TPS43060: Inconsistent start-up issues with the boost IC and Low Side MOSFET

Hi Ashwini,

Thank you for designing with the TPS43060.  There are 4.7 Ohm gate resistor which may reduce the dead time between the high side and low side MOSFETs.  Could you please monitor the two MOSFET's Vgs voltage waveforms and see if there is enough dead time?  

Thanks,

Youhao

Hi Youhao,

Please find the waveforms there are issues that we can see. Dead band is definitely less which is a major concern but there is a 6.6V drop in HDRV waveform which is not understandable.

We could not get anything out of damaged boards but the waveforms on working boards are also not fine. Please find the waveforms below from a working board

Loading conditions: Vin = 15V, Iout = 0.5A from a DC Electronic load in CC Mode

HDRV Vs VCC

HDRV Vs LDRV at MOSFETS pin

Deadband Issue

HDRV Rise Deadband which is fine

Hi Ashwini,

I would recommend decreasing the impedances of the low and high-side FET in order to force a faster turn-on/off. As for the 6V drop, I believe what you are seeing in the first waveform is after the high-side MOSFET conducts, the boot-strap capacitor is mostly discharged, so the voltage across it is small. Therefore, the high-side gate voltage just sees the voltage across the boot-strap capacitor and SW pin, which is at 24V. You can check this by probing the difference between HDRV-SW, alongside with LO to check for the actual dead-time.

A question I have is that you mentioned the IC is damaged during EN turning on. How high are you pulling up this pin?

Regards,

Richard

Hi Richard,

1. By decreasing the impedance, I believe you mean decreasing the total gate charge (Qg) and gate resistor of both the MOSFET. Is that correct?

2. So is the waveform I am seeing on HDRV a valid waveform or do I need to increase the value of bootstrap capacitor to hold the charge for a while longer?

3. The EN pin is maximum 3V3 as it is driven by a microcontroller

Hi Richard,

Please find the waveforms below. The deadband looks about 60ns which is good enough. Definitely not the cause of failure

Yellow: SW

Green: HDRV

Blue: Math C1 - C2

Orange: LDRV

Hi Ashwini,

1) I just meant the gate resistor of the MOSFET.

2) Thanks for providing the waveform below. For future reference, when checking the dead-time it is important to check LDRV, and HDRV-SW to get the actual waveform to check the dead-time.

3)  Your EN condition seems correct here, so can you replicate the failure on a working board using an ABA test? 

Regards,

Richard

1. The dead time recorded in the waveform HDRV-SW and LDRV is 58ns - 60ns which is ok. Deadband doesnt seem to be the problem as per the waveforms attached earlier. Do we still need to reduce the gate resistor? It increases the ringing

2. The ABA test the boards failed at B

Board #1: All short in VCC, HDRV and LDRV

The IC was changed. There is no short now but there is no internal VCC of the IC generated

Board#2: Eventual failure of the low side MOSFET

The low side MOSFET was changed but there was no VCC of the IC. Then the IC was changed. Now we can see the internal VCC of the IC but it still stands disabled. The IC doesnt drive any of the MOS

Will look more into those board this week.

Hi Haroon,

Thanks for providing the information. When you say there is no internal VCC, are you saying that the VCC pin reads 0V?

Regards,

Richard 

Yes it reads a flat 0V but there is no short with the gnd 

Hi Ashwini,

Can you check the continuity of the VCC pin to the VIN pin and the BOOT pin? It seems strange that you could see VCC on the board and now it is not apparent. It makes me think that this is a connection issue. 

Regards,

Richard