Trying to design a circuit

[carlostex]

Hy there guys,

here's a complete newb, who knows little of electronics and is trying to design something for a retro computer. So the setting is this:

I managed to overclock a Tandy 1000 SL computer. This machine came originally with a 8086 CPU running at 8MHz. I replced the CPU with a pin compatible NEC V30HL which can run up to 16MHz. i noticed by taking a look in the SLtech manual that the OSC 24MHz goes into the buffer blue chip which divides the signal by 3 to supply the CPU with 8MHz. I took a look to see where the trace with the signal was, i i cut the trace, got a 48MHZ Oscillator, patched a wire into the spliced trace and basically i got a overclocked machine. If you don't believe it here it is:




But i want to do this better. Basically i want to design a dual oscillator circuit that i control via latching switch. It would work like a turbo switch although i could maybe only switch it when the machine was off. This would be great as i would be able to use the Tandy in a variety of speed ranges, which would help with the games i want to play.

I took my liberty to design a pseudo circuit. You know how when programmers write an idea in pseudo code? I did the same for my idea so i (poorly) made this drawing. This is where i need the help and knowledge to know if this is possible or not:

[Paul Price]

You can easily switch between the two clock sources in many ways, here are some suggstions:
1)Use a SPDT mechanical toggle switch mounted on the front panel to select outputs of a  14-pin bus driver (74HC126) that has  three state output enables you can select with a SPDT toggle switch. You would use pullup resistors on the individual output enable pins(1k-ohm).
2) You could even use a tiny 5V SPDT relay to control which oscillator source feeds the CPU.
3)A bubba solution is just using a SPDT switch carrying the clock signals to a a front panel mounted toggle switch. Use twisted pairs to the switch contacts and it might  work just fine. You might add a 330-ohm pullup resistor at the clock input to +5.
4)A slightly less bubba solution that you can build with just about any genpurp NPN transistors you might find laying around, for example (2) 2N3904 transistors, both in common-base configuration, with the collectors connected tied together at the clock input on the CPU, using a 470-ohm pull-up resistor to +5, and a front panel SPDT switch grounding either of the 2N3904 bases. You bias both the bases each with a 1K resistor to +5 and a LED (anode to base, cathode to ground, both LED's bypassed with a .1uf cap from base to ground).  Connect one emitter to each clock source.

[glarsson]

Some circuits don't function well if you switch between two unsynchronized oscillators (as in the example) as this will sometimes generate short pulses (the active oscillator just went high and you switch to the other oscillator that is low).