VOGONS


Vectra 4/66 cpu upgrade problems

Topic actions

First post, by jm8881

User metadata
Rank Newbie
Rank
Newbie

I was trying initially to overclock the dx2 66 in the system to 100mhz and none of the jumpers seemed to increase the speed of the processor. Changing the multiplier didn't have any effect but I could clock it down to 50mhz. I then ordered a dx4 100 and although it "works" nothing detects it as the correct cpu. When the system does its initial tests it shows up as a dx2 66. Cpu id in dosbench lists it as a 100mhz processor with no cpu id. I can change the speed of it via jumpers to be 50, 66, 75, 100 mhz but that's it. Changing jumper jp16 (dx2 or dx4) does seemingly nothing. And with the dx4 100 installed when you get into windows 95 you immediately get an error about not being able to launch explorer.exe and windows needs to be reinstalled.

Should I upgrade the bios from gp.05.05 to gp.05.07? I'm afraid of anything failing and bricking the box.

The dx2 66 is write back and the dx4 100 is write through. Would that be a problem?

Finding information on this pc is really tough 🙁

Reply 1 of 31, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

The DX2 is, in modern terminology, multiplier-locked. It always runs at 2x. From what you're saying it sounds like your board only supports bus speeds of 25 MHz or 33 MHz. Had there been a 40 MHz setting, you could have used that to overclock the DX2. The DX4, I believe, can either run at 2x or 3x depending on the jumper which would explain the speeds you see.

It isn't surprising that the BIOS mis-identifies it. These old BIOSes determine the CPU type through a combination of a value that is in the DX register at startup, and by benchmarking the cpu to determine its frequency. If the max value it knows about is 66 MHz that is what it will report the speed as. One of my boards reports a 5x86-133 as an Enhanced DX4-100 for example. The CPU still runs at the correct speed.

Are you sure the DX4 you put in is one that can handle 5 volts? i.e., it is one on a regulator board. If not, you might be feeding 5 volts to a 3.3 volt CPU, so it isn't surprising there would be severe stability problems.

Maybe you can see the release notes for the later BIOS? Or link to the old and new versions and we can look at the cpu table perhaps. Could you also take a picture of your board.

Reply 2 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie

Both the dx2 66 and dx4 100 are amd cpus and both say 3volt on them.

The release notes for the new bios are:

BIOS REVISION INFORMATION
*************************

Platform: HP VECTRA VE 486
Version: GP.05.07
Release Date: August, 1996
Previous Release: GP.05.06

This BIOS is available in US language only

Differences from the GP.05.06 version
=====================================

Bug fixes:

- Fix issue for memory allocation conflict between ISA Legacy
and ISA PnP or PCI card.
- Change value of read data port to prevent problems with PCI
cards with a joystick port.
- Implement final fix to support Western Digital Sirocco2 hard
disk drives
- Implement fix for second level cache.
- Added software workaround to cope with minor hardware bug on
specific AMD processor chips. Chips are CPU A80486DXx-xxxSV8B
date code C9529xxx (and later) and S80486DXx-xxxSV8B date
code C9530xxx (and later).
- Fix improper DMA type recognition algorithm.
- Improves handling of ISA legacy option ROMs.
- Fix detection and configuration issues for DEC205 ISA card.
- Fix ressource conflict between video and add-on ISA cards
by changing the minimal address for the high dos area address
(c0000 instead of a0000). This reserves the area a0000-bffff
for the video.
- Fix to properly initiate and support Intel EtherExpress card
- Fix the year 2000 transition bug.
- Fix error detection with P24T/83MHz processor. The 83MHz
overdrive processor is now properly identified at power-on.

New features
============

- Added power-on test of the NVS (memory storing ISA cards
configurations).
- Add a feature to detect the S3 video chipset revision in order
to display at the POST screen "Integrated (PnP) VGA"

To flash this BIOS you must use version of HPINIT AP.04.02 or later.

******************************************************************************


I cannot find the .05.06 or .05.05 bios release notes.

Reply 3 of 31, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

ok. Looks like JP8 is 1-2 for write-back and 2-3 for write-through; did you change it to 2-3? Write-through L1 is a downgrade BTW.
I do see references to 486DX4 CPUs in that latest BIOS. So perhaps try it.
I bet you could also run an Am5x86-133 on this board if you get adventurous. Those are all Write-back.

Reply 4 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie

I did change the cache jumper. It hangs at the initial tests if you leave it on 1-2.

I've read a few articles that say write through is faster than write back? I can run the full gambit of dosbench on the dx4 chip through dos and it scores higher in everything than the dx2 chip.

I'm currently getting ready to flash the bios to the latest one. I'll let you know how it goes. Thanks for the help so far.

Reply 5 of 31, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

Well L1 (internal cache) and L2 (external cache) can both be write-back or write-through.
It's true (and I have such a board) that sometimes the write-back implementation of L2 isn't good.

But WB L1 got added to the last batch of 486es for a reason, as a performance boost,... by the way, I see in past posts you were running into a roadblock with trying to add any L2 cache, so I will again plug the Am5x86 as it doubles the L1 cache.

You do have a heatsink right? Otherwise yeah, it's mysterious how it crashes going into windows, especially with no L2 cache which is often where problems are. I wonder if the DX4 could be bad?

Reply 6 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie

It does have a heatsink.

I flashed the bios and everything went well except for the fact that it doesn't recognize my cf card to ide adapter anymore and it still detects the cpu as a dx2 66.

It does detect an ide hdd though.

Reply 7 of 31, by Horun

User metadata
Rank l33t++
Rank
l33t++

Curious...Is your HP a Vectra VE 4/66 or a Vectra N2 4/66 ? or some other ?
Here are a couple manuals that might help...but think you maybe already have more info than they contain...

Attachments

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. https://archive.org/details/@horun

Reply 8 of 31, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie

Maybe try clearing your CMOS since you upgraded the bios?
Would you also have an "Enhanced" AMD 486DX4-100? In this process you're switching from an "enhanced" DX2-66 to a "regular" DX4-100. I don't know if the "enhanced" version would do you any better (or, just get the 5x86 😜 )

Reply 10 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie
jakethompson1 wrote on 2020-11-01, 04:03:

Maybe try clearing your CMOS since you upgraded the bios?
Would you also have an "Enhanced" AMD 486DX4-100? In this process you're switching from an "enhanced" DX2-66 to a "regular" DX4-100. I don't know if the "enhanced" version would do you any better (or, just get the 5x86 😜 )

I have cleared cmos already.

I only have this one chip. I'm looking to find a copy of the gp.05.05 bios now since my cf card adapter stopped working after the bios "upgrade". sigh. I just wanted to play some duke nukem 3d.

Reply 12 of 31, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie
jm8881 wrote on 2020-11-01, 19:20:

Small update, the cf card adapter only works as a slave now. It works fine as master in my other computers.

That's odd. A lot of BIOS Flashing programs make a backup of the old bios but unfortunately it sounds like this HP one doesn't. How big is the CF card you are using? I wonder if the later BIOS adds broken support for bigger hard drives that the old one was oblivious to. Did the Win95 problems go away? I also wonder if the BIOS would work better with a CPU that supports CPUID.

Reply 13 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie

I have been using the dx2 66 since the cf card stopped working. Also the cf card works as a slave about 1 out of 10 times. Usually after posting i just get a blinking cursor on a black screen.

Also I have had no luck finding the older bios online, even through the way back machine.

Reply 14 of 31, by jakethompson1

User metadata
Rank Oldbie
Rank
Oldbie
jm8881 wrote on 2020-11-01, 20:24:

I have been using the dx2 66 since the cf card stopped working. Also the cf card works as a slave about 1 out of 10 times. Usually after posting i just get a blinking cursor on a black screen.

Also I have had no luck finding the older bios online, even through the way back machine.

Weird. If repartitioning, reformatting and reinstalling doesn't work (as the new BIOS may have changed the translated disk geometry) you may be able to get out of this mess with XT-IDE BIOS which would also lift all hard drive size limits.

Reply 15 of 31, by rmay635703

User metadata
Rank Oldbie
Rank
Oldbie
jm8881 wrote on 2020-11-01, 04:36:
Horun wrote on 2020-11-01, 03:53:

Curious...Is your HP a Vectra VE 4/66 or a Vectra N2 4/66 ? or some other ?

It's the vectra ve 4/66

It’s very strange that a 486 made that late in the game can’t detect a DX4

HP Vectra VE 4/66

I guess your not the only one with problems with this model, oddly there are some manuals and other blurbs on those irritating Russian sites and they list jumpers for DX4, so really not sure what’s going on

Reply 16 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie
jakethompson1 wrote on 2020-11-01, 22:14:
jm8881 wrote on 2020-11-01, 20:24:

I have been using the dx2 66 since the cf card stopped working. Also the cf card works as a slave about 1 out of 10 times. Usually after posting i just get a blinking cursor on a black screen.

Also I have had no luck finding the older bios online, even through the way back machine.

Weird. If repartitioning, reformatting and reinstalling doesn't work (as the new BIOS may have changed the translated disk geometry) you may be able to get out of this mess with XT-IDE BIOS which would also lift all hard drive size limits.

Are you talking about this?

https://www.ebay.com/itm/NEW-XTIDE-XT-IDE-OPT … ZfQ9ie&LH_BIN=1

Ok I bought it. Should have it by friday.

Reply 17 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie

Just a little update, I've been using the dx4 100 since shortly after this post and it does work. Everything detects it as a 100mhz cpu but nothing can detect it's cpu id. It is measurably faster than the dx2 66. But something about it not detecting it's cpu id has gotten to me and I've started looking into it more.

jakethompson1 mentioned about the cpu being the "enhanced" version. The dx2 66 was model number Am486DX2-66SV8B which is enhanced while the dx4 is Am486DX4-100NV8T which isn't enhanced. I've ordered a Am486DX4-100SV8B so maybe being the enhanced version will solve my issues. These vectras seem to have a very specific upgrade path and compatibility suffers.

edit: also TimmermanV, if you read this I cannot reply to private messages since my post count is so low and it's probably too late since your pm is a couple months old, but I would be interested if you still have it.

Reply 18 of 31, by Intel486dx33

User metadata
Rank l33t
Rank
l33t

On a Windows-10 computer download and extract this zip file.
https://www.philscomputerlab.com/dos-benchmark-pack.html

Burn contents to a CD-R DIsk.

Put in your 486 and and run these benchmarks.
Speedsys
CheckCPU
TopBench

Report what they read ?

Reply 19 of 31, by jm8881

User metadata
Rank Newbie
Rank
Newbie

Speedsys:

Processor: 80486dx 100mhz
cpuid not support

Processor benchmark: 34.77

Topbench: 164/165 (does it ever end?)

Checkcpu:

Cpu vendor and model: 'classic' 486
Internal cpu speed: 100.7mhz
Cpu-id vendor string: -no cpuid-
Current cpu mode: real
Internal (l1) cache: enabled in write through mode

During startup its built in test lists the CPU as a 486 dx2 66.