🚀 go-pugleaf

kirk@beta.tricity.wsu.edu (James Kirk) writes:

 >I have one of the old 16Mhz 386 boxes. It has two of the 1MB
 >memory boards in it. I seem to recall some time ago that 
 >somebody posted a way to make these take four meg. I know the
 >holes are all there for the 20 pin sockets but there was some
 >other stuff two. Does anybody recall what the process is our can 
 >you e-mail that posting to me? 

 >thnx in advance

Zenith 386/16 Memory Board Conversion:  1 MB to 4 MB (Z505-Z515 ??)
 
Cost $230 (Fall '91), PLUS lots of time
 
Parts list: 36 18-pin DISP IC sockets, 36 M1000-10 1MBit x 80 ns DRAM chips,
        a 444-529 PAL and a 444-530 PAL
 
***
Initial Caveat: DO NOT start this project unless you have lots of experience 
soldering and desoldering components from printed circuit boards. It requires 
sucessful completion of around 1500 soldered connections.
***
 
Source of Procedure and parts: Henry Fale, Quildata, Inc, 2618 Penn Circle, 
        Sheboygan, WI 53081, (414) 452-4172. Published in the Zenith Users 
        Group magazine, Remark, June 1991.
 
Step 1: Remove all 36 256KB chips from their sockets
Step 2: Remove all 36 256KB chip sockets from the board by desoldering and
	removing the sockets. If desoldering all pins at once, the socket can 
	be popped off of the pins first, and the pins removed separately. 
 
DANGER: do not simply shove a screwdriver under the sockets to pop them off.
	This WILL scar the board and damage the traces, thus renering the board
	useless without wire patches to correct the broken traces. There is no
	commonly available circuit diagram - you have to figure it out
	yourself. 
 
Step 3: Solder 36 1MB chip sockets into the previously unused pin holes which 
	are adjacent to and overlapping the 256KB pin holes from which you 
	just removed the sockets.
 
Step 4: Move 3 resistors into different positions on the board: 
        R405 and R406 go to positions R404 and R407 (marked on board already)
        R410 goes to R414.
 
Step 5: the PAL in socket U464 (444-558) is removed and replaced with PAL
        444-529
 
Step 6: the PAL in socket U470 (444-559) is removed and replaced with PAL
        444-530
 
Step 7: Get out your owners manual and adjust the switches for the correct 
        amount and location of the memory.
 
Notes: I had a friend taking a course at the local community college who
        originally volunteered for this as part of the learning process. This
        is a nice idea, but not smart on my part. Removing the 256KB sockets
        is a lot of desoldering, and if the equipement is not in good shape
        the sockets are hard to remove. Considerable damage to the traces on
        the board resulted. Another friend, who has considerable experience
        at board-level diagnostics was able to recreate a circuit diagram of 
        the memory chip addressing, given that I had access to the pin-outs
        for the 1 MB chips. This took quite a while, as the board is 4-layer.
        We jointly recommend extreme care in removing the 256KB sockets!!!
 
End Result: I now have a 386/16 with 8MB memory (I already had a 4 MB board
        installed that I paid $2000 for when I bought the PC in 1988).
 
        Windows 3 runs much faster! (it has a good solitaire game)
 
        Using Word Perfect Office Shell 3.0, I can load all of the software
        I want to, and as long as I'm somewhat careful about the order of
        program exit (not swapping), I have no problems. Multiple copies of 
        Advanced Revelation database, Lotus 123, compilers, and Word Perfect
        all load into memory at once for context switching which is much 
        faster that it could possibly be with swap to disk.
 
        Now if only IBM is not lying about the capabilities of OS/2 2.0
        (sigh - fat chance)
 
Fall '92 Addendum:
	I abandoned Windows 3.0 for a "Real" operating system, and put OS/2 2.0
	in it this spring. Except for the slowness (mostly of the disk drives)
	it works fine.

	
Signed: a happy OS/2 user (still not running Windoze software)

chad price
price@helios.unl.edu
price@cse.unl.edu

--
chad
price@helios.unl.edu
cprice@molecular.unmc.edu