What hurt is the cap doing in your estimation?

To be clear I concede that the schmitt trigger is not a switch de-bouncer. OK, with that said I tested the circuit in #14 and
measure the full ON voltage vs OFF voltage and the difference or hysteresis is less then 100mv. The reason the output on Q2 is .8 volt when conducting is because of the lower resistor values for R1 and R3. It's the increased base current also flowing through R5 causing the higher voltage.
The circuit I'm posting below has an appx. ON voltage of 2.13V and OFF of 1.65V with a hysteresis of about 500mv. Low output voltage on Q2 is below 500mv.
As far as de-bouncing the switch I rearrange the input configuration and increased C1 to slow the rise and fall voltage at the input to Q1 to take advantage of the hysteresis.
Would be interested if this is an improvement.

 

I don't have your R7 here but I think I'm going to order my PCBs so I can at least play around with different component values without having to deal with the mad house that is my protoboard.

This also also piqued my curiosity enough that I want to try an independent trigger circuit where I will try having the additional series resistor on the input, though I would move R6 to the same input node for faster discharging.

The whole point of my circuit has been using logic gates to generate a pulse that is approximately set by the propagation time of a single gate. So I've been after a single rising edge, I can afford the input being heigh forever and the logic only generates one pulse, but it has a bad time with slow rises or bounces.

 

Not a direct answers to your requests but when implementing a counter with a rather noisy input signal, in the need of a Schmitt trigger I went to wikipedia and implemented the topology shown there which seems probably similar to yours. Worked at the first try.

Should have to search to retrieve the actual details.

 

Try putting C1 across R4 in your simulation. That may change it into a debouncer circuit.