LM74930-Q1: Inrush Current / Voltage Instability During Startup

Hi Nicolas,

Let me check this.

Regards,

Shiven Dhir

Hi Nicolas,

Can you try adding some CGS capacitor on HGATE and test again?

1nF should be good starting point.

Regards,

Shiven Dhir

Hi  Shiven,

We will do it and let you know

Thank you

Hi again Shiven Dhir.
We tried adding 1nF between  gate and source of HGATE mosfet and we got the same result. We're running out of ideas here, any other suggestions?
Regards

Hi Nicolas,

Did you try increasing the value as well? From 1nF to 10nF as so on.

Regards,

Shiven Dhir

Hi Shiven,

We tested the circuit using a 10nF capacitor, but the IC did not start up and the MOSFET did not turn on, so we removed it.

Additionally, we want to share some important details:

• As you may have seen in the schematic we posted in the first message, we have a 1kΩ resistor on the Vs pin. We did this because with a 10kΩ resistor, the Vs voltage stayed around 3V and the circuit wouldn't start up, for a reason that we don't know.
  
  
• We took the evaluation board and modified it to replicate our custom circuit. Specifically, we made the following changes to the evaluation board:
  
  
  • Rvs = 1kΩ (R10)
    
    
  • Ctmr = 100pF (C9)
    
    
  • Cgate = 330nF (C6)
    
    
  • Series resistor with Cgate = 47Ω (R6)
    
    

Configuration jumpers were set as follows:

• J5: VREG
  
  
• J6: VREG
  
  
• J8: GND
  
  
• J10: C9
  
  
• J11: 6A
  
  
• J13: VREG
  
  
• J14: VIN
  
  

We also connected approximately 7.8mF at the output of the evaluation board.

After applying 28V to the evaluation board, we observed the same instability issues that we encountered on our custom PCB. This can be seen in the following waveforms (yellow trace shows input current in all graphs):

(zoom in)

• When measuring Vgs, the instability worsens. This leads us to believe that the oscilloscope probe capacitance is adding extra load to the MOSFET's Vgs, and increasing this capacitance appears to make the situation worse rather than better. (Blue = Vgs, Yellow = input current)

• We discovered the problem is even more severe: in the graph below, you can see that the instability propagates to the input supply, causing voltage spikes of up to 60V. (Blue = input voltage, Yellow = input current)
  

• We've observed that increasing the output hold-up capacitance makes the instability worse. For a particular case of 1.5mF of hold up capacitor, there is not instability.
  
  

• In the following graph, we show measurements of Vs from our custom PCB:

(Green = Vs voltage, Yellow = input current, Blue = input voltage)

• We have tried different power supplies with various ramp-up profiles, but saw no improvement. And also:
  
  
  • We also reduced the gate capacitance to 47nF.
    
    
  • We changed the gate resistor from 0Ω to 10Ω.
    
    
  • We tried a Ctimer value of 47nF.
    
    

In all cases, the MOSFET appears to behave unstably.

We strongly suggest that you replicate this test on your evaluation board using the exact modifications listed above, with 7.8mF or more hold-up capacitor. If you identify any issues in our configuration, please let us know so we can test alternative setups in parallel while you investigate the root cause.

We would greatly appreciate your assistance in resolving this issue. We are concerned that this instability also occurs on your evaluation board. We are currently using this circuit with the LM74930 in multiple products for a customer, and this issue is critical for us.

At the moment, we are not compliant with DO-160G Section 16.7.5 Category I, and it is extremely important for us to resolve this and fulfill our commitments to the customer.

Regards

Hi Nicolas,

Thanks for all the details. I will replicate on EVM and suggest. Please expect my EVM results by Tuesday.

Regards,

Shiven Dhir

Hi Nicolas,

When replicated on EVM, did you change the FET? or was it the same FET as the EVM comes with.

Regards,

Shiven Dhir

Hi Shiven.
We didn't change the FET of EVM, it's the same FET as the EVM comes.
These are the modifications, as we mentioned in previous message:

• Rvs = 1kΩ (R10)
  
  
• Ctmr = 100pF (C9)
  
  
• Cgate = 330nF (C6)
  
  
• Series resistor with Cgate = 47Ω (R6)

Okay Nicolas,

I will test it.

Regards,

Shiven Dhir

Hi Nicolas,

I tested same conditions on my bench and I dont see the issue.

COUT = 7.9mF

Cdvdt = 310nF

Maybe the issue is related to the power supply.

Regards,

Shiven Dhir

Hi Shive,
First thank you for testing it.
I can see you’re experiencing the same initial current spike/s of approximately 9A that we are. That indicates instability:

That behavior doesn't seem right; the MOSFET is allowing too much current through and not controlling properly.

On the other hand, have you set the current limit to 6A like we did?  After the spike, the current protection kicks in for us and cuts the current.

Our configuration jumpers were set as follows:

• J5: VREG
• J6: VREG
• J8: GND
• J10: C9
• J11: 6A
• J13: VREG
• J14: VIN

Regards

Hi Nicolas,

This seems to be a fundamental issue and not related to controller. This is due to some parasitic and changing the FET to lower Gm is the viable solution.

AKX00067-1

Regards,

Shiven Dhir

Hi Shiven,

Thanks for your response.

If the root cause is indeed related to parasitics and Gm of mosfet, wouldn’t it be possible to mitigate the issue by adding a resistor at the gate? We think this should help dampen the oscillation.
Could you please run tests on your end using a gate resistor and let us know the results?

(we have already tested this approach on our side by adding a resistor in Rg=10 ohm, but unfortunately, it did not resolve the issue)

Best regards,
Nicolas

Hi Nicolas,

Since you are testing a EVM too, our FET remains the same and Rg will have the same effect. We need to test with the FET with lower Gm.

I can try increasing resistor further. For impactful difference, Gm needs to be lowered.

Regards,

Shiven Dhir

Hi Shiven,

We would appreciate it if you could test with larger gate resistors to see if the issue can be mitigated that way. Our preference is to avoid switching to a different MOSFET if the problem can be resolved with resistor adjustments.

That said, if the results are not satisfactory and you  try also with another MOSFET, please let us know the result. We really appreciate your involvement in addressing this issue.

Regards,
Nicolas

Hi Nicolas,

I will test and update.

Regards,

Shiven Dhir

Hi Nicolas,

100ohms in GATE path : 

The resistance isnt helping much either as it is much more dependent on gm.

Please change the FET to lower gm and test again. 

Regards,

Shiven Dhir

Hi Shiven,

Thank you for the feedback, we appreciate it.

Would it be possible for me to demonstrate that this solution is valid so we can consider applying it?
We are not entirely sure at this point, we have already tried changing several components and parameters without much improvement. And its a issue that appears in the official evaluation board

Best regards,
Nicolás

Hi Nicolas, 

I understand its appears on an EVM as well but it is a fundamental issue of MOSFETs when turned-on very slowly.

I may not have different MOSFETs with same package to be tested on EVM.

Regards,

Shiven Dhir