Hi -
In my Apple IIe, which I did acquire used, there is this slot card I can't identify:
Looks like some weird sort of serial (RS232 ?) card to me (the rainbow ribbon cable hides the word "serial", and due to the many discrete components which could be level shifters), but I'm stuck with this conjecture. If it is a serial card, it must be 100% software driven. No ACIA IC to be found. But there are two PROMs on it.
It does not look like an Apple product.
Any hints to what it might be, and if there is a manual (pdf) for this sort of card would me much appreciated.
- Uncle Bernie
Looks like it could be a clone of Apple's serial card: https://www.cs.grinnell.edu/old/drupal6/Apple%20Computer%20Apple%20serial%20interface%20board%20for%20the%20Apple%20II.html
I agree with Dr. Webster. Looks like a pretty close clone of Apple's original serial card. Unfortunately that card isn't great. Ones using a 6850 UART are better and ones using a 6551 like the Apple Super Serial Card are better yet. There were also a few cards that used the Zilog Z8530 SCC which is a good dual port chip that was used in later Apple products like the IIgs and toaster Macs.
The original Apple Serial Card is pretty much only good for runing slow printers at 110 or 300 baud. I don't remember anyone I knew ever using one for a modem, in fact I don't think very many terminal programs even support it. I remember it mostly being used with Daisy Wheel printers like the Qume Sprint, Diablo 630 or NEC Spinwriter. Apple sold a version of the Qume printer as the "Apple Daisy Wheel" at one time. These printers normally maxed out at 30 characters per second.
Agreed, the "670-x005" part number and the labels "PROM P7" and "PROM P8" absolutely identify it as a copy of Apple's first generation serial card. Apple didn't use those number again on any other components.
Those schottky PROMs could consume a lot of power, and this card has two of them so Apple devised a circuit to switch-on their power supply only while they are actually being accessed. Luckily, schottky PROMs are fast enough for that. Curiously, Apple's schematic specifies lower-power LS471 and LS472 PROMs in those locations but they usually used the higher-power ones.
Selective ROM power.png
PS: Apple used similar power-supply control circuits in other cards that used schottky PROMs, such as the Centronics Card, Parallel Printer Card, and Disk II Controller. The Disk II burns the most power because it continuously powers both its schottky PROMs whenever ENABLE is active. Watch for cards that employ LS471 or LS472 PROMs...they're not common, but they do exist!
It's this one:
http://mirrors.apple2.org.za/Apple%20II%20Documentation%20Project/Interface%20Cards/Serial/Apple%20II%20Serial%20Interface%20Card/
The docs: (temporary):
http://www.ralf-kiefer.de/A2/Serial%20Interface%20Card.pdf
Regards
Ralf
Thanks for all the responses and the links. It's helpful for me to decide what to do with this card.
Most likely, I will take it out.
- Bernie
Yeah, it is an intereting item (I love clone hardware). But unless you've got a low speed serial device like a printer you want to drive, it isn't terribly useful. If you want to do serious serial communications an SSC is usually the way to go. Even it is a little limited for high speed use but in the day it was about as good as it got. Late in the Apple II run there were a few cards that used newer UART chips like the 16550B, but they are extremely rare. Sometimes you can find 16450 based cards that can be upgraded with a good 16550B chip. Those chips were common in IBM PC clones, and the 16 byte FIFOs in them are pretty much a necessity to run higher speed modems, 14.4k bps or above.
Damn, that's an understatement!
This circuit is laughably weak. And disassembling the firmware calls attention to additional flaws that weren't conspicuous in the circuit:
Circuit Schematic Serial Board.png
In short, this serial interface isn't suitable for most common applications:
But surely there must be some other application. Why else would Apple produce such an apparently-useless serial interface...and why did someone copy it??
Wikipedia talks a little about this card and the similar Apple Serial Interface Card.
https://en.wikipedia.org/wiki/Apple_II_serial_cards
The only times I actually knew of anyone using one of those two boards was for hooking up to the devices you mention. Acoustic coupler or 300 baud external moden or a low speed printer. In one case it was a Decwriter printing terminal (which also included an acoustic coupler), but usually it was as I mentioned in a previous post a Daisy Wheel printer. In the printer case it was usually they used a specific cable that was wired crossover like a null modem adapter if they were using an Apple Communications Card.
As you've found these are really pretty lousy cards. They were some of the fairly early cards released for the Apple II. If you look at the date the schematic was drawn in mid 1978 but probably the design work was done earlier than that.
I'd almost think that whoever designed thsee cards probably didn't know a lot about how serial communications usually work or maybe even didn't know of the existance of UART chips like even the fairly primitive Motorola 6850. Or maybe they were just cutting corners on cost as much as possible. Back in those days chips like the 6850 were considered pretty expensive by a lot of people. But when you consider Apple charged over $200 for these cards back in 1978 that seems excessively cheap, not "frugal".
The question of why someone would clone such an awful card is a good one because it doesn't make much sense. Even more so you have to wonder why Apple kept those two cards in their product line as long as they did. If memory serves they still offered them well past when the Super Serial Card was released. And the biggest question of all is why would someone have put such a card into a much more recent machine like a //e? By the time the //e was out the Super Serial Card and clones as well as tons of 3rd party cards using the 6850 or 6551 chips were available.
Before we bash the poor chap who did the cheap design of this card too much: are we sure the original card wasn't one of Woz's designs/ideas? :)
A simple interface card to connect an accoustic coupler would certainly be something that Woz would have wanted early on. And replacing an "expensive" UART chip with just a flip flop and some transistors, leaving everything else to be done by software, no timing support, except for relying on counting instruction cycles, sound pretty much like basic Woz design concepts. And if it was good enough to connect an accoustic coupler (or a current loop teletype, as mentioned in the manual), then, hey, it was good enough then.
The guy drawing the schematics ("B.K.") is the same who drew the Disk II Interface schematics. The latter clearly was Woz's design, and "B.K." just did the drawing. Could have been the same for this serial card. The revision entry "corrected schematics to match circuitry" is another hint that B.K. was just documenting someone else's work.
Don't let the spectre of "bashing" hold you back. The circuit's quirks and limitations reveal so much about its origins. Sometimes it's amusing to point and laugh at how little capability the device really had, but it's also really instructive...especially if you try to observe how they worked around those limitations (or failed to do so).
Look at this card's RS-232 IO section and notice that it has only two connections to the rest of the computer. You could build exactly the same capability by connecting it through the game port, albeit with a lot of power-supply connections too. In fact, a game-port hookup would be a little more convenient than the slot card because the input and output are fully independent -- your communications routine could read the input without worrying about affecting the output.
We could even cut down that schematic to a simple game-port implementation that's just as capable as the serial card's hardware. It could use game annunciator output AN2 for the sending channel and paddle button PB2 for a receive channel.
Use these transistors to send AN2 to XMIT and receive PB2 from RECV. Just copy the PROM firmware and patch it to read from $C063 instead of $C080,Y and to transmit bits via $C05C,Y instead of $C080,Y. It would be just as capable as Apple's serial interface without installing a single IC!
Serial via game port.png
Another benefit to poking fun at the circuit is that it reveals Woz's distinctive shortcuts.
If you study the PROM-enable circuits, it would initially appear that the schematic is faulty. Or wrong. Surely it won't work because the card always enables at least one of the PROMs:
But that 'obviously' wouldn't work, would it? If PROM P8 is enabled by default, why doesn't it fight with every other device on the data bus? Because it exploits the fact that the Vcc power supply for PROM P8 is switched off except for bus cycles when either IO SELECT or IO STROBE are active. So the PROM enable-logic is apparently wrong, but when both PROMs are powered-off it doesn't matter which one has an active "enable" signal.
P7 P8 enable and power logic.png
While you're at it, notice all the optimizations that were achieved through quirky use of J-K flip flops to switch the expansion ROM on and off via IO STRB and address bit A10. The circuit chooses to use the postiive output (Q) as an active-low signal, so that RESET initializes Q high, and IO SELECT clears Q low. That means IO SELECT always clears the flip-flop, but that's the ON-state for the expansion ROM. Every access to an expansion-ROM address ($C800-$CFFF) causes the flip-flop to latch the state of the A10 address bit, so you can turn off the expansion ROM by accessing any address (in that range) that has A10 high such as $CFFF or $CA81 or $CEEE. That gives you a thousand redundant addresses that will switch-off the expansion ROM, but what matters is that $CFFF does it. Without any extra gates.
Breathtakingly minimalist.
If this was Woz's design, then it clearly wasn't one of his best days. I could forgive cheaping out a design if the savings was getting passed along... but look at what Apple chearged for these cards in 1978-1982ish... $200+ Crazy. For that kind of price they could easily have used at least a 6850. Of course, that kind of pricing would probably be the result of the other Steve.