cycle clock. One could fashion a daughter board with a one shot or some other type of clock shaper between the PC clock and the V20 and that should do the trick. If you have any further questions or would like a data book to prove this yourself, just let me know. - Eric Hanselman eric@necntc.nec.com harvard!necntc.nec.com!eric