H-Bridge: Locked-Antiphase or Sign-Magnitude?

[meteor]

Most common H-bridge motor control implementations with PIC chips use the common Sign-Magnitude implementation, where there is a PWM signal and a direction bit.

The problem is that changing the direction bit inverts the PWM signal, making the programming a bit of a pain, and you need four control lines for two motors.


Apparently Locked-Antiphase uses just the PWM signal, and offers forward/reverse with a 50% duty-cycle neutral point.

Other than the obvious loss of resolution on the PWM signal, are there any other compromises/caveats to using this method? (power/heat, etc?)



Thanks.

[Comatose]

Most common H-bridge motor control implementations with PIC chips use the common Sign-Magnitude implementation, where there is a PWM signal and a direction bit.

The problem is that changing the direction bit inverts the PWM signal, making the programming a bit of a pain, and you need four control lines for two motors.


Apparently Locked-Antiphase uses just the PWM signal, and offers forward/reverse with a 50% duty-cycle neutral point.

Other than the obvious loss of resolution on the PWM signal, are there any other compromises/caveats to using this method? (power/heat, etc?)



Thanks.

Can't give too much away, as our drivers are a variation on what you're calling locked-antiphase. http://www.dimensionengineering.com/SyRen10.htm some of the feature set comes inherent from the topology and some of it comes from things we've done to modify it. I'll leave it as an excercise for the reader to figure out which.

Signal-direction is flat-out the easier topology to work with. With the other, you have all your transistors switching on and off every switching cycle. You have to have some sort of very reliable anti-cross conduction scheme in place, or they'll popcorn quickly. With the one-wire implementation you can't actually turn the bridge off. Switching losses will be at least 4x what you'd have with signal-direction (one transistor switching instead of four)

Unless you're working below 2 amps or so, you'll have to go with a two or four wire implementation to get good results with either topology. If you're working under 2 amps, use an integrated driver chip and call it a day.

I really like the multiple-transistors-switch driver topologies. But realize that our SyRen series took about 500 man-hours of development to bring to market, while a PWM and direction driver would have taken about 100. If all you're trying to do is save yourself hassles, its not the best road to go down.

[meteor]

C... If you're working under 2 amps, use an integrated driver chip and call it a day.


I guess I should have mentioned that I'm focusing on low-power H-bridge driver chips, such as the L293D and SN74410.

Does the "one-wire per motor" system work better/easier with these devices?



Thanks.