hi , interfacing (PC and DC motor)

Use a Solid State Relay (Hosfelt.com or similar) I would still buffer that with a small silicon transistor in emitter follower configuration. 2N2222 or a generic will work fine. The SSR works from +3v - +30vDC so it is real easy to interface and they will control a motor up to several HP if properly selected. If you are not familiar with the transistor arrangement I am talking about... Simply connect the 5VDC from the power supply to the collector, connect the parallel port you are using to the base and connect the emitter to the + side of the SSR. Minus on the SSR to power supply ground. The AC terminals of the SSR are in series with the load. Bear in mind the SSR will probably need a heat sink.

Solid state relays are pretty expensive, but are the only ones that will work OK with the setup you just described. A coil relay would be much cheaper, but you really should sink the coil current instead of sourcing it. I always use optoisolators, too, after I managed to burn out a parallel port during an accidental short a long time ago. The following site has a couple circuits that should work for both isolated and non-isolated setups:

(you might also be interested in his section titled "bad circuit example" -- though with the solid-state relay it *probably* wouldn't be as much of an issue.)

You can drive an SSR directly from CMOS but I usually use the good old 2N2222. The advantage is they are optoisolated from the load and from a guy like Hosfelt they are in the $12 range for 10 amp and $25 for 50 amps. You also avoid the dust problems and the inductive "kick" from a coil. I use the hell out of them. My 4xxx CMOS spa controller uses SSRs to run both pumps (3/4 and 2.5HP) and an 11kw heater. I also use them for lighting controls all around the house.

I looked at the site and I see what you are talking about. It is somewhat moot with the SSR since they "spec" at 3v and actually work a bit below that but I can see where a regular relay driver could be a problem. I agree the optoisolator is the safest choice and they are pretty cheap too if you mail order. I got a bag of them for about a dime each.

Regular relays are isolated from the load too, it's the external power supply that powers your 2n2222 that is the problem. Your relays already have an optoisolator in them with a 1.5K resistance (I checked the datasheets to see how it's hooked up). This would be great for hooking directly to the parallel port. Basically, by introducing your transistor, you've nullified all the advantages of the optoisolation, and it's much more likely to damage the parallel port.

Why wouldn't you simply use the PC supply to pick the SSR? Certainly it won't see another 5MA load?

I'm not talking about damaging the PC power supply, I'm talking about your parallel port no longer being optoisolated. The parallel port can supply enough current to drive an optoisolator LED with a 1.5K resistor. This is good because the parallel port is only connected to the LED and resistor, and nothing else...in other words, isolated. When you have an additional supply electrically connected to the parallel port pin, even through a transistor, there is a greater risk of something going wrong and current going where it's not supposed to. It's how I blew out my first parallel port.

You can take a calculated risk and consider it not worth the effort to optoisolate; but in this case, your relays are already set up for optoisolation. By introducing that transistor, you are throwing away a feature of your relays that could enhance reliability and lower the risk of damage to your computer.