Microelectronics | AdLib Clone


I have always wanted to try an AdLib Music Synthesizer Card since the very early 1990s. But for some reason, I have never heard one and neither have I seen one. Maybe because Sound Blaster clones were already superseding AdLib cards back in 1993? I cannot say for sure. But I clearly remember seeing the AdLib option in the various Setup programs of the MS-DOS games back then.

I got the PCB free of charge from Martin of the Czech Republic. A big thank you since if it were for me, the AdLib clone project would've waited for a longtime to reach an upper position on my electronics TODO list. But Martin helped to speed things up a bit. So here I am again making parts lists and searching left and right for components in my spare parts bin.

The construction of this card is straightforward, and it took me only a couple of hours to solder every component. Instant gratification, if I may say it like that.

I plan to sporadically use this card with my 32-bit 80386DX ISA Single Board Microcomputer which is presented in System Design Assembled Machines.

Construction and Pictures

This 2-layer printed circuit board was made in China and has an ugly Factory identification code just above the ISA card edge connector. Also, it is slightly bent on both x-axis and y-axis. Thus, if you look at it into the light, from the edge, it appears to be slightly concave. Like a parabolic dish. Other than that, it is rugged and well-made. It also has some small scratches and dents but nothing bothering, anyway.

My PCB variant does not have an AdLib logo and is just simply labeled as Replica of 1990 Ad Lib. Which is great. I like it like that.

The PCB layout is kind of brutal for my taste. But I know the author tried as much as possible to replicate the original design. In my opinion, compared to other ISA cards out there, the AdLib design was outdated even back in 1987 in terms of layout and overall execution. And the routing of the circuit traces... It appears that the original layout is a combination of human and computer auto-router.

But in the end the card looks nice and the PCB has a nostalgic feeling.

I decided to use whatever part I had lying around in my spare parts bin. I only bought some of the ICs, the potentiometer, the mounting bracket and a couple of capacitors. All the rest is salvaged from other computer or industrial PCBs.

Chances are you have never seen a Romanian built 1N4148 diode in a yellow-black color scheme. Or gray augat-class IC sockets. Or even film capacitors used for decoupling purposes. I bet you saw WIMA film capacitors, but have you seen Siemens counterparts? They look almost like the WIMA parts. I like all electronic components that are quite uncommon, and I try to use them whenever I can.

I remember that I recovered the gray IC sockets back in 2008 from an industrial computer PCB assembly. But I cannot really remember what that card was supposed to do anyway.

For now, I am making visible progress with the soldering of the components.

I even installed an IRQ2 jumper albeit AdLib cards do not use any interrupt request lines at all. I will have to check whether there is any software out there making use of interrupts generated by an AdLib card. But in any case, I will test whether the Yamaha YM3812 chip is even asserting the /IRQ line.

The other jumper sets the base address for the card, in this case 0x388h.

I would just like to mention that position D over the base address selector is already hardwired on the PCB. That means the jumper is redundant and the AdLib card will always be located at address 0x388h regardless of which address jumper I might set. In addition, setting any other position would short output 6 of the U3 address decoder to the respective output selected by the installed jumper. The original designer wanted to have some flexibility in base address selection. But he (or she) ended up just adding unfinished functionality to the AdLib PCB assembly. And apparently the clone PCBs carry this flaw nonetheless.

This design flaw can easily be corrected by cutting the hardwire trace to position D. But that would look ugly and most probably the software is written for probing the AdLib card at address 0x388h only. For now, the jumper in position D will remain set.

I have soldered almost all the remaining components on the card.

Next, I soldered the electrolytic capacitor, the audio output connector, and the potentiometer.

I inserted almost all the integrated circuits into their respective sockets. I am still missing the Yamaha OPL2 chip and its associated digital to analog converter (DAC) IC.

A picture of the assembled AdLib card from a different angle. It sure does look retro!

Detailed view. I don't know who the manufacturer is for the yellow film decoupling capacitors, but the logo is interesting. I have a lot of them, all recovered from old industrial equipment.

Also, at some point, WIMA dropped the use of colored plastic cases for their film capacitors. Now they only offer them in red cases. In the past, I have seen green (FKP - polypropylene film) and blue (FKS - polyester film) as well.

Then I installed the steel bracket with two machine screws. Also, the Yamaha chips are in position.

I bought the YM3812 chip from eBay. It was deemed as a new, unused part. But the reality proved that it was recovered from existing electronics. It was then sanded and remarked with laser but I am positive that it is an original Yamaha part. The DAC chip was new, as in never soldered.

Well, how does it sound? Very similar to an OPL3. In fact almost identical. I like it.

However, that output stage based on the LM386 chip is awful. It introduces large amounts of hiss. While it's true that I tested the card with the headphones connected directly to the output jack, I am pretty sure that if I introduce this hiss in the electro-acoustical chain it will get amplified by the power amplifier, and it will propagate through the loudspeakers. The hiss is just like that of an old cassette tape machine without any noise-reduction circuitry. I tried to play Wolfenstein 3-D for a couple of minutes, but my high-sensitivity headphones brought the hiss right in the middle of my head. I quickly found out that I had issues concentrating on the gameplay. Thus, I abandoned the idea of using headphones with the Adlib clone card.

I don't have any immediate use for this Adlib card. As a matter of fact, my ISA Audio Interface card is a vastly better option for an 80386DX machine. In addition, the output noise floor of the AIF is minimal compared to that of the Adlib clone card. But the Adlib is a nice project nonetheless. I am pretty sure that I can add a simple RC network on the output of the LM386 chip so that I can filter out the hiss.

Many thanks to Eric Schlaepfer (TubeTime), who put a lot of work into replicating the original design. The project is also on GitHub. And thanks again, Martin, for the free PCB.

Copyright © 2004- Alexandru Groza
All rights reserved.
VER. 1.0 | REV. A