DRV8835 shutting down

Hi Than,

Is there a capacitor on the load? This can cause over current conditions, which typical cause the DRV8835 to stop driving for 1ms.

If there is no capacitor can you provide a scope capture of the outputs shutting down and restarting? Also can you provide a scope capture zoomed in when the device is re-enabling after the shut down? These can sometimes provide additional information.

In general, capacitance and resistive loads can create overcurrent conditions because the current can change quickly and the rate of current change is much higher than inductors.

Hi Rick,

There is no capacitor on this load. All I did was slowly incrementing Vm, by turning the knob of a lab supply. The chip is restored in working condition when I reduce Vm. So the trip point is Vm ~1V, load current ~ 500mA.

How does OCP work ? It doesn't look like it triggers on a level but on a rate of change ?

Than

Hi Than,

Straight from the datasheet:

Overcurrent Protection (OCP)

An analog current limit circuit on each FET limits the current through the FET by removing the gate drive. If this analog current limit persists for longer than the OCP time, all FETs in the H-bridge will be disabled. After approximately 1 ms, the bridge will be re-enabled automatically.

Overcurrent conditions on both high and low side devices; i.e., a short to ground, supply, or across the motor winding will all result in an overcurrent shutdown.

This does not appear to be an overcurrent, but knowing how long the bridge is disabled helps to rule it out.

Other protection is overtemp which usually takes longer to cycle and clear, and UVLO on VCC. Is it possible when driving current through the VM side, VCC is dropping below 2V?

Hi Rick,

The bridge is disabled indefinitely as long as Vm > 0.95 to 1V, and re-enabled indefinetely when Vm is below such level.  

I used a lab supply for Vcc, which is rock solid at 3.3V.

Than

310-642-7975 x309

Hi Than,

This sounds like it may be overcurrent event, although it is confusing why. The outputs should be attempting to activate for approximately 1us, then disable for 1ms. Because of this duty cycle, the power supply probably will show little change in the current.

If you have a scope, please look at the outputs. You will want to look for one pair of outputs going to 1V for ~1us or longer, then shutting down for ~1ms. If you can provide a scope capture, this will help.

Are you certain the inputs and outputs are connected properly? If not, this could be the cause of the overcurrent event.

DRV8835 Troubleshooting.pptHi Rick,  

Attached is a set of scope capture.

Than

Hi Than,

It appears like the output is trying to turn on, but is unable to sustain the voltage.

Can you zoom in after the bridge shuts down and provide the scope capture? Trigger on the rising edge of Lsr_Tec+ and measure the time it is high and the time it is low.

Another thing to try is adding a large (47 to 100uF) cap at TP24 to GND. Depending on the length of wires from the VM supply to TP24, C76 may not have enough capacitor to maintain the voltage.

Hi Rick,

I added pictures on pages 4-5. 

Another observation:  As Tp24 increases monotonically , the bridge got disabled and then re-enabled.  See pictures on page 5.

Than0385.DRV8835 Troubleshooting.ppt

Hi Than,

Thanks for the additional information. I am puzzled why the bridge is disabling at .75V and re-enabling at 1.25V, but once it disables it appears to go into overcurrent protection. Please note the repetitive pattern of approximately 1ms on page 4. Also please note the output is on for approximately 1us, which is the ocp deglitch.

Of note, there appears to be some capacitance on the Lsr_Tec+ output. If purely resistive, the output rise time time should be faster.

What is the part number of the 1.4Ohm resistor?
What does the device do when ramping the power supply down from 1.5V to 0V? Does the output disable and re-enable at the same voltages?
What is the part number of the power supply?

Hi Rick,

I don't have the part number for the 1.4 Ohm resistor.  It looks like a heavy duty power resistor from the lab.

Yes, the device behaves somewhat the same on the way down.

At this point in time, is there a way to disable the OCP, or is there any equivalent chip that is not so sensitive to load type ?

Thanks,

Than

Hi Than,

OCP cannot be disabled. It is there to protect the device.

The DRV8835 should work with a 1.4 Ohm resistor in parallel mode.
Can you provide a layout of the board? Perhaps there is something about the layout that is affecting operation.

On our side, we will attempt to recreate this behavior using our EVM in parallel mode with 1.4Ohm across the outputs as you described. This may take a few days.

Hi Than,

I was able to recreate your observation in our lab.

Although the datasheet states the recommended minimum VM range is 0V, it is referring to the fact that there is no undervoltage lockout to interfere with the device operation. For details of undervoltage operation, please see page 8 of the datasheet.

The DRV8835 will attempt to drive the outputs all the way down to 0V. The OCP (overcurrent protection) logic may not work properly as VM droops. If VM droops to 0V and returns to a voltage high enough to drive a load, the DRV8835 should begin driving no later than 1 OCP cycle after VM returns to the higher voltage.

Hi Rick,

Thank you much for the clarification and support !

Than

I've received similar results, the bridge does not work when Vm is within 0.9-1.4 V range. From the discussion above it looks that it is inherent feature of this driver, though not mentioned in the datasheet.
Is there any driver that covers 0-8 V range flawlessly?

BTW, in distinction from the datasheet, minimum voltage specified in DRV8835 web-page is 2 V, not 0.

Kestutis

Hi Kestutis,

Would you please provide a link to the web-page that specifies 2V?

The datasheet www.ti.com/.../drv8835.pdf , shows 0V as the minimimum VM voltage

Hi, Rick,

The page is - http://www.ti.com/product/DRV8835

Kestutis

Hi Kestutis,

On the page you have pointed to, it appears 2V is listed in one place. Vs min is listed as 2V.

Is this what you are referring to? It is confusing because there is no Vs; there is Vcc and VM on the DRV8835.

Thank you for pointing this out. We will clarify in the future.

Hi, Rick,

Just now tested DRV8838 - no "dead zone" throughout entire Vm range, starting from 0.

Yet something is wrong with DRV8835.

Kestutis

This is an interesting thread, very similar to the problem I have right now with the DRV8835 driving a TEC at 1A and below.
Higher currents work fine, but below about 0.4A the DRV8835 starts triggering its protection.
Adding up to 100uf hasn't helped.
I used this part because the H-bridge we normally use ZHB6718 requires more components (transistors), and is less efficient, but has never shut down either.

Was there a solution ?
Mike

Hi Mike,

The problem in the thread was related to a low VM voltage.

What is the VCC and VM voltage used? How are you controlling the current? Are you PWM'ing the outputs or adjusting the VM voltage?

When the protection circuit triggers, does it re-enable the outputs in approximately 1ms? This would be an indication of overcurrent.

Hi Mike,

My original intention was the same, using the DRV8835 for a TEC. However, I think this chip is designed for PWM motor applications, which are not the same as for a TEC.

I ended up building up my own bridge using a 4 N-channel mosfets. A mosfet has an Rds_on around 20-30 mOhm, so the efficiency is not to bad !

Than

A partial schematic is in my other thread I started few days ago.
Using a current source to drive VM. Since the TEC is < 2 ohms, the net voltage can be quite low.
Works ok above 0.5A, but protection kicking in when low.

Darn, hope you found a workaround.
Oh well, looks like I'll follow suit.