Still baffled by this. Anyone have any suggestions?

I'm not really sure, but having had a few problems that illicit a response of deafening silence myself, I feel for ya. So here goes nothing...

First, I would make sure all the power pins on the ISA slots have continuity. It could be as simple as some cards using a ground pin that isn't tied to anything on the motherboard. Print out an ISA slot reference and probe between the AT power connector and the slot pins that carry that rail. I read on here that the PAS16 uses the slot bracket as an analog ground point. That would certainly explain why MINE didn't work right. (I removed the bracket because I had to replace the corroded audio jacks, and none of the usual parts houses carried the jack variant with the exact height of the original.)

You said you have a scope. Try probing that OSC pin on the motherboard -- if nothing else, just to settle that theory once and for all.

There's something about that board, and given the problem you're having, it seems like a "layer 1" issue. Can't be THAT many things that would cause it, so let's knock those out and see what happens.

SirNickity wrote:

I'm not really sure, but having had a few problems that illicit a response of deafening silence myself, I feel for ya. So here […]
Show full quote

I'm not really sure, but having had a few problems that illicit a response of deafening silence myself, I feel for ya. So here goes nothing...

First, I would make sure all the power pins on the ISA slots have continuity. It could be as simple as some cards using a ground pin that isn't tied to anything on the motherboard. Print out an ISA slot reference and probe between the AT power connector and the slot pins that carry that rail. I read on here that the PAS16 uses the slot bracket as an analog ground point. That would certainly explain why MINE didn't work right. (I removed the bracket because I had to replace the corroded audio jacks, and none of the usual parts houses carried the jack variant with the exact height of the original.)

You said you have a scope. Try probing that OSC pin on the motherboard -- if nothing else, just to settle that theory once and for all.

There's something about that board, and given the problem you're having, it seems like a "layer 1" issue. Can't be THAT many things that would cause it, so let's knock those out and see what happens.

Thank you for the suggestions! I haven't had much time to work on this lately, but I was going through my sound card collection and realized that one of my SB Pro 2 cards is a revision 3 without its own clock crystal (my other three all have crystals). And what DOES it have? It has the OSC pin connected. And... it doesn't work! I can swap between a revision 6 with crystal and a revision 3 without and one will have sound, the other won't. The clock circuit HAS to be the cause. I'll try messing around with the scope today to see if I can find any fault with the crystal.

This is a dumb question, but where do I actually put the leads from the scope if I want to test a crystal in circuit? I have almost no experience with scopes. The ONLY things I have ever used my scope for are checking an old laptop for a dead inverter board (you can measure the high frequency it gives off through the bezel without dismantling anything) and for playing with the neat little laser beam lines on the display while making "pew pew" noises to myself.

Well... I think my oscilloscope has a problem. Its quite ancient (1990 I believe). A friend of mine who is an engineer gave this to me when he moved about 5 years ago. He had replaced several capacitors in it at some point, and it was working. Now, I can't get it to really do much of anything. When it starts up sometimes the display is blank, other times the light will draw a random squiggle on the screen and then vanish. I've monkeyed with every single switch and knob and cannot get that steady line or any moving dot at all. The closest I've gotten is a stationary dot that I can move with the vertical and horizontal adjustments.

This is certainly way over my head. I wonder if there's some other way I can test this. Maybe I should just swap out the crystal with a spare...

You may use a ~47pF capacitor connected in series with two antiparallelly connected red LEDs. Just connect this probe between the TTL/HCMOS oscillator's output and ground and it should light up. If it doesn't, try again with slightly larger capacity.
Another low-cost option is to use a cheap Chinese frequency counter.

suntac wrote:

You may use a ~47pF capacitor connected in series with two antiparallelly connected red LEDs. Just connect this probe between the TTL/HCMOS oscillator's output and ground and it should light up. If it doesn't, try again with slightly larger capacity.

I can't even begin to visualize what you just described. 😵

suntac wrote:

Another low-cost option is to use a cheap Chinese frequency counter.

I just ordered one of the cheap $10 kits from amazon. I'll try putting one together so I can test the crystals. Hopefully the tester works. 🤣

In the mean time, I only had a couple of similar 14.31Mhz crystals, and one had leads clipped too short to use. The other didn't look so great (a bit of rust on the leads). I tried putting the crusty looking one in place of the one on the board and it didn't make any difference. I guess I can try borrowing a crystal from a working device in the mean time... or I can wait until I build the tester.

I've tried tracing out from the OSC pin and I can't even find the next solder connection for it... I have no idea where it leads to. Anyone know what kind of ICs may be connected between the crystal and OSC pin?

Okay, I've made a rather interesting discovery. I've attempted to trace out the OSC pin from the ISA slots several times and using the continuity tester on my DMM I haven't found a single lead on this entire board that even gives a short beep indicating any kind of circuit at all. I've even tried the "foil" method that I just read about online. Basically, I've attached an alligator clip from my DMM to a sheet of aluminum foil and laid it over most of the back of the board, with the other lead connected to the OSC pin on the ISA slot. I get absolutely no response from the DMM (no beep, no circuit at all) from the entire board... only the other OSC pins on the other slots have continuity. There are only 3 components (the Opti chipset and the CPU) on the entire board that are not through-hole, so every single solder joint aside from the chipset itself (which I have tested every lead on) should be exposed on the back of the board for continuity testing.

The 14.31Mhz crystal does connect to a nearby DM74LS04N IC. According to that datasheet, the board's 5v circuit is directly connected to the IC's Vcc, one lead of the crystal is connected to A1 (input 1?), while the other lead is connected to Y1 (output 1?) which has a trace directly connecting it to neighboring A2 (input 2?). Y2 (output 2?) has a small trace that goes behind the chip to a via but I can't figure out where it goes from there. I'll keep investigating that... probably have to desolder the chip to figure it out.

But, is it possible that this board was actually not designed with the OSC pin connected? Is that normal? It seems really odd that it would actually have the crystal but it wouldn't be connected to the slots that would make use of it. Maybe it's a design error that simply wasn't caught at the time because no one was using sound cards or other devices that needed OSC back in 1988.

This of course begs the question... can I simply add in a connection to the OSC pin? I assume it requires some logic other than the crystal itself for it to work, but presumably that DM74 above would be that logic, or at least part of it. If I wasn't afraid of frying something old and useful, I'd simply run a wire from Y2 on that IC to one of the OSC pins, but I have no idea what that'd do, so I'm not going to try it... yet. However, when my crystal tester arrives and I get it built, it does have a simple frequency counter function too, so I could potentially test a few places on the board for that elusive 14.31Mhz clock signal...

Maybe you have a very early version of even a pre-release version of the motherboard.

Back then it was quite common for there to be jumper wires and also added components to fix hardware issues they found after the fact.

It could even be that somebody at some point removed a hardware fix.

Maybe you could find pictures of the same board you have and compare.

High quality pics of the front and back of your board might also lead to being able to spot the appearance of a removed fix. It is pretty easy to tell if something was removed if you know what to look for.

Edit, I noticed in one of your other threads that you said you fixed the solder joints on the oscillator.. It is very strange for solder joints for a single thing on a board to look "sad" like that unless there was something else there at some point.. aka.. a fix that was wired to something else on the same side of the board that somebody removed and did a poor job of removing it.

Thanks for the suggestions.

The solder points on the crystal that looked weird before were just kind of different looking than the rest... I'd describe it as older looking, tarnished. I think my tester should be delivered soon so I'll try to get it put together as soon as possible so I can test the frequencies and voltages coming across the OSC pin and the 74LS04's output and I'll compare it to some other boards.

I've read that measuring a crystal or connecting it to anything at all will modify the oscillating frequency of the circuit. Is it safe to assume that passing in and out of the 74LS04 twice prevents anything connected to the output from affecting the crystal directly?

The IC is a buffer, or inverter. It specifically isolates the crystal from the load it's connected to. You can safely measure its output without affecting the oscillation of the crystal. Even if you did probe the crystal, you might cause it to stop oscillating, but at worst, power-cycling the board should resolve that. Not a big deal.

If you're not seeing continuity to anything on the board, most likely there's a broken trace or something that disconnected the OSC pin from the output of that buffer. You could try tacking a jumper wire between them to see if it changes anything.

IT'S ALIVE!

I built my cheapo frequency counter from Amazon (impressive... ordered Saturday, got it Monday), and the crystal measuring circuit doesn't work, but the separate one for measuring frequencies directly DOES. I've read lots of comments about differing parts being sent with some of these clone kits and they often break the crystal tester, but I'll have to sort that out at a later date.

Long story short, I measured the OSC pin on a working board and it reported 14.32. I measured the OSC pin on my DTK board and it showed nothing. I measured the 4th pin (Y2) on the 74SL04 and it showed 14.32! I ran a wire from that pin to one of the OSC pins on the first ISA slot... voila! So far the SoundGalaxy BXII and CT1350B with CMS are both working! I haven't tested the PAS16 cards yet, but I'm hopeful.

I'm still somewhat curious as to where that mysterious via goes after it leaves Y2 on the 74SL04, but I'm guessing it goes nowhere. There are no damaged traces from battery corrosion anywhere near this side of the board and I've already fixed the 5 or-so traces that did have corrosion damage near the battery. None of them have any connection to the clock circuit. I really think they just left it out, possibly by accident.

Thanks everyone who offered input. I'm glad I can finally move on with this build. I will hopefully have a chance to do a full writeup soon. 😀

Ah good! I love a happy ending. 😄

Just wanted to add, I had problems getting a SIIG Enhanced IDE SC-JEE012 card working in this board before figuring out the OSC problem. Basically, the card's BIOS would load but it would never detect any IDE devices, no matter what I did. A different but similar SIIG Enhanced IDE card worked fine so I thought the other was defective.

Well... I just thought about this card again (because it is in pristine condition and has the manual) and realized I should check to see if it uses the OSC pin. Sure enough, it does! I just tested the card again in my 386sx and it now works fine after I did the OSC pin fix mentioned earlier. Yay! 😀

The unconnected OSC pin has turned out to a pretty frustrating issue. I hope this thread saves other people some time.

Hey! I guess they put that pin on the ISA slot for some reason after all, eh? 😉 Good to hear. I love it when I find out hardware that I thought was dead is still in fact working fine. It's like getting something for free.