My Windows 95 (Socket 7) PC Build

Windows, Mac, DOS, and all those-other personal computing platforms
User avatar
Anapan
Next-Gen
Posts: 3945
Joined: Mon Nov 17, 2008 11:15 am
Location: BC, Canada

Re: I Finally Got a Socket 7 Motherboard !

Post by Anapan »

Compatibility, ram freeing, and reports of the long filename feature causing corruption in some situations.
It's still better than others IMO, just needs a little customization as a full feature install was unstable.
I went with a boot partition manager - PLOP, but always liked the idea of swapping oses out on CF cards. I think I have some spares; If so I'll send them your way. I always wanted some of those tiny mechanical hard drives in CF form. They are as cool as PSP UMDs. While I'm wishing, I was hoping those were going to be available in RW and we had drives for them in sony's ultra-portable handheld computers too.
I haven't gotten around to it yet, but I planned to install a SD to IDE adapter I found out about (I think it was in a LGR video). All this dos talk bumped my tower up in queue of unfinished projects that need attention, so I'll let you know if it works soon.

Still need to boot it up and try for the first time Crusader: No Remorse. How did I never know this existed?
ImageImageImageImage
ImageImageImageImage
User avatar
marurun
Moderator
Posts: 12169
Joined: Sat May 06, 2006 8:51 am
Location: Cleveland, OH
Contact:

Re: I Finally Got a Socket 7 Motherboard !

Post by marurun »

What about FreeDOS. What's the compatibility like on that?
User avatar
Anapan
Next-Gen
Posts: 3945
Joined: Mon Nov 17, 2008 11:15 am
Location: BC, Canada

Re: I Finally Got a Socket 7 Motherboard !

Post by Anapan »

Probably good. Tried it and it seemed adequate, but barebones. I usually like to run MSD on a new system. It is not present.
ImageImageImageImage
ImageImageImageImage
User avatar
Ziggy
Moderator
Posts: 14866
Joined: Mon Jun 09, 2008 5:12 pm
Location: NY

Re: I Finally Got a Socket 7 Motherboard !

Post by Ziggy »

Both cases have come in. I had a good laugh at the C-3P0 color of the crappy PSU that came in that cheap weird case.

Image

Image

Looks like someone replaced one of the fans at some point. Hey, it's an Antec! I pulled it out and it look like it has LEDs wrapped around it. Look like a very typical mid-2000's fan.

Being that I have a case now to put this motherboard into, it's my last chance to get a few things done. The PS/2 voltage issue was caused by a bad inductor. I soldered in a jumper wire to get it working, but I wanted to replace it with a working inductor before installed the board in a case. Also, I wanted to install a 20-pin ATX connector.

Image

Image

Image


Unfortunately, that lead me to discover another potential issue with the board. Up until now, I've been using the AT power connectors on the motherboard, with an AT-to-ATX adapter. Now that I have the 20-pin connector installed, this is the first time I'm using a power supply in that way. There seems to be an issue with the soft power circuitry. When I flip the switch on the back of the PSU, the board instantly turns on. Pressing or holding the power button does not turn it back off. So I'm thinking the power-on signal is being permanently held low. I don't think there's many components that make up the soft power circuit, but I'm having a hard time tracing it out due to the nature of a multi layer board with traces that run under components on the top side. I'd like to try and figure out and/or solve the issue. But if I can't, I think I have a backup plan.

Image

This is the PSU that came in the case I'm going to use. I opened it up to inspect it. No bulging caps, and they're mostly (or all) Rubycon brand. The only thing I don't like about it is that there's no standby switch on the back of it. That, coupled with the soft power issue I'm having with the mobo, means that I can't use it right now.

Image

And here's just a quick shot of the front of the case with the black drives installed. I rather like the way it looks. I think it might look even better if I spray the power and reset buttons black. But I don't think I want to commit to changing the color of them. I could always swap beige drives into this case. Although then I would have to spray the CF card adapter beige. :lol:


Anapan wrote:I went with a boot partition manager - PLOP, but always liked the idea of swapping oses out on CF cards.


I'll have to check out PLOP. I used a boot manager once in the past, I cannot recall the name off the top of my head but I know it wasn't PLOP.

I also like the idea of swapping OSes by swapping CF cards, it makes it real easy to try out different OSes without any commitment or fear of messing up your system. I also plan to have a real HDD in there, mostly for nostalgia's sake. I don't know if this will work, but I'm going to give it a shot: I'm going to set the CF card up as the master, and a real HDD as slave. So I'm hoping if a CF card is preset, the computer will boot from that. But if I eject the card and leave it empty, the computer will default and boot from the HDD. Not sure if it'll work out that way, but it would be real neat if it does!

Anapan wrote:All this dos talk bumped my tower up in queue of unfinished projects that need attention, so I'll let you know if it works soon.


Nice! This forum could use some more PC talk.

marurun wrote:What about FreeDOS. What's the compatibility like on that?


Every time I hear about it, it's when something doesn't work. :lol: I'm not saying it's good or bad, it's just that I only hear the bad stuff. I plan to give it a try eventually, if only out of curiosity. That's the benefit of having the CF adapter. I can just put a blank card in there and install FreeDOS on it. Don't like it? Then take the card out and put MS-DOS back in!




edit: I forgot to mention a few things about case #1...

It was shipped from the UK, and was basically just wrapped in cardboard. Luckily, it arrived without any major damage. But the front bezel is held by a few clips as well as those long screw posts (common for this era of case) and all of the posts have broken off the bezel. Now, this could have happened in transit, or before, they do break easily. I think I can fix it though without much issue.

The black DVD-ROM drive turned out to be PATA interface. Since the HDD was SATA, I was kind of assuming since it was a black drive that it would also be SATA. It wont be horrible to keep as a spare, but I was planning on using it in a system that requires a SATA drive. No biggie, but kind of a bummer. I also realized that since this shipped from the UK that the drive is probably set to region 2. Now, you can set the region on PC DVD drives, but it only lets you do it a hand full of times. I only own one or two region 2 DVD movies. I use to own a DVD player that I was able to set to region free. And back in the day I used less than reputable apps to rip those discs and then burn them as region free. But it would be kind of nice to have a dedicated region 2 DVD-ROM drive. So I may just keep it that way, and use it in a future PC project I have planned (but have not yet posted about).

Now the HDD that it came with is a Seagate 80 GB SATA drive, which I also plan to use in a future PC project. You know when mounting a HDD in a case like this, how you can put 2 screws on each side? Well there was only 1 screw on each side, and not opposing each other, and they weren't especially tight so the HDD had some seesaw action going on. So there's no telling how much this drive flopped around inside the case while in transit. You know how those couriers just throw packages like these around. I had low expectations of this drive being in good condition. But surprisingly, I have yet to find anything wrong with it. I put it in another future PC project and installed XP on it, doing a full format (not quick) which I believe will check for bad sectors. I also used SeaTools to perform a quick test, and it's currently in the middle of the long test. No errors found so far!
Image
User avatar
Ziggy
Moderator
Posts: 14866
Joined: Mon Jun 09, 2008 5:12 pm
Location: NY

Re: I Finally Got a Socket 7 Motherboard !

Post by Ziggy »

If you are enjoying this thread... Please let me know by posting a reply! This thread has over 4,000 views but only a 3 people making replies. That makes it hard for me to gauge if people are actually enjoying it or not (for all I know, 4,000 people would give it a thumbs down).




I haven't had much progress since my last update, just a couple of small things. I don't expect any major progress until after the holidays are over, as the holidays usually suck up all of your time.




In my last post, I mentioned that there was an issue using the newly installed ATX 20-pin connector. I think I've determined what the issue is. I'm going to spoiler tag this since some people might find this boring...

There's some missing circuitry relating to the soft power on feature. There's a footprint for a missing IC, looks like an SO14 package, probably a logic chip. There's a missing resistor and cap (probably for decoupling), a missing diode, and a diode footprint that's jumpered. I could try and figure out what the missing components are to restore the soft power switch feature of the board, but it would be a lot of trial and error. To be honest, there would be some uneducated guess work involved. So I think I may fall back to plan B.

In short, my backup plan is to create my own soft power on circuit. It would be a small PCB that would sit between the case's power switch and the ATX power-on signal.

Now since I'll be going back to using the AT power connector (I'll remove the ATX 20-pin connector I installed, it can be used for another project), that means I'll be using one of those P8/P9 to ATX adapters like this one. AT computer cases have an actual latching power switch. So this type of adapter breaks out the power-on signal from the ATX power supply, as well as a ground wire, so you can connect it to the latching switch on an AT case. The problem with that is I'm using an ATX case with a momentary power switch. Now, one option would be to get a latching SPST switch and mount it on my case, but I'd really rather not modify the case. Another option would be to try and find a latching SPST switch that is the same shape and size as the momentary switch on my case, so that I could swap them, and the mod would be reversible. But it would have to be an exact drop-in replacement. So my idea is to just leave the case intact and create my own latching circuitry. I suppose another option would be to find another power supply. The one that I lucked into has no standby switch on the back of it, so as soon as I plug the power cable in the board powers on. If I got another PSU with a standby switch, I could use the standby switch as my main power switch. But that seems a little clunky.

So the way an ATX power supply works is that it has +5v standby power and a power-on signal which is usually labeled PS_ON. The PS_ON line needs to remain grounded for the power supply to be on. When you turn the standby switch on the back of the PSU on, it supplies +5v to the motherboard. The motherboard contains circuitry for the PS_ON signal. The PS_ON line is held high. When you press the power button, this circuitry latches the PS_ON signal low (to ground) and holds it there until something happens. The point of this is now the power button is out of the equation, your computer can turn itself off. Or if the PSU wants to go into protect mode and shut off, you wont still have a power switch that is latched on.

There are many latching circuits that I've already found online. There's two in particular that I like, and I think I have all the parts I need on hand already. One uses two inverters, the other uses a 555 timer. I think I'll breadboard both and try them out. See which of them I like, if any.

I have also thought about using a microcontroller. It might be the easiest way to get exactly what I want. Ideally, I would like to press the power button once to turn it on, but have to hold it for something like 3-5 seconds for it to turn off. That way you can't accidentally turn it off. Also, I believe a microcontroller would also solve the problem of switch bounce. Although switch bounce can also be taken care of in an analog circuit.

For those that don't know what switch bounce is... When you press a momentary switch, the contacts inside can actually bounce and make and break contact many times in a fraction of a second. So if you have a circuit that will latch to ground when the switch is pressed, then disconnect when pressed again, if the switch bounces then your circuit could actually latch and then unlatch over and over.





Another minor problem... The case that I got, the rear I/O plate is not the standard size. I bought a blank I/O plate so I could cut out a hole for just the AT keyboard and PS/2 mouse ports, but I can't use it. It's too small, the opening on this case is slightly larger in both the X and Y axis. I think I can come up with a solution that is easy enough and effective. It's just annoying, is all.




So you all might think I'm nuts, but I've actually acquired another Socket 7 motherboard.

Image

I got it for $25, which is very little money for a Socket 7 motherboard on eBay. It's an MSI. The condition was described as "powers on, keyboard doesn't work." Well, the Biostar Socket 7 motherboard (the original one that I posted this thread about) also had a keyboard issue that turned out to be a pretty simple fix. I'm not banking on the MSI board having the same exact problem, but for the price I don't mind rolling the dice on it. I've been learning a little bit about diagnosing and repairing old motherboards. Not that I think I'm an expect all of a sudden, rather, getting another board with issues will be practice and another learning experience.

This MSI board is actually pretty neat. It doesn't have an AGP slot, so it's not a Super Socket 7. But it's a full fledged ATX form factor. Most of the Socket 7 boards I've seen are Baby AT. Sure, some of them have ATX power connectors, but they're still Baby AT's.

So if I can get this MSI board to 100% working condition, I'm not sure what my plans for it are. I do have that other empty case now. And I did find a 233 Mhz Pentium MMX CPU in my stash of parts, which I believe is compatible with this board. So I could actually cobble together another working system. But I mainly got it because of the cheap price tag, I thought it would be a good way to practice diagnosing and repairing more. And if I can get it working, I wouldn't mind having a backup Socket 7 board since these old boards seem like they can be finicky.


edit: Hmmm, so I've thought of a solution for now. Since I have no I/O plate, I'll just have the PS_ON and ground signal wired to a latching switch, and I'll have it hanging out the back of the computer since there's a big gaping hole. I suppose when I create a proper I/O shield, since it'll be a blank plate, I could just mount a toggle switch there and that wouldn't be modifying the case. But then I'd have to reach around the back of the computer to turn it on and off. It wouldn't be so bad if it was a switch on the PSU, because that's easier to blindly locate, but I think having to find a switch on the I/O shield would get annoying. And also, I'd really like to use the power button on the front of the case. So I'll deal with it like this for now, and that'll allow me to come up with a solution at my leisure.


another edit: I made this post late last night. It's 3:30 PM now, and this thread has a little over 200 more views than it did last night. It's hard to tell what that means!
Image
User avatar
Ziggy
Moderator
Posts: 14866
Joined: Mon Jun 09, 2008 5:12 pm
Location: NY

Re: I Finally Got a Socket 7 Motherboard !

Post by Ziggy »

I bet you all thought I lost interest in this! :lol: I didn't, it's just that life has kept me from it for a while. I still have a few things to do before I consider this build "finished." Although, a computer build is never really finished, since you always end up tweaking things and swapping out parts. But by "finished," I mean it will generally be out of the decision making phase.

My last post in this thread was from mid December, and the view count was around 4,200. Now at the time of this post, the view count is 11,476. Interesting. I doubt there's 7,200 views from RB regulars (those who post as well as lurk) so they must be mostly coming from random search results. Well, hopefully some info in here is helping someone. But I'm mainly posting in these threads for entertainment value, in a blog sort of fashion, and not a guide or anything like that.

Anyways, I've made some decisions...


Firstly, I'm going to stick with the Biostar board (the original one that I bought for this). I had also acquired an MSI socket 7 board for a steal, and since it was a true ATX board I was considering using it. That would solve the on/off problem I have. A refresher: The Biostar board is AT, and although I added the ATX 20-pin power connector it's missing circuitry for the soft power on feature. The Sparkle PSU that I lucked into doesn't have a standby switch, which would be used as the main power switch since I'm using an ATX case and this is an AT board. That's why I was considering using the MSI board, since it's ATX it has the soft power feature so I'd be able to turn the computer on/off from the power switch on the front of the case. But I've ultimately decided to stick with the Biostar board.

As for the power switch issue, I could simply solve that by swapping out the PSU for one with a standby switch. But where's the fun in that? Although I'm not super keen on this PSU not having a standby switch and being ~30 years old (I would have peace of mind by keeping it on the off position when not in use - I might keep it unplugged actually) I'm going to keep it in this build for now. My work around for the power switch will be making my own soft power circuit.

Basically, the way an ATX PSU powers on is by grounding a single pin.

Image

If you ground the PS_ON pin, the PSU turns itself on. So the way this normally works in a PC is that the +5VSB pin is always supplying +5v to the board, even when the computer is powered off (but when the standby switch on the back of the PSU is on). So yeah, that's what I meant above talking about leaving a ~30 year old PSU plugged in without the standby switch. It's essentially always on standby, supplying +5VSB. Anyways, the +5VSB signal is supplied to the soft power circuitry on the motherboard. When you press the power button on the front of the case, it latches PS_ON to ground which causes the power supply to turn on. So to recreate that behavior, all I need is a latching circuit. So that's what my plan is, to create a latching circuit and use that to power the computer on/off by the switch on the front of the case.

Since this is an AT board, and I'm using an ATX PSU, I'm using an adapter such as this one...

Image

You can see it breaks out the green (PS_ON) and a black (ground) wire. Normally you would hook those up to the power switch on an AT case. But since I'm using an ATX case, I don't have that kind of switch (ATX cases have a momentary switch, hence the need for a latching circuit). But being that I'm using this adapter, I already conviniently have those two signals broken out.

But, I'm missing the +5VSB signal! My adapter actually doesn't have that pin included. And I don't think my Dupont crimping tool will work for this type of Molex pin. So I'll have to get one. But that's OK, because then I'll have the ability to make my own adapters going forward.

For right now, I just have those green and black wires connected to a latching toggle switch. That's how I'm currently turning the computer on and off. It works fine, but there's no where on the case I can mount the toggle switch, and I don't want to drill any holes. I'd rather use the case's actual power button. Since I'm missing the I/O shield, I just have that toggle switch hanging out the back of the computer. Which leads me to the next issue I have to solve...

I love this case, but it turns out that the I/O shield is a non-standard size. It's slightly larger. Which is actually a big reason why I decided to stick with the Biostar board. Since it's an AT board, it has only and AT keyboard and PS/2 connector. Meaning, whatever I/O shield I fabricate, I'll only have to worry about cutting holes for those two ports. Compared to the several that an ATX board would have. I'm not sure exactly how I'll make it yet, or how I'll attach it. But I have decided that I'm not going to drive myself nuts with it.


marurun wrote:So, the sound card appears to be an SB32 PNP (CT3600). Both RAM slots are populated.


I finally had a chance to test this card out, and... it works! :D I did a few quick tests with it, but plan to do a lot more in-depth testing.

I have acquired a few more sounds cards since my last post. So I got that CT1740 last year, but the left channel is not working. I still think it's fixable, but I haven't had the time to mess with it. A more experienced person could diagnose it, but I have to interleave researching and diagnosing. I just haven't had the time. I also have a few electronics projects that I took on for RB members, so those take the priority for any time I get to spend working on electronics. Anyway, I was kind of bummed that my only card with an OPL chip isn't working, so I ended up getting a couple more sound cards.

Image

Sound Blaster Vibra 16 (CT2260) - Got it for a decent-sih price. From what I've read, the Vibra 16's are cost (feature) reduced versions of the Sound Blaster 16. For that reason, they can usually be had for less. I found one on eBay that looked good, and the seller made me an offer that I decided to accept. It has a wave table header, and a real OPL3 YMF262 chip. Also, it looks to be in near mint condition. Looking at it in person, if you told me it was new old stock I would believe you. I haven't had a chance to mess with it yet though, because I first wanted to test...

Sound Blaster 16 (1740) - I ended up getting another CT1740 LOL. It is identical to my first CT1740 in every way, except being broken. It has the same production code, and thus the same DSP version with the hanging note bug (oh well). It had a buy it now price comparable to others on eBay, but the seller sent me an offer that I couldn't refuse. I just got the card in today, and wanted to test it straight away (as to not end up with two problematic CT1740's). The card looks to be in very good condition and is very clean. My first CT1740 had scratched and bent caps. This second one is in much better condition. So I installed it today and it seems to work perfectly!

To be honest, I wouldn't have purchased both cards if the seller of the CT1740 sent me the offer before I purchased the Vibra. But now that I have both, as well as the CT3600 from marurun, and some others, I want to do some comparisons. The CT1740 and Vibra both have YMF-262 chips. The SB32 and my SB16 WavEffects cards are both CQM, which is basically the OPL chip emulated. A lot of DOS gamers say to stay away from cards with CQM, and stick with an authentic OPL chip. But I've heard the SB32 and the WavEffects cards in person now, and they don't sound horrible. They definitely sound different, for sure. But CQM doesn't seem to sound has horrible as the internet has made it out to be. At least, not from my two cards. So I want to record samples from each card in order to compare them all together. I'll probably upload them here so you guys can give a listen as well. I also have an FM801-AU PCI card that I got from Anapan. I haven't tested it out yet, but if I can get it working then I might include that in the comparisons as well.

As for which games I'm going to use for sound samples, just some of my favorites. Mostly stuff that I'm familiar with, and that I feel is good to make a comparison with. But I only have a few games installed so far. DOOM, of course. I need to get DOOM II on it as well. Then I got Lemmings, which the first few levels are great for testing out music with. I really wanted to install Monkey Island, but I got the dreaded not enough memory error message. Why am I using DOS again? LOL. I'm not sure just yet what else I'll use for sampling. Any suggestions?
Image
User avatar
Anapan
Next-Gen
Posts: 3945
Joined: Mon Nov 17, 2008 11:15 am
Location: BC, Canada

Re: I Finally Got a Socket 7 Motherboard !

Post by Anapan »

I like this thread.
When are you going to install win98SE with the unofficial service pack and fire up the voodoo card?
ImageImageImageImage
ImageImageImageImage
User avatar
Ziggy
Moderator
Posts: 14866
Joined: Mon Jun 09, 2008 5:12 pm
Location: NY

Re: I Finally Got a Socket 7 Motherboard !

Post by Ziggy »

Anapan wrote:I like this thread.
When are you going to install win98SE with the unofficial service pack and fire up the voodoo card?


Thanks :D Glad you're enjoying it!

Honestly, it's so awesome that you sent me that Voodoo card that I have something special panned for it. That is, an entire build focused around that card! I'm going to post a dedicated thread for it.
Image
User avatar
Note
Next-Gen
Posts: 1396
Joined: Thu Feb 05, 2009 4:39 pm
Location: NYC

Re: I Finally Got a Socket 7 Motherboard !

Post by Note »

Just stumbled on this thread more recently, but also wanted to say I'm enjoying it! Definitely continue with posting updates on the computer build.

Would also enjoy seeing some photos or vids of the computer in action!
User avatar
Ziggy
Moderator
Posts: 14866
Joined: Mon Jun 09, 2008 5:12 pm
Location: NY

Re: I Finally Got a Socket 7 Motherboard !

Post by Ziggy »

Note wrote:Just stumbled on this thread more recently, but also wanted to say I'm enjoying it! Definitely continue with posting updates on the computer build.

Would also enjoy seeing some photos or vids of the computer in action!


Awesome, thanks for the feedback!




Not much progress on anything, but I did get the PCB in for the soft power switch.

Such a simple circuit could have been easily made on perf board, but since PCBs are so cheap these days (this 1 square inch board was $5 shipped from OSH Park) it would be silly NOT to get one. I already had the resistors and caps needed, as well as pin headers, I just had to grab the IC. I hate having to order a part that's like 50 cents and pay $7 for the cheapest shipping option. Usually I like to make a list of parts I need, and then order them all at once to make the shipping worth it. But this time around there was nothing I needed. So I instead got the IC from eBay, a 5 pack for a cheap price including shipping.

I haven't installed it in the PC yet, but now that I finally have a bench power supply I decided to at least bench test it.

Image

It works! The only thing I noticed is that when I cut the power, it remembers the latched state for about 8 seconds. Meaning, if the power was to cut out for a second or two, the computer would turn off and then come back on. I'd prefer it to stay off, but it's no big deal. It's kind of surprising that those tiny caps can keep it latched for 8 seconds, but I suppose that little IC uses hardly any power at all. I'm really not gonna worry about it though, it works well enough as-is.

Image

With the size of the IC, I could have used 0603 parts and made it pathetically small. But I decided to make it 1 square inch and use 0805 parts to make it easier for a human to assemble and handle. Also, the mounting hole is sized for a standard 6-32 screw. So since the motherboard is a baby AT, and this is an ATX case, I have a bunch of unused mounting holes above the motherboard. I'll just use a standoff in one of those unused holes to mount this PCB.

I got the circuit from here: https://electronics.stackexchange.com/q ... -as-switch

I found another using a 555 timer, found here: https://www.instructables.com/Latching- ... onversion/


I also picked up a power cable with a switch in it: https://www.amazon.com/dp/B07QJWLZFJ?psc=1

Image

The Sparkle power supply seems to be well made and in great condition, but it doesn't have a standby switch on the back of it. I had thought about adding one, I actually have a 120v rated panel mount switch that I salvaged recently. But I just don't feel like hacking up the PSU. So I'm not sure why I didn't think of this earlier, but I decided to get one of these switched IEC C13 cables. So now I can leave the computer plugged in, and flip this switch to turn it off. Otherwise the 5v standby rail will be powered 24/7. And I think this is more convenient than hitting the switch on a surge protector.

And yes, I realize that I could have just used this switched IEC cable to turn the computer on and off, but where's the fun in that? LOL. If I wired my AT to ATX adapter so that the PS_ON wire is always grounded, flipping the switch on the cable would turn the computer on and off. And that would give me the "It's now safe to turn off your computer" screen. Which, I'm not going to lie, I thought about doing it this way just so I could see that screen for nostalgia's sake. But I wanted to make the latching circuit just for the fun on tinkering with things.

You can order the board from here: https://oshpark.com/shared_projects/h6LgDD3k

I have two spares, if any Racketboy regulars need one just let me know!

12781
Image
Post Reply