MAMEWorld Forums - EmuChat - Re: Analog emulation

It's all about solving differential equations. Does [url=http://en.wikipedia.org/wiki/Rungeâ€“Kutta_methods
]Runge-Kutta[/url] mean anything to you? If it doesn't, you probably need to study some more.

> How I should treat a capacitor is already a little more difficult. Should I care at
> all? Probably, because a capacitor while being charged is conductive.

Not just that, but you need to model the effective impedance of whatever's driving it. TTL circuits have far lower output impedance when the output is low than when it's high. Classic 4000 series CMOS outputs' current capacity varies with the input state. You also have to consider the characteristics of the input it's connected to. TTL inputs sense current flow - pulling them down switches on the multi-emitter transistor. CMOS inputs sense voltage, and look like a small capacitor to the driving circuit.

> Diodes are conductive for one polarity, depending on how anode and cathode are
> connected to other parts.

No, they allow a small forward current to flow up to the bias voltage, at which point they allow far more current to flow. They allow a smaller reverse current to flow up to the Zener voltage, at which point they allow a much larger current to flow.

> Now comes a difficult part: a network of transistors, resistors, diodes and
> capacitors. Would it be feasible at all to try to evaluate such a network's digital
> equivalent, if there is any?

No, it wouldn't. You need to be able to deduce the differential equations for the voltages and currents, and solve them using numeric methods.

> The other Pong emulations treated that specific section of the schematics as being
> non-existent (mine) or as a special black-box digital circuit (Dan Boris'). I would,
> however, prefer to try to emulate such a circuit at the simplest level, using only
> two voltage levels of 0V and 5V and just time delays due to capacitor values.

Time delays won't cut it - you either black box it, or simulate it properly.

> Is there someone out there who would be able to tell me how to tackle this job?
> That'd be fine, because otherwise I see no way out but using replacement black-box
> function generators in the schematics.

Seriously, go with black boxes, and consider studying circuit theory if you want to move to a more realistic model in the future.