Gmlb256 wrote on 2022-07-05, 19:57:

CalamityLime wrote on 2022-07-05, 16:50:

Quick question: […]
Show full quote

Quick question:

In regards to what works well on both both DOS and Win98

Sound Blaster Audigy 2
or
Sound Blaster Live! SB0220

I sound someone selling both for cheap but I'm just wondering how they are with dos. Philscomputer lab has a video on both (technically the audigy 2zs) but only talked about dos with the live.
In an ideal world I'd use ISA card for dos but isa cards are getting to be more pricy than they are worth to me.

I did search on vogons and from what I see the live sounds bad as a sb16 and for the audigy I did find this: sound blaster audigy 2 zs and dos games?
but what do you guys think?

-Lime

The DOS drivers for both sound cards will require EMM386 loaded to work, making impossible to get sound with several DOS games that refuses to run with EMM. Apart from that, FM synth emulation is mediocre and in some situations it can slow down the computer (holding spacebar in Wolf3D for example with AdLib sound effects enable, Phil doesn't mention that issue).

Regardless, Joseph_Joestar has excellent guides for installing both SB Live! (using Audigy 2 ZS drivers) and Audigy 2 ZS sound cards. Note that the Audigy 2 ZS will use two IRQs, one for the sound card itself and the other one for the FireWire interface it has onboard.

I wonder if there is a mod that could be done to add a enable/disable switch for the firewire.

Of course you could just disable it in software also.

Sphere478 wrote on 2022-07-05, 20:07:

I wonder if there is a mod that could be done to add a enable/disable switch for the firewire.

Of course you could just disable it in software also.

Fortunately, this isn't much of a problem unless one has a lot of cards installed. Serial and LPT ports can be disabled in the motherboard BIOS setup if they aren't needed, freeing some IRQs.

Hey there

In the end I went with the Audigy2, it was cheaper.

I'm having a problem that I'm struggling to solve on my end. the stock windows serial mouse drivers don't like the floppy drives?

So heres the thing,
When windows tries to access the floppy drives, windows hangs, leaving the seek LED on the floppy drives and I have to hard reset it.
Changing the generic serial mouse drivers for MS serial mouse drivers, windows doesn't hang but it takes a long time before it gives up with a "drive not ready" error.
Disabling the serial ports solves the issue, but then I don't have serial ports. It sounds like an IRQ conflict but windows doesn't show an IRQ conflict between the floppy disk controller and the com ports.

I can use the mouse in windows and access the floppy drives in an msdos promt, it handles that no problem, they just don't work in windows while I have a serial mouse driver loaded.

I ram memtest and that came up clean, I disabled other items in the bios and that didn't help. It seems to be just the serial mouse driver and the floppy drives. The issue exists across win 95, 98 and ME.

I'm at a loss, anyone got any ideas?

9x has a bad habit of being basically unusable while trying to read a fdd. It’s like the entire system waits on the floppy drive to report back. And refuses to multi task at all 🤣.

After a bit more experimenting I got the machine to crash when accessing the floppy drive in dos prompt and using the mouse.

It also crashes in safe mode.

I'm now thinking that i have a hardware problem where at least irq 3,4,6 are getting shorted somewhere.

Does anyone have any clue where to short looking?

I did decide to reflow the PCI chipset and that very quickly turned into a rescue but i think thats fine now, the legs in the PCI chip set are really really soft.

Try different cards and default bios settings. Return to bare basics. Nothing harder to diagnose than a 100% new setup

I got it

It was hardware, Oh boy was it hardware!

So a quick overview of what happened. I remembered that IRQ lines are run to the isa slots which was handy, I was hoping for a short across the pins but nope, they were fine. I did note that they went to a resistor array pulling the irq lines up to 5v via a 6.5k resistor. I also noted that the 8bit irq's were grouped separately to the 16bit irq's. My issues were in the 8bit irqs, 3, 4 and 6. So following them it brought me to the multiio chip at the edge of the board. I probed out the IRQ lines and the 8 bit IRQ's went to that chip just fine. Knowing that there's no way that the IRQ's go to that chip and stop there I checked the datasheet for a out line. The Lines that sounded promising didn't really lead me anywhere, so I followed out the IRQ lines.
Each IRQ of the chip went somewhere I could follow except IRQ_b as it was called. I traced it to the Parallel and IDE connector which didn't make any sense.
Eventually I figured out that IRQ_b was grounded and it looked like it went to the chipset.

I reflowed around the multi-io chip and no change, it stayed grounded. I saw nothing obvious and with the trace going in and around the ram slots I couldn't follow it, with it grounded I couldn't trace randomly hoping for the best.

Then I remembered something I had said already.

CalamityLime wrote on 2022-05-31, 17:38:

Just as well I went poking around the board. […]
Show full quote

Just as well I went poking around the board.

I spotted a flake of solder under the chipset, I didn't rework that part of the board except to remove one cap that was in the way of a trace I thought was tio10 for a bit.
I have noticed that the board is missing two caps that I did not remove. So this board got some re-work before me. The two caps are part of the voltage regulation circuits.

Not sure what this really means for the board yet but I should replace those before powering it at least. The board uses only 2 values of caps so it's easy to guess the values of the missing caps.

-Lime

.....

CalamityLime wrote on 2022-05-31, 17:38:

I spotted a flake of solder under the chipset

Well, I tapped the chipset with the ass of my screwdriver and the short went away.😶

So I sprayed some compressed air under the chipset and some contact cleaner, followed by more compressed air.
It works now but it doesn't like to be touched, I think I might need to clean the ram contacts.

-Lime

Some of my boards have a handy hole in the board under the chips handy for blowing air into.

Glad you are on the right track. Lucky find. Good sleuthing.

Sphere478 wrote on 2022-07-07, 06:10:

Some of my boards have a handy hole in the board under the chips handy for blowing air into.

Glad you are on the right track. Lucky find. Good sleuthing.

Honestly I don't know how to feel about it.

Following datasheets, probing around with a multimeter, following my instinct accurately and feeling like "wow, about of year of tinkering and I can do this, something I never thought I could do" just for it to end with "just hit it and it'll be fine"
Feels a bit anti-climatic

Sounds like another blob under there. Can you get it out?

Sphere478 wrote on 2022-07-07, 06:28:

Sounds like another blob under there. Can you get it out?

I think you misunderstood. I wasn't saying that I was having the issue again or something, I was commenting on my experience in the first place.
I hope everything is out now.

I also polished the contacts of the ram with some metal polish, lots of dirt came off of the sticks.

​Thinking about it, did i use percussive maintainance to solve an IRQ conflict?

majestyk wrote on 2022-06-30, 22:34:

"Cachechk" and other tools won´t tell you the cacheable RAM size for L1 and L2. You need "ctcm7" for that. But as long as you ha […]
Show full quote

"Cachechk" and other tools won´t tell you the cacheable RAM size for L1 and L2. You need "ctcm7" for that. But as long as you have only 64 MB RAM populated it will just tell you the cacheable size is 64MB.
You´ll have to go for more RAM!

In case you don´t have it - here you go:

ctcm17a.zip

My extra ram arrived today

Bingo!
... and congratulations.

Woohoo!!!

Order 4x 128mb now :p

I cleaned up my wiring a bit.
We'll have to make a thread for this 👍

EDIT: I also changed the underlaying socket since it was a bit iffy and I didn't trust it.

you can use runny superglue to bond those wires down like traces and rout them from below

Sphere478 wrote on 2022-07-13, 00:56:

https://www.vogons.org/thumbs/45320_5b34962f9581f4c35512297b51dd884b/E42C21DB-94C4-4362-A0F5-7788F7943521.jpeg […]
Show full quote

you can use runny superglue to bond those wires down like traces and rout them from below

This time around I did use a bit of super glue to "tack" the wires down to the board as well as each other. On top of the wires on the chip is some curable solder mask. Hopefully that will help the mod survive a snag.
About wiring the wires underneath the board, the contacts are not available on the underside of the board. I would expect most boards that would benefit from this mod would not have the contacts available either since if the board has only 1 tag ram chip, two out of the three connections we need are not routed away from the chip.

I did test the s-ram chip I bought from mouser and it didn't work for me. I was guessing that it would work but it didn't work, maybe two of them would but I couldn't tell you.
My muRata modules did arrive so I was very happy about that, very small things.

-Lime

CalamityLime wrote on 2022-07-13, 08:59:

This time around I did use a bit of super glue to "tack" the wires down to the board as well as each other. On top of the wires […]
Show full quote

Sphere478 wrote on 2022-07-13, 00:56:

https://www.vogons.org/thumbs/45320_5b34962f9581f4c35512297b51dd884b/E42C21DB-94C4-4362-A0F5-7788F7943521.jpeg […]
Show full quote

you can use runny superglue to bond those wires down like traces and rout them from below

This time around I did use a bit of super glue to "tack" the wires down to the board as well as each other. On top of the wires on the chip is some curable solder mask. Hopefully that will help the mod survive a snag.
About wiring the wires underneath the board, the contacts are not available on the underside of the board. I would expect most boards that would benefit from this mod would not have the contacts available either since if the board has only 1 tag ram chip, two out of the three connections we need are not routed away from the chip.

I did test the s-ram chip I bought from mouser and it didn't work for me. I was guessing that it would work but it didn't work, maybe two of them would but I couldn't tell you.
My muRata modules did arrive so I was very happy about that, very small things.

-Lime

Under the board you made

Hey there!

So I got the muRata modules, pcb and bits of LCSC that arrived really damn quick.

Installing the murata module resulted in it looking as good as I hoped it would and it seemed my hastily made reference model was about right. Unfortunately I got the pinout wrong on the foot print, I swapped the vout and the trim pads.
I had to desolder the muRata module and in the process I ripped 4 pads off of it, thankfully none were important! I altered the PCB to route a bodge wire for the vout to the correct pin and a thing bodge wire to the dip switch.

I'm also using different values for the resistors giving me very granular control of the output voltage but it does mean that the dip switch table on the PCB is wrong. So it needs updating but it's fine for me.

Now on to the installation!