Boost Converter Control Error Amplifier. configurations ?

[mengghee]

error amplifier

hie,

I am just wondering what kind of configuration is the error amplifier for the control of boost converter ? thank you.

regards,
mengghee

 

[v_c]

boost converter control

What control IC are you intending to use for your application? Usually, they will
have some kind of application circuit that you can work from. Of course, you will have to pick your own component values based on your specifications.

Take a look at this datasheet for the LM2731 from National to give you some
idea. **broken link removed**

Best regards,
v_c

 

[mengghee]

boost converter controller

hello,

i am not using any boost converter ic, currently i am building a feedforward control of a boost converter. i am to take the input voltage and compare it to the sawtooth. but i think i need a error amplifier in order to control the pulse width gain. my aim is, say i want 0.5 duty cycle when there is no variation in input voltage. so the output of the amplifier should give a 2.5v if my sawtooth peak is 5v. but when there is an error. i need the vref to change to a controlable value. so i am just wondering what kind of error amplifier is used. thank you very much.

regards,
mengghee

 

[v_c]

how to build boost conveter controller

In order to generate open-loop pulse-width modulation, I think what you need is a comparator. You feed the sawtooth at one input and the some variable input (actually might need a scaled-down version) to the other input. The output should now be a pulsed waveform whose frequency is the frequency of the sawtooth and the duty ratio is controlled by the *ratio* of the input magnitude (variable) to the sawtooth magnitude (constant).

Best regards,
v_c

 

[mengghee]

control loop cookbook

I have build my sawtooth waveform generator, i have also used a comparator to generate the pulsewidth. the Vref generated from the potential divider from the input has been calculated. say if i want 0.5 as duty cycle. and my input voltage is 5v. so, vref should be 2.5 and potential divider r1 and r2 should be both equal 100k. when my input voltage drop to 4v. my vref will be at 2v. that will make the duty cycle to be 0.6 but what if. i want a variation of duty cycle greater than that ? say when vin=4v i want a duty cycle of 0.7. thus i think i need a error amplifier.

regards,
mengghee

 

[v_c]

error amplifier pdf

Well here is what I think ... Say you have a sawtooth with a 5V magnitude (0 to 5V) ramp at a given frequency. Then what you need to look at is the range of input voltages you have. Say it is 0V to 15V and you you would like to translate this input voltage to duty ratios. So 0V --> 0% duty and 15V --> 100% duty. Since your sawtooth is only 0 to 5V you should divide your input by a factor of 3. So have a divider that has two resistors with 2R and R and take the voltage across R as one of your inputs to the comparator (it will be 1/3 of the actual input voltage). So you have to choose the resistors to cover the range of actual input voltages that you have so they lie within the magnitude range of the sawtooth. Does this make sense?

Best regards,
v_c

 

[Hughes]

soft start boost converter

You do need a error amplifier in the boost converter. It may have two non-inverting input terminals if you want soft-start function for the converter.

 

[mengghee]

control loop design of boost converter

You do need a error amplifier in the boost converter. It may have two non-inverting input terminals if you want soft-start function for the converter.

Hughes,

thank you, if you don't mind, can you explain to me how do i achieve it ? error amplifier ? and also softstart ? can you explain to me what does a softstart do ? and how to basicly design it. and what do you mean by having two non inverting terminal ? having to non inverting amplifier ? thank you very much.

regards,
mengghee

 

[Hughes]

closed loop control of boost converter

Error amplifier is nothing but an operational amplifier. Maybe a single-pole amplifier with pmos input pair (if your design is targeted to cmos) is what you need. To implement two non-inverting input terminals, we can just copy the non-inverting input brunch in a general purpose input stage. One of the non-inverting input is applied to a fixed reference voltage, the other to a ramping-up voltage. Thus make the output voltage ramp from zero to the final value -- so-called soft-start.

I don't have much experience on dcdc converter. Please let me know if I am wrong.

Regards,
Hughes

 

[v_c]

slup173.pdf

I agree -- you do need an error amp if you are going to do closed-loop control, but the original post said feedforward control. In feedforward control, you do not change the dynamics of the sytem, you just take one of the system parameters (the input voltage in this case) and use it to adjust the duty ratio. Of course, for a real system you do need to have a closed-loop system and you will need to have a error amplifier as well. You have to combine the error amplifier output and the feedforward to compare to the ramp to generate the duty ratio.

Best regards,
v_c

 

[mengghee]

error amplifier in control system?

V_C, I have looked at the method you mentioned and done several calculations, i have to say the method i was trying to developed the feed forward control was very immature. thanks for reminding me and the results i obtained was satisfactory. I have seen in the book 'fundamental of electronics', there was a feed forward block diagram. and there is an error amplifier as well. but that is not an important issue. and if i would like to make a feedback loop, i am wondering how i should do it. i have done some researches previously and i noticed we do need error amplifier in all of the cases that i have seen.

Hughes, I am a newbie in many areas in electronics, i don't understand what u mean by single pole amplifier and pmos input pair. brunch ? do you mean have a non inverting amplifier and the output of the non inverting amplifier is connected to the non inverting input of the comparator ? how does this refer to softstart ? thank you.

thanks,
mengghee

 

[v_c]

boost converter duty cycle control

I think Hughes was trying to give you the transistor-level implementation of the error amplifier. And I think "brunch" referred to a "branch" and not the meal between breakfast and lunch

Getting back to business ... please take a look at the application note for the
UCC3570 here **broken link removed**
This is a low-power PWM chip and it has a dedicated pin for voltage feedforward called VFWD. Voltage feedforward allows faster and more accurate response to input voltage variations. The actual input voltage is usually scaled down by using a resistive divider before connecting to this pin. Read the datasheet for more information on the IC and in particular this pin. This IC obviously has an Error Amp and you do put a compensation network around it but it had the feedforward pin as well so I thought that you would want to see the feedforward control in action in a real product!

For generic information, the Texas Instruments has archived many of their past seminars on power. These are generic topics and I think you will probably benefit by browsing and studying these documents which are stress the practical implementation of the dc/dc converters. The archive is here **broken link removed**
Look at the bottom of the page for the Unitrode Design Seminars (Unitrode was bought by TI some time ago). Click on the archive and see which topics interest you.

Below are some that I picked out for you but take a look around this and other sites -- especially search the application notes related to the integrated circuits that you are using (or are intending to use).


From 1984
Closing the Loop
https://focus.ti.com/lit/ml/slup068/slup068.pdf
Appendix A: Error Amplifier and Compensation Network Design
**broken link removed**
Appendix B: Bode Plots
https://focus.ti.com/lit/ml/slup070/slup070.pdf

From 1991
Control Loop Design
https://focus.ti.com/lit/ml/slup098/slup098.pdf

From 1996
Control Loop Cookbook
**broken link removed**

From 2001
An analytical comparison of Alternative Control Techniques
https://focus.ti.com/lit/ml/slup168/slup168.pdf
Designing Stable Control Loops
**broken link removed**

Best regards,
v_c

 

[mengghee]

ic error amp

V_C,

thank you very much, i have browse through the papers and have yet to properly read it. but the papers looks like what i am really looking for. never found any sites like that. thank you very much.


regards,
mengghee

Added after 10 minutes:

I have read some papers, and several websites on closed loop control. i have noticed that all of them have a bode plot referring to open loop gain. and frequency response etc. can somebody enlighten me with the meaning of open loop gain ? and also frequency response ? i have read about this several times, i have even tried to read my past year uni text book on frequency response. i still have no idea what does the frequency has to do with the gain of a boost converter. thank you.

 

[v_c]

boost converter open loop bode plot

Well, this forum is probably not the best venue to give you a "lecture" on Bode plots and stability, but I will just say a few things.

First of all, the boost converter itself constitutes the "plant" (in control system terminology) -- that is, that is the system to be controller. The compensation network with the error amplifier, the reference voltage, the resistor dividers to sense the output voltage -- these form the controller. Now the thing to note is that in switching power supplies, the plant is not linear -- but we usually linearize it about an operating point. So you will see lots of references to "small signal transfer functions". These are s-domain transfer functions giving the transfer characteristics from the duty ratio input to the output voltage. The controller (which we pick ourselves) has a transfer function from the output voltage to the duty ratio input. So that is the loop thas is formed -- and we use negative feedback. You can see this in standard control system diagrams with G(s) representing the plant and H(s) as the controller.

The open loop gain of the system G(s)H(s) tells us something about the closed loop properties of our system. The important thing to note is that the closed-loop system has its roots (sometimes called poles) where 1 + G(s)H(s) = 0 or G(s)H(s)= -1. So we get the Bode plot of G(s)H(s) and see at what frequency the gain crosses the 0dB point -- this is called the crossover frequency. We then measure the phase at that frequency. The further away from -180 degrees the phase is -- better the stability. The idea is this -- the plant and the compensation network give use some phase delay and we add another 180 due to negative feedback. so any noise in the system can go around the loop and add constructively if the total phase around the loop is 360 degrees. I suggest you take a look at any control systems book for this. The difference in phase away from -180 degree is called the phase margin. In practical systems we require a phase margin of at least 45-60 degrees.

If you get a chance please take a look at the book by Erickson and Maksimovice called "Fundamental of Power Electronics" -- it has excellent explanations about all of this stuff.

Best regards,
v_c

 

[wolf001]

how error amplifier work

Can anyone have some paper about loop stability?