GALs are reprogrammable unlike PALs, so treat this as a firmware update. Downloading the fuse map into such IC is an insignicant effort. Redisigning the equations is the smarter thing to do, especially if it can save your buffer. Is there a proof the rumored problems were due to the lack of that WE signal buffering? It may not only be buffer loading capability that could be beneficial, it may be the delay it introduces.

It seems a bit like trolling to me. On the one hand you have a problem with SMT components and would like THT, on the other hand you don't see a problem with reprogramming the GAL which requires a programmer and not everyone has one. In my upgraded PCB the original GAL can be used, that was the intention.

The routing of the #WE signal was clearly problematic in the original design.

Quoting from the document "Memory upgrade module specifications Revision 1.1" by Creative Labs: To minimise capacitive loading seen by EMU8000, a 74F244 buffer with 33ohm damping resistor at the output is needed on the address and control signals.

So connecting a #WE output from EMU8000 only via damping resistor directly to a total of sixteen #WE inputs on double-sided 32MB module means 127pF* capacitive loading. That's quite a lot.

*according to Micron datasheet for 4Mx32 DRAM SIMM

Trolling you say. Problem you say. I would say your thinking is not very logical. There is no connection between using/or just soldering SMDs and having a GAL programmer. In fact historically these PLDs first appeared in DIL/TH packages and just years later PLCC variants emerged. The default programmer's sockets are in 99% of programmers DIL/ZIF. I hardly can imagine people building your PCB without programming GALs themselves! It is nonsense to propose to throw away already built SIMCONNs and transfer/use GAL chips in your PCB. These days GAL programmers are dirt cheap, some are even homebrew. On the other hand I have designed numerous devices and their PCBs that use such simple PLDs. And I can tell you -- programming of PLDs and PCB routing go hand in hand. Because just by swapping pins in the fusemap one can simplify the PCB routing significantly, reduce the number of vias, etc. Last but not least for homebrew projects it is best to use TH components. One day you will get older too and you will start having problems with SMDs, etc.

georgel reappears on the forums and instant argument with somone. Best to ignore him HMX

badmojo wrote on 2025-11-30, 22:22:

georgel reappears on the forums and instant argument with somone. Best to ignore him HMX

When you have no arguments, or the conversation is above your level of competency, sure you exactly can ignore me. Otherise most likely silently, sometimes unaware, you are using my products and knowledge.

badmojo wrote on 2025-11-30, 22:22:

georgel reappears on the forums and instant argument with somone. Best to ignore him HMX

this tbh

HMX wrote on 2025-11-27, 09:29:

Took a while but finally done!
https://github.com/Blaster-HMX/SIMMConn2

I'm planning to build this, but what physical size SMD capacitors are? I don't have eagle 7 installed, and BOM doesn't list the capacitors.

Because it's difficult to find memory slots that can be installed at a 90-degree angle, I used this method to change the orientation of the memory stick so it's parallel to the sound card. However, it's a bit cumbersome. Here's how I did it:

This is the final result:

I am using a PCB board based on the original LuckStar design.

Production Process:

1. Install the jumper pins on the back, shortening them to avoid contact with the sound card. They should be roughly the same height as a jumper cap or slightly shorter. Don't let too many pins protrude on the other side to avoid obstructing the memory slot installation.

2. Install the Socket Header. Shorten the pins on the top one to prevent them from protruding too far on the other side. Keep the solder joints as flat as possible to avoid obstructing the memory slot installation.

zwrr wrote on 2026-03-26, 08:33:

Because it's difficult to find memory slots that can be installed at a 90-degree angle, I used this method to change the orienta […]
Show full quote

Because it's difficult to find memory slots that can be installed at a 90-degree angle, I used this method to change the orientation of the memory stick so it's parallel to the sound card. However, it's a bit cumbersome. Here's how I did it:

This is the final result:

The attachment 0.jpg is no longer available

Production Process:

1. Install the jumper pins on the back, shortening them to avoid contact with the sound card. They should be roughly the same height as a jumper cap or slightly shorter. Don't let too many pins protrude on the other side to avoid obstructing the memory slot installation.

The attachment 1.jpg is no longer available

The attachment 2.jpg is no longer available

2. Install the header pins. Shorten the pins on the top one to prevent them from protruding too far on the other side. Keep the solder joints as flat as possible to avoid obstructing the memory slot installation.

The attachment 3.jpg is no longer available

3. Bend one row of pins in the memory slot at a 90-degree angle, and shorten the other row. Install them on the board as shown in the picture. First, solder the bent pins. Then prepare some 90-degree bent pins, shorten them appropriately, and solder them to the memory slot pins on the board. Shorten the pins on the other side of the board as well to avoid contact with the sound card.

zwrr wrote on 2026-03-26, 08:34:

3. Bend one row of pins in the memory slot at a 90-degree angle, and shorten the other row. Install them on the board as shown i […]
Show full quote

zwrr wrote on 2026-03-26, 08:33:

Because it's difficult to find memory slots that can be installed at a 90-degree angle, I used this method to change the orienta […]
Show full quote

Because it's difficult to find memory slots that can be installed at a 90-degree angle, I used this method to change the orientation of the memory stick so it's parallel to the sound card. However, it's a bit cumbersome. Here's how I did it:

This is the final result:

The attachment 0.jpg is no longer available

Production Process:

1. Install the jumper pins on the back, shortening them to avoid contact with the sound card. They should be roughly the same height as a jumper cap or slightly shorter. Don't let too many pins protrude on the other side to avoid obstructing the memory slot installation.

The attachment 1.jpg is no longer available

The attachment 2.jpg is no longer available

2. Install the header pins. Shorten the pins on the top one to prevent them from protruding too far on the other side. Keep the solder joints as flat as possible to avoid obstructing the memory slot installation.

The attachment 3.jpg is no longer available

3. Bend one row of pins in the memory slot at a 90-degree angle, and shorten the other row. Install them on the board as shown in the picture. First, solder the bent pins. Then prepare some 90-degree bent pins, shorten them appropriately, and solder them to the memory slot pins on the board. Shorten the pins on the other side of the board as well to avoid contact with the sound card.

The attachment 4.jpg is no longer available

The attachment 5.jpg is no longer available

The attachment 6.jpg is no longer available

The attachment 7.jpg is no longer available

The attachment 8.jpg is no longer available

4. Now it's basically finished, but the memory slot isn't very secure. I used a simple and crude method: soldering it to the board, as shown in the picture.

Mike_ wrote on 2026-03-25, 10:05:

HMX wrote on 2025-11-27, 09:29:

Took a while but finally done!
https://github.com/Blaster-HMX/SIMMConn2

I'm planning to build this, but what physical size SMD capacitors are? I don't have eagle 7 installed, and BOM doesn't list the capacitors.

Hi,
there are 3 x 100nF ceramic capacitors in 0805 size and one 6,8uF/10V tantalum capacitor B size. Bus drivers are TSSOP20

HMX wrote on 2026-04-01, 15:31:

Hi,
there are 3 x 100nF ceramic capacitors in 0805 size and one 6,8uF/10V tantalum capacitor B size. Bus drivers are TSSOP20

Hello, I built this and tested it with a bunch of SIMMs. It worked fine with all of them, so based on my testing compatibility seems to be fairly good. However, one of the 8MB SIMMs was recognized as 4MB.

The SIMMs starting from left are the following:
-4MB FPM, this worked even though it's 70ns
-8MB FPM with parity
-8MB FPM, double sided
-8MB FPM, single sided, this was recognized as 4MB
-8MB FPM, double sided
-16MB FPM with parity
-32MB EDO, double sided

Some additional notes:
-SMD PLCC socket is a bit of an pain in the ass to solder. If you do a second revision of the PCB in future, a through-hole version of the socket would be a nice quality of life improvement.
-Apparently ceramic capacitors are available in 10µF capacity in 1206 and even in 0805 size, so there's probably no need to use a tantalum capacitor.

EDIT: Please disregard my earlier report that several of the SIMMs would not work. This was caused by Simmconn being poorly connected to AWE64...