Front Panel

Memory

—

Serial output (88-2SIO, port 21)

—

What this is

A working MITS Altair 8800 front panel. The processor is a full Intel 8080 written for this page. It passes the four diagnostics the retrocomputing world uses to settle arguments about 8080 emulators: 8080PRE, TST8080, CPUTEST, and 8080EXM. The last of those checks a CRC of the result and every flag for each instruction group. All twenty-five groups match.

Why the lamps blur

On a real Altair the address and data lamps are wired to the processor's buses. While a program runs they do not display a tidy value, they show whatever the bus was doing millions of times a second. This panel samples the bus every machine cycle and lights each lamp by how often that bit was set, so a running program looks the way it looks in the photographs.

What is faithful, and what is not

• The 8080, the front panel logic, memory protect, and the sense switches at port 377 are faithful.
• Cutting power clears memory, as the original did.
• The WO lamp is active low on real hardware, printed with a bar over it. It is lit during reads and goes dark on a write, which is why it looks permanently on until a program stores something. That is correct, not a bug.
• SLOW throttles the clock so you can watch single cycles. The earliest machines left that switch position blank; later ones used it this way, and it is too useful for showing someone to leave out.
• On the real machine the switches sit in one row of sixteen. On a narrow screen this panel splits them into the two banks of eight that people already think in: sense switches on top, data switches below.
• The original needed a STOP then RESET after power-on to reach a sane state. This one resets cleanly, because that quirk is a fault, not a feature.

Sources

• Altair 8800 Operator's Manual, Part 3 — switch and lamp semantics.
• Kill the Bit — Dean McDaniel, 15 May 1975, original listing.
• 8080 CPU diagnostics — the test suite used to verify the processor.

Built by Cisco Caceres. If you want the other end of the story, train a language model in your browser.