LM5066: Power cut off issue

Thanks for reaching out to us. Is the Connector B at higher or lower potential compared to the Connector A? Design calculator looks fine. However, there should be RC filter across the sense resistor to avoid false turn off during this connector removal or insertions. 

Hi Avishek,

Thanks for your reply.

The customer tried adding RC filters (C (1000pF), R (7.68 ohm) ) as you recommended, but the symptom did not change.

1. How should R, C values be calculated?

2. Is there a layout for arranging R and C that is recommended?

3. After setting bit3 of the GATE_MASK (D7h) register to ‘1’, the symptom disappears. What are the side effects of using this setting?

Regards.

JH

I will get back to you by today end of the day. 

Hi JH,

This circuit should be implemented. R1 = 10 Ω and C1 = 10 to 100 nF. 

Looking at your schematic, it looks there is no provision to add R1 and C1. May I know how you are adding these filter components in the PCB?

Please capture and share the following waveforms at the instants the connector B is inserted and removed.

• LM5066 pin 5 to GND
• LM5066 pin 1 to GND
• LM5066 pin 2 to GND
• LM5066 pin 25 to GND

If bit3 of the GATE_MASK (D7h) register is set to ‘1’, the controller won't turn off the MOSFETs while there is a real overcurrent event. This is risky. That's why this note is mentioned in the datasheet.

Hi Avishek,

Below is the captured image you requested.

LM5066_capture_image.pdf

Since there is no RC filter circuit in the customer's circuit, they tested it by constructing a temporary circuit as shown below.

Connector A is already supplying power, and when connector B is inserted, when measuring the current on the connector A side, it is confirmed that about 10A flows instantaneously as shown below. Can this cause the LM5066's OCP to work?

Thank you.

JH

Hi JH,

I am suspecting this is causing the fast trip to turn-off the Hot Swap controller. 

Hi Avishek,

What is the current at which the OCP operates? Can it be blocked by OCP even though it only increases by 10A?

Could you elaborate more on why you doubt it?

Is there any way the user can avoid this?

Thank you.

JH

Please give one day to think about what's going on inside the device. 

Hi Avishek,

Is there an update?

Thanks,

JH

Hi JH,

I am not sure what is happening in this design. Looking at this waveform, the timer duration corresponds to the insertion delay. The insertion delay triggers when the device is getting power on reset. I don't know reason for power on reset seeing this waveform.