1
\$\begingroup\$

I'm in need of an H-Bridge capable of driving moderate-high current for motors on a small car (7A peak, 2-3A nominal, 12-14 VDC), and I'm hoping to control it with two (2) digital pins (5 VDC).

I've come up with the following controlling circuit which, based on the 2 inputs (FWD, REV), I think/hope are capable of the 4 desired states:

  1. Coast: FWD=Low, REV=Low
  2. Forward: FWD=High, REV=Low
  3. Reverse: FWD=Low, REV=High
  4. Braking: FWD=High, REV=High

So my questions are:

  • Are there any actual op-amps that will work in the circuit? I only have a few ICs in my stock (LM339N) and it doesn't work when prototyped (even with pull-ups since the outputs are OC).
  • I hope to be able to drive it using PWM (probably from an Arduino). With one pin Low and PWM on the other it should alternate between states 1-2 or states 1-3. Is this possible, and safe? Or are there timing issues at the transitions that I should concern me?
  • Ultimately, am I deluding myself or can the circuit work as I describe above, either as-is or with some tweaks?

enter image description here

Schematic intents: R1-R4 limit current to the MOSFETs, R5-R6 give a ~2.5V reference voltage (Vref) for comparison with the input signals. R7-R9-D2 (R8-R10-D1) shift reference voltage on pin 5 (pin 8) so Q3 (Q4) cannot be on with Q1 (Q2) and short/drain the battery. R11,R12 pull the inputs low to disable the motors when the digital inputs are disconnected.

Edit: This is for a single, hobby project at home that I expect to wire up on a simple proto-board, not even a small-run fabricated PCB.

I added emphasis on the 2nd half of the first sentence (2-wire control) - which was my primary purpose in posting: I wanted a 2-wire controller to make it impervious to software bugs.

\$\endgroup\$
1
\$\begingroup\$

I recommend using a pre-built motor bridge controller chip. You have lots to choose from; Digikey has 4280 options.

I say this for several reasons:

  1. Non-overlap times must be enforced, or else you'll release the magic smoke.
  2. Driving MOSFET gates through large resistors makes them sensitive to Miller turn-on.
  3. Purpose-made ICs will be generally less noise-sensitive.
  4. These purpose-made circuits often have internal charge pumps, permitting you to use the same transistor (i.e., all-NMOS) for all of your switches.
  5. In the end, it'll take less board space and be cheaper.
\$\endgroup\$
  • \$\begingroup\$ This is good advise. \$\endgroup\$ – gsills Apr 19 '15 at 23:20
  • \$\begingroup\$ I appreciate the suggestion, but from the Digi-Key link with: (through-hole, 10+A, full H-Bridge, brushed DC) -- the cheapest part is over $240.00. That seems rediculous, so I'll look around for other pre-fab boards (eBay, etc). Originally, my intent was more to learn.... \$\endgroup\$ – NVRAM Apr 21 '15 at 4:14
  • \$\begingroup\$ Why insist on through-hole? At the power levels you're dealing with, you don't need a heat sink; you can use the PCB for heat removal, which means surface-mount components will do fine (better). Allegro makes a fine line of motor driver products. \$\endgroup\$ – Zulu Apr 21 '15 at 17:09
  • \$\begingroup\$ @NVRAM: And if you want to build your own H-bridge, you don't want to be using comparators. A circuit built out of discrete transistors, and perhaps opto-isolators, will do much better. Or if you want to go for a bit more integration, many companies make gate driver ICs for half-bridges -- you just pick the power MOSFETs, and they handle the level shifting and non-overlap times. Additionally, Silicon Labs and Analog Devices make some great isolated gate driver products. There are many, many awesome options, but using LM339's or op-amps isn't one of them. \$\endgroup\$ – Zulu Apr 21 '15 at 17:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.