TPS546B24A: VOUT STABLITY ISSUE

Hi Rajat,

It's not very clear to me what you are showing in the scope capture. Are you turning on/off the output, or is the output shutting down as as you apply the load transient? Can you show the output current in the same waveform? 

The GUI is showing the part triggered an IOUT_UC_FAULT, indicating the low-side FET current flowing from PGND to SW is more negative than it's threshold of -10A. In the scope shot, can you also capture the inductor current and the SW node? 

Thank you,
Tomoya

Tomoya Ide said:

It's not very clear to me what you are showing in the scope capture. Are you turning on/off the output, or is the output shutting down as as you apply the load transient? Can you show the output current in the same waveform? 

we are not turing on/off .

Output is Shutting down at the Transient.

Output current reading you can see in the TI fusion GUI Screenshot. If its not clear then I will capture it tomorrow.

Tomoya Ide said:

The GUI is showing the part triggered an IOUT_UC_FAULT, indicating the low-side FET current flowing from PGND to SW is more negative than it's threshold of -10A. In the scope shot, can you also capture the inductor current and the SW node? 

Yeah Sure I will capture it and share with you.

Many Thanks

Rajat

Thanks, Rajat.

Could this be how you are applying the load transient? Yes, if you could send us the waveform, it would help us understand what's going on. 

Hi Tom,

I have used Electrnoic load(N3300A) at the output capacitor of regulator.

VIN 12V(Max)

VOUT 1.5

FSW 650Khz

I have captured the waveform at No Load Condition and it seems like from waveform that Minumum on Time is violationg that's why i can see below obseration from the waveform

1.Frequency is decreased

2.Regulation is not stable

3.Pulse also skip

One more thing i notice number of pulses at SW node is 6-7-6 .Is this beacause of I have set 7D

I have checked the datasheet but i did not find miimum on time of Mosfet.

Please let me know your thoughts.

Many Thanks

Rajat

Hi Rajat,

Can you increase the IOUT_OC_WARN and FAULT limit to something higher because it is hitting the current limit, shutting down, and restarting is what seems to be happening.

Thank you,
Tomoya

Hi Tom,

its already set to IOUT_OC_WARN -6.5A and Fault is 7.5A

Attched is the GUI screenshot with No load confition.

Hi Rajat,

Can you increase threshold (>10A) and see if the problem goes away? Even though at no load condition, the inductor has a positive and negative current ripple and it may exceed the current limit during transient, startup, etc., to charge the output cap and maintain the regulation voltage. 

Thank you,
Tomoya

Rajat Singh 

Helping out Tomoya Ide 

The minimum controllable pulse-width for the TPS546B24A is listed in the electrical specifications table on Page 9 under the sub-section "PWM" it is 20ns.

Looking at your schematic and waveforms, the 55kHz bursting that you are seeing is most-commonly a result of excessive phase delay in the forward power path from the inductor to the remote sensed monitoring point, which appears to be after your 1mΩ current sense device R33023.

Is there additional capacitance on the the GTM_AVCCAUX output after R33023?

 the schematic shows 1 un-stuffed capacitor location, but is there other capacitance on that rail, possibly on another page of the schematic, such as input capacitance?

What is the PCB routing from the inductor to R33023 and from R33023 to the remote sense point?

Is the ESR of the 330uF polymer capacitors really 3mΩ?

Have you tested the lower VLOOP gain options through PMBus to check to see if reducing the gain achieves stability?

Code 29 @ 650kHz is 1340421C84  (VLOOP = 4)

Code 28 @ 650kHz is 1320841C84 (VLOOP = 2)

The loop is predicted to be stable with about 57kHz of bandwidth, which is the frequency we are seeing the loop oscillate at, so we are looking for something that's not in the model, such as a secondary filter from the current sense resistor and the remote capacitance, un modeled layout resistance or inductance, or something else.

Peter James Miller said:

Code 28 @ 650kHz is 1320841C84 (VLOOP = 2)

Peter,Vout is stable with these changes.

So Question is how the designer can assure based on the excel sheet calculation Vout will be stable in Actual Board.

Note- As you see the excel sheet calculation doens't give any design warning.

Many Thanks

Rajat

Rajat Singh 

What about the other questions from my message?

Output capacitor ESR?  (Do you have the part number?)

Additional Capacitance after the R33023 current sense resistor?

Routing after the R33023 current sense resistor to the remote sense point?

The design tool assumes the user enters the component values correctly and does not model a secondary filter past the Cout capacitors and before the remote sense point, so those can cause the tool issues.

The design included a local feed-forward capacitors from GTM_AVCC_AUX, but with no resistance from the remote sense point, the capacitor wouldn't actually be effective to compensate for the added filter after R33023