MITS Altair 8800 front panel operation of one-shot 74123 Herb Johnson May 26 2026 This note was inspired by a close exchange to diagnose by email, a MITS Altair 8800 front panel which wasn't operating correctly. Here's the link to that May 2026 discussion: https://forum.vcfed.org/index.php?threads/mits-altair-8800-strange-front-panel-address-bus-issue-need-help.1257536/page-2 This note discusses the role of 74123 ICs as pulse-producing signal generators, which drive the careful timing of stepped operation of the 8080 processor on the MITS Altair 8800. This logic also applies to other 8080 based front-panels of the era. I provide means to calculate the intended pulse widths, as MITS (et al) didn't provide those timing values. Thnaks to Dave daver2 for his encouragements to these considerations. - Herb --------------------- A common 21st century problem with the MITS Altair 8800 front panel (also with the IMSAI 8080 and other S-100 front panels of similar design) is accummulated damage to the 74123 one-shots (monostable multivibrator IC's and their timing resistors and capacitors. These generate pulses of specific widths, to operate the stepping functions of the front panel. The front panel forces the 8080 processor to run by cycle or by instruction. This is NOT done only by hard wired logic, but also by "soft" pulses that are timed to "match" the timing of the 1MHz 8080 clock and 8080 instruction cycle. When that pulse timing falls out of alignment of processor timing, the front panel fails to operate reliably if at all. https://www.retrotechnology.com/herbs_stuff/sn74123.pdf Review a 74123 data sheet for the detailed opeation of those devices. The link above is to a TI brand 74123 sheet. These are not simple flip-flops that either work or not work; they are *analog* devices whose operation degrades over half a century (1976-2026 as of this writing). The IC, the resistor, and the capacitor degrade, experience shows. The ultimate solution is to replace the parts. But don't fix first and test later - test & confirm before repair. Please also consider, that the 74123, a "TTL" device, has different timing from the 74L123, 74S123, 74LS123, 74ALS23, 74C123 and so on - again, these are *analog* devices that depend on measuring a voltage level from a resistor and capacitor! DO. NOT. SUBSTITUTE among those different classes of logic devices. Dave also pointed out to me, the timing calculation varies slightly from *brand to brand* of 74123 - so check among brands to see how that matters. Operation of the IC can be measured with care using an oscilloscope to show function and most important *the pulse width generated*. The intended duration can be calculated, from the designed (schematic, marked) values of the resistor and capacitor. Point being: the *actual* duration is what you actually measure. Likewise the *actual* reistance and capacitance are what you measure. What you can't measure, is degrading of the 74123 - you can only measure the resulting pulse width. Following are details of measurement and calculation of pulse width. Operation of the resistor can generally be measured in-circuit using an ohmmeter. But measuring the capacitor often means removing one pin from the circuit, because the circuit (the 74123) adds capacitance or increases capacitor current during in-circuit measurement. Also, use a reasonable accurate DVM with capacitance scale. Also consider the tolerance of the resistor and capacitor. Resistors are marked as 5%, 2%, 1% and so vary even new from marked value. Capacitors are often low-tolerances like 20% or even more - usually too high rather than too low. https://deramp.com/downloads/altair/hardware/altair_8800_computer/Altair%208800%20Theory%20of%20Operation.pdf includes schematics for the Altair 8800 front panel. The monostable (one-shot) ICs 74123 are dual units, and they are at positions F and G and K. Here's descritions of their R and C values from the MITS schematic, and an estimate of pulse width as described later below: 74123 section cap resistor pulse calculated Measured on MITS F pins 14-15-3 0.1uF 47K 1.336 millisecond 1.35-1.4 ms F pins 6-7-11 20pf 7.5K ~150ns data sheet ~150ns G pins 14-15-3 0.001uf 30K 8.59 microseconds* G pins 6-7-11 0.01uf 30K 85.9 microseconds* K pins 14-15-3 0.1uf 30K .859 milliseconds K pins 6-7-11 20pf 7.5K ~150ns data sheet ~150ns M pins 6-7-99 20pf 20K ~300ns data sheet ~250ns L pins 6-7-11 20pf 25K ~400ns data sheet ~300ns * an Aug 8 1975 errata sheet in the MITS 8800 Assembly Manual, says IC G, C7 (pins 14-15) should now be 0.01uf, and C8 (pins 6-7) 0.1uf. Increasing the capacitance by 10, increases the pulse width by 10. - Herb Data sheets for the TI brand 74123 (not the LS or ALS or other variations) from 1983 say, the pulse width is defined based on R and C. The duration calculations is below, with C is in pF, R in kohms, Tw in nanoseconds. BUT this calculation is only good for C above 1000 pF ( .001 uf); below that, the data sheet has a graph to read. Tw = K * R * C * (1 + (0.7 / R) ) where K is a constant for the 74123 of 0.28 (.32 for the 74122, .28 for the 74130). There's another calculation which is simpler but should generate very close to the same result. There may be other modifications of this calculation for other brands of 74123, and also for the 74LS123 and other varients. I said, this is an *analog* part. for .1uF (100,000pf) and 47 kilo ohms, the calculation is Tw = .28 * 47 * 100,000 * (1 + (0.7 / 47)) = .28 * 47 * 100,000 * 1.015 = 1,336,000 nanoseconds = 1.336 milliseconds (This happens to match the timing of a 2026 oscilloscope measurement on a MITS Altair front panel, of a signal with R and C of those values, also 1300 useconds.) for cap values below 1000 pf, the data sheet shows a graphic chart. for 20pf and 7.5K, the chart shows a duration of around 150ns. (This happens to match the timing of a 2026 oscilloscope measurement on a MITS Altair front panel, of a signal with those R and C values, of roughly 150ns.)