I’ve been struggling all summer with half-baked ideas to try to improve the class-D power amp lab in the applied electronics course. Since the first run of the course, there have been too many concepts packed into that one week, and too much stuff to do for students to understand it all. It also has been the one that been most difficult to convert to a home-lab exercise, because it used a triple power supply.
One idea that I have implemented, though I’m a bit worried about whether it will work, is to make the first amplifier lab, the microphone pre-amplifier, require soldering, so that the class-D lab can use the microphone pre-amp as a pre-built module, without having to rebuild it. I’ll need to think about the possibility of reordering the labs, to do a breadboard lab before the soldered pre-amplifier. Perhaps the transimpedance amplifier can come before the microphone amplifier??
The other idea I’ve been toying with, and that I’ll have to build and test this week, is to use a single power supply and an H-bridge for the loudspeaker. That way I could keep the voltage down to 6V (in spec for the op-amp in the preamplifier) and still get enough power to the speaker (4.5W for an 8Ω speaker). Because the voltage range is now ok for many cMOS parts, I can eliminate the open-collector output comparators, and use a rail-to-rail comparator like the TLC3702. I’ll need an inverter for controlling the two sides of the H-bridge, so I could use a hex inverter package and use a separate inverter as a driver for each of the four FETs. I could even have students get a second 74HC14N Schmitt trigger chip, as they have the same drive capability as 74HCU04 inverter chips, and increasing the number lowers the cost below buying one of each for the students. I estimate that the gate rise and fall times will be faster than what we’ve been getting with the open-collector designs, so shoot-through current should not be a major problem.
H-bridges are probably more useful to the bioengineers than open-collector circuits, because many of them will be taking the mechatronics class and building motor controllers. Having seen and designed H-bridges before will make their motor control seem more natural.
For the argument in favor of open-collector circuits: Many students will be learning how to do I2C interfacing in the sensors class, which uses open-collector (or open-drain) wiring, but the reasoning for the pullups needed in those designs is somewhat different from what is needed in the class-D amplifier anyway.
I’ve drawn up a schematic for a possible design and will try building it tonight or tomorrow. The last time I tried using a TLC3702 comparator in a class-D amplifier, I did not have much success, but I was trying to swing a much larger voltage then and did not have inverters as extra amplifiers for each gate—the inverters seem to have about twice the drive capability of the TLC3702 output. The combination of smaller voltage swing and greater current should make the transitions on the gates fast enough that shoot-through current should not be a problem (fingers crossed).