Author Topic: H-bridge output short circuit protection  (Read 8353 times)

0 Members and 1 Guest are viewing this topic.

Offline jyrgenTopic starter

  • Contributor
  • Posts: 14
  • Country: se
H-bridge output short circuit protection
« on: November 03, 2011, 07:40:54 am »
I'm using the Allegro A3959 H-bridge to power a motor and that works fine. I want to add short circuit protection to the motor output but I'm not sure how to do that. The short circuit protection should ideally handle all possible shorts which means OUTA<->OUTB, OUTX<->GND, OUTX<->Vss.

A fuse does not work because it is to slow. The H-bridge can handle 3A continuously and a 6A spike that is <3us.

 

Offline codeboy2k

  • Super Contributor
  • ***
  • Posts: 1836
  • Country: ca
Re: H-bridge output short circuit protection
« Reply #1 on: November 04, 2011, 03:59:46 am »
Don't try to sense current on the OUTA or OUTB lines, it's too complex because of the need to sense current in both directions on both lines and shutdown both lines.

It's better to sense the current on the supply line going into the top of the H-bridge and shut off the supply if there is any current greater than 3A for longer than 3us.

An N-channel MOSFET on the supply into the H-bridge, a current sense resistor, op-amp and RC circuit to shut it off if you want to craft it yourself.
It's basically a high-side load switch with current sensing for shutdown.

You can also use one of many integrated high-side current switches, like the IO AUIR3315.  This particular device is just an example, it doesn't fit your requirements,
but you can search for one that does.  This one uses a single resistor to set the over-current  protection and latches off in 40-100us .. That's too slow for you, which
is why I said you'll need to search, but you get the idea.

 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3802
  • Country: au
Re: H-bridge output short circuit protection
« Reply #2 on: November 04, 2011, 04:24:21 am »
The chip has current sense built in already, so you don't have to add anything external.
 

Offline codeboy2k

  • Super Contributor
  • ***
  • Posts: 1836
  • Country: ca
Re: H-bridge output short circuit protection
« Reply #3 on: November 04, 2011, 10:09:23 pm »
The chip has current sense built in already, so you don't have to add anything external.

I didn't even check the spec on the chip , I just assumed he needed it, but you're right, and on closer inspection it seems that the
Allegro A3959 already has low side current sensing and current control, which seems to try to PWM the supply voltage to limit the current.
I don't know what it would do in the event of a full short on the OUT terminals

The high-side load switches with current sense would shutdown the supply completey, if that's what is desired, and would
require an external reset.

I guess it all depends on what the OP really wants to achieve.
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: H-bridge output short circuit protection
« Reply #4 on: November 05, 2011, 12:34:26 am »
Honestly it looks like this IC has only some kind of current limitation ("current control" with a fixed off time), but no real overcurrent/short circuit protection. The overtemperature protection won't help either, as it is most probably too slow to protect the FETs in a short circuit condition. Indeed the datasheet states that the existence of the overtemperature detection "should not imply that output short circuits are permitted"...

A safe H-bridge needs overcurrent sensing on all 4 FETs. With a common current sensing on the lowside, you can only detect overcurrent flowing through the lowsides, i.e. short circuit of the load or a short circuit to power on the outputs. Actually I think that the FETs will be destroyed in these fault conditions, even if the overcurrent sensing should be fast enough (not sure about this), as the bridge is not completely disabled, but just switched off temporarily. That's probably why they call it current regulation, not overcurrent protection.

Furthermore in some ("decay mode") configurations, the bridge seems to enter freewheeling on the lowside after sensing an overcurrent, so the lowside transistors are switched on for a certain time. This is ok if you assume in a normal operation condition (current blocked by the highsides), but with a short circuit to power on one of the output terminals, this could destroy the lowside FETs. And of course a common current sensing on the lowside makes it impossible to detect a short circuit to GND on the outputs, so the highside transistors will be most probably destroyed when shorting the outputs to GND.

BTW: even with OC sensing on all the 4 FETs, the highsides can still be destroyed in a reverse current condition. I.e. the bridge supply voltage is low and the outputs are connected to power, as then the current can flow in reverse direction through the highsides.
Trying is the first step towards failure - Homer J. Simpson
 

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3802
  • Country: au
Re: H-bridge output short circuit protection
« Reply #5 on: November 05, 2011, 02:07:25 am »
Honestly it looks like this IC has only some kind of current limitation ("current control" with a fixed off time), but no real overcurrent/short circuit protection. The overtemperature protection won't help either, as it is most probably too slow to protect the FETs in a short circuit condition. Indeed the datasheet states that the existence of the overtemperature detection "should not imply that output short circuits are permitted"...

A safe H-bridge needs overcurrent sensing on all 4 FETs. With a common current sensing on the lowside, you can only detect overcurrent flowing through the lowsides, i.e. short circuit of the load or a short circuit to power on the outputs. Actually I think that the FETs will be destroyed in these fault conditions, even if the overcurrent sensing should be fast enough (not sure about this), as the bridge is not completely disabled, but just switched off temporarily. That's probably why they call it current regulation, not overcurrent protection.

I think you are wrong here - the current protection is adequate.  I have not looked at the data sheet, but I think the capacitor across the current sense is just big enough to stop the current sense tripping during the fet switching transients. Otherwise on overload, it will interrupt the current ON cycle.  Since all output power goes through the current sense resistor, it is a genuine short circuit current protection.  On application of a short, there will always be a high transient power that the MOSFETS have to absorb, but they are very tough, so as long as the transient power is within the transient specs for the MOSFETS, then the supply is protected.
Quote

BTW: even with OC sensing on all the 4 FETs, the highsides can still be destroyed in a reverse current condition. I.e. the bridge supply voltage is low and the outputs are connected to power, as then the current can flow in reverse direction through the highsides.

This is very true. Unless you design a 4 quadrant switching circuit that can both sink and source power, and you have something like a battery to dump excess power into, then reverse power will cause overvoltage to the bridge and component failure. Under these conditions, the MOSFETS just act like rectifiers dumping the power onto the supply rails.

So
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: H-bridge output short circuit protection
« Reply #6 on: November 05, 2011, 09:20:47 am »
Since all output power goes through the current sense resistor, it is a genuine short circuit current protection.
Not in case of a short circuit to GND. In this case the current flows through the highside, but not through the current sensing circuit. As I pointed out, this circuit could only protect against a short circuit of load or a short circuit to Power.

Besides, as remarked before, the lowsides are in danger, too:
"Inputs PFD1 and PFD2 determine the current-decay method after an overcurrent event is detected at the SENSE input. In slowdecay mode, both sink drivers are turned on for the fixed off-time period."

Turning on both highsides or both lowsides in an OFF condition is usually called freewheeling on high/low side. I understand that the "current control" mode turns off the highsides, but turns on the lowsides in parallel (at least in "slowdecay" mode).
This is ok in a non-short circuit condition, but with a short circuit to power, this means that the lowsides are most probably destroyed since they are turned on instead of shutting them off.

Anyway, as I quoted, the datasheet quite explicitly says that short circuits are not permitted.
Trying is the first step towards failure - Homer J. Simpson
 

Offline jyrgenTopic starter

  • Contributor
  • Posts: 14
  • Country: se
Re: H-bridge output short circuit protection
« Reply #7 on: November 09, 2011, 07:53:09 pm »
Thanks for the input on this!

In my case I think that the current sense suggestion on the input power might be the solution I have to try. The most likely shorts are OUTA<->OUTB or OUTX<->GND.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf