@smashedProtonYes... and no.
The uA741 should be kept to an output below ~10mA (for the LM324, about 40mA). My rule-of-thumb is to use an output transistor if it's above 5mA, so as not to overdrive the 741. Since this circuit is just a simple comparator, that output transistor will operate in saturation mode (on or off). So, the four main parameters I consider are: max current (Icmax), max voltage (Vce), max power (Pdiss), and Hfe (or Beta).
If we're just driving an LED, then the opamp can pretty much handle that without issue. But, we're not driving one, are we? What we've got is an unspecified load - we only know it's a motor with a 6V rating (most motors can handle running down to about half the voltage rating up to about 1.25x). This means we have no idea how much current is going to run through it, nor do we know if that will vary based on startup surge, torque changes, or internal circuitry to maintain a stable RPM (Hall effect, etc.).
Now, a typical brushless DC fan motor will pull about 100-500mA @6V. So, if we look at that 2N2222, we'll find that:
Icmax = 800mA, Vce = 30V, Pdiss = 500mW, and Hfe = 100.
So, on paper, this all looks good. Since we're operating in saturation mode, the power on this thing will be the voltage drop (typically .7V) times the current being run through it. Assuming the full 500mA fan, this means we'll see about 350mW on that little TO-92, which is just a shade under it's max. Great. Now, all we have to do is figure for the Hfe...
Hfe, or Beta, for those who are following this and don't know that, is the amplification (gain) factor of the transistor. Shariar over at TheSignalPathBlog on YouTube does a great job explaining some of these parameters while designing a simple amplifier.
TL;DR version: Whatever you shove into the base will be amplified, by the Hfe factor, across the collector-emitter of the transistor.
So, if we send that 5mA from the 741 into the 2N2222's base, it should bump it up by a factor of 100, to give us about - by golly! - 500mA to drive the motor. In this case, it looks as if the 2N2222 is all we'll need...
Sort-of. Again, this is great on paper. In the real world, as ignator points out, that part may not match the datasheet. That motor may
run at 500mA, but the startup might be a real bitch. This thing is designed to cool things down, so we can guess it's going to be hot in there, and may very well have enough heat already on it, that when the extra 350mW gets to it, she burns up. Since that motor really has no known value other than "6V", we, as designers, have to design for a worst-case scenario.
Assume 1 or 2 amps for the motor, a watt or two on the transistor, and an opamp that can only spit out a milliamp or two. When I see that, I immediately think "Darlington". It doesn't matter if it's two individual transistors (as shown in "741motor6.png"), or the all-in-one TIP-120 (741motor4.png), the effect is the same:
TIP-120: Icmax = 5A, Vce = 60V, Pdiss = 2W (65W with heatsink), and Hfe = 1000.
Now we're cookin' with gas!Punching the numbers back into this: 1mA (741) x 1000 (Hfe) = 1A. The expected ~5mA will allow the full 5A to the load. However, pulling 5A across that .7V drop will put 3.5W into the package, so we'll need a heatsink. If we plan on only 2 amps, then power will be 1.4W, no heatsink required.
^The above is a simple (believe it or not!) run-through-your-head or back-of-envelope calculation. So, if the load is light, then just the 2N2222 can be used (as shown in "741motor7.png"). If the load is heavier, or you need the extra power dissipation, then go straight to the Darlington.
Another note: Since we're using relatively low voltages here, then dropping the value of the base resistor (R5) from 1K to 100-510R may help if you can't get the 2N2222 to pull the current for a light load. However, this may draw too much from a uA741, so only do this with the LM324.
@mrflibble, ignatorWhen using my TI uA741CN (1982, 14th week), the output was 1.8V when low (!), and about 1.6V below Vcc (I used a 9V battery, so it tended to drop voltage as the motor kicked in). Of course, this resulted in the damn thing not shutting off. Once I switched to my LM324 (Raytheon Corp, 1980, 12th week -
damn, I've got some old-ass mother-suckling chips!), the output went down to (IIRC) ~30mV low, and about 1.2V from Vcc.
They (LMX24) really make a stupid proof design
Yup. Except for that whole "Let's put the V+ at the bottom row, and the ground at the top!" thing. Note to chip designers: If you're dead-set on swapping simple things like where the traditional power rails go, then for Pasta's sake, put some damn reverse-protection on the forking pins!
I did lots of 6502 code assembly in the day, but this now is "obsolete for new design" and non procurable, at least that's what I've been told.
What's a 6502?
Just kidding. And you'd better tell the folks at Western Design - they just got done launching a newer series of 65xx chips designed for higher speeds (14MHz). Man, it's impressive seeing what my C64s can do at 1MHz. It'd be darn near pornographic to re-engineer one with the new WD chips, y'think?
nop
Boy, that got off-topic in a hurry, huh?