Maker Pro
Maker Pro

determining minimum current loop bandwidth

M

Mook Johnson

Jan 1, 1970
0
I'm designing a current mode switching power supply with a load that will
have a very strong load current harmonic at 50Hz (repeating sine wave from
100mA -> 1A -> 100mA in 20mS). I need the current control loop to have
enough bandwidth to regulate this so it has less than 5% effect on the
output voltage. There is not sufficient bulk capacitance on the output to
supply the 50Hz current due to packaging so the control loop will see a
large portion of it (assume all of it).

How do i determime the bandwidth and loop gain required to achieve this
level of performance?
 
V

Vladimir Vassilevsky

Jan 1, 1970
0
Mook said:
I'm designing a current mode switching power supply with a load that will
have a very strong load current harmonic at 50Hz (repeating sine wave from
100mA -> 1A -> 100mA in 20mS). I need the current control loop to have
enough bandwidth to regulate this so it has less than 5% effect on the
output voltage. There is not sufficient bulk capacitance on the output to
supply the 50Hz current due to packaging so the control loop will see a
large portion of it (assume all of it).

How do i determime the bandwidth and loop gain required to achieve this
level of performance?

Here is the simplest estimate:

Start with how much of the ripple attenuation the loop should provide at
the given frequency. This depends on your topology. Then assume that the
open loop gain rolls down towards the high frequencies as 6dB/oct.
Figure out the frequency where the open loop gain will be 0dB - this is
the loop bandwidth.

Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com
 
J

John Larkin

Jan 1, 1970
0
I'm designing a current mode switching power supply with a load that will
have a very strong load current harmonic at 50Hz (repeating sine wave from
100mA -> 1A -> 100mA in 20mS). I need the current control loop to have
enough bandwidth to regulate this so it has less than 5% effect on the
output voltage. There is not sufficient bulk capacitance on the output to
supply the 50Hz current due to packaging so the control loop will see a
large portion of it (assume all of it).

How do i determime the bandwidth and loop gain required to achieve this
level of performance?

A few KHz should do it. Post a pic of the schematic.

John
 
J

Jamie

Jan 1, 1970
0
Mook said:
I'm designing a current mode switching power supply with a load that will
have a very strong load current harmonic at 50Hz (repeating sine wave from
100mA -> 1A -> 100mA in 20mS). I need the current control loop to have
enough bandwidth to regulate this so it has less than 5% effect on the
output voltage. There is not sufficient bulk capacitance on the output to
supply the 50Hz current due to packaging so the control loop will see a
large portion of it (assume all of it).

How do i determime the bandwidth and loop gain required to achieve this
level of performance?
Off the top of my head, I would say the switching frequency is the
first factor that governs the regulation response. And the
next factor would be the amount of capacitance being used on that
switching carrier frequency for the filtering process.
I suppose one could go as far as putting a regulating shunt load
on the output to increase the Skew rate.

Lots of factors.. But in my opinion, the switching frequency is
the major contributor here for the response time.
 
J

JosephKK

Jan 1, 1970
0
Mook Johnson [email protected] posted to sci.electronics.design:
I'm designing a current mode switching power supply with a load that
will have a very strong load current harmonic at 50Hz (repeating
sine wave from
100mA -> 1A -> 100mA in 20mS). I need the current control loop to
have enough bandwidth to regulate this so it has less than 5% effect
on the output voltage. There is not sufficient bulk capacitance on
the output to supply the 50Hz current due to packaging so the
control loop will see a large portion of it (assume all of it).

How do i determime the bandwidth and loop gain required to achieve
this level of performance?

A combination of SPICE and experimenting sounds like the quickest way
to a solution to me.
 
Top