On the Cyrix chips, I bought old “386” PC motherboards and just pulled the chip. The 33MHz’s were/are common, the 40’s are rare. I got mine from a guy on e-bay in Poland!
Never tried the TI chip. Does it even work with the IOPORT read test in the monitor? If you are up to it you may have to adjust the wait state settings and pWR* signal with a patch/jumper.
From: Gary Kaufman [mailto:geka...@gmail.com]
Sent: Thursday, February 19, 2015 8:18 AM
Subject: Re: [N8VEM-S100:6242] Re: 80386 CPU Board - A possible solution
As you suspected the 74Fxx's didn't work either.
I removed K4 and I used the parallel port board jumpered directly to the 1-2 jumper. I was able to switch both Z80 -->80386 and 80386-->Z80 with a single bit toggle. It also works fine with the SMBv2 completely removed from the system. I'm pretty convinced that the timing for the EEh port on the SMBv2 is the issue. I have two SMBv2's and both have the same issue.
I suppose I'll just use the QI,ED method to switch in the 80386 for now and see if others have the same issue as more 80386 boards are built. I also have a second bare 80386 board I'll populate and see if that has the same problem.
One last observation - I wasn't able to locate any Cyrix 80486DLC/E's at a reasonable cost, but did find some TI 80486DLC/E's which are similar with a larger cache. Unfortunately they seem to hang right after the 80386 monitor signs on.
Thanks as always for all of the help!
On Monday, February 9, 2015 at 12:08:26 AM UTC-5, monahanz wrote:
If you don’t get back to the Z80 with 74Sxx’s the 74Fxx’s probably will not work. S’s are faster, just sink a lot of input 2-3 TTL’s I think.
With your parallel port setup, remove the K4 jumper completely on the SMB-V2. Be sure P36 has 1-2 & 3-4. With a logic probe show that you can lower the 1-2 jumper by outputting a 1 on bit 0 with the Z80. The 80386 should switch in. Them within your 80386 monitor output a 0 on bit 1. Check the P36, 1-2 jumper now returns high. The Z80 should come back on. Do you get this?