Viper265K Buck Converter Design - 0.5% duty cycle ?

[shrikantB]

Hi,

I am presently planning to use Viper265K part for High Voltage Buck converter design with below requirements:

	Min	Max
Vin	200	900	V
Vout	4.5		V	+/-5%
Iout	0.44		A
Pout	2		W

With above requirements the duty cycle for the device is observed to be 0.5%. Please can you 

help me to know if the device can be used for such low duty cycle and what would be recommendation from your side 

for achieving above design requirement.

[Peter BENSCH]

Welcome @shrikantB, to the community!

As you write, you want to implement a high-voltage buck, but you have not specified whether the input voltage is DC or AC.

I will initially assume you mean DC. For your requirements, the buck topology is not suitable, because you would need a duty cycle in the range of 0.005 to 0.0225. A flyback could mitigate the duty-cycle problem through the transformer turns ratio, however you would have to implement it with a PWM device plus an HV MOSFET, because the VIPER265K would not be able to withstand it.

With AC, it would look even worse, because the peak voltage would be about 41% higher.

Hope that helps?

Regards
/Peter

[shrikantB]

Hi Peter,

Thank you for your reply. For the above requirement share we have decided to go for the 2 Buck.

Wherein first stage would implement 900VDC to 20VDC with Viper267K which seems to be feasible.

Other stage would be 20VDC to 4.5VDC.

Now after doing some simulations with Viper267K switching device there is one query that during start up with 680uH of Buck inductor the initial peak is observed till 2Amps. Later the current is as per the load.

 

Please can you suggest if this peak can be sustained by the Viper267K as the max current limit seems to be 0.75Amps. Please help for this query.

Thanks and Regards,

Shrikant Bhairavkar

[shrikantB]

Hi Peter,

One request I was checking the Viper267 simulation on SIMPLIS tool but the simulation file is showing below error:

Please can you tell me how I should do it in correct way.

Regards,

Shrikant

[Peter BENSCH]

You spoke of a simulation with the VIPER267K, but the VIPER265K is in the schematics. However, I also simulated it again with the VIPER267K and, for the most part, get the same results.

By the way, your simulation refers to an input voltage of around 300Vac, i.e. 424Vdc. As the input voltage increases, the peak current also increases (theoretically), but is limited by the internal current limiting, which makes the peak last longer in order to charge the output capacitance of the system. If the inductor reaches its limit at high start-up currents, the effective inductance drops, so it is inversely proportional to di/dt, which significantly increases the peak currents.

I would advise you that there is really no point in pursuing the idea of a buck for a conversion from an input voltage of up to 900V further, regardless of the output voltage. At that point, you are already too close to the maximum drain voltage of the internal MOSFET. But for a reliable design the limiting factor is not only the MOSFET voltage rating itself, but the combination of:

• very steep inductor current ramp at high input voltage
• start-up dynamics with discharged output capacitor
• current limit behaviour
• inductor saturation or component stress

If you already have two regulators planned, you can also put this material and financial effort into a flyback, which will likely work out cheaper overall. In addition, you won’t be tempted to accidentally connect or mix up the separate input and output GNDs used in the buck converter.

Regards
/Peter

[Peter BENSCH]

This is a question that is only indirectly related to the original question and would be better placed in the Power management > eDesignSuite forum.

If you want to continue pursuing this, I can gladly move the question there for you and create it as a new thread.

Regards
/Peter

[shrikantB]

Hi Peter,

Thank you for your reply and suggestions. The plan is to reduce the power supply size and provide the compact solution. However, the risks you mentioned we will check during the testing and can take the decision on same.

But if we go for the non-isolated Flyback option whether the above issues will be reduced with Viper267K or you are suggesting some other part with external MOSFET.

Please suggest if you have any high voltage solution device with external MOSFET also we can check on the same.

Regards,

Shrikant Bhairavkar

[Peter BENSCH]

The question now is: what are your actual required parameters for continuous operation?

• Does it need to be able to handle 900V, or would 870V also be sufficient?
• Does it need to deliver 2W, or would 1.5W be enough?

The slightly weaker version is offered by the eval board STEVAL-VP26K01B.

Otherwise, the only real option is a flyback with an external MOSFET, for example as implemented with the STEVAL-ISA034V1 (already obsolete, based on L6565, the documentation attached for future reference), which with appropriate design (HV capacitors, a matched transformer, 1.7kV MOSFET = e.g. STW3N170) can also handle higher input voltages.

Regards
/Peter

[shrikantB]

Hi Peter,

Thank you for your reply. I would be ok for 870V input also that is not an issue, but I would require at least 2.5~3W output. What parameter of the Viper267K would limit its capability to deliver this power?

Please can you help us know so we can try to mitigate this risk factor either through Inductor sizing or capacitor sizing.

Regards,

Shrikant Bhairavkar

[Peter BENSCH]

With a higher output power, you exceed one or more of the reasons that prevent the use of a VIPER26xK in such a buck converter: regulation, minimum duty cycle, current limiting, start-up behaviour, external inductance and parasitic effects all become limiting factors long before the MOSFET breakdown voltage is reached. For the 2.5–3 W you mentioned, with such a high input voltage you can no longer use VIPER in buck mode and would need to consider a flyback converter.

For a more in-depth analysis, I recommend contacting your local distributor, who can provide detailed assistance with their technical resources.

Good luck!

Regards
/Peter