l6235 brake issue

[ross_atkinson]

i have a design using the L6235 motor controller. when the motor is running at 4800 I applied the brake and disabled the enable pin of the IC, all worked fine, But i was getting the motor overrun because of the inertia of the gears.
I tried to applied the brake with the enable pin high of the IC, but when i did this the IC blow. Is this because of my BEMF?
I believe the L6235 BRAKE pin can be used to quickly stop the motor while it is running: providing a low logic level to this pin all the high-side DMOS switch on, making a short-circuit across the motor windings. tried and it blow IC
While the motor is braking, both Thermal and Over Current protections still work? avoiding BEMF to cause a current exceeding the device's maximum ratings?
how can I change my design to avoid this?

regards

Ross

[EThom.3]

Hi Ross,

I haven't used this particular motor driver, but I rather doubt that the BRAKE pin is designed to stop a motor quickly. Especially if there is significant inertia involved. When you pull BRAKE low, the high-side FETs will short-circuit the motor windings, which can result in a tremendous current, depending on motor, inertia, wiring etc. The overcurrent protection won't help you, because the EMF current doesn't flow through the sense resistors when you do this. And the FETs may be dead long before the thermal sensor discovers that something is wrong. So, yes, I would definitely guess that Back EMF killed your chip.

The BRAKE feature is probably designed to (more or less) hold the motor in place, once it is no longer rotating. If you short the wires on a DC motor, and try to rotate it, you can probably feel what I am talking about.

I can think of only one relatively simple thing to try: Gradually reduce the speed by reducing the voltage on VREF when you need to stop.

As this device only has slow decay mode, I don't think it will increase the supply voltage while braking. Otherwise, there are simple methods of dumping the excess power in a resistor.

A more clunky hard-core method could be to use an old-fashioned relay to connect three power resistors to the motor windings when EN is low.

[EThom.3]

Correction: I had missed this:

"... the L6235 device features a non-dissipative overcurrent protection on the high-side power MOSFETs"

However, it will only work if you have connected the DIAG output to EN as in the datasheet, figure 19.

[ross_atkinson]

Hi EThom.3

thanks for the reply, yes i have connected DIAG pin to EN as per solution and my sense resistor is 0.3r. I thought if anything that it would go into current limit and protect the chip. I was running ok if i disable the EN at the same time as the Appling the brake, but this seems to not give me dynamic braking. I have a chopper back emf circuit, feeding the motor controller, but only has a 2w resistor. I have followed the dev kit design, tweeking some values for my motor and gears

cheers

Ross

[EThom.3]

Hi Ross,

I have read parts of the datasheet again, in particular section 10. Figure 19 is too simplified for the circuitry to make sense to me. All they show is an extra N-channel FET on each output, which happens to have the same drain-gate voltage as the power FETs. How that is ever going to deliver "a small but precise fraction of the output current", I have no idea. I believe that they have effectively omitted the "sensing element" from figure 19, which might really be measuring the drain-source voltage on the high-side FETs. Even though more than one page has been spent on it, very little is actually explained.

Even less is explained about the brake function. For one, I can't see anywhere if the non-dissipative overcurrent protection is active when the braking function is used. It might be, or it might not. It should be possible to test this, though, by connecting a big, fat load resistor between an output and ground, and then activate the brake. (No motor windings needed.) I would use a resistor value that limits the current to about 7 - 8 A. If the DIAG output does nothing when you set BRAKE low and EN high, then you have your answer.

I rather doubt that I can be of any help, and wish you good luck with your project. And, if you live on the northern hemisphere, a nice spring as well.

[ross_atkinson]

Hi ETom,

thanks for the help and the suggestion, I have spare ICs that I can fit to check this out. The datasheet and app note suggest the current would be ok as it states not to put extra diodes in. would i need a 3-4R resistor to on each outs to simulate the 3 windings?

my system is very basic and i have a static brake also, so I could rely on that and disable the EN. but would like to know how the dynamic brake works and have a solution going forward

cheers

[EThom.3]

Hmmm...

I actually don't know.

Well... thinking again, I would probably leave the motor windings connected, even though no current should flow through them (as the motor isn't rotating). And then just add the load resistor between one of the outputs and ground.

I don't know what supply voltage you have, but remember to calculate the dissipated power in the resistor, so you don't accidentally start a small fire.

[ross_atkinson]

Hi Ethom

my supply is 28v, the only thing i can thing to do is not use the brake and tie it high and use my the brake signal, to control a fet switch a load onto the h's lines via a diode and power resistors. enable the brake coms on the l6263 will be disabled also. so no over current is seen into the motor controller

[EThom.3]

Hi Ross,

Yes, that sounds like a good approach.

In any case, even if the overcurrent protection did work while braking, the FETs would just be switched off again every time they were engaged, probably not braking the motor much. The energy from the rotating motor needs to go somewhere – feeding back to the energy grid, charge a battery, or turn it into heat in a resistor.