Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!zaphod.mps.ohio-state.edu!usc!apple!uokmax!d.cs.okstate.edu!ong From: ong@d.cs.okstate.edu (ONG ENG TENG) Newsgroups: comp.sys.ibm.pc.hardware Subject: Speedup 386SX from 16MHz to 20MHz: SUCCESS! (for now) Message-ID: <1991Jan21.215159.25118@d.cs.okstate.edu> Date: 21 Jan 91 21:51:59 GMT Organization: Oklahoma State University Lines: 41 Well, folks. Just put the 40MHz oscillator (for 20MHz operation) into my 16MHz 386SX, and it is working just fine. This was 10 minutes ago. I use Norton SI to check the computing power: it increased proportionally as expected. My only concern is the heat on the 386SX chip. The other chips are not emitting much heat. I bought a heatsink, but it don't seem to fit the cpu (0.75" by 0.75"). The cpu has emitted slightly more heat than when on 16MHz operation (after 15 minutes from OFF). I have the casing off, hence exposing the cpu, and I am going to leave it this way for 4 hours. If the cpu is still at acceptable heat level, I will close up the casing and give it similar burn time. Then maybe I will burn it in for 24 hours. Let you all know the results (you will hear me screaming if my cpu burns out). I have a 386SX NEATsx 16MHz motherboard, which has an optional oscillator slot, and a shorting block JP2 to switch to the optional oscillator. Hence what I did was: 1) buy oscillator part# X120 from Digi-Key (call 800 directory) for less than $5 2) open up casing 3) put X120 into optional oscillator slot in proper manner 4) switch JP2 shorting block 5) close casing (yet to be done in my case) To do the above, I have the DRAM at 80ns INTERLEAVE (4meg in my case). If I cannot interleave, then I must have 60ns RAM for 20MHz operation. You can interleave at 1meg, 4meg, 5meg, and 8meg, if you have the proper RAM configuration. IMPORTANT: If you burn out or damage your CPU in any way copying what I have just done, I take NO responsibility! Just in case the heat is too much, I also bought a 35MHz oscillator (for 17.5MHz operation) from Digi-Key. Thanks in advance for any comments/warnings you might have. E. Teng Ong (ong@d.cs.okstate.edu)