• Please review our updated Terms and Rules here

How to adjust the speed of an 8088-2 CPU?

T-Squared

T-Squared

Veteran Member
Joined
May 29, 2011
Messages
657
Location
San Antonio, TX
Since I have gotten my Sanyo MBC-775 working, I had considered running 8088MPH to see if it would work, despite the CPU in the system not being an 8088, but an 8088-2.

(I'm putting the IRCjr attempt on hold, for now, if you read my previous topic about the Sanyo.)

Indeed, the pre-demo test says that the measurement is 40% above the normal 4.77mhz, and I didn't want to risk burning out the internal picture tube from overdrive, given the warning in the pre-test.

Is there a way, either through hardware jumpers or software configuration, to change the speed from 8MHz/7.16MHz to 4.77?
 
Pardon my ignorance, but how can one damage the CRT by running the CPU too fast?

On the old 8088 systems, the old Landmark Speed test is a decent "quick and dirty" check on CPU speed.

Most 8088 "turbo" systems had some sort of multi-keystroke capability to reduce the speed to 4.77MHz for compatibility reasons.
 
I think it's more the timing of the video than anything else. Here's what the disclaimer states in the pre-test:
 

Attachments

  • 20201009_132229.jpg
    20201009_132229.jpg
    254.8 KB · Views: 2
Several of the blog posts about it reiterate that claim, but I'm not smart enough to say exactly what they might be referring to. There are several parts of the demo where they "race the beam" and bang on the 6845's registers in various ways that might do "something". My impression was that damaging a monitor by sending it wonky sync signals was more a problem with MDA cards than CGA because the IBM 5151 monitor strictly speaking doesn't "sync", it's more of a direct drive, but maybe it's possible to murder a CGA monitor too, no idea.

Strictly speaking I don't think it's the tube that would be at risk, but the high voltage power supply that handles the sweep.
 
The PC Magazine review indicates that a compatibility slow down was supposed to be added to the MBC-775. Are there any readmes on the disks that came with the system? I would suspect if Sanyo provided the ability, it would be using software so checking the disks and seeing if any unusual programs are provided would be a necessary step. While using names like SLOW or TURBO was common, Sanyo did have a tendency to do things their own way.
 
I tried Wikipedia's suggestion to use Ctrl-Shift-(+) or Ctrl-Shift-(-) or Ctrl-Shift-(?), but they didn't seem to work.

Furthermore, the Landmark benchmark shows a CPU speed of 7.995MHz.

krebizfan said:
Are there any readmes on the disks that came with the system?
Click to expand...

These systems were picked up from a lot sale, and the casings were in pretty rough shape. They didn't come with any disks, unfortunately. I've been looking for Sanyo MBC manuals, but I haven't found any yet. (Plus I'm using a different 8088 BIOS, a Phoenix Technologies BIOS, for compatibility's sake.)
 
Swapping to a third party BIOS would almost certainly neuter any kind of key combination speed switching unless it’d been specifically built for the machine. Presumably speed is controlled by twiddling a bit on an I/O port, but without the original software or a tech manual...
 
Here's some teledisk-format archives of some Sanyo disks:

http://www.retroarchive.org/maslin/disks/sanyo/

Maybe there's some kind of speed setting program on them.

Out of curiosity, how compatible does the machine seem to be with other graphics software? There's an old review of the system out there that says it wouldn't run MS Flight Simulator (a common target for PC compatibility) and mentions vaguely there may have been other cases where software could be heard running in the background but with no video. If that indicates that the CGA implementation isn't register compatible then 8088MPH will almost certainly faceplant right out of the gate.
 
Eudimorphodon said:
Swapping to a third party BIOS would almost certainly neuter any kind of key combination speed switching unless it’d been specifically built for the machine.
Click to expand...

I had considered that too. It's a bit difficult to use Sanyo's v2.33 BIOS, though, with how buggy Sanyo's attempt at the IBM BIOS was; at least how I've heard it.
 
According to my 885 manual, there's a small alternate-action pushbutton switch near the right lower area of the expansion card area (looking at the rear head-on). There's also the CLKCHK command in the Sanyo DOS distribution that indicates the speed of the CPU.

Sorry for chiming in; I know you wanted to ignore me. :)
 
Chuck(G) said:
According to my 885 manual, there's a small alternate-action pushbutton switch near the right lower area of the expansion card area (looking at the rear head-on).
Click to expand...

Is it blue? Because the other board I got with the non-working Sanyo has JUST that!

Edit: AHA! IT IS BLUE! I have a clock reading of 4.655 MHz on the other one! I was suspecting that!

I got two versions of the same board! One has the switch, the other does not, but has a place FOR the switch!
 
Last edited:
FWIW, the 885 is the 10MB hard disk version of your system. :) As far as I'm aware, there's no software way to change the system speed. That button is it.
 
Since you were going to run 8088 MPH: I don't know much about the MBC-775, but it looks like the internal CRT shows RGB video - does this thing have composite video out?

If so, it'd be nice to know how close the colors are (or aren't) to the expected IBM CGA output. :)
 
Yes, it does have composite out. And for games like Space Quest I & II, and Burgertime, the composite artifact colors work like a dream. (Space Quest clued me in when I noticed the vertical patterns of lines)

Check back here later. I'll post some pictures to show the colors.
 
Eudimorphodon said:
Several of the blog posts about it reiterate that claim, but I'm not smart enough to say exactly what they might be referring to. There are several parts of the demo where they "race the beam" and bang on the 6845's registers in various ways that might do "something". My impression was that damaging a monitor by sending it wonky sync signals was more a problem with MDA cards than CGA because the IBM 5151 monitor strictly speaking doesn't "sync", it's more of a direct drive, but maybe it's possible to murder a CGA monitor too, no idea.

Strictly speaking I don't think it's the tube that would be at risk, but the high voltage power supply that handles the sweep.
Click to expand...

When we wrote 8088 MPH, we weren't sure that wonky sync signals couldn't damage a CGA monitor. I've since learned that (as you said) it's really not an issue for CGA monitors so I think it's perfectly safe for T-Square to go ahead and try running the demo at the higher speed. It'll probably just lose vertical sync in some parts (I don't think it will even lose horizontal sync).
 
No, it will read the same. The -2 simply means that the CPU has passed the selection criteria for operation at 4.77MHz. It's still the same die. There can be other implementation details that confuse the speed computation. If you really want to know the CPU clock, stick a good 'scope on the clock input to the CPU and measure it directly. If you're relying some software, you're just goofing around.
 
T-Squared said:
Here's how the demo turned out. I wonder if an original 8088 CPU would work properly. The 8088-2 doesn't seem to lock on the proper 4.77MHz signal in slow mode.

8088 on a Sanyo MBC-775
Click to expand...

Possible clock discrepancies aside: did you try playing with the settings on the initial calibration screen some more? in particular changing "CGA model" to "new"?
 
You know, I may have found my problems. Would bad RAM cause the discrepancies found in the demo?

In fact, both 8088MPH (Look at the text) and Space Quest had some cut-out portions in the graphics:

20201010_105626 copy.jpg

and several programs would either not start, or cause the system to hang, despite their size being able to fit within the 256 KB of memory.

RAMTEST even found a bad chip in one of the memory banks of the processor board I'm using at the moment.
 
You must log in or register to reply here.
Back
Top