Boost converter, continuous-conduction mode, regulation

N

Nobody

Guest
Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

In discontinous-conduction mode, simple negative feedback works fine:
enable or disable PWM depending upon whether the output voltage is above
or below the target voltage.

In continuous mode, the inductor current presents a problem, as
the immediate effect of reducing the duty cycle is to increase the mean
current to the output capacitor until the inductor current actually drops.
And once that happens and the output voltage drops, it will continue to
drop until you've built up the inductor current again.

So, how do you deal with this? A more complex control scheme, or a
different topology?
 
On 03/27/2016 09:55 PM, Nobody wrote:
Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

In discontinous-conduction mode, simple negative feedback works fine:
enable or disable PWM depending upon whether the output voltage is above
or below the target voltage.

In continuous mode, the inductor current presents a problem, as
the immediate effect of reducing the duty cycle is to increase the mean
current to the output capacitor until the inductor current actually drops.
And once that happens and the output voltage drops, it will continue to
drop until you've built up the inductor current again.

So, how do you deal with this? A more complex control scheme, or a
different topology?
Click to expand...

Never done a CCM boost myself, but AIUI you use current-mode control
with slope compensation. TI has a bunch of app notes on that, e.g.

http://www.ti.com/cn/lit/pdf/slva636
http://www.ti.com/download/trng/docs/seminar/Topic_3_Lynch.pdf
http://www.ti.com/lit/pdf/slva452

If you're comfortable with complex numbers and the frequency domain,
they're not too complicated.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net
 
On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

In discontinous-conduction mode, simple negative feedback works fine:
enable or disable PWM depending upon whether the output voltage is above
or below the target voltage.

In continuous mode, the inductor current presents a problem, as the
immediate effect of reducing the duty cycle is to increase the mean
current to the output capacitor until the inductor current actually
drops.
And once that happens and the output voltage drops, it will continue to
drop until you've built up the inductor current again.

So, how do you deal with this? A more complex control scheme, or a
different topology?
Click to expand...

Given a step increase in duty cycle (and assuming that the thing stays in
continuous conduction) the output will first dip, then settle out to a
higher value than before you changed duty cycle.

In control systems terms, it has a stable pole and an unstable zero.

I don't know how you define "simple", but a PID controller would do it,
probably with all sorts of possible problems to deal with because system
nonlinearities probably matter in this case. I just erased a whole bunch
of bloviating on this because IIRC the converter is a nonlinear system
which means that my general-purpose control systems knowledge isn't a lot
of help.

--
www.wescottdesign.com
 
On Mon, 28 Mar 2016 10:14:20 -0500, Tim Wescott wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

In discontinous-conduction mode, simple negative feedback works fine:
enable or disable PWM depending upon whether the output voltage is
above or below the target voltage.

In continuous mode, the inductor current presents a problem, as the
immediate effect of reducing the duty cycle is to increase the mean
current to the output capacitor until the inductor current actually
drops.
And once that happens and the output voltage drops, it will continue to
drop until you've built up the inductor current again.

So, how do you deal with this? A more complex control scheme, or a
different topology?

Given a step increase in duty cycle (and assuming that the thing stays
in continuous conduction) the output will first dip, then settle out to
a higher value than before you changed duty cycle.

In control systems terms, it has a stable pole and an unstable zero.

I don't know how you define "simple", but a PID controller would do it,
probably with all sorts of possible problems to deal with because system
nonlinearities probably matter in this case. I just erased a whole
bunch of bloviating on this because IIRC the converter is a nonlinear
system which means that my general-purpose control systems knowledge
isn't a lot of help.
Click to expand...

Three things occurred to me after I'd had my morning caffeine:

First, if you have the circuit-analysis chops to work out ON PAPER what
happens when you start at steady-state at one duty cycle and change duty
cycle, that should at least provide a lot of material for figuring this
out on your own, and thus internalizing the answers.

Second, there are entire books written on switching power supply design;
if you're serious about this you may want to get at least one -- I'd
start by asking the group for a recommendation (I may do that myself at
some point: I seem to be designing a lot of switchers these days).

Third, if you use a part with a built-in current limit, and run close to
that limit, you'll see some rather interesting interactions between the
control loop and the current limit. "Skrody" comes to mind as a
description of some of the resulting waveforms.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
 
On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?
Click to expand...

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I need
to stabilise involves matching the rate of buying books to the time
available for reading them).
 
On Mon, 28 Mar 2016 23:12:57 +0100, Nobody <nobody@nowhere.invalid>
wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I need
to stabilise involves matching the rate of buying books to the time
available for reading them).
Click to expand...

You also have to limit the mosfet duty cycle, or the ON state current,
or you get into the crisis of increasing mosfet duty cycle causing
decreasing output duty cycle approaching zero, and it blows up.

I've built fairly mindless voltage-feedback continuous-conduction
boosters which worked when the loop was fairly slow and max duty cycle
was limited.

Spice it!


--

John Larkin Highland Technology, Inc

lunatic fringe electronics
 
On Tuesday, March 29, 2016 at 12:11:39 PM UTC-7, Tim Wescott wrote:

Why does Linear make so many really good little switcher chips, but not
very many (or any) current-controlled ones?
Click to expand...

They also sell currrent-in/v-out monitors (LTC6101 etc.).
Sell a chip, get $. Sell two chips, get $$.
 
On Mon, 28 Mar 2016 10:21:00 -0400, Phil Hobbs wrote:

On 03/27/2016 09:55 PM, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

In discontinous-conduction mode, simple negative feedback works fine:
enable or disable PWM depending upon whether the output voltage is
above or below the target voltage.

In continuous mode, the inductor current presents a problem, as the
immediate effect of reducing the duty cycle is to increase the mean
current to the output capacitor until the inductor current actually
drops.
And once that happens and the output voltage drops, it will continue to
drop until you've built up the inductor current again.

So, how do you deal with this? A more complex control scheme, or a
different topology?


Never done a CCM boost myself, but AIUI you use current-mode control
with slope compensation. TI has a bunch of app notes on that, e.g.

http://www.ti.com/cn/lit/pdf/slva636
http://www.ti.com/download/trng/docs/seminar/Topic_3_Lynch.pdf
http://www.ti.com/lit/pdf/slva452

If you're comfortable with complex numbers and the frequency domain,
they're not too complicated.

Cheers

Phil Hobbs
Click to expand...

Current-mode control would still leave you with an unstable zero. It's
probably the best approach -- but you still have that unstable zero.

Why does Linear make so many really good little switcher chips, but not
very many (or any) current-controlled ones? Is it against their
religion, or does TI hold all the patents, or what?

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
 
On Mon, 28 Mar 2016 23:12:57 +0100, Nobody wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I need
to stabilise involves matching the rate of buying books to the time
available for reading them).
Click to expand...

A paper that I'm thinking of writing is a piece entitled, basically,
"Change the Plant!". It pertains to the fact that you can't make a silk
purse out of a sow's ear, meaning, in this context, that given a specific
arrangement of mechanism, structure, sensors and actuators, there's a
limit to the performance that it's possible to achieve with closed-loop
control. To go beyond that limit, you need to change something for
better inherent regulation, bandwidth, linearity, etc.

So learn to read faster!

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
 
On Tuesday, March 29, 2016 at 3:15:03 PM UTC-4, Tim Wescott wrote:
On Mon, 28 Mar 2016 23:12:57 +0100, Nobody wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I need
to stabilise involves matching the rate of buying books to the time
available for reading them).

A paper that I'm thinking of writing is a piece entitled, basically,
"Change the Plant!". It pertains to the fact that you can't make a silk
purse out of a sow's ear, meaning, in this context, that given a specific
arrangement of mechanism, structure, sensors and actuators, there's a
limit to the performance that it's possible to achieve with closed-loop
control. To go beyond that limit, you need to change something for
better inherent regulation, bandwidth, linearity, etc.

So learn to read faster!

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Click to expand...

Grin, "change the plant" is what I believe Phil calls,
"turd polishing".

We all get stuck there when the plant or turd is an old friend.

George H.
 
On Tue, 29 Mar 2016 14:15:00 -0500, Tim Wescott
<seemywebsite@myfooter.really> wrote:

On Mon, 28 Mar 2016 23:12:57 +0100, Nobody wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I need
to stabilise involves matching the rate of buying books to the time
available for reading them).

A paper that I'm thinking of writing is a piece entitled, basically,
"Change the Plant!". It pertains to the fact that you can't make a silk
purse out of a sow's ear, meaning, in this context, that given a specific
arrangement of mechanism, structure, sensors and actuators, there's a
limit to the performance that it's possible to achieve with closed-loop
control. To go beyond that limit, you need to change something for
better inherent regulation, bandwidth, linearity, etc.
Click to expand...

Possibly sense "internal" plant nodes, or break the plant into smaller
bits that can be individually controlled, which is sort of similar.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics
 
On 03/29/2016 03:15 PM, Tim Wescott wrote:
On Mon, 28 Mar 2016 23:12:57 +0100, Nobody wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I need
to stabilise involves matching the rate of buying books to the time
available for reading them).

A paper that I'm thinking of writing is a piece entitled, basically,
"Change the Plant!". It pertains to the fact that you can't make a silk
purse out of a sow's ear, meaning, in this context, that given a specific
arrangement of mechanism, structure, sensors and actuators, there's a
limit to the performance that it's possible to achieve with closed-loop
control. To go beyond that limit, you need to change something for
better inherent regulation, bandwidth, linearity, etc.
Click to expand...

Local feedback is good too. I've done several diode laser temperature
controllers like that. One of them used local feedback to make the
Peltier/thermistor response simple, then just wired it in parallel with
the current-tuning control. That way I got an automatic coarse/fine
wavelength adjustment for a cavity stabilized laser. Worked great.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net
 
On 03/29/2016 10:15 PM, George Herold wrote:
On Tuesday, March 29, 2016 at 3:15:03 PM UTC-4, Tim Wescott wrote:
On Mon, 28 Mar 2016 23:12:57 +0100, Nobody wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I need
to stabilise involves matching the rate of buying books to the time
available for reading them).

A paper that I'm thinking of writing is a piece entitled, basically,
"Change the Plant!". It pertains to the fact that you can't make a silk
purse out of a sow's ear, meaning, in this context, that given a specific
arrangement of mechanism, structure, sensors and actuators, there's a
limit to the performance that it's possible to achieve with closed-loop
control. To go beyond that limit, you need to change something for
better inherent regulation, bandwidth, linearity, etc.

So learn to read faster!

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Grin, "change the plant" is what I believe Phil calls,
"turd polishing".
Click to expand...

Turd polishing is _not_ changing the plant when you really need to. ;)

We all get stuck there when the plant or turd is an old friend.
Click to expand...

As I often say to my wife when she brings in some new orchid or
something, "Death to house plants!"

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net
 
On Wed, 30 Mar 2016 08:41:36 -0700, John Larkin wrote:

On Tue, 29 Mar 2016 14:15:00 -0500, Tim Wescott
seemywebsite@myfooter.really> wrote:

On Mon, 28 Mar 2016 23:12:57 +0100, Nobody wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I
need to stabilise involves matching the rate of buying books to the
time available for reading them).

A paper that I'm thinking of writing is a piece entitled, basically,
"Change the Plant!". It pertains to the fact that you can't make a silk
purse out of a sow's ear, meaning, in this context, that given a
specific arrangement of mechanism, structure, sensors and actuators,
there's a limit to the performance that it's possible to achieve with
closed-loop control. To go beyond that limit, you need to change
something for better inherent regulation, bandwidth, linearity, etc.


Possibly sense "internal" plant nodes, or break the plant into smaller
bits that can be individually controlled, which is sort of similar.
Click to expand...

I was making a joke about the speed at which he's getting through the
reading list, actually.

Maybe take some of those "smart pills" that are being advertised on the
Internet these days?

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
 
On Wed, 30 Mar 2016 12:15:47 -0400, Phil Hobbs wrote:

On 03/29/2016 03:15 PM, Tim Wescott wrote:
On Mon, 28 Mar 2016 23:12:57 +0100, Nobody wrote:

On Mon, 28 Mar 2016 02:55:22 +0100, Nobody wrote:

Is there any simple way to regulate the output voltage of a boost
converter in continuous-conduction mode?

Thanks to Phil and Tim for the replies.

I think that the meta-answer is basically to elevate Tim's book up my
reading queue, then come back to this (another control loop which I
need to stabilise involves matching the rate of buying books to the
time available for reading them).

A paper that I'm thinking of writing is a piece entitled, basically,
"Change the Plant!". It pertains to the fact that you can't make a
silk purse out of a sow's ear, meaning, in this context, that given a
specific arrangement of mechanism, structure, sensors and actuators,
there's a limit to the performance that it's possible to achieve with
closed-loop control. To go beyond that limit, you need to change
something for better inherent regulation, bandwidth, linearity, etc.

Local feedback is good too. I've done several diode laser temperature
controllers like that. One of them used local feedback to make the
Peltier/thermistor response simple, then just wired it in parallel with
the current-tuning control. That way I got an automatic coarse/fine
wavelength adjustment for a cavity stabilized laser. Worked great.
Click to expand...

Jokes about reading speed aside, yes, inner loops can do wonders. One
(generally) still has to change the plant to the extent of adding
sensors, but it's often a mild change that can give great results.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
 
On Wed, 30 Mar 2016 15:05:38 -0500, Tim Wescott
<seemywebsite@myfooter.really> wrote:

[snip]
Jokes about reading speed aside, yes, inner loops can do wonders. One
(generally) still has to change the plant to the extent of adding
sensors, but it's often a mild change that can give great results.
Click to expand...

Inner loops are life-savers in ASIC's... local linearization avoids
gross swings in loop gain with signal swing.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

The touchstone of liberalism is intolerance
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top