Simple question about H-bridge motor controller / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Simple question about H-bridge motor controller

1 2 next

Move Thread

LAN_403

cjk2

Mon Jul 21 2008, 10:36PM

Registered Member #51 Joined: Thu Feb 09 2006, 04:17AM
Location:
Posts: 263

In an H-bridge using 4 switches, and being driven by PWM, is there a significant difference in efficiency if the motor is short circuited for the off part of each cycle vs. the drivers doing tristate?

In other words:

On part of PWM wave:
{Output 1 is high, Output 2 is low = motor spins}
Off part of PWM wave:
{Output 1 is low, Output 2 is low = motor braking/off/shorted}

VS.

On part of PWM wave:
{Output 1 is high, Output 2 is low = motor spins}
Off part of PWM wave:
{Both outputs tristate, no switches conduct = motor costing}

It would seem to be that the second option is best, but the inductance of the motor will need to be clamped. If i go with the first option, I imagine the switching noise will be reduced.

Anyone have and opinion either way?

Steve Ward

Mon Jul 21 2008, 11:56PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

I would typically suggest letting the output tri-state during the "off" times, though this can be more abuse on the internal diodes of the MOSFETs, which might become an issue if you are running high currents. Also, most boot-strapping drivers rely on the output from each leg of the bridge to momentarily pull to 0V to refresh the bootstrap capacitor. If you simply let the motor free-wheel, this wont happen as it will drive the output to the rails.

What i ended up doing with my DC motor drives was having a 1uS period where the outputs both pulled low to refresh the bootstrap cap, but then the rest of the time was tri-state. This seemed to give less motor heating since the regenerative currents were lower, and in this case were going back to the power supply instead of just being shorted out though the fets.

cjk2

Tue Jul 22 2008, 01:08AM

Registered Member #51 Joined: Thu Feb 09 2006, 04:17AM
Location:
Posts: 263

Alright, i think ill go for choice one, which breaks the motor during the off cycle. I also intend to use this driver for electromagnet control so there will be alot of inductance/ high powers involved. In the past, I had always put a diode across the coil anyway to short it during the off cycle, so I think I will be fine.

Thanks for the input.

cbfull

Tue Jul 22 2008, 12:47PM

Registered Member #187 Joined: Thu Feb 16 2006, 02:54PM
Location: Central Ohio
Posts: 140

cjk2 wrote ...

In an H-bridge using 4 switches, and being driven by PWM...

Curious, are you using the H-bridge as a variable frequency amplifier for a single phase motor? Sorry if this seems obvious to others, I am familiar with H-bridges, just not VFD.

Craig

Steve Conner

Tue Jul 22 2008, 01:03PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

You also have to consider whether you want regenerative braking. Roughly speaking, this works by shorting the motor and letting current ramp up in its internal inductance, and then letting it kick back into the battery, just like a boost converter. If you always tristate, you can't do that, in fact you can't brake at all: it's a 2-quadrant drive.

It is possible to have a full 4-quadrant drive by just making output 2 a mirror image of output 1. At idle, both outputs have a duty cycle of 50% so the average voltage across the motor is zero. You can motor or regen brake in either direction with very simple control logic (you never tristate, so you don't have to make a circuit that decides when to tristate) but the price you pay is high losses.

cjk2

Tue Jul 22 2008, 06:36PM

Registered Member #51 Joined: Thu Feb 09 2006, 04:17AM
Location:
Posts: 263

Cbfull: This will be a constant frequency PWM drive for a brushed dc motor / large electromagnet at about 10khz or so

Thanks Steve. I have attached a simple diagram of how I currently plan on doing it. I had considered the drive you suggested where the motor/inductor actually sees AC at 50% duty cycle when its not spinning("locked anti-phase PWM"), but I assumed that this would be Inefficient. Perhaps shorting the motor for some time every cycle as I had planned to do is also very(more?) inefficient?

If i do do it the way you suggested, the motor is always regenerative breaking, right?

GeordieBoy

Tue Jul 22 2008, 09:55PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

> I had considered the drive you suggested where the motor/inductor actually sees AC at 50% duty cycle when its not spinning("locked anti-phase PWM"), but I assumed that this would be Inefficient.

Think of the motor as a voltage source (the back EMF) in series with a big inductor (the winding inductance.) If the switching frequency is high enough then the ripple current is triangular in shape and the amplitude is arbitarily low. Whilst it is true that the peak-to-peak AC ripple is highest in the resting state, it should not result in excessive heating with a big DC motor and high Fs. (If you application is battery-life critical you might consider suspending all switching when not motoring or generating to eliminate this loss and device switching losses whilst at rest.)

> Perhaps shorting the motor for some time every cycle as I had planned to do is also very(more?) inefficient?

Short-circuiting the motor just causes the motor current to free-wheel through the motor's inductance. As Dr Conner said, this automatically has a regenerative breaking effect if you program the PWM for a slow speed when the motor is being driven by an external prime-mover at a faster speed.

In the motoring mode the H-bridge can be considered as operating like a synchronous buck converter stepping down the battery voltage to drive the motor. And in the generating mode (regen breaking) the H-bridge can be considered to be operating like a boost converter stepping up the motor's back EMF to charge the battery.

The 4qd website might be worth a look if you haven't already checked it out.

Your diagram doesn't seem to show any provision for dead-time in the switching sequence. It would be wise to introduce a microsecond or two of deadtime either in software or in hardware. This is also one of those applications where reverse recovery time of the free-wheel diodes can be important.

-Richie,

Sulaiman

Tue Jul 22 2008, 10:33PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

Go to National website

and look at the datasheet for LMD18245
It's old/obsolete but the theory is worth reading.
Then look at related application notes for more info.

P.S. I think I have a couple of LMD18245 (or similar) lying around - paypal p&p and they're yours.
NOTE: 55 V dc maximum for the LMD18245

cjk2

Thu Jul 24 2008, 03:09AM

Registered Member #51 Joined: Thu Feb 09 2006, 04:17AM
Location:
Posts: 263

Thanks for all the help guys, I ended up designing a driver that can be switched between locked anti-phase, and forward/reverse/tristate modes simply by reprogramming the ATMEGA88 that runs the board.

GeordieBoy

Thu Jul 24 2008, 10:06AM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

> I ended up designing a driver that can be switched between...

That's the best option - Maintain control of everything in software. Then you can play with different switching configurations to suit your needs.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob