Are you sure that's safe? While they run from 5V, the input thresholds move with the Vcc voltage and so don't look to be TTL compatible. I can't think of a family other than HCT that's got TTL thresholds while powered from 5V (normally we use all sorts of other families powered from 3.3V which conveniently puts the CMOS input thresholds roughly in line with the 5V TTL thresholds).[*]It's worth noting that the two devices I do power down are also LVC devices. However, these are single gate devices (74LVCxGxx), and this single gate series does support 5V operation, so work fine without any additional voltage regulator.
So at a quick glance, the ADLC_IRQ is OK as it has a pull-up to 5V (and isn't that fast in the first place). But the inputs to IC17 are TTL signals from the main board.
I'm also not quite happy with the diode in the 5V supply to the main module - though if you choose a really huge one it might be OK. The current here isn't trivial (~60mA from the line driver into the cable when enabled, plus all the static consumption of the various parts, so probably at least 100mA to be considered).
All in all, I'd incline towards just powering everything from the 5V. The statid buffer will be fine - a disabled HCT doesn't impose much load on the bus nor consume much power, the length of tracks are probably more significant than the chip itself. But the unconnected inputs of IC17 will need high-value pullups to stop them floating.
Statistics: Posted by arg — Sat Mar 22, 2025 12:09 pm