Dave-H

Well I compared the INF folders from the working Win98 and the non-working Win98. As there was over 1000 files in each, I restricted the search to *.inf files (ignoring the *.pnf files) which contained the text "storage". I identified a few possible culprits. There were quite a few inf files from my HP photo printer, which contains card readers. There was jmusbsto.inf, which came from a USB-SATA interface. There was UFD.inf, a Microsoft mass storage driver. Also USBZIP.inf, which goes with my USB Iomega ZIP drive. I disabled all of those by renaming them, but no difference, still a system lock-up on the installation of the second drive. I also tried using the autoexec.bat and config.sys files from the working system in the non-working system, and that made no difference either.

@herbalist Thanks very much for that feedback. I think I've now confirmed that there is actually nothing wrong with NUSB. It's something peculiar to my system. @dencorso Well, I tried installing NUSB on my "clean" Windows 98, and it worked perfectly! I installed both my card readers, the original and the new one, and both the HP and Integral memory sticks. All worked with no problem at all. So, at least I now know that the problem isn't with NUSB. I did save the USBSTOR, USB, and complete ENUM registry keys from the working system. I'm wondering whether to try importing at least the first two into the faulty system to see what happens......

That's good to know! Done all that. The second update (the MS one) didn't seem to replace any system files, so I may well have had it installed already. The first one updated and added quite a few. Unfortunately, none of it made a scrap of difference. First "USB Disk" mounted fine, second one, immediate system freeze. Anyway, here's what Chip Genius made of the card reader. I'm wondering if there would be anything to gain by firing up with the clean new Windows 98 install that I did on a separate drive, and try installing NUSB on that. If the card reader works on that, at least we'll know that it's not an intrinsic problem with NUSB and my card reader. What do you think of that idea Den?

OK, I've done all that. Installed NUSB, rebooted and it found my USB 2 motherboard hardware and successfully installed drivers for it. Restarted OK. Backed up the registry again. Checked in HP System Diagnostics for any unconnected devices and removed them. Still restarted OK. Backed up the registry again. Put my HP single drive memory stick in. Drivers installed correctly and drive appeared in "My Computer". Restarted, still OK. Backed up the registry again. Put my five drive card reader in. Mass storage device found and installed OK. First USB Drive found and apparently installed OK. Second USB drive found, and..........immediate system lock-up! So, I'm back exactly as before, after all that! NUSB will install one drive, but freezes the system if it tries to install any more. One clue which might be important. On one occasion when the system froze, before the keyboard stopped responding I did manage to hit Ctrl-Alt-Delete and the close program box did come up. It showed that msgsvr32 was not responding. Looking at what this file does, I'm wondering if that's what's crashing and bringing the system down on the second drive install. The version I have is 4.10.0.2222, which seems to be reasonable. I don't need a later version for this to work do I, the Windows ME version perhaps? What I would like is confirmation from someone with NUSB installed that it will indeed mount more than one drive from the same device. If it's an intrinsic problem with NUSB then I'll have to abandon it and try something else.

Thanks Den! Here's my Device Manager. It does look the same in Safe Mode, so there are no hidden devices lurking. The "PCI Universal Serial Bus" in the Unused Devices section is the USB 2 hub, which NUSB will provide support for. If you're happy with this, I'll do a final backup, go ahead with the NUSB install, and we'll see what happens!

OK, I think the system's as clean as it's ever going to be of spurious junk, both in the registry and the files. I'm thinking of just reinstalling NUSB now and hoping for the best. Is it safe to leave my other USB devices (listed in the previous post) or should I disconnect, uninstall, and clean the system of their files and registry entries before I try NUSB again? None of them are mass storage devices, so my gut feeling is that they will make no difference. Any advice appreciated.

@stuck Ah, I thought that would probably be the case, but thanks very much for the thought anyway! @dencorso Thanks yet again Den. It appears the the waters were extremely muddied by the fact that Realtek posted a driver that doesn't actually work in Windows 98 (not on my system anyway) even though the download page clearly says that it supports all Windows versions from 98 Gold to XP! The second fault condition, with the error in the Device Manager resources tab, was caused by the faulty driver. The original fault, where there was no error in DM, but just no sound, seems to have gone away now I have cleaned out my system. For anyone else with a Realtek AC'97 sound system and Windows 98, beware! Version 4.00 works, version 4.06 (the latest posted) doesn't. I have informed Realtek of this, I'm not holding my breath while waiting for a reply! So, I am now going through my system cleaning the registry of all the obsolete junk left by the "dummy" clean Windows 98 installation. I now have very many fewer devices registered in the system. The only USB devices there are the ones that are actually part of the motherboard, plus my mouse, which is connected via a hub built into the keyboard, my Epson printer, and my Wacom drawing tablet. They are all working fine. All other devices, including all the pen drives and card readers, are gone, and I have been through and deleted all their associated files too. So, when I'm happy that the system is as clean as possible (and fully backed up!), are we ready to try installing NUSB again (it seems so long ago.........)

Well, I'm pleased to be able to report significant progress! I did as you suggested Den, and ran up the system with the System32\Drivers folder files and system.dat file from the working installation. The system started OK, and the sound worked! I then tried just restoring the original system.dat from the non-working installation. The resource allocation error on the sound hardware in Device Manager did not reappear, all seemed to be well, but the sound was not working. This was the interim condition that I had before, where the sound failed but with no apparent device errors. I then tried reinstalling the driver for the sound, using the latest version, and the error reappeared in Device Manager! I suspect this is what happened before, and the device error actually first appeared when I tried reinstalling the driver. Uninstalled the driver, even used Driver Cleaner on it, reinstalled the old driver from the motherboard installation CD (which I had used on the dummy Windows 98 install) still no joy. Went back to the "working" system.dat, and it all came good again. This is still using the very old version of the sound driver from the motherboard installation CD. I then wondered if the latest driver in fact doesn't work with Windows 98, despite the fact that the download page says that it does. http://218.210.127.131/downloads/downloads...p;GetDown=false Anyway, I'll look at that later. I then merged the SOFTWARE key from the non-working system into the working system. That still worked, apart from a few error messages on start-up about missing files. That I fixed by merging the [HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\VxD] key from the non-working to the working system. So, I think I'm nearly back to a system which is getting close to as it was before, but with working sound! I've still got a few drivers to sort out, and I'll have to check that all my software applictaions are still working properly, but given time that can and will be sorted. Of course, that still hasn't answered the question as to what caused the problem in the first place! Until I know that, I'll always be worrying that it will happen again. At least if it does, I will have a full backup next time, I promise! I can't believe that it is a problem with the sound driver. I only went to the latest version after the problem had already appeared, and the installed version had been working for at least a year without any problems. Anyway, I'll plug on getting the system back to "normal" and then we can go back to trying to get my card reader working (remember that?)! Thank you all so much.

Thanks Den! Tried all that. The first attempt was a non-starter because I'd installed the working Windows 98 to a different folder to my normal installation (C:\Windows instead of C:\WIN-98). I therefore had to do the whole re-installation of Windows 98 to the new drive again. Did I say "D'oh!" before? Well I'll say it again! Anyway, after all that, and making the working installation as identical as possible to the faulty installation, I tried what you suggested. It all worked, and the system did start OK with the system.dat file from the working installation, but, exactly the same result! Still a resource error on the sound hardware. Nothing had changed at all in Device Manager. So, what now? I think this just confirms that it isn't a problem with anything in the registry.

I'm afraid I didn't have a lot of luck with regedit in DOS. It seemed to work OK to delete a key, but I couldn't import the key I wanted to import from the other Windows 98 installation with working sound. It went through the motions of importing it, with a count from zero up to 100%, but as soon as it got to 100% a message came up saying - "Couldn't import enum.reg: Error accessing the registry." I made sure that system.dat and user.dat had their read-only and hidden attributes off, but it made no difference. It certainly deleted the enum key OK, judging by the mess I was greeted with when I rebooted, but it would not import the replacement key that I wanted it to. Presumably it will import registry data from another Windows 98 installation? The only thing I could think of that might have made the new data incompatible was that the new installation was installed to the default C:\Windows folder, and my original is in C:\WIN-98. I can't see any path specific data in that registry key though. So, I gave up on that, and tried again just removing the enum and config keys and letting everything install again. I also this time deleted the two .bin files from the INF folder. After a huge amount of messing around I was back where I was before. The system absolutely refuses to allocate a working configuration to the sound hardware. I even un-ticked the "automatic" option on its resources tab, and then found that it was now apparently on IRQ 11. I had written down the resources being used on the working system, which was also IRQ 11, but some of the Memory Ranges and I/O Ranges were different. I changed them so the information in the resources tab was absolutely identical to that on the working system, and still the drivers refused to load and no sound! All the other hardware can be made to work fine. I really am puzzled completely by this now. I think that we're perhaps looking in the wrong place by playing with the registry. If you remember, I did restore a very old copy of the registry, from long before this problem appeared, and nothing at all changed. I suspect that the problem isn't being caused by anything in the registry in that case. We know now that it's not an actual hardware fault, as it works fine with the new installation. So where does that leave us? Surely it can only now be a configuration error somewhere that isn't stored in the registry, or a system file problem. Are there any files associated with NUSB that could conceivably cause this sort of problem? I think I've got rid of them all now and restored the originals of any that NUSB replaced, but you never know, I could well have missed one or more!

Thanks Den! I will do that thing. Just one quick question. The links you gave say that the version of regedit.exe that came with Windows 98 is buggy and should be replaced with a "fixed" version. I actually have the version which came from ME, 4.90.0.3000. I can't remember exactly where it came from, possibly soporific's Auto-Patcher. Is that OK to use for this exercise? BTW, I completely forgot that Partition Manager would have done the necessary formatting for me!

It's gone very quiet............. Anyway, I have news to report! I finally bit the bullet and tried making a completely new install of Windows 98SE. This was not without its problems, needless to say. Fortunately I have IDE drives in removable cradles, so that bit was quite easy physically. I disconnected my system drives (which are SCSI) and booted up using a boot disk. I then found that Windows wouldn't install on the drive, because of course it wasn't a primary active partition. D'oh! Once I'd sorted that out with Fdisk, and then spent ages reformatting the drive, it worked. It went through the setup process until the last restart, and then said "not enough memory". Of course what it actually meant was "too much memory"! I have 4GB of RAM fitted. Double d'oh! Fortunately I still had the original two 512MB DIMMs from before I upgraded, so I had to take the machine even more apart, and put them back in, praying that setup would complete with 1GB of RAM. Fortunately, it did. I've now told it to limit the memory to 512MB, so hopefully it will still start now without me having to physically swap the memory again. I then put the Supermicro setup CD in, and installed just the sound drivers (an ancient version, but would prove the point). The sound worked! I then installed the Intel chipset drivers, which includes the USB hardware (I made sure that no USB devices were actually connected during all this of course.) They installed fine, and the sound still worked. I then installed the drivers for the ethernet controller. That eventually also worked fine after a bit of messing around, and the sound still worked! So, the only things on the PCI bus without drivers were the modem, my Canopus video capture card, and the USB 2 controller, which has no Windows 98 drivers of course. I didn't think it was worth installing the first two's drivers, as I think the point had been proved. At least I now know that this is a configuration problem, there is nothing wrong on the motherboard, which is a relief. I have a few files to post, if anyone would like to look at them. This contains the Device Manager "printout" plus a save of the System Information from the working Windows 98. I also saved the Enum key from the registry of the working system. Dare I put it into the non-working system?

Dave, in case you decide to follow Multibooter's idea, which may in fact help, I've decided to quote myself from elsewhere, just to point you to the right version of it: However, I'm begining to think that you ought to image your system or even put in a spare HDD in place and do a clean install of 98, using RegShot before and after the audio driver's installation, for us to find out which entries are created in the register and which files are needed and where do they go. It would also require the creation of a full file list before and after the sound drivers installation, so it must be planned carefully beforehand to be effective. This is a really time consuming task, but might be the way to go in case all else fails. Thanks for that Den and Multibooter. Tried using Driver Cleaner, but it made no difference. The only thing I didn't do was actually physically disabling the sound hardware, as that involves taking the machine to bits again. I will try again when I've got it apart, if it comes to that again! Swapping drives and installing Windows 98 from scratch are my very last resort. I really hope that it doesn't come to that!

Tried that, and the COM1 port and the LPT1 parallel printer port were put back by the plug and play system even though they were disabled in the BIOS! Didn't make any difference to the sound problem. I did update the BIOS version ages ago acting on the the instructions of Supermicro technical support to resolve a problem with Seagate hard drives. I ran the update again recently just in case it had any bearing on this problem, but it made no difference. Don't worry! That's exactly how I resolved the problem! Tried again, but still no joy! Created a bootable floppy, copied HWINFO.EXE and HWINFO.DAT to the floppy (I assume they're the only necessary files, any more would have been too big for the disk!) When I booted to the floppy and ran HWINFO, it just said "Program is too big to fit in memory". Tried copying himem.sys to the disk, and creating a config.sys on the disk just with "device=a:\himem.sys /testmem:off" in it, and it still threw up the same error message. Yes, well I might try that next! I'd like to know too! That option doesn't appear in my BIOS settings. I have the latest version (1.3b)

Good thought Multibooter, but even with the audio drivers uninstalled, the audio hardware still appears in the Device Manager as an unknown device, with the same resource problem, i.e. no IRQ (IRQ 00). I will give it a try though once I get to the end of all the other avenues I'm currently following!

Thanks stuck, I appreciate that. Are you sure that the Jessops card reader you have is the same as the one I'm trying to get working in Windows 98? The reason I ask is because I also have an older Jessops multi-card reader, which they did supply Windows 98 drivers for. That is a four port box with "Universal Card Reader" on it, and it has to be connected with a separate USB cable. The new one is a five port box and has "All in One" on it, and the USB connector is built into the box and swivels out. The difference is that the the older reader won't read SD HC cards, and the new one will. I'm pretty sure that no manufacturer's driver for Windows 98 exists for the new reader, at least not downloadable from the Jessops site where I downloaded the driver for the older reader. Cheers, Dave.

@dencorso Thanks Den! That's actually a later version than the printed manual I've got (which is revision 1.1a)! I will follow up your suggestions when I've done all the analysis with Charlotte, as I'm hoping that might show some way forward without having to open the machine up again! @CharlotteTheHarlot Well, I managed to get the information out of Device Manager. PRN files, I never want to see one ever again! I managed to convert it to plain text (don't ask!) and here it is. I'm still having no joy with HWinfo. I'm a bit puzzled when you talk about floppies. I've been running it in true DOS from the hard drive. Is that not the way to use it (the documentation is silent as to how to actually deploy the program)? You mention copying himem.sys to the floppy. Are you talking about a bootable floppy? Which config.sys do I modify, the one on the hard drive, in which case why copy himem.sys to a floppy, or a copy of it on a bootable floppy? Sorry, I'm confused here.......... I tried disabling the USB functions in the BIOS, which you can do (as Den says, the only way to disable the sound hardware is by physically pulling a jumper.) This didn't bring the sound hardware to life. I did notice that in this condition I now had 4 disabled "APCI IRQ Holder for PCI IRQ Steering" devices, number 3, 4, 7 and 8. With the USB enabled only one (7) is disabled. At least that has shown me that they are disabled because of a problem elsewhere, not that they are causing the problem by being disabled (if you follow me!) If you can point me to getting HWinfo running properly, I'll let you know its findings so you can compare with what Device Manager says. If we don't find anything, then it's off with the covers again and we'll follow Den's suggestions.

Ah yes, I realised after I posted that I'd forgotten about the modem! I don't actually use it now of course, but I've always kept it there as a backup in case my broadband fails (it actually did once, and I had to use the modem again, I'd forgotten how painfully slow dial-up was!) There is no on-board video hardware on the motherboard, just ethernet and audio. There has always been a large number of things sharing IRQ11 on Windows 98, but it's always worked OK. I will try HWinfo again, using the settings you suggest. I have updated the BIOS version to the latest, but I assume that doesn't change the routing tables. I think I may have discovered what caused that BIOS settings reset. I used my USB ZIP drive yesterday for the first time in ages, and it worked fine, but I rebooted without unplugging it. When the system booted up, the BIOS correctly detected the drive after a bit of a pause, and when I rebooted again, the BIOS had reset again! I think what may be happening is that every time the BIOS detects a new USB device connected to it, it resets itself. I've no idea why it should do that, but the first reset was probably caused by my rebooting with my USB flash drive plugged in. It therefore probably has no bearing on the sound problem. I will try booting with all USB functions disabled in the BIOS, to see if that has any effect. There are no reserved resources. I'll report back on all this later, and I'll attach that readout from Device Manager. Thanks Charlotte!

I've run the Craig Hart utility successfully, and I've attached the logs as a zip file. There are three logs, one is PCI logging in true DOS, one is PCI logging using a DOS prompt with Windows 98 running, and one is PCI32 logging at a Windows 2000 command prompt. I was glad to see that to my untrained eye there seemed to be no obvious problems recorded. I didn't have so much luck with the HWinfo utility. I just kept crashing and/or throwing up "not enough memory" messages (this was in true DOS). I hope these logs may show something up. Certainly it seems to be happy with the IRQ tables, which is a relief!

Thanks again Charlotte! I tried your first suggestion, saving and deleting the two .bin files in the inf folder. When I re-booted the system started up normally, which surprised me. I assume that I should have removed the problem device first, which you didn't mention! Anyway, I did that, and it was put back exactly as before. I then tried removing all the devices on the PCI bus, and the steering devices and the PCI bus itself. When I restarted it went through all the motions of installing everything again, but the sound card still has no IRQ. There are now two new bin files in the inf folder. Should I just keep them, or restore the ones I saved? Incidentally, in case you didn't pick this up from earlier, the sound "card" and the ethernet controller (which is working fine) are part of the motherboard, so swapping cards around, at least as far as they are concerned, isn't possible. The other things actually in the slots are the AGP graphics card, a SCSI card, and a DV video capture card. They are all working perfectly. So, does this leave us with a hardware fault? I find that very hard to believe as the sound on Windows 2000 still works perfectly, using the same hardware. I believe that the IRQ steering system on Windows 98 (I assume 2000 is completely different) uses a thing called an IRQ table, which it gets from the motherboard BIOS or the devices themselves. The settings can be changed using the PCI Bus properties in Device Manager. I have tried different settings here (it's on the default at the moment) but nothing seems to help. Is it possible that the system is not getting the information for the sound device, or is getting incorrect information? Could some sort or corruption on the motherboard cause this (I'm still very suspicious of that BIOS re-set which I've never had happen before)? Maybe there could be something wrong in this area which doesn't affect Windows 2000 because it detects devices in a different way? I would phone Supermicro technical support about this, but I'm sure they won't want to know as soon as I say the words "Windows 98"! They did tell me ages ago that the board was never tested with Windows 98, but it's worked for years without any problem. Are there any diagnostics programs which will allow me to see what's on the PCI bus at a very basic BIOS level? So many questions.......

I've already put the new MSHTML.DLL file from KB976749 into my Windows 98 system, and I can confirm that it still works!

Thanks yet again guys! Well, I've had the crate apart, replaced the CMOS backup battery just to be on the safe side, although the one in there seemed to be perfectly OK. I cleared and re-flashed the BIOS, just to completely re-initialise it. System's still working thank goodness, but the Windows 98 sound problem hasn't gone away. I must say I was deeply surprised when restoring a months' old registry backup had no effect on the problem. All the registry backups in the world wouldn't have helped with this it seems. I'm usually pretty good at backing up the registry, even though I got caught this time, but what I don't do is back up my entire Windows folder regularly. I think I may well review that policy now......... I would have bet the farm after the registry restore didn't work that that the cause of the problem was on the motherboard, either a BIOS corruption of some sort (especially after it spontaneously reset on me) or an actual hardware failure of some sort (although why that wouldn't affect Windows 2000 as well is beyond me!) It would seem that's not the case. So, what are we left with if it's not something in the registry or the hardware/BIOS? It can only be a wrong file or files sitting somewhere. I'm pretty sure it can't be incorrect sound driver files as I've uninstalled them and reinstalled them several times, so I can only assume that it's a Windows system file or files that have been replaced perhaps with problem versions. Time for the Windows System File Checker I think...........

Thanks "Charlotte" that's really useful diagnostic information! I checked the registry parameters you referred to. I too have eight "APCI IRQ Holder for PCI IRQ Steering" devices. They are all listed under the {9b4e7760-3196-11cf-97ea-00aa0034319d} with "00" data. However under the {cf2524c0-29ae-11cf-97ea-00aa0034319d} key, number 00000007 was missing. That is the problem one. I manually restored the key for that device and gave it "00" data. When I rebooted nothing had changed, but the key was still there. I tried removing the sound hardware in Device Manager, and it put it back as before when I rebooted, and those registry entries still hadn't changed. I them removed the sound hardware and the problem steering device using the HP System Diagnostics program. (This does tell you which devices have a problem, which Device Manager in Safe Mode doesn't). On a reboot everything was put back exactly as before, and the registry entry under{cf2524c0-29ae-11cf-97ea-00aa0034319d} for that device had disappeared again. I am now pretty sure that the sound problem must have appeared when I had the problem with the BIOS resetting itself. I think I'm going to have to open the machine up, check the CMOS backup battery (I'll probably replace it anyway even if it seems OK) and then decide whether I'm going to bite the bullet and clear the CMOS and try reloading it again. I do have a motherboard BIOS recovery boot disk, so I hope I'm covered in case of disaster. If I disappear from here for a while, it will be because I've rendered my PC completely useless and am tearing my hair out trying to get it up and running again! I'm really sorry that this thread has been so much off topic, I hope the MSFN mods understand why it has rambled away from the original subject so much. At least the memory stick problem was vaguely relevant, this problem is almost certainly just an awkward and annoying coincidence which might well have happened anyway but has nothing whatsoever to do with any USB issues!

Tried changing those settings already. As I mentioned in a previous post, disabling IRQ steering made the sound hardware show sensible settings under its resources, but didn't bring it to life, and disabled my ethernet controller instead! Using BIOS instead of Hardware seemed to make no difference at all. Anyway, surely all these settings are stored in the registry, so if they were wrong restoring an old version should have made things come good, which it didn't! I'm convinced that the problem is more fundamental than an incorrect registry setting. I think something has gone wrong on the motherboard itself. I assume it's not an actual hardware failure, or it wouldn't work in Windows 2000, but I think that something has happened in the BIOS. For that reason, although I have downloaded it and am willing to try it, I don't think that Driver Magic will show anything that will solve the problem either. I think I need a piece of software that will analyse the motherboard resources at a very basic OS independent level, to see if the sound hardware is actually being recognised and configured correctly by the BIOS.

Well, just tried a set of old files. I had a set of files called system.els and user.els, which from their sizes looked like they might actually be backed up registry files. I've no idea what generated them, and a web search didn't reveal anything either. They are date stamped some time in last March. Renamed them to system.dat and user.dat, and they worked! The system booted OK, with registry data from eight months ago. Unfortunately, the sound still doesn't work, and the entries in Device Manager are exactly the same! Also the disabled "APCI IRQ Holder for PCI IRQ Steering" is there in the device list exactly the same, and I'm sure it was never like that before. Well, this at least proves that the messing about with the registry wasn't what caused the sound problem. It must be something more fundamental to do with the hardware or the motherboard BIOS, as I feared. Still doesn't explain why the sound still works fine in Windows 2000 though, but maybe Windows 2000's methods of enumerating devices isn't affected by the problem, but Windows 98's is. Don't know what to do now. Dare I re-flash the motherboard BIOS? I did it once years ago to install an update, and it did work but it scared the life out of me! I suppose I need to open her up and change that battery too, just to be sure.