SoundBlaster 16 PnP VS USB on Win95B

[bizzybody]

I have an old Epox EP-440BX motherboard and an ISA SoundBlaster 16 PnP. To have the USB ports working USB must have an IRQ enabled in BIOS setup. The "Plug and Play" sound card tries to grab the USB controller's IRQ. USB wins.

 

Even booted right to command prompt only, the DOS CTCU program says it can't be run "under Windows 95".

 

It's been around 15 years since I last set up one of these Soundblaster cards and I remember they were a big PITA to get them to do the play part of PnP. I've forgotten whatever tricks there were to make them work correctly.

 

This PC is going to be a standalone system for a CNC mill. Doesn't have a modem or network controller, just the built in USB, two RS232 and one LPT plus an AGP video card and the ISA sound card.

[Drugwash]

Some ISA cards used to have jumpers that allowed changing some parameters. You'd need the manual to find out how to operate them.

If (and only if) you have an AWARD BIOS, you may try to press Ctrl+F1 at the main BIOS screen. If the screen flickers for a split second, you'll see a few hidden settings in certain sections and maybe - just maybe - it will allow you to change the USB and/or ISA IRQ so that it won't be shared anymore.

The motherboard manual may also show which IRQs are shared and how, so that moving the card(s) to other slot(s) may fix such issues.

Sound cards also used to come with (DOS) configuration applications where one could set up Memory address, IRQ, DMA. Such application would automatically configure autoexec.bat with the BLASTER environment variable required by DOS applications/games.

 

if you don't have the configuration apps and/or drivers, please offer more details on the sound card so that someone here could provide them for you.

As a last resort, try to find another sound card that would (hopefully) use other IRQ or would at least have the configuration application.

[bizzybody]

I'm going to pull this ISA Sound Blaster 16 PnP. Would likely work fine in an all ISA system or one without built in USB, or one where the USB IRQ can be manually assigned.

I picked up a free PCI Sound Blaster 128, model CT5801 with original software CD. It's 'only' 15 years old, what kind of trouble could it cause... At least it has no useless IDE controller on it. Also going to wipe the hard drive and start over so there won't be any digital debris left from the previous sound blaster.

 

No jumpers on the card. It's an Award BIOS but nothing new shows with Ctrl+F1. I tried the Creative Labs DOS configuration utility and the only IRQ it will allow the card to be set to is 5, which the BIOS also helpfully assigns (without giving any choice except to enable or disable) to the built in USB controller. Selecting the setting to disable the card's IDE controller doesn't work, if it did then IRQ 10 would be freed up for the Sound Blaster.

 

The problem is the buggy early implementation of Plug n Play, specifically IRQ Holder for PCI Steering, which also gets IRQ5, or if a PnP ISA requests a different IRQ, the system will assign the same IRQ to PCI Steering. Also apparently buggy is the inability of this sound card to be set to a 16 bit IRQ or DMA.

 

Another piece of the problem is apparently the BIOS cannot find the sound card during POST, but it does list the IDE controller on the card. I did update the BIOS to the last release from 2000. Could be Epox and/or Award didn't obtain one of this particular Sound Blaster model to ensure it was properly detected, or it could be Creative Labs not correctly following the PnP specifications.

 

If I still had any PCI USB 1.1 cards I could try disabling the built in and manually assigning a non-5 IRQ to its PCI slot. (More half-arsedness common to that era, let you manually set *some* things so the ones you can't set can be automatically set to conflict with each other.)

 

I did get both the sound card and USB working by enabling manual configuration in BIOS then assigning IRQ 5 to Legacy ISA. The problem is now Windows 95 will not shut down, it gets stuck at the shutting down screen and until it will complete a proper shutdown it will not save settings to the Registry to re-enable the gameport and the Soundblaster's IDE controller in Device Manager. Being an OEM model, the IDE controller (using IRQ 10) cannot be disabled so BIOS and the OS can't see it. If I leave it all on Auto in BIOS, Windows will shut down, but of course both sound and USB won't work. Yup, more of the same crap I had to deal with from the early PnP hardware lo these many years ago. (And here we are with Windows 8.1 and hardware still cannot be disabled *and ignored/hidden* in Device Manager. It just sticks out there with a red X on it.)

 

That's why I preferred ISA cards with jumpers, because I could set the resources, then if Windows decided "No, I'm going to assign conflicting resources!" it was possible to force it into compliance via the manual configuration options. I called that "Jumper n Stay". Even better were cards that used a DOS program to change settings in an EEPROM. I don't recall Windows ever digging in its heels like a tantrum throwing child over any of those. "Oh, you want this IRQ and that DMA? Yes Master User, whatever you say!" It was much easier and quicker to manually set jumpers on everything (set them *correctly*) then install Win 95 and have it work the first time and stay working because there was nothing in the hardware to allow the software to ignore the jumper settings.

 

Trying to get this old thing to work has been a trip down memory lane, one of the sections that detours through a dark alley...

Edited April 13, 2015 by bizzybody

[Drugwash]

Oh yeah, I remember those days... Original manuals were hard to impossible to find out here and we were looking at those jumpers wondering what they do and how to set them properly if anything failed. There used to be a site, mirrored in a few other places, that listed many old ISA cards of various types with their jumper settings - that was gold, at the time! I forgot the links long time ago and I doubt they'd still be online.

 

I used to have an ISA ESS1869 sound card back in the day, on my 486. It did have the IDE controller which saved me big time. Considering the BIOS limited the native hard drives to 528MiB, there wasn't much space left for data when Win95 was installed. With that controller enabled, I could use a 1,2GB Quantum hard drive for data storage while the system was booting off the smaller, native drive. I still have that 486, it's still working, but that sound card got fried long ago. How I wish I had another one like that! It used to work with most (if not all) DOS games at the time and no issues with it in Win95 either.

 

Your particular issue may well come from the fact that the card is an OEM. They like to tinker with the settings in such a way that the device most likely wouldn't work (properly or at all) on other systems but their own. Won't work with other drivers but their own too, many times. I had such issue with a Creative Labs sound card manufactured for Dell, which - for the life of me - I couldn't find a driver for.

 

Anyway, for old OS versions such as Win95, the setting 'Plug'n'Play OS installed' (or similar) in BIOS should be disabled and resources be manually assigned. But it greatly depends on how well the BIOS implements resource management. It may take a long time to tinker with all settings in order to get the system to work properly, but hard work gives great satisfactions. And it may also depend on certain OS updates being installed (or not). I know there is an update for Win98 that fixes such hang on shutdown so maybe there is one for Win95 too. It's a common hardware and software effort. But it's fun... or at least used to be, back in the day.

 

Well, whatever you choose to do, I wish you good luck.

[bizzybody]

That would be MicroHouse. They used to have a CD-ROM set and a printed book. The CD set had a hardware copy protection dongle. The content of the discs was uploaded here and there on the web a while back. 'Twas the WWW that killed that company after so much of their pricey content was available for free from manufacturers and other sources.

 

Here's part of it, hard drives and controller cards. All obsolete of course. http://alasir.com/books/hards/

More nostalgia https://web.archive.org/web/20051223080930/http://public.planetmirror.com.au/pub/jumpers/All three volumes, in case you ever need to know the jumper settings for some old 486 or early Pentium or an RLL drive controller.

 

There was another company with a similar product, but IIRC they started quite a bit after MicroHouse and had a smaller collection of information. Went under about the same time as MicroHouse. Ah! Now I remember, Total Hardware 99, the gzip file on the link above.

 

The absolute best pre-Web support was Western Digital's. They had a dialup BBS where document numbers could be looked up and some drivers downloaded. After finding the document numbers you disconnected from the BBS then dialed a toll free number to an automated FAX back system. Had to use a touch tone phone to navigate the menues and enter the doc numbers. IIRC the max was 4 or 5 documents. Then you entered your phone number, hung up and waited for the system to FAX you the files and Western Digital paid the toll. The only cost to the user was if it was a long distance call to the BBS.

 

Even better was WD had docs for many product lines they'd sold off to other companies. I got info on an Orchid VESA Local Bus video card from WD some years after WD was out of the video card business. IIRC I was able to get the drivers from whichever company currently owned the Orchid name at the time.

 

In the years since, with old documentation and software so much easier to provide, many companies have or have had a policy of destroying everything the instant a product is discontinued. No docs, no software, no nuttin. "Buy our new stuff." Uhh, nope. Not when you won't even provide the specifications for this dingus that's only been EOL for a month and you'll do the same on the new model.

I've even encountered a few that disavowed ever making a product and even the very existence of the product. One of the craziest was a monitor I wanted the specification on and the manufacturer said they would not provide the specs "because someone might want to buy one" of that model. Can't remember what it was but once I missed *by one day* being able to download drivers for a device. The manufacturer had just killed the product the day before and had removed all the info on it from their site, except a note that it had been discontinued.

 

When it costs essentially nothing to throw old software and docs onto a website with a "Here's our old stuff, don't ask us anything about it. Seriously, we mean it." note, there's really no reason not to do that, especially when doing that shows old customers and potential new customers that your company isn't going to do the "Pi$$ off. Buy our latest stuff!" routine. It helps create good feelings (and $ale$) when people know that the info and software they need to use your products will always be available, even if it won't work with newer operating systems.

Edited April 13, 2015 by bizzybody

[Drugwash]

Yep, Total Hardware 99 it is! I have just found one of my old bookmarks from the early 9x systems and - expectedly - the original site doesn't exist anymore, but there's traces of it at Web Archive here. Apparently it's a mirror which also lists the original here where there's a longer list of mirrors.

 

These company guys don't think of a simple fact: people need choices! They want to ponder themselves whether to use an old thing or a new thing, based on their own needs, personal preferences and other factors. But choice seems to have been deemed dangerous by "someone" up there.

 

I still have a couple of 286 motherboards and related hardware (IDE controllers, video cards etc.) and at some point I may just want to play with them, just for fun, for old times' sake or whatever crosses my mind. Is that a crime? They sure think so, because information regarding that hardware is long gone (or would've been if it wasn't for the Wayback Machine, but still we don't know how much exactly of the original content is cached in there).

 

I also still have four IBM Model M keyboards, massive, heavy and clunky, but they never miss a keypress. All built around 1989-1991. Luckily they don't need some weird fancy drivers because definitely they would've pulled those loooong time ago, tu push the use of "modern" keyboards that break within a few months and are way too unreliable.

 

What I call my main machine is a 667MHz Soyo SY-6VBA 133 board with Win98SE that's been running 24/7 (!!!) since 2006. It's doing all I need it to except for browsing, because they purposely changed standards and everything to push the 9x systems out of the Internet. But I use that machine because it does what I want it to do, not what IT wants to do behind my back, as the new systems do.

 

Well now, a month or so ago I bought a second-hand LCD monitor (Hanns-G Hi221) to replace the heavy 21" CRT I had. Thought it was a bargain. But there you go trying to get the native resolution on the digital input - wouldn't budge. Went to the manufacturer's site, or at least I tried - it wouldn't open for me, probably because they didn't like my region. I managed to open it through a proxy and took me a while to find my particular model. Of course, it was in the 'legacy' (read 'forget about it') section. But funny enough, while a very similar model had a driver and manual to download, this one had nothing available (or at least wouldn't display it for me)! So besides shipping it with a bad EDID EEPROM they wouldn't even offer something helpful for the !d!ots like me that already bought it!

I had to waste two days, tinkering with different driver versions and custom resolutions just to get that darn thing to reach its native 1680x1050 resolution. So it's like you said, that manufactures simply hide under the carpet any bad/old products, ceasing to provide minimal required help or completely denying having manufactured those products. And I'm not including here forgeries.

 

However, on a few, very few occasions, I managed to find drivers, manuals or other related information on very old products on their respective manufacturers' sites. "Here you are, no guarantee or warranty, don't come crying if your box explodes, blah-blah", all fine by me and I got what I wanted, thank you. I respect those guys. Gives you a feeling of trust. Just as you said. Sadly the world forgot what trust means...

 

(fixed a minor typo...)

Edited April 13, 2015 by Drugwash

[jaclaz]

I also still have four IBM Model M keyboards, massive, heavy and clunky, but they never miss a keypress. All built around 1989-1991. Luckily they don't need some weird fancy drivers because definitely they would've pulled those loooong time ago, tu push the use of "modern" keyboards that break within a few months and are way too unreliable.

Which by themselves qualify you as a "real computer expert" (and NOT a "low end computer user" ), JFYI :

http://www.msfn.org/board/topic/155361-good-mechanical-pc-keyboard-amigaatari-xl-feeling/?p=989233

 

Clickity, clickity click....

 

jaclaz

[Drugwash]

Ah, finally I get the recognition I deserve!

 

Oh and I've also been given recently a used creamy-white Dell keyboard model AT102  (internal board says 'Silitek') with keys that resemble the 'Cherry' ones in first picture here. Unfortunately some of the keys are defective, most likely oxidated. I've managed to fix one of them, the Escape key, but it's darn hard to open them and clean the contacts, especially with my bad eyes.

[rloew]

It's getting even worse. Now they give you little or no documentation on even their brand new stuff.

I bought a brand new Laptop. No manual comes with it. I found the online Manual. There was no description of the BIOS Setup.

After finding that the SATA Mode had exactly one choice, namely AHCI, and "Secuire Boot" could not be disabled without killing UEFI, I returned it.

A decent manual would have prevented this.

[bizzybody]

Raise your hand if you've ever taken apart 3 or 4 Model M keyboards to obtain enough good coil springs to make one working keyboard. :-) It's easy to tell the bad ones, they either make no sound when the key is pressed or the *twing* sound of the spring buckling is lower pitch and quieter than a good one that snaps properly.

 

I'd like to see how well one of those with its steel bottom and layers of steel inside would fare against various sizes of lead shot fired from a 12 gauge shotgun. :-)

 

The shotgun packing goons chased Frank into an electronics recycling warehouse. Things were decidedly looking down, way down, for his continuing to stay alive and upright. Then he spotted his salvation, a box of old IBM Model M keyboards and a roll of duct tape! "Heh. Body armor with buttons. The keys to the kingdom."

Edited April 14, 2015 by bizzybody

[Drugwash]

Current business model is "my way or the highway". They know people are [insert your preferred bad word(s)] enough to accept any compromise since they've already been "convinced" through aggresive (and subliminal) advertising that they want and need the latest, the newest, the best-est-est-est... Too much to discuss on this subject.

 

Nice movie scene there, bizzybody! However I wouldn't purposely destroy such a gem as the Model M keyboard. It hurts enough that I have another two of them defective that I can't fix - one has had a certain amount of vinegar from a pickle jar spilled inside and many of the the tracks are destroyed, the other one just got old and pieces of tracks got faded out.

A few of those steel plates inside could make a decent body armor though. One'd feel like... Iron Man!

[jaclaz]

JFYI, out of the 4 I used to have, actually 3 of the "real" ones, as one - while still a model M is *somehow* different/less chunky/less clicky - and doesn't count, only 2 remain , as I had a similar incident (it was wine instead of vinegar, but the result is pretty much the same).

At least that model that I actually opened up was -if I recall correctly - very difficult to open to inspect the actual tracks, and since I needed anyway some key caps (to replace the ones that *someone* had cracked by stepping on a few of them after having made them fall as they had been removed to be washed) I never pursued the repair.

 

But I remember repairing a Microsoft keyboard, where some of the tracks had been wiped in yet another similar spilling incident using one of those "silver conductive paint" that you can find in some car parts shop, they are used to repair rear windows defrost panels, something like:

http://uk.rs-online.com/web/p/conductive-adhesives/1015621/

 

jaclaz

[Drugwash]

Actually I do have such silver conductive paint (quite expensive IMHO) and I did try to fix some tracks using it but the operation failed. Maybe it's too thick (what could I use to dillute it?) or maybe I don't have the proper tools.

 

It is indeed difficult to repair because the plastic panel that holds the keys is bolted riveted (the actual term slips my mind thank you jaclaz for the heads up!) to the metal plate in a lot of places and once they're cut they'd need to be tightened back using very small screws or something else that could hold the plastic panel tight.

 

The key caps can be replaced, there used to be (and maybe still is) a company that manufactured such Model M keyboards and also offered key caps for sale. Shouldn't be a problem for standard English alphabet but what should other people do, when their (preferred/native) keyboad layout contains special characters (such as Romanian, for example, where we have ŞŢÂĂÎ) and such keys are not available? I've always used a (not so) permanent marker for those special keys but the narrow vision of manufactures around the world, now in 21st century, astonishes me.

Edited April 14, 2015 by Drugwash

[jaclaz]

The plates, if I recall correctly , are "riveted":

http://en.wikipedia.org/wiki/Rivet

using some plastic "prongs" (or *whatever* you call them) that look like being melted on spot.

 

jaclaz

[Drugwash]

Yes, that'd be the right term, thank you very much!