I have recently refreshed a pair of Type 114 Pulse Generators as follows:
Both have serial numbers > 1680 and so are the latest revision in the manuals that I have
Both had leaky PS filter capacitors (originals from the Apollo era)
Both had carbon composition resistors which had drifted upward out of specification
Both had output amplifier transistors that showed slow rise time and very slow fall time. This caused pulse stretching and rendered the units unfit for calibration. Measured rise/fall times (as a unit) were ~15/55 ns using a 2 GHz scope.
After all the usual component replacements and adjustments, they are both working well and have pulse and square wave rise/fall times well below the 10 ns specification (< 7ns on a 2 GHz scope) given in the Factory Calibration Document. I swapped out NOS 2N2501 output amplifier transistors until I found pairs which gave fast, closely matched rise and fall times.
The as received o/p transistors in these units were not likely original as they either ECG 123A or unmarked/unreadable. I can only conclude that the OEM selected 2N2501’s were replaced at some point just to get the units running without regard to wave shape or rise/fall time.