L293 vs L298

Morning all....

I see that the 293 bills itself as a "Quadruple half-H driver" while the 298 is styled "Dual full-bridge driver". Voltage and current aspects aside, what are the things one would consider to choose between 293 or a 298? My simple arithmetic tells me that 4 x 0.5 = 2 so "quad half" = "dual" :~

The TI version of the 293 datasheet shows how to set up a reversible motor on 2 of its channels, so seems to me that the usefulness is much the same?

With the 293 you can drive a stepper motor with 4 windings, etc, so the quad channel comes in handy. As I see it, and I've used both bridges, as well as the 754410, the really big advantage of the 298 is, you can attach a large heatsink and pull 1.5A average with a larger battery voltage. OTOH, the 293 has very minimal heatsinking, and will get especially hot with higher battery voltages.

Eg, on a robot, you cannot drive very big motors with only 600 mA, but 1.5A gives you a much more useful range. Some of the appnotes talk about parallelling the 293 channels, and Acroname [IIRC] actually talked about soldering one 293 chip on top of another, but I never really liked those ideas all that well. Better to choose a more competent h-bridge from the get-go.

So apart from current, the usefulness of the 293 is that the quad setup manages a 4-wire stepper. But for DC motors, it's basically the same to run 2x DCs on a 293 or 298, again, current aside.

That's how I always ended up using these things on my robots. The 293 has 4 channels, but I only ever needed 2 for robot motors, so 2 channels normally went unused. And the 298 has exactly the 2 channels needed for the robots. OTOH, the extra channels on the 293 could be used for driving other things, like relays, solenoids, lamps, etc, provided they use the same Vcc as the motors.

Now, there is also another use for the 293. Since it can handle voltages as high as 36V, you could use it as the "first" stage of a high-powered h-bridge that uses 4 high-current MOSFETs, with 2 p-channel above and 2 n-channel below. The 293 could drive the gate voltages. Also, since the 293 has 4 channels, you've got the makings for controlling 2 of the MOSFET h-bridges side by side with a single 293 chip.

Thanks.... I'm going to point another thread to this reply, where someone is looking to drive some serious motors.

The 293 -> MOSFET h-bridge idea I mentioned as mainly an experimental thing, since you seem to be playing with these things. However, for "serious" motor control I would go with a more established route, where you have a lot of safeguards built in, eg, over voltage, over current, over temperature checks, etc. With the above, you would need to add those in yourself to have a safer circuit.

For your friend, I'd suggest something like the following, good up to 20A or so, http://www.pololu.com/catalog/category/11

For even more current, you could check the OSMC project, http://www.robotpower.com/osmc_info/ http://www.robotpower.com/products/osmc_info.html

High-current motor controllers aren't especially cheap, but you get what you pay for.

The other issue with the L298 vs the L293 is the pinout footprint of the L298; one row of pins is staggered by 0.05" so it won't fit a standard 0.1" spacing of a protoboard or a breadboard, without bending the pins (which can easily snap them off if not done carefully). If you are developing your own PCB, this isn't that big of a deal, but if you are prototyping on a breadboard or protoboard, it can be an issue.

I've found that these adaptors are very useful, though:


They are easy to solder up, and if you notice, the motor pins are kept opposite of the signal pins, which makes interfacing a lot easier. This guy also sells a PCB for mounting your own parts in place to build an L298 h-bridge module, if you already have such parts handy.

Yesterday, I was at my favorite junkyard when I found an interesting stepper driver board that looked like it used either a pair of L293 or SN754410 ICs on it; the board only drove a single stepper, though, at up to 2 amps (supposedly, based on what I read on the board). The chips were arranged side-by-side, and had a large heatsink mounted (bolted down) on top of them, so I couldn't see the parts to know for sure what chips they really were; I suspect the SN754410 because of the larger current rating (1 A vs 600 mA). It might have been some other kind of part, though (but the tale-tell weird middle ground/heatsink pins indicated that they were probably L293 or SN754410 chips).

I probably should've bought those boards - maybe next week...


Nice heavy traces.