ray%othervax.uucp@BRL.ARPA (Raymond D. Dunn) (10/14/85)
Although many, including myself, are successfully using an NEC V20 as a direct replacement for the Intel 8088 in our PC's, it looks as if this may not be a good idea in all but a hacking environment. I am indebted to my colleague Humphrey Brown, for the following information:- "The following caution should be issued in connection with the widely touted NEC V20 microprocessor (P/N 70108) as a plug-in upgrade replacement for the 8088 in personal computers. The specification for the V20 calls for a clock with a 50% duty cycle. The clock driver for the 8088 delivers a clock of 33% duty cycle. NEC's data sheet for the V20 specifies a minimum clock high time of 85nS and a low time of 90nS. Intel gives a minimum clock high time for the 8284 of 1/3 tclcl+2nS, which for a clock frequency of 4.77Mhz, yields 71nS. While the chances are that this will work anyway, those who wish to try it should note that there are no guarantees. I have'nt got the data sheet for the 8Mhz part yet, but by judicious extrapolation, that part will require a clock high time of about 48nS, thus that part of the clock requirement will be met. However, NEC are quite explicit in their data sheets as to the need for a 50% duty cycle. The solution to the mismatch between the Intel clock generator and the NEC processor might be to also plug in the NEC clock driver (P/N 71011) in place of the 8284, since it appears to be pin-compatible, but note that the 71011 divides the external frequency by 2, not 3 as for the 8284." Ray Dunn. ..philabs!micomvax!othervax!ray