VOGONS


First post, by PTherapist

User metadata
Rank Oldbie
Rank
Oldbie

I've been working on my 486 and am constantly running into an error trying to boot Windows 95. I thought I had it sorted, but turning it back on this morning yielded the same result. I've tried googling the issue, but not really getting anywhere so thought I'd try posting here.

Basically the system is setup like this:

XTIDE Universal BIOS installed on an EEPROM on a 3Com Ethernet card (version 2.0.0 beta 3, tried both WITH & WITHOUT the Win95 patches)
Only 1 IDE channel with:
IDE Master = 1GB CF Card
IDE Slave = 120GB HDD

CF card is formatted as a single partition FAT32. The HDD partitions are irrelevant to this issue, as the problem occurs even with the 120GB HDD disconnected.

After doing a clean install of Windows 95 C (also tried with B and had the same issue) it will work fine for some time. Then I'll switch off the PC and come back to it later to be greeted with the following error upon boot:

While initializing device IFSMGR: The Microsoft Installable File System Manager found an incompatible version of the helper driv […]
Show full quote

While initializing device IFSMGR:
The Microsoft Installable File System Manager found an incompatible version
of the helper driver. Please ensure that the correct version of
IFSHLP.SYS has been installed.

Press any key to continue...

And then it shuts down.

If I'm lucky after a few dozen ALT, CTRL, DEL & switch off/on routines it will sometimes decide to boot up fine. But the problem will reoccur at random intervals.

Anybody experienced this with Windows 95 before? I'm suspecting XTIDE to blame so am going to experiment with different versions of the XTIDE BIOS. But just wondering if anybody has any other ideas?

Reply 1 of 7, by Jo22

User metadata
Rank l33t++
Rank
l33t++

CF card is formatted as a single partition FAT32

Okay, so the PC is running one of these new-ish copies of Win95 with DOS 7.1.. 😀
Hm. I pretty much got in conntact with FAT32 on Win98SE first, so I don't know what to suggest, to be honest.
My father usually skipped OEM versions at the time and bought the full retail versions.
That's why I was using Win95ABC not in the 386/486 times really, but rather later on when we got rusty old PCs
from friends and family members, along with scratchy OEM CDs as part of the bundle.
Anyway, I'm going off-topic.. 😅 Maybe the original release of Win95 (RTM) runs better on your PC ?
It's so limited (by comparison) that it lacks advanced features that became popular in the Pentium era.

Edit:

PTherapist wrote on 2020-02-16, 11:37:

Anybody experienced this with Windows 95 before?
I'm suspecting XTIDE to blame so am going to experiment with different versions of the XTIDE BIOS.
But just wondering if anybody has any other ideas?

Hm. XTIDE Universal BIOS switched to another LBA calculation matrix, I suppose.
In the past, that was one of the big issues of Win95 with Dynamic Drive Overlays (DDOs) I remember.
BIOS, DDO and Win95 had different expectations on actual drive geometry which led to data corruption during installation of Win95.
The DOS part was still using the DDO's fake paramerters, while Win95 thought it was smart and thougt it knew better by using its Protected-Mode IDE driver.
Re: SD/CF to ide CHS setup

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 2 of 7, by PTherapist

User metadata
Rank Oldbie
Rank
Oldbie

Well looks like I might have figured out the issue anyway. Won't know for definite until I try it out again tomorrow.

The problem immediately went away when I disabled some RAM caching in the BIOS, so I decided to remove 3x 4MB RAM chips leaving just a single 16MB stick installed. Sure enough, the problem has disappeared for now. The problem reappeared again when I reinstalled 1 of the 4MB SIMMs for testing, so I took it back out.

Seems like this motherboard doesn't like the RAM configuration I was using and it was weirdly only showing up with XTIDE + Win 95.

Reply 4 of 7, by PTherapist

User metadata
Rank Oldbie
Rank
Oldbie

I haven't tested memory properly, other than booting 95 with all but 1 of the SIMMs removed.

I booted the system up this morning with the exact same single RAM stick as I left it working with last night and it wouldn't POST complaining of a memory read error. Switched it off and on and then it did POST but Windows 95 gave me the same ifshlp error as before. Played around with different memory configurations and got it back working again same as last night.

So I've put all the RAM back in and am going to let memtest run for several hours. If there are no errors after a couple of passes, I'm going to just assume that Windows 95 is a piece of crap 🤣.

I had NT 4.0 running on this PC in the past with this exact configuration and it never gave me a problem, so it may just be 95 being flakey.

Reply 5 of 7, by pewpewpew

User metadata
Rank Oldbie
Rank
Oldbie

Sanity check: "it wouldn't POST complaining of a memory read error"

I've hated W95 as much as most, but it can't do that. You've got hardware issues.

Back to What Could Possibly Go Wrong in general, another cause of persistent install errors on known-good hardware is faulty install media. Your W95 floppies or disk may not be as fit as twenty years ago. I'm certainly not.

Reply 6 of 7, by PTherapist

User metadata
Rank Oldbie
Rank
Oldbie
pewpewpew wrote on 2020-02-18, 14:58:

Sanity check: "it wouldn't POST complaining of a memory read error"

I've hated W95 as much as most, but it can't do that. You've got hardware issues.

Back to What Could Possibly Go Wrong in general, another cause of persistent install errors on known-good hardware is faulty install media. Your W95 floppies or disk may not be as fit as twenty years ago. I'm certainly not.

Yeah I wasn't blaming 95 for that particular one. That POST error only happened once so it's possible that the RAM just wasn't seated properly at the time as I haven't seen it do that again since.

But yeah I'm not ruling out issues with the motherboard on this PC as the latest development is that the error will always happen after the PC has been switched off for several hours (or I'll occasionally get no error and it'll either reboot or freeze). Then once it's been switched on for a few minutes and warmed up, I can get 95 to boot perfectly with just 1 single 16MB RAM stick. Any more RAM slots filled and it's a nightmare to get 95 to boot at all.

I disconnected the external lithium battery before I started testing, so I obviously lose all BIOS settings when switched off. But I've noticed that occasionally the BIOS will have some random settings, such as Floppy disabled & Type 10 set for the Secondary Slave. So I might try replacing the old battery and see if it's a lack of working battery that is causing this weirdness.

Reply 7 of 7, by Jo22

User metadata
Rank l33t++
Rank
l33t++

Since the issue seems to be memory related.. :What order does Windows 3.11 use memory?
- Anyway, not sure if it is of any relevance here. It just came to mind. 😐

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//