DRV8844: How to prevent Overcurrent Protection being tripped due to inrush current

Hello,

The OCP shutdown will be triggered if the output current is higher than OCP threshold (5A) for longer than the deglitch time. The deglitch time for the DRV8844 is 5us and with 30% tolerance, you need to control the turn-on time to be shorter than 3.5us to avoid OCP trigger.

A problem that could arise is that the 3.5us turn-on time and the 20us PWM cycle time (DRV8844 PWM frequency less than 50kHz) might not be able to turn on the LED.

Let me know if a PWM signal of around 17.5% (3.5us/20us) is large enough to turn on the LED and operate correctly.

Hello,

I want to add a few more things to my previous reply.

In order for me to better understand your problem and find a solution, can you answer the following questions:

• How many LEDs are you driving with the driver?
• How are you connecting the LEDs to the outputs of the driver
• Can you send me a schematic of the LED load at the output of the driver

Hi Pablo,

I am only driving 1 LED FLOODLAMP during my testing and here is the LED I was using:

https://www.aps-supply.com/post-product/speaker-a6700-par36-12-48v-led-work-lamp-with-rubber-housing/

@12V VBAT, the LED turns on fine without tripping the OCP and current draw of the LED @ 12V is around 1.2A

@ 48V VBAT, the led just flashed for a brief milliseconds and remains shutdown. We are sensing an input in our MCU and turn on the LED when the input is triggered. The LED is connected only on one of the output. See below for the schematic with LED connected on OUTPUT8 to GND. Steady current @ 48V is around 0.4A. I think the LED Lamp has big capacitors inside that creates the huge inrush current.

I have not tried the 3.5uS/20uS duty cycle yet. But even if the 17% duty cycle works, we would need the full brightness of LED @ 100% duty cycle. I was thinking of cycling the reset pin of the DRV8844 but that would reset the whole chip and I dont want that to happen either.

ts of DRV8844 to GND. 

Hello Tor,

I agree with you that the big capacitance of the LED is causing the large inrush current. In this case, a possible solution is to set the INx high and PWM the ENx to slowly ramp the voltage up to the desire value. You want to keep the Ton time below ~3.5us at the start to ensure OCP is not triggered. Once the voltage has risen to a  level where the inrush is not large enough to trigger OCP, you can increase the ENx duty cycle until the voltage is at desired value. 

You will have to experiment in order to figure out when the inrush current is small enough to increase the ENx duty cycle and avoid OCP shutdown. 

If you can spare another output from this driver, you can connect two of the outputs (OUT3 and OUT4) in parallel to double the current and speed up the voltage ramping in the capacitor. 

Let me know if this solution works or if we need to think of something else.

Hello Tor,

Has your issue been resolved?

Hi Pablo,

The PWM duty cycle tweak did not resolve our problem. Even at 3.5uS/20uS Duty cycle still causing the OCP to trip. I was monitoring the Fault output and I could see it toggles several times before it completely shutdown. Is this normal? I was under the impression that the fault output will latch once it is triggered but in my case it is not.

I tried connecting 2 outputs in parallel, since the 2 inputs are driven separately with an MCU in our hardware, I have to use the EN pins to turn off/on the outputs. If the EN pins are active and then toggling the 2 inputs in my C- code somehow shutdown the device since it sees that the 2 outputs are shorted while the 2 inputs are toggled. And since all 4 EN pins in our hardware are tied together, that means all outputs are affected by this scenario.

The 2 outputs in parallel were able to run the LED without tripping the OCP but disabling all 4 outputs together is not what we intend to do in our application. It looks like a hardware change is needed to make it work.

Is the 5A OCP per channel or for all 4 outputs? I am guessing it's per channel since 2 outputs did not trigger the OCP.

Thanks.

Hi Tor,

I am sorry for late response. Here are my responses to your questions:

"I was monitoring the Fault output and I could see it toggles several times before it completely shutdown. Is this normal?"

-No, for OCP the nFAULT pin will remain low until RESET is asserted or VM power is cycled. There might be something else toggling the nFAULT.

"Is the 5A OCP per channel or for all 4 outputs?"

-5A OCP Per channel.

Here are some questions I have:

-Can you send me the Fault output if you can?

-When you connected the 2 outputs in parallel and were able to run the LED, did you toggle the EN pin?

Hi Tor,

Do you have any updates?

Hi Tor,

I will assume that your issue has been resolved. I will close this thread but you can still reply. In case the thread is locked, click on the "Ask a related question" button.