Relation between filters and controllers

mesmart <98908@dsprelated> wrote:
> You are right Greg. In spite of the very insightful comments by various
> members, I don't think my query has been resolved.
 
> Let me refine the question a bit and see if it helps others understand what
> I am actually seeking,
 
> Consider a 2 pole 2 zero IIR filter that has well defined frequency
> response characteristics.It would "reject" a certain band of frequencies
> and "allow" others. 
(snip)

> So, it seems the same 2P2Z block which in certain applications can be used
> as a frequency filter, is now being used in the power converter application
> as a kind of PID controller. While being used as a filter, it is fed with a
> signal that possibly has a broad frequency spectrum, whereas when used as a
> controller, it's fed with an "error" signal.

The "error" signal depends on how far the output is from the desired
output, and that can depend in a complicated way on the load.  But
assuming a reasonable load, it is important for the combination of the
filter and the rest of the circuit to have certain nice properties.
It shouldn't, for example, start oscillating for reasonable loads.

The original IBM PC power supply would fail destructively under zero
load conditions. I don't know why, but I did see a brand X power supply
do it once. Yes, including the magic smoke leaking out.

On the other hand, power supplies for laptop computers pretty much have
to be stable under zero load, and also the transitions between zero load
and full load, in both directions. That is the way they are often used.

Either filter or controller should not get into an oscillating state
under normal conditions. 

-- glen

On Sunday, November 17, 2013 3:58:35 AM UTC+13, mesmart wrote:
> You are right Greg. In spite of the very insightful comments by various
> 
> members, I don't think my query has been resolved.
> 
> 
> 
> Let me refine the question a bit and see if it helps others understand what
> 
> I am actually seeking,
> 
> 
> 
> Consider a 2 pole 2 zero IIR filter that has well defined frequency
> 
> response characteristics.It would "reject" a certain band of frequencies
> 
> and "allow" others. 
> 
> Now, I have come across several instances in my study of digital power
> 
> converters where a particular CONTROLLER is called as a 2 pole 2 zero
> 
> controller.  The typical context is when in a given DC-DC converter, the
> 
> output voltage is compared against a reference voltage, The error signal is
> 
> then fed to a 2P2Z "controller" which then calculates a duty cycle for the
> 
> power switches to regulate the output at the desired set point.
> 
> 
> 
> So, it seems the same 2P2Z block which in certain applications can be used
> 
> as a frequency filter, is now being used in the power converter application
> 
> as a kind of PID controller. While being used as a filter, it is fed with a
> 
> signal that possibly has a broad frequency spectrum, whereas when used as a
> 
> controller, it's fed with an "error" signal.
> 
> 
> 
> I am not able to see how a block that in one context rejects a certain band
> 
> of frequencies, is being used in another application to generate a control
> 
> signal from an error input.
> 
> 
> 
> I hope this should help us narrow down the discussion from a generalized
> 
> scenario to a specific instance which is also in common use.
> 
> 
> 
> Thanks everyone for your time and inputs!
> 
> 	 
> 
> 
> 
> _____________________________		
> 
> Posted through www.DSPRelated.com

Look, they are ALL filters, end of story. The only caveat is that controllers tend to be lower order filters. eg a phase lead or lag is only first order as is a PI. The composite cascaded controller order of course can be much bigger when you put a lag-lead and a P-I together of course. Also for dc control systems the frequency response always goes down to dc unlike some filters which may well be ac coupled. The controller is shaping or altering the plant (system) transfer function. This is not rocket science.

On Sat, 16 Nov 2013 13:30:38 -0800, gyansorova wrote:

> On Sunday, November 17, 2013 3:58:35 AM UTC+13, mesmart wrote:

>> snip <<

> 
> Look, they are ALL filters, end of story. The only caveat is that
> controllers tend to be lower order filters. eg a phase lead or lag is
> only first order as is a PI. The composite cascaded controller order of
> course can be much bigger when you put a lag-lead and a P-I together of
> course. Also for dc control systems the frequency response always goes
> down to dc unlike some filters which may well be ac coupled. The
> controller is shaping or altering the plant (system) transfer function.
> This is not rocket science.

I would like to say "please keep this attitude, it can make me lots of 
money when products designed with it crash and burn".  Except that 
usually products that are designed with the attitude of "it's not rocket 
science, we'll just build a <whatever> and slap a PID on it and succeed" 
end up crashing and burning for more reasons than just the controller 
software needing help.

I've come in at the last minute on projects like this twice.  In both 
cases I managed to realize great improvements, but in both cases both the 
client and I were frustrated, because there were limitations elsewhere in 
the control system that could have been fixed cheaply given more time, 
but for which there wasn't enough room in the schedule to fix by the time 
I was called in.

They _are_ all filters, and that is most of the story.  But it isn't all, 
by any means.  There are exceptions to just about everything you say 
above, and if you act like they aren't then you're just rolling a set of 
dice that are loaded against you.  If you treat a controller design like 
any other filter then sometimes it'll work great and sometimes it won't, 
and the only way you'll know for sure is if you verify _before hand_ that 
treating your controller as just a filter will be good enough.

-- 

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

On Monday, November 18, 2013 11:01:32 AM UTC+13, Tim Wescott wrote:
> On Sat, 16 Nov 2013 13:30:38 -0800, gyansorova wrote:
> 
> 
> 
> > On Sunday, November 17, 2013 3:58:35 AM UTC+13, mesmart wrote:
> 
> 
> 
> >> snip <<
> 
> 
> 
> > 
> 
> > Look, they are ALL filters, end of story. The only caveat is that
> 
> > controllers tend to be lower order filters. eg a phase lead or lag is
> 
> > only first order as is a PI. The composite cascaded controller order of
> 
> > course can be much bigger when you put a lag-lead and a P-I together of
> 
> > course. Also for dc control systems the frequency response always goes
> 
> > down to dc unlike some filters which may well be ac coupled. The
> 
> > controller is shaping or altering the plant (system) transfer function.
> 
> > This is not rocket science.
> 
> 
> 
> I would like to say "please keep this attitude, it can make me lots of 
> 
> money when products designed with it crash and burn".  Except that 
> 
> usually products that are designed with the attitude of "it's not rocket 
> 
> science, we'll just build a <whatever> and slap a PID on it and succeed" 
> 
> end up crashing and burning for more reasons than just the controller 
> 
> software needing help.
> 
> 
> 
> I've come in at the last minute on projects like this twice.  In both 
> 
> cases I managed to realize great improvements, but in both cases both the 
> 
> client and I were frustrated, because there were limitations elsewhere in 
> 
> the control system that could have been fixed cheaply given more time, 
> 
> but for which there wasn't enough room in the schedule to fix by the time 
> 
> I was called in.
> 
> 
> 
> They _are_ all filters, and that is most of the story.  But it isn't all, 
> 
> by any means.  There are exceptions to just about everything you say 
> 
> above, and if you act like they aren't then you're just rolling a set of 
> 
> dice that are loaded against you.  If you treat a controller design like 
> 
> any other filter then sometimes it'll work great and sometimes it won't, 
> 
> and the only way you'll know for sure is if you verify _before hand_ that 
> 
> treating your controller as just a filter will be good enough.
> 
> 
> 
> -- 
> 
> 
They won't crash and burn, they'll work properly unlike fiddlers with PID type approaches. I don't treat them just as filters, but that is what they are (in the linear case at least). Of course there is always an exception to every rule but you don't make the rule up based on the exceptions.

Hello All,

I think I get it now. Kindly let me know if you agree to the way I see it:

Continuing with the same power converter example, its evident that the
output voltage signal shall have some average (DC) error and also some
ripple due to high frequency switching. Now, we need to develop the control
loop in such a way that it doesn't try to react to the high frequency
switching ripple as it is an integral part of the system that simply cannot
be eliminated. Rather we would be more interested in reacting to any other
variations (which would be necessarily of frequencies considerably lesser
than the switching frequency). For this, we first need to separate out
these low frequency variations (including any DC error as well) and then
make the control loop react to only these variations. This is where the
filtering nature of the 2P2Z block becomes important as it would be
designed to reject the high frequency variations(caused by switching
action, and its harmonics) in the output signal.

Thanks everyone!
	 

_____________________________		
Posted through www.DSPRelated.com

On Sun, 17 Nov 2013 23:24:50 -0600, mesmart wrote:

> Hello All,
> 
> I think I get it now. Kindly let me know if you agree to the way I see
> it:
> 
> Continuing with the same power converter example, its evident that the
> output voltage signal shall have some average (DC) error and also some
> ripple due to high frequency switching. Now, we need to develop the
> control loop in such a way that it doesn't try to react to the high
> frequency switching ripple as it is an integral part of the system that
> simply cannot be eliminated. Rather we would be more interested in
> reacting to any other variations (which would be necessarily of
> frequencies considerably lesser than the switching frequency). For this,
> we first need to separate out these low frequency variations (including
> any DC error as well) and then make the control loop react to only these
> variations. This is where the filtering nature of the 2P2Z block becomes
> important as it would be designed to reject the high frequency
> variations(caused by switching action, and its harmonics) in the output
> signal.

You still don't get it.  Even without the ripple, you'd still need a 
controller, you could still use a filter block to implement it, and your 
"controller" would still be a "filter".

If you learn control theory this will be obvious.  If you don't, you'll 
just be stumbling around in the dark.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com