XTs and large hard disks

[Jim Leonard]

Wolfe, Julian wrote:
> 1. Does anyone know a good solution for putting a large drive in an XT,
> while still leaving it an XT?

IDE with "540MB" (actual capacity is less) drives or less, if you can find an 
IDE card that works.  I have two (different brands) but sadly I haven't found 
time to try them yet in my model 5150.  Only other solution is SCSI, something 
I have had extremely unlucky progress with (boards with BIOSes won't POST; 
boards that need drivers don't have the damn drivers bundled with the board; 
etc.)  However, I believe my experience with 8-bit SCSI on XT is uncommon 
compared with the other gentlemen here.

> 2. Should I leave the V20 in?  I've heard it has compatibility issues with
> some programs, but I'm not solid on my information.

The V20 can enable some ill-behaved programs to work as it implements the "SHL 
register,immediate" 80186 opcode, and this opcode is sometimes the *only* used 
opcode used in "286-only" programs.  I was able to run programs claiming they 
needed a 286 on my AT&T PC 6300 (Olivetti M24 8086 clone) after I upgraded it 
to an NEC V30.

HOWEVER, the V20 and V30 expand the prefetch queue from 4 bytes to 6 bytes (the 
source of their speedup) and this screws with timing-sensitive programs such as 
games, copy-protected diskette copiers, and turnkey stuff like robot 
controllers, etc. from the early 1980s.

I'm going to go out on a limb and assume that you're expanding your convertable 
so that it can play most IBM PC games from the 1980s.  :-)  Based on my 
experience in that context, the V20 will cause the following behavior compared 
to a factory original 8088:

- 88% of games will run just fine, unaffected by the V20
- 5% of games will run smoother/better due to the small speedup of the V20
- 5% of games will run too fast to be enjoyable
- 2% of games will not run at all

Games are a good example to flush out this kind of behavior because they were 
usually programmed as close to the hardware as possible to achieve the best 
performance; also, many copy-protection schemes are timing-based and may fail 
if the CPU speed is not what is expected.

If I have incorrectly assumed your intentions, please let me know as I'm dying 
with curiousity :-)
-- 
Jim Leonard (trixter at oldskool.org)                    http://www.oldskool.org/
Want to help an ambitious games project?             http://www.mobygames.com/
Or check out some trippy MindCandy at             http://www.mindcandydvd.com/