TPS62147: Making the buck output voltage proof

Hi 

Can you let know which where 100V is?

Shuai

The load (RL) is outside the device. In the worst case (i.e. isolation of a cable fails) 100V can be applied to this connector pin, where the load is connected. In this case, 100V are applied at the cathode of the diode, which blocks the voltage so that the TPS62147 is not damaged.
We are talking about a fail case, not normal operation, the diode is for protection. During normal operation, 5V are generated at the load. Because I want to compensate the current dependent voltage drop of the diode I want to connect the FB voltage devider to the diode cathode.
Thanks for your help

Hi Klemens!

I don't think it is a available solution.

1. The diode has the reverse recover behavior. When 100V voltage comes, it will keep to conducting for a short time, this will cause the IC damage.

2. Even through there is no reverse recovery, the voltage of VOUT pin is determined by divide ratio of the impedance of VOUT-GND and the reverse bias impedance of the diode. Both of them are high impedance, so I still think the Vout-GND can get a high voltage.

Shuai

Hi, thanks for the answer, I am not sure if I am understanding your second point. Do you mean the VOS pin in the circuit when talking about the VOUT pin? I can't see how any pin can get a voltage above abs. max. rating.

• The SW and VOS pin are blocked by a 22uF MLCC at the anode of the diode so they are actually low impedance
• The FB voltage cannot rise above VIN according to the feedback voltage divider ratio and even if it did get higher than VIN I doubt that the little current inijected via a 220k resistor would damage anything.

Regarding the first point: It would need a massive diode to lift the voltage across the 22uF cap by ~5V during reverse recovery time. I plan on using a high speed pn diode with 4ns reverse recovery time. Instead I could also fancy a schottky diode...

My main concern is the stability of the control loop when an additional component is inserted. Will it be stable as described in the datasheet?

Hi 

I mean the impedance between VOS and GND is also very high, at least same level as diode. The 22uF cap is opened with low frequency or DC value. We don't recommend you do as that.

Shuai

Are you an expert in control engeneering? Can you tell me how inserting a nonlinear component like a diode influences the phase/gain margin of the loop and overall stability? 

Hi Klemens!

Sorry for late because of weekends.

The small signal model of diode is a voltage source so as short circuit, so it will not affect the stability theoretically.

Shuai

Hi,

talking about a load current jumping from 1A to 1.1A I agree that applying a small signal model is valid and the stability is not affected. What about a load current transient from ~0A to 0.1A though? Is such a case you cannot model a diode as a voltage source, it is nonlinear. This is the question I am interested in.

Can you provide your input and output voltage?

Vin: 12-13V

Vout: 5.37V

I also don't think there will be big issue on the DCM stability. But we can double check the bench. How much voltage drop of your diode you want to use?

Shuai

Hi,

thank you for checking, let's assume a high voltage high speed si diode with 1V drop at 100mA. Are you going to check it using an evaluation board? If so can you send me the output voltage scope picture when applying a load transient from 0 to 100mA?

Thanks for the info, can you also provide the circuit diagram/setup you used for the measurement? Can you try inserting different feed forward capacitors (maybe start with 100p ??) to see if the transient response improves?

Hi Klemens,

My colleague will check your requirement and give you feedback soon!

BRs

Zixu