Sad news! My most recent purchase on eBay was an Asus P5A-B (for parts)! This board was missing both NFETs for the low voltage. See pic...

It seemed odd that someone would salvage the FETs off a working board. It had some signs of wear but was otherwise clean. Maybe this board was undesirable and ended up a parts board at some time in the past...? Anyway, after finding only a few leaky caps, I decided to swap in some replacement FETs I had that were "close" IRFZ44N, the originals being PBYL1025B as far as I can find.

To my surprise, the board came to life!

I decided to order some replacements that were a better fit as well as some fresh caps. UNFORTUNATELY, I ordered something that would spell DOOM for this board that was saved from the dumpster. I mistakenly order these, IXTA100N04T2 being a 100A continuous vs 10A from the originals. The originals have a 100A peak current rating, but this was an obvious mistake on my part. After the recap, I did a quick thermal check to see if everything was okay and I noticed the buck chip near the RAM running nearly 70C after a few minutes. The IRFZ44N was close enough to get the board running, but that high temp made me think that FET was a bit out of spec from the original. The other near the CPU was just fine. At this point, the board was still functional and posted fine. So... I decided to swap in one of the IXTA100N04T2 and see if that would run cooler.

Here's a thermal pic of the HIP6013CB chip near the RAM. It did hit 70C while in the BIOS screen. I don't know if that's normal for these, but the datasheet lists 70C as an upper limit operating temp.

Had I double-checked my purchase I might have given this move a second thought, and maybe if I wasn't in a hurry I would have just asked if that was a suitable replacement. Of course, this board could have had a deeper issue lurking that didn't manifest until this moment, but at this point, I'm blaming myself for this one.

I decided to power up the board with no RAM or CPU and test the voltage. I wasn't even able to reach for my meter when I heard a loud pop and crackle! 😞😞😭😭😭 The damage is subtle but fatal in this case!

I should have left well enough alone... I consider myself lucky to have gotten one of these boards for under $60 and it looks like a working replacement is going to cost far more as the prices keep going up... 😡😡

I'm feeling a little depressed so I think I'll just go to bed for now.

Question for anyone who reads this...?

I see many postings for the IC that blew up, "M1543C" on eBay for about $12 from China. What are the chances those chips are "genuine" for a replacement and that everything downstream of Q10 NMOS on the board wasn't also cooked?

Given the high price of these boards and the possibility of chasing many dead chips, what would you do? The board was powered on for about 5 seconds before the M1543C chip popped with a little smoke. I see no burns anywhere, but that doesn't mean much.

I guess I'm looking for encouragement or assurance that it's simply a fool's errand... 🤷‍♂️

Oh... Made a mistake on the post. I gave the part number for the double diode instead of the original FET which is a HUF76121S.

Only you can answer that question! It comes down to what your time is worth, and/or if you have a strong desire to have a go at fixing it, and if the $12 part is an inconsequential extra cost. If it ends up needing more replacement parts then it's a case of how far down that road do you go...

Weigh all that up against the cost and avavailbility of a working board and go from there.

I guess I'd feel better if I had a boardview file or schematic... Something that would show where on the board Q10 was touching. That's probably impossible I'm guessing...

Got super impulsive and bought another P5A-B. I had my heart set on that board and... I'm sorry, everyone! I'm now partially responsible for raising the prices of these on eBay. I did order a replacement Southbridge chip and will try to fix the old board though.

This is the setup so far:

P5A-B 1.04
AMD K6-2+ 570mhz modded to 256k L2
ATI Rage Fury AGP main graphics
2 x Voodoo 2 12MB accelerators
ESS-1868F + wavetable MT-32 Pi

And it's going to fit into this. The goal is to make a "portable" rig that's kind of all-in-one. Mainly because after the nostalgia of retro gaming wares off I can easily store this in my closet and forget about it until I remember all the time and effort I put in and decide to pull it back out. 😄🤪

The sound card portion ESS 1868F + wavetable MT-32 Pi has been a fun side project! I only recently found the MT-32 Pi project and based on a YT vid - Wavetable Pi - https://www.youtube.com/watch?v=xC-qhcPUeRs, I decided to try and replicate it. I had to order the Pi Zero 2 but was able to assemble a working prototype quickly. I'm happy to say it all worked the first try and it sounds super cool. So much different from the card onboard synth and especially Windows GM.

The emulator also allows for sound fonts! I don't have a favorite yet, but I really wanted to try the SC-55 or SC-88 on DOS games without actually having to buy one. It all works pretty transparently and I'm grateful for the folks who put in the time and effort to make the MT-32 Pi project happen.

Beyond my prototype, I'm going to try my hand at a genuine PCB version that takes the Pi Zero 2 similar to the Wavetable Pi in the YT vid. His footprint was very optimized for size having the Pi on the reverse side of the PCB. I'm going to make one that's a bit more friendly to a wider variety of sound cards that have components around that area making it difficult or impossible to attach a card in that way. Case in point, I have an OPTi 82C925 card with a wavetable header, but they decided to place 10mm high capacitors in that area making it impossible for even a single-sided PCB to fit.

Many cards also have audio headers directly above the wavetable header limiting how high the PCB can be above that area. After looking at several Sound Blaster cards, I think a left-justified arrangement would fit most cards. Although the outliers would need a right-justified arrangement. Especially those with wavetable headers in the vertical orientation.

Anyway, I'll still post the plans on github for anyone interested. I'm also aware of the Rasp Pi compute module version, but that's a bit expensive! I was able to make this prototype for under $50 so I consider that a win!

Who does this!!!

Tevian wrote on 2024-04-21, 08:53:

And it's going to fit into this. The goal is to make a "portable" rig that's kind of all-in-one. Mainly because after the nostalgia of retro gaming wares off I can easily store this in my closet and forget about it until I remember all the time and effort I put in and decide to pull it back out. 😄🤪

20240420_234313.jpg
20240420_234332.jpg

Now that "case" is kinda a stroke of genius. That's gonna inspire me in my test bench endeavors ...

Well... It's mostly put together. The wiring management is a bit clumsy and I need a proper cage for the drives. This setup actually needs a 6 socket powerstrip internal! I need to rearrange a bit but I'd like to be able to store a keyboard, mouse, and controllers inside.

A few inclusions are a Rasp Pi as portable network storage, my Mister FPGA, and a 5 port network switch. The Mister is a fun bonus for playing the older console stuff and I was having a hard time finding a good place for it to live. I'll need to get a monitor switch though... My Dell LCD only has one VGA input. Or I might look for a 17" 4:3 that has VGA + DVI.

It's rather heavy but the idea of being mostly self-contained is coming together nicely.

Tevian wrote on 2024-04-28, 11:14:

Well... It's mostly put together. The wiring management is a bit clumsy and I need a proper cage for the drives. This setup actually needs a 6 socket powerstrip internal! I need to rearrange a bit but I'd like to be able to store a keyboard, mouse, and controllers inside.

A few inclusions are a Rasp Pi as portable network storage, my Mister FPGA, and a 5 port network switch. The Mister is a fun bonus for playing the older console stuff and I was having a hard time finding a good place for it to live. I'll need to get a monitor switch though... My Dell LCD only has one VGA input. Or I might look for a 17" 4:3 that has VGA + DVI.

It's rather heavy but the idea of being mostly self-contained is coming together nicely.

This is awe inspiring!
Can't wait to see the finished product!

There are people the DO buy broken voodoos for $100 and then we complain about jacked up prices…

To imagine I sold one locally for literally $4 🤣
But to be fair, I didn't want to be a jackass and sell a broken board for so high.
It went in the hands of someone who loved retro stuff, and he found loads of stuff wrong with it. IIRC it was a Powercolor card.

Updating design for the WaveTable Pi. My first version definitely sticks out too far and will cause problems with the SBC I'm including in the Brick. This version uses low-profile SMTs allowing the board to sit closer to the sound card. This also makes adding and removing the RaspPi solderless.

One change from the original and my own, (MT32-Pi Wavetable Board), I've decided to craft a carrier board that integrates the PCM5102 DAC IC and removed the display. This removes two extra PCBs! Having the display on the board is cool, but a little impractical at the end of the day. However, I'm including two buttons and the header for an external display/control. One more walk through but I should be ordering these soon. This retains the same functionality that I'm enjoying now while allowing the sound card to sit closer to the nearby SBC without touching. The original one had the idea of folding the RaspPi to the under side to make it more low-profile. This version should allow for the RaspPi to be installed on either side depending on which direction you install the pin header.

Since my recent adventure into industrial SBCs, I've decided to include a fourth system to the Brick. Instead of installing a second full-sized socket 3 motherboard for my older DOS games, I went with a half-size socket 3 SBC (Advantech PCA-6145B) and a small backplane. The board came with an AMD DX5 133MHZ, but I've opted for a much slower 486 DX2 66MHZ. Ya, the cable management sucks. I'll work on that another time.

This backplane is designed for PCISA systems, however, the EISA slot is compatible with standard ISA cards. This means I can include an ISA soundcard and disregard the PCI slots.

Now with dual power supplies, this rig is getting quite heavy!

Update to WAVEE, wavetable Pi thingy.

The goal for this addon is to make a daughterboard that can take a pre-soldered Rasp Pi Zero 2 and also allow for removal of the Pi for whatever need may arise. Like a future revision maybe...

One major issue is the height. As the PCB stack moves further away from the sound card it comes closer to making contact with a neighboring card. That would be bad! Most "real" wavetable addons are a simple PCB with all the components needed. The challenge is to make the PCB stack fit in less than 15mm which is just at the limit of the safe space between cards. Consideration needs to be given for through-hole components that might reduce that space.

Another issue is the wavetable footprint standard or lack thereof! There is in fact a footprint standard that aligns the 2 x 13 pin header with some mounting holes in the sound card itself. Some older cards have a few of these holes while others reluctantly included one, usually at the bottom of the full-height cards. "Newer" ISA sound cards simply removed these holes for convenience as I assume the wavetable cards started to become smaller or less popular over time. Even worse is the lack of discipline in the design and component choice, which results in partial contact or prevents the addition of the wavetable card at all!

See this example of this Crystal Audio 7268VPE-A/OPTi 82c929A card with a 10mm tall cap as well as the rather tall codec IC about 25mm south of the pin header. 🙄

Here's another OPTi 82C925 example:

There are a few examples that place tall components near the pin header making it "nearly" impossible to install a larger wavetable card. There seems to be a common placement of audio headers directly above the pin header as well causing an issue if the wavetable card extends above the header. Either way, unless the card is attached with a ribbon cable, soundcards with the taller caps in the way will probably need to be replaced especially since these caps are not in any standard arrangement.

Of the three different ISA cards I have, both OPTi variants needed to have these caps replaced with something shorter while the ESS had plenty of clearance.

This first version of the new design has some unfortunate issues that rendered it useless but allowed for a visual inspection of the clearances. Using the low-profile surface mount header allowed the PCB to sit much closer to the sound card. However, this could also be an issue as the possibility of the through-hole pins on the Rasp Pi could make contact with components on the sound card. Getting the Rasp Pi as close as possible to the card will need the surface mount header to be reversed. This allows for the Pi to sit closer to the adapter board by almost 2mm.

While the card seems reasonably stable, there is an issue with the surface mount through-hole header and the card leaning downwards. This means to could easily push the card too close to the sound card and make contact if not careful. I've decided to add an M3 hole for a 6mm tall nylon spacer that would ensure the PCB stays level. This is only a spacer and would not attach to the sound card itself. The placement of the hole puts the spacer in a clear area on the three card types I have. I'm sure some cards might have an IC or other small component in the way, but I imagine you'd simply choose a spacer that would allow for that. It's not foolproof, but it's at least something... 😅 It would be possible to simply extend the PCB down and to the right in order to align with the wavetable standard mounting hole, but this would make the PCB exclusive to the full-height sound cards only. Some shorter cards like the ESS ones have a wavetable header but lack the hole because it would be below the edge connecter. And that assumes all full-height cards have that mounting hole in the first place. So, non-conductive spacer it is.

So, this is V0.2. I fixed all the major issues I had as well as a different layout for the PCM5102A DAC. Hopefully, I'll actually be able to test this version! 😁

V0.2 WAVEE card turned out pretty good. One minor issue with the design is that I mistakenly grounded the DAC status LED to an island in the ground plane. 🤬 Totally my fault but an easy fix. Sound quality is identical to the first version so that's a plus. I did spec for some rather expensive acrylic poly caps as the audio couplers, and you could probably sub with quality ceramics. It's a bit of audiophile flare so, ya... 😁 I'm sure the noise characteristics of this card are far better than the best ISA Sound Blasters, so I guess it's not a huge deal. They do look cool though, and that's all that matters.

The other minor issue is the heat build-up from the Rasp Pi CPU. In the inverted installation, the heat radiates directly to the PCB and in turn, heats the DAC a bit. The temps are probably within all operating specs, but I might make a small change for V0.3 by cutting a hole into the PCB that will allow the heat to not build up as much, or allow for a very low profile 10mm square heat sink on the CPU. I'm operating a completely open case and any amount of airflow would probably compensate for this. I'd like to stick to this orientation to allow for a pre-soldered Pi to be used. I'm sure I could reverse mount the Pi under the PCB, but that would put the Pi in a position to be almost touching the sound card. For instance, the Crystal sound card I have has the wavetable header in a location that puts the Pi directly over the mixer IC. I haven't tested if the Pi would cause interference in the sound, but if I don't have to reverse mount, I won't worry about that.

The distances look good as far as clearance with the other cards. There "might" be an issue if the neighboring card had excessively tall components on the backside, but they would need to be rather tall!

This is what V0.3 looks like. The cutout should allow for the radiant heat to not build-up on the DAC PCB, but I'll do some better testing when it arrives.