TPS50601-SP: Shutdown during current inrush

Hey John,

What is the soft start you are using for the TPS50601-SP?

Sometimes using low values for this can cause high inrush current.

I would be interesting of seeing this 10% dip at start up if you have waveforms.

Thanks,

Daniel

Hi Daniel,

We're using a 0.027uF cap for the soft start. However, I should clarify that it is not the initial inrush during startup that is creating the problem. The regulator's pwrgd enables a linear VTT regulator downstream. When that load kicks in, it's inrush is what is pulling down the 1.8V regulator. This also causes the pwrgd line to drop which disables the linear VTT regulator. The 1.8V appears to then enter a shutdown and then tries to come back up. In some cases this can lead to a limit cycle of starts and stops before the 1.8V finally enters into regulation.  We've observed that this limit cycle can be sensitive to the choice in compensation components and can persist for many cycles, .e.g >10 in some cases.  We have been able to prevent the limit cycle from occurring by modifying how the VTT regulator is enabled, thereby greatly reducing it's inrush current. However, in trying to get to root cause, I'm being asked to explain the initial mechanism that leads to this limit cycle occurring.The concern is that prior to the VTT modification, the nominal compensation values chosen exacerbated this issue. However, these values clearly provide better dynamics in a small signal sense as we've observed less overshoot during load changes. My primary goal is to understand what mechanisms could lead to this behavior to ensure we now have a robust design.

Thanks,

John

Hey John,

It would be more useful to see the inrush current as a load step on the TPS50601-SP.

When thinking of it like this it becomes easier to understand why this is happening.

In this case, this means that the main contributors to the droop will the the compensation and output capacitance.

Increasing the crossover frequency/output capacitance will help immensely.

Our PGOOD will drop when the output is outside of regulation.

This can be for many different reasons, but here its because the transient caused the device output to drop.

Thanks,

Daniel

Hi Daniel,

I understand why this load transient causes the output to droop and lose regulation. That's not really the question I'm trying to answer. I don't understand why the regulator is entering shutdown, you can clearly see the soft start cap being discharged before the regulator comes out of shutdown as described in the datasheet. The datasheet seems to only cite 3 reasons for entering a shutdown. 1.) pulling the enable below 1.032V. 2.) When the UVLO is triggered. 3.) Thermal Shutdown. We've ruled out 1 and 2 through direct measurement. 3 seems rather unlikely given the duration (80us) of the load current, plus locally cooling the part makes the problem worse (increases the length of the limit cycle).  We've also seen this behavior on multiple parts so it's likely not a part issue. Am I missing another intentional protection feature for entering shutdown?

Thanks,

John

Hey John,

The only three things that would shut the device off in this mode would be the current limit, thermal shutdown, or enable pin.

My guess if you were correct with the assumption it was the current limits.

The current limit of the device can change over temperature as well which would explain why it got "worse" at colder temperatures.

Thanks,

Daniel

Hey Daniel,

I thought that the current limit was used to determine cycle-by-cycle limiting as described in section 8.3.17 not for initiating a shutdown? There's no mention of this in the datasheet. If what you say is correct, this would definitely explain what I'm seeing. However, Fig 8 shows that the current limit (at least for the high side FET) goes up at lower temperatures. So it would trip less, which is not what we observed. It also doesn't explain why the limit cycle phenomena is so sensitive to the value of the compensation lead resistor.  Can you please double check that the current limit circuitry can indeed cause a shutdown? 

Thanks,

John

Hey John,

There are quite a few things that would be considered current limits in the device.

The section you are talking about is talking about the cycle by cycle current limiting done by comparing  COMP and the current.

There is another over current protection that happens when COMP reaches 2 Vs.

This current limit will shutdown the device and cause a restart.

Thanks,

Daniel

Hey John,

Its certainly not explained the the application section, but it is actually noted in the electrical characteristics.

The 3 current limits are noted there.

Thanks,

Daniel

Hi Daniel,

I was fairly sure those thresholds referred to the cycle-by-cycle current limiting. The 2V limit on the COMP for overcurrent shutdown does not appear to be...

Thanks,

John

Hey John,

The 2 V on COMP is how the high side one is set.

I think your right that I am getting a little confused.

I was thinking of the COMP voltage limiting the current as the error voltage is compared to the current rather than the max COMP voltage saturating and limiting the current.

In any case, you would be able to check by measuring the COMP voltage if you are hitting the high side current limit.

Thanks,

Daniel