Boost converter output voltage spike on startup

[TheRedDevil18]

I'm trying to simulate a boost converter using multisim. I get this waveform

Why does it spike like that at the start ?, is it because of the inrush current ?

 

[berkeman]

I'm trying to simulate a boost converter using multisim. I get this waveform
View attachment 114189

Why does it spike like that at the start ?, is it because of the inrush current ?

Could you post the schematic and where your probe point is? Can you also post your full startup simulation waveforms? I'm still trying to figure out what that is that you posted.

 

[TheRedDevil18]

It's the output voltage waveform from the oscilloscope

 

[berkeman]

It's the output voltage waveform from the oscilloscope
View attachment 114194

I don't see any startup waveform there. It should go vrom 0V up to whatever your target voltage is, and it should take dozens of cycles at least to get there. Can you show the whole waveform from t=0 and V=0?

 

[Integrand]

I don't see any startup waveform there. It should go vrom 0V up to whatever your target voltage is, and it should take dozens of cycles at least to get there. Can you show the whole waveform from t=0 and V=0?

His measurement is ac coupled, but as you say, it's odd that the initial rise is missing.

OP, what you are seeing is typical of the inrush current caused by charging the dc bus capacitor. For this reason, most boost converters have a controlled startup sequence. You can find plenty of application notes about solutions.

 

[berkeman]

Also, you don't show any closed-loop regulation control of the output voltage. Is that on purpose? You want to run this circuit open-loop and expect no overshoot or line regulation or load regulation issues?

EDIT/ADD -- @TheRedDevil18 -- I didn't mean that to sound aggressive. I'm just trying to understand what you are wanting to do.

 

[berkeman]

This is more like what I'm used to seeing...

http://www.ti.com/ods/images/SNVS346F/wvfrm_startup_boost_snvs346.png

 

[Rive]

Set your 'oscilloscope' in DC mode.
As the converter starts, the output voltage moves from 0 to ??V. This move has an AC component: with the oscilloscope in AC mode you can measure only this.
In DC mode you will see something like Berkeman linked in #7.

 

[TheRedDevil18]

Thanks for the replies, this is my startup waveform with oscilloscope in DC mode

It rises from 8.2 V (not sure how, maybe its the simulator) to 38 V and then stabilizes to roughly 29.5 V. This converter is a 9 to 30 volt. I am planning on adding feedback using an mcu. Will that help ?

 

[Rive]

It rises from 8.2 V (not sure how, maybe its the simulator) to 38 V and then stabilizes to roughly 29.5 V. This converter is a 9 to 30 volt. I am planning on adding feedback using an mcu. Will that help ?

Check the schematics: if no control for the FET, then you have closed current loop for the 9V source. It's expected to have a DC current => voltage drop on the resistor.

About the 29.5V: I think you forgot to design for the forward voltage of the diode.

To design an acceptable feedback is not easy.

 

[berkeman]

I am planning on adding feedback using an mcu. Will that help ?

That should help, but as Rive says, it's not trivial. Still, it's very fun to learn about and use in practical circuits.

What is the switching frequency of the FET, BTW?

 

[TheRedDevil18]

That should help, but as Rive says, it's not trivial. Still, it's very fun to learn about and use in practical circuits.

What is the switching frequency of the FET, BTW?

Its 200kHz

What would be the simplest way to control the duty cycle ?, I am thinking of using a microcontroller as I think this would be much easier

 

[Rive]

What would be the simplest way to control the duty cycle ?, I am thinking of using a microcontroller as I think this would be much easier

I think it's more like a nightmare than fun, but Berkeman is right, you have to see it for yourself.
Since it's already a simulation, why don't you give a try to a classic sawtooth-and-comparator style analog PWM control? You can check the whole thing with that oscilloscope. Quite demonstrative.
Also, it's a good experience to modify the load while it's still running and see what it does.

Using a microcontroller for control is jut no good for beginners. You have to be familiar with applied control theory first.

 

[NascentOxygen]

Why does it spike like that at the start ?, is it because of the inrush current ?

Is the frequency you see in this transient equal to the calculated resonance of your L and C? (Maybe only roughly equal, because the capacitance of ordinary electrolytics has a very high tolerance, possibly a factor of 2.)

 

[TheRedDevil18]

Is the frequency you see in this transient equal to the calculated resonance of your L and C? (Maybe only roughly equal, because the capacitance of ordinary electrolytics has a very high tolerance, possibly a factor of 2.)

Not too sure. I didnt calculate any resonance values. The inductor and capacitor in the simulation are ideal

 

[TheRedDevil18]

I think it's more like a nightmare than fun, but Berkeman is right, you have to see it for yourself.
Since it's already a simulation, why don't you give a try to a classic sawtooth-and-comparator style analog PWM control? You can check the whole thing with that oscilloscope. Quite demonstrative.
Also, it's a good experience to modify the load while it's still running and see what it does.

Using a microcontroller for control is jut no good for beginners. You have to be familiar with applied control theory first.

I was thinking the microcontroller would be much easier. I am familiar with coding, so I would just use the MCU to generate a PWM and then use the ADC to sense the voltage and if this voltage goes above or below 30V then it would increase or decrease the duty cycle ?

 

[Svein]

I was thinking the microcontroller would be much easier. I am familiar with coding, so I would just use the MCU to generate a PWM and then use the ADC to sense the voltage and if this voltage goes above or below 30V then it would increase or decrease the duty cycle ?

Or use a specialized power controller like the one documented here: http://www.linear.com/product/LT3757.