dencorso

The specific bug I saw mentioned in those links seems to only apply to a USB1 modem under Win2K. [...] There's really no need to risk your main system right now, since the issue in question is limited to USB 1.x (because USBHUB.SYS 5.0.2195.7006 is part of the USB 1.x stack), so that a machine having only USB 1.1 is the ideal test bed for problems with USBHUB.SYS. That said, I'd never risk my main system without first creating at least one known-good partition image, so as to be able to recover fast, in case anything goes wrong. In my reckoning, there isn't such a thing as "too many backups".

That's a really tricky question... IMO, 2k3 has some advantages over XP, mainly in how much RAM it can access officially, so that for the right hardware it may be clearly advantageous. However, it's much more time-consuming to set a daily-use Desktop starting from 2k3 (there is a how-to in MSFN for that), so I'd never recommend it for anyone not already very familiar with XP, let alone someone migrating from 9x/ME, with no previous experience of the NT-family OSes. Of course, that's just my 2¢.

Naah, that is the way to COMPLETELY disable it, simpler (and IMHO enough) is just empty sfcfiles.dll : http://www.vorck.com/windows/xpsp4.html http://www.vorck.com/windows/software.html no hacked files, a single filed replaced, SFC continues running and check just an empty list. jaclaz The empty sfcfiles.dll is one of the files I recommended. It's precisely the same file creatd by Damian Bakowski & fdv, with a version property tag added, that identifies it as v. 5.1.2600.5515. The other file uses jdoe's patch. Either can disable WFP completely. So using both is bulletproof. And making no change in the registry ensures easy reversibility, if ever needed. So, at the end of the day, you're recommending just half of the procedure I indicated. It works. But twice as sure is better.

+1! I do it like this: I) regsvr32 /u zipfldr.dll II) ren zipfldr.dll zipfldr.not (step II is not really necessary, but helps prevent it from being reregistered silently, when WFP is disabled). And, since we're on the subject, IMO, the best way to disable WFP is the following: Get this, but don't run it. Open it, instead, with 7-zip, and grab inside sfc_os.dll and sfcfiles.dll, and put them in a temp folder. Now, from true DOS or from Win 9x, rename those original two files in the Win XP SYSTEM32 folder And then drop the replacements from the temp folder into the Win XP SYSTEM32 folder. When you next start Win XP, WFP will be permanently disabled. No registry mod is required, and if you ever decide to reverse, it's just a matter of using the original files again.

Hi dencorso,Last call to stock up on Win98-compatible hardware, the shelves will be empty next year. You bet I'm doing it already. Just bought a pair of never-used A7V600-X motherboards (I love those boards) for a trifle (DDR, no SATAII, who, but one of us, would want them?)! @duffy98: Download and save Virtual PC 2004 SP1, while it's still available. It's one of the real pearls MS offers for free!

When I faced that decision, I found out I just could not bear the idea of never more using Win 98SE... so I moved on to my present multiboot configuration (described in the > 1 GiB list). It took me some time to make it work in a really confortable way, but now I have Win XP SP3, Win 98SE, Slax, TinHat and two different PEs as boot options: it solved my problems and let me keep using Win 9x. I'm saying all this just to point out there's a third option, that is, going multiboot. It also helped me learn how to use XP effectively, because every time I was stuck with something, I could easily fall back to Win 98SE and get it done first, and then figure it out how I might do it with XP at easy, instead of in a hurry. The day I end up with no hardware whatsoever capable of running Win 9x will be the day I'll move my 9x to a virtual machine. By now, I really do intend to keep going, no matter what, because I'm sure it's possible if I add virtual machines to the equation, when it proves to be inevitable.

The Sisters of Mercy - Marian http://www.youtube.com/watch?v=n8i8pP3lcAU

If you want the best, get either Zalman or Noctua.

The Mission - Like A Child Again http://www.youtube.com/watch?v=Py6RfOyLH38

No. RLoew managed to get UDF.VxD (v. 4.10.0.2223 from a Win 98SE hotfix) to work with Win 95 (it adds support to UDF on CD/DVD Disks in Windows 95), but it requires custom patching. It's not simply a question of downversioning... And since the Win98SE versions of those files work OK, with no obvious bug, and the Win ME versions add no new functionality, that isn't worth the effort.

No, you did it right, as far as I can see. Problem is: probably the USBHUB20.SYS, which starts being the MS one, seems not to have been correctly replaced by the VIA one. If anyone else experiences this issue, please do check and report what is the File Version and Manufacturer (from the Copyright string) of the USBHUB20.SYS that is present, before copying the correct one over it. But since your fix worked, it seems it was the MS one all along, till you fixed it by hand. BTW, do you still have a VSBHUB20.SYS in WINDOWS\SYSTEM32\DRIVERS?

A simple question you may already have answered elsewhere, Dave, but bear with me: Does your pendrive get recognized right either by ASPIEHCI.SYS/GUEST.EXE or by USBASPI.EXE/DI1000DD.SYS and work OK?

@RLoew: This collection may interest you. And this is (pre-RENESAS) NEC v. 1.0.14.0 driver, the earliest one (09/30/09) I was able to lay hands on, up to now. I think no Win 2k driver was ever created.

Wow! That's probably a Thoroughbred B (CPUID 681), but it can also be the *rarest* variant of all Thoroughbred A (CPUID 680)! Be sure to find its CPUID with, say, CPU-Z, as soon as you lay hands on it.

That's a 2500+ Barton with FSB=333 MHz (running at 11x System Bus, which is 166 MHz, of course). It's a good CPU. But, for your system, if you can get hold of one of them, I'd suggest as a drop-in replacement, an Athlon XP-M, the AXMJ2800FHQ4C. It's a 2800+ Barton, with Vcore =1.55 V (but it'll run OK at 1.65 or 1.60V, in case your board doesn't allow lowering the Vcore). All the AXMJ2800FHQ4Cs have unlocked multiplier, so you can get it to experiment different multiplier/system bus combinations... It's default is 16 x 133 MHz so it would allow you to run 1:1 at FSB 266 MHz, or, what I think will really work, the CPU @ 133 MHz and the DDR RAM @ 100MHz. Or run everithyng at 100 MHz and set the CPU multiplier to 21, which it'll accept gladly. Of course, that's just my 2¢.

Use a floppy. Until you fix your hardware memory speed problems and get a clean 24h memtest86+ run, its useless to try anything more permanent. BTW, are you using BIOS v. 1.90 of Sept 5, 2003, right? What Athlon exactly do you use?

Try the free PDFill PDFTools. Take care to install the FREE GhostScript before use, there's a link for the correct version way down below the same page whence you can download the PDFTools.

Try: http://www.msfn.org/board/index.php?act=findpost&pid=895875 to the post itself or: http://www.msfn.org/board/index.php?showtopic=88734 to the whole thread. They just work.

Sure. Deal. But I cannot do it right now. After this weekend?

Problem is aurgathor posted all over the place. I saw that and reacted to it in yet another thread. I don't think it's bad RAM, really. My A7V600-X uses VIA KT600, which is one of the successors of the KM266. For the KT600, it's well documented that it can control 3 GiB DDR RAM at 333 MHz (which is how my board runs), but it will hardly POST at 400 MHz, although it'll run up to 2 GiB at 400 MHz. Then again, some KT600 sold are below par, so once I had another A7V600-X that needed to be set to 328 MHz at the most, in order to run stably (at 333 MHz it would work for some days, then crash randomly). But the RAM sticks were OK in that case, too, the problem is the Northbridge chokes when controlling so much RAM so fast. I'm not aware that a similar problem has been officially documented for the KM266, but since it's the same family of Northbriges, I believe we're seeing a manifestation of the same design flaw. Of course, I may be wrong.

Well, you were right in ask about hardware, from the start. We should wait some more, to see whether further reports come in, so then, maybe, we'll be able to find a common factor, especially if it really is hardware related, as I suspect. BTW, I shall look into those .INFs soon. But the key is to select by VEN value (= vendor ID) VIA is VEN_1106. However, some file renaming will also be needed because we cannot have two different USBHUB20.SYSs in any single folder.

I use USBPORT.SYS version 5.00.2195.5652, so I maybe should test the later version when I have time. But Tihiy, as he does sometimes when he makes vague remarks, must be relaying experience gathered from the Russian forums he also participates in. And now we have a report of problems from an MSFN member. It cannot be ignored. But it may be yet another case of hardware-dependent issues. Let's wait and see. BTW, I understand WDMCHECK does not find any missing dependencies for USBPORT.SYS 5.0.2195.6941, is that right?

That confirms the latest safe USBPORT.SYS really is version 5.00.2195.5652, which is the version that should be used. Now USBSTOR.SYS version 5.00.2195.6773 works, but *requires* WDMSTUB.SYS version 5.0.0.6 for working. This is taken care of in U98SESP3.EXE, so there is not a problem here, at least for U98SESP3.

There are no drivers for USB 3.0 for Win 9x/ME that I know of. There is a USB 3.0 PCI (not PCI-e) card available: StarTech PCIUSB3S2, which, of course, limit maximum transfers to 133 MB/s (the 32-bit/33 MHz PCI limit), but that's good enough. Now, what is interesting is that fast USB 3.0 pendrives, working in USB 2.0 compatibility mode (which identifies itself as "USB 2.1") are the fastest pendrives there are. The Patriot Supersonic Magnum 64 GB and the less expensive Kingston Data Traveler Ultimate G2 32 GB leave any and all USB 2.0 pendrives in the dust by much, and work out of the box with NUSB. The Kingston Data Traveler HyperX should be good also, but I still didn't test one of them (although I sure intend to).

The manual says: "North Bridge: VIA KM266, FSB@200/266 MHz" if you set the bus speed to 166 MHz, the Northbridge won't be able to cope with it. Then again, it's highly doubtful whether the KM266 can really control 2 GiB at DDR 266 (= bus speed: 133 MHz). Try setting bus speed to 100 MHz (DDR 200) and it'll probably become stable. Moreover you don't need to run 1:1, so you still can set the CPU clock at 133 or maybe even 166 MHz, provided you run the RAM at DDR 200. Do give it a try. That is a typical symptom of Northbridge choking... try setting them to DDR 200 and there's a good chance they'll run OK together.