Designer’s Guide Community

feiyue

Junior Member

Offline

Posts: 15

strange simulation result
Jul 10^th, 2007, 10:42am

hi all,

I am designing a voltage-mode switching regulator,it needs a compensation network with 2 zeros and 3 poles.
I simulate the whole regulator without adding the compensation network,simulaiton result is ok.but when the compensation network is added,the
result is weird,it seems that regulaor unstable,error amplifier output also saturate and oscillate between vdd and 0.
I am not sure it is because my unconsiderable setting of initial condition for simulation or it is because the compensation network is unstable.
Any comments and suggestion are greatly appreciated.

feiyue

Back to top

ideal_EA_with_and_without_compensation_network.pdf

IP Logged

feiyue Junior Member Offline Posts: 15	Re: strange simulation result Reply #1 - Jul 10^th, 2007, 3:24pm nobody could help? :'( it is a type 3 error amplifer,I think the component values should be true,but i don't know whether i need to do something for simulation initiation, i really appreciate if anyone can help!
Back to top	IP Logged

Eugene

Senior Member

Offline

Posts: 262

Re: strange simulation result
Reply #2 - Jul 10^th, 2007, 6:13pm

It would be nice to know what each waveform represents but your "compensated" system indeed looks unstable. I would recommend constructing a state space averaged model and looking at the loop gain with and without the compensation. If your compensator has three poles, it is quite possible for your compensated system to be unstable. How did you design your compensation? i.e. how did you choose the compensation poles and zeros?

Back to top

IP Logged

feiyue

Junior Member

Offline

Posts: 15

Re: strange simulation result
Reply #3 - Jul 10^th, 2007, 7:05pm

Thank you very much,Mr. Eugune!
I design a type 3 error amplifier ,it seems very classical and is described in detail in the book of "switching power supply design".
I choose two pole at the same frequency and is 5 time higher than defined unit-gain bandwidth and two zero at 1/5 unit-gain bandwidth.
I check it using matlab,the phase margin is 60deg. But if i delete the capacitor which is series connected to the integrating resistor R2,then regulator works ok,so i don't know what is the exact problem. is it because the error amplifer without a dc feedback? but i think for a whole regulator,the dc feedback is automatic established.Is it right?
also i don't know whether there are some specific attention i need to pay for the correct simulation.
I don't know how to construct a stata space averaged model.is there any reference?

Back to top

IP Logged

Eugene Senior Member Offline Posts: 262	Re: strange simulation result Reply #4 - Jul 11^th, 2007, 12:47pm How did you model the power supply to compute the phase margin?
Back to top	IP Logged

feiyue

Junior Member

Offline

Posts: 15

Re: strange simulation result
Reply #5 - Jul 12^th, 2007, 10:48am

I model the gain of PWM=Dvin/Vramp, gain of resistor string and then gain of error amplifer
check the open loop gain margin and phase margin of the whole system with LC fitler.

Acutally,I found it is the problem of compensation,when i increase the phase margin to almost 90deg,then the systyem work.But when PM=60deg, (calculate at matlab),it doesn't work.
I don't know what is the problem of this.Is it because the model is not correct?
thanks Eugune!

Back to top

IP Logged

Eugene Senior Member Offline Posts: 262	Re: strange simulation result Reply #6 - Jul 12^th, 2007, 1:52pm Yes, I strongly suspect an error in the model. Did your brute force simulation include an input filter? If so, did your Matlab model also include the input filter?
Back to top	IP Logged

feiyue Junior Member Offline Posts: 15	Re: strange simulation result Reply #7 - Jul 12^th, 2007, 4:39pm do you mean the capacitor in the input battery? I have not added it and modeling it.
Back to top	IP Logged

Eugene

Senior Member

Offline

Posts: 262

Re: strange simulation result
Reply #8 - Jul 12^th, 2007, 8:44pm

I mean an impedance between the ideal power source and the input to the switch mode power supply. In space craft power supplies for example, one commonly places a filter between the main power bus and the power supply drawing power from the bus. This is to keep switching noise off the main power bus. Such a filter can destabilize your power supply if it's not designed properly. However, it does not sound like that is your problem. It sounds like you are driving your power supply model from an ideal voltage source.

I'd have to see your model and the circuit you are simulating to be sure but I still suspect your model is not a sufficiently accurate representation of the circuit you are simulating.

Back to top

IP Logged

hedonist81

New Member

Offline

Ghost In The Machine

Posts: 1
Hyderabad

Re: strange simulation result
Reply #9 - Oct 16^th, 2007, 8:23pm

Hi Guys

Just joined the forum and saw through the entire discussion. You have 3 poles and 2 zeros. Can you please indicate where each of these occur. Even though you have a good phase margin, it is possible to see and oscillation if there is a pole-zero couplet close by. Such a pole zero couplet gives rise to oscillations.

Also, what is the frequency of oscillations you observe ??

Back to top

IP Logged