Forums

Tuning PDF loop (pseudo derivative feedback)

Started by Unknown January 15, 2014
On Wednesday, January 15, 2014 11:20:49 AM UTC-6, Tim Wescott wrote:

> Somehow I remember Phelan doing something different -- didn't he have a > > PD controller whose output was integrated? At any rate, if you get your > > scaling right then in the linear case it all works the same. >
From what I can tell, he maintains the integral term just as it is in PID. Then he subtracts away the (input * Kd). Not the error, but the raw input. Seems to work ok but again, that's with guesses for coefficients. I don't know if anyone has seen this, but this is how I first learned of it. http://www.billnye.com/richard-phelan/
George Herold wrote:
> On Wednesday, January 15, 2014 8:23:09 AM UTC-5, hon...@yahoo.com wrote: >> I know about the auto-tuning algorithm for PID: cycle the temperature up and down, measure the peaks and distance between the peaks, and calculate Kp, Ki, and Kd. >> >> >> >> But recently I've been experimenting with what Phelan called PDF, and am intrigued by it. There isn't a lot of information out there about it, but nearly everyone that's used it has good things to say. >> >> >> >> The problem is, how is it tuned? It only has a Ki and Kd, so that makes it slightly simpler. An auto-tune algo like the one I just mentioned would be great, or at the very least, some general guide for manual tuning. A method that doesn't mention poles, zeros, or transforms would be nice because even though I have an EE degree, my strong point isn't controls. >> >> >> >> Either I forgot everything in my controls class or never learned it, I don't know which. > > Excellent! (I'm very interested also and will try this in the next control task.) I think the PDF term is a bit of a confusing as a name. What he really does is just use negative feedback around the 'plant'. So (though I haven't done it and don't know for sure.) You are still going to have to gain up the error signal, then integrate it with some time constant, and apply negative feedback as the damping term. I'd start with near the same gain and integral time as in the 'normal' PI controller. and then add the negative feedback. > > So 'normal' PI looks like > > error->gain--+--P(gain=1)----+->to plant--- > +--intergral----+ > (where the P and I terms are summed.) > > and the PDF will look like, > > error->gain--+--intergral---+->to plant---+-- > +-<-g=-1)-<---+ > > That's my understanding, Perhaps someone else can correct or refine it. > > George H.
I do not know anything about PID or control "systems". But i have a fancy MicroOmega PID controller as a part of a toaster oven temperature controlled system. PID systems are billed as being the best thing since Langendorf sliced bread). Well, then why the heck is it that they all seem to take forever to stabilize, ponderously ring after the "step" turn-on, and take five times forever to stabilize at a temperature DIFFERENT than the set-point? Seems to me, if proper parameters were used in a proper control system (most likely _NOT_ PID), the controlled environment would see the turn-on step with only a slight overshoot, NO ringing, followed with an almost immediate settling TO the set-point. Am guessing a damping factor of 0.95; there is a paper somewhere that gets into that and calculates the ideal under-damped value.
On Wednesday, January 15, 2014 11:55:51 AM UTC-6, Robert Baer wrote:
> > I do not know anything about PID or control "systems". > > But i have a fancy MicroOmega PID controller as a part of a toaster > > oven temperature controlled system. > > PID systems are billed as being the best thing since Langendorf > > sliced bread). > > Well, then why the heck is it that they all seem to take forever to > > stabilize, ponderously ring after the "step" turn-on, and take five > > times forever to stabilize at a temperature DIFFERENT than the set-point? > > > > Seems to me, if proper parameters were used in a proper control > > system (most likely _NOT_ PID), the controlled environment would see the > > turn-on step with only a slight overshoot, NO ringing, followed with an > > almost immediate settling TO the set-point. > > Am guessing a damping factor of 0.95; there is a paper somewhere that > > gets into that and calculates the ideal under-damped value.
Exactly. PID isn't a very good control algo, but everyone knows of it so they keep using it.
On Wed, 15 Jan 2014 10:08:04 -0800, hondgm wrote:

> On Wednesday, January 15, 2014 11:55:51 AM UTC-6, Robert Baer wrote: >> >> I do not know anything about PID or control "systems". >> >> But i have a fancy MicroOmega PID controller as a part of a toaster >> >> oven temperature controlled system. >> >> PID systems are billed as being the best thing since Langendorf >> >> sliced bread). >> >> Well, then why the heck is it that they all seem to take forever to >> >> stabilize, ponderously ring after the "step" turn-on, and take five >> >> times forever to stabilize at a temperature DIFFERENT than the >> set-point? >> >> >> >> Seems to me, if proper parameters were used in a proper control >> >> system (most likely _NOT_ PID), the controlled environment would see >> the >> >> turn-on step with only a slight overshoot, NO ringing, followed with an >> >> almost immediate settling TO the set-point. >> >> Am guessing a damping factor of 0.95; there is a paper somewhere >> that >> >> gets into that and calculates the ideal under-damped value. > > Exactly. PID isn't a very good control algo, but everyone knows of it > so they keep using it.
I beg to differ. Ripping some PID algorithm off of a web page and implementing it without knowing what you're doing isn't going to bring you joy. Most control loops use some form of PI or PID control loop, with various decorations. Choosing the right decorations to your PID loop, and tuning it correctly, make a huge difference. "I don't know anything about making XXX's work, and my XXX works like crap" is more a statement about one's level of knowledge about the thing in question than about the thing itself. Following that with "so therefore all XXX's are crap" says a lot more about one's own ability to think things through than it does about whatever it is that you're slamming. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
On 1/15/2014 11:25 AM, George Herold wrote:
> On Wednesday, January 15, 2014 10:59:34 AM UTC-5, hon...@yahoo.com > wrote: >> On Wednesday, January 15, 2014 9:46:23 AM UTC-6, Tim Wescott >> wrote: >> >> >> >>> >> >>> Look around and see if you can find a Z-N method for just Ki and >>> Kp. >> >>> >> >> >> >> Hmmm, I already have that implemented for PI control. But P is not >> exactly the same as the D term in PDF, although close. I'm >> terrible at the math surrounding this. I was thinking that at >> minimum, there was some tuning methodology for PDF, but it's >> apparently not common enough. > > I've got a copy of Phelans book.. but it's at home... And I never > understood his notation (in depth) There is no D (derivative term) > in PDF. Just an integral and damping. Hey you might want to read > Bob Pease's "What's all this PID stuff anyway" So what's a bit weird > is that if you do a PI loop the I term becomes the 'gain' and the P > term is the damping.. Then for PDF he keeps the integral (gain) but > uses negative feed back for the damping. And (for the way I think > about it) in PI the P term is one. And for PDF the negative feed > back term is -1 (I think.) > > George H. >
I have it too. The main nugget I took away was that you win by wrapping a P loop around the plant to make it fast and simple, and then putting an integrating loop around the whole works. The trick works great, but his "pseudo-derivative" nomenclature is claptrap. 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, 15 Jan 2014 17:40:29 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>On 1/15/2014 11:25 AM, George Herold wrote: >> On Wednesday, January 15, 2014 10:59:34 AM UTC-5, hon...@yahoo.com >> wrote: >>> On Wednesday, January 15, 2014 9:46:23 AM UTC-6, Tim Wescott >>> wrote: >>> >>> >>> >>>> >>> >>>> Look around and see if you can find a Z-N method for just Ki and >>>> Kp. >>> >>>> >>> >>> >>> >>> Hmmm, I already have that implemented for PI control. But P is not >>> exactly the same as the D term in PDF, although close. I'm >>> terrible at the math surrounding this. I was thinking that at >>> minimum, there was some tuning methodology for PDF, but it's >>> apparently not common enough. >> >> I've got a copy of Phelans book.. but it's at home... And I never >> understood his notation (in depth) There is no D (derivative term) >> in PDF. Just an integral and damping. Hey you might want to read >> Bob Pease's "What's all this PID stuff anyway" So what's a bit weird >> is that if you do a PI loop the I term becomes the 'gain' and the P >> term is the damping.. Then for PDF he keeps the integral (gain) but >> uses negative feed back for the damping. And (for the way I think >> about it) in PI the P term is one. And for PDF the negative feed >> back term is -1 (I think.) >> >> George H. >> > >I have it too. The main nugget I took away was that you win by wrapping >a P loop around the plant to make it fast and simple, and then putting >an integrating loop around the whole works. The trick works great, but >his "pseudo-derivative" nomenclature is claptrap. > >Cheers > >Phil Hobbs
Using local loops has the advantage of improving the bandwidth (and phase) of the individual "chunks". ...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 | I love to cook with wine. Sometimes I even put it in the food.
On Wednesday, January 15, 2014 5:40:29 PM UTC-5, Phil Hobbs wrote:
> On 1/15/2014 11:25 AM, George Herold wrote: > > > On Wednesday, January 15, 2014 10:59:34 AM UTC-5, hon...@yahoo.com > > wrote: > >> On Wednesday, January 15, 2014 9:46:23 AM UTC-6, Tim Wescott > > > I've got a copy of Phelans book.. but it's at home... And I never > > > understood his notation (in depth) There is no D (derivative term) > > > in PDF. Just an integral and damping. Hey you might want to read > > > Bob Pease's "What's all this PID stuff anyway" So what's a bit weird > > is that if you do a PI loop the I term becomes the 'gain' and the P > > term is the damping.. Then for PDF he keeps the integral (gain) but > > uses negative feed back for the damping. And (for the way I think > > about it) in PI the P term is one. And for PDF the negative feed > > back term is -1 (I think.) > > > > George H. > > > > I have it too. The main nugget I took away was that you win by wrapping > a P loop around the plant to make it fast and simple, and then putting > an integrating loop around the whole works. The trick works great, but > his "pseudo-derivative" nomenclature is claptrap.
Which is too bad, If he'd just called it negative feedback it might have had more impact. As Tim said PI vs PDF is basically the same with respect to 'outside' changes. But in response to setpoint changes the PDF is a bit better. and then Phelan makes some semi-outrageous claims about how much better PDF is where he compares the exponential 'power' response of the PI to a more delayed PDF, with the cavet that they both stay in the linear region. Anyhow I'll be doing one soon, I guess I'll tune it by looking at the step response. (to answer the OP's question) George H.
> > > > 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 Wednesday, January 15, 2014 5:40:29 PM UTC-5, Phil Hobbs wrote:

> I have it too. The main nugget I took away was that you win by wrapping=
=20
>=20 > a P loop around the plant to make it fast and simple, and then putting=20 >=20 > an integrating loop around the whole works. The trick works great, but=
=20
>=20 > his "pseudo-derivative" nomenclature is claptrap. >=20 >=20 >=20 > Cheers >=20 >=20 >=20 > Phil Hobbs >=20 >=20 >=20 >=20 >=20 > --=20 >=20 > Dr Philip C D Hobbs >=20 > Principal Consultant >=20 > ElectroOptical Innovations LLC >=20 > Optics, Electro-optics, Photonics, Analog Electronics >=20 >=20 >=20 > 160 North State Road #203 >=20 > Briarcliff Manor NY 10510 >=20 >=20 >=20 > hobbs at electrooptical dot net >=20 > http://electrooptical.net
How come everyone hates Phelan so much? Adding proportional feedback to a s= umming node after the integrator is equivalent to differentiating and summi= ng before the integrator, without actually building a differentiator, hence= the pseudo. Supposedly PDF is more immune to nose disturbances in the feed= back sensors. Wasn't Phelan a mechanical engineer, so he's going to be doin= g things that seem odd to an EE.
On 1/15/2014 8:06 PM, bloggs.fredbloggs.fred@gmail.com wrote:
> On Wednesday, January 15, 2014 5:40:29 PM UTC-5, Phil Hobbs wrote: > >> I have it too. The main nugget I took away was that you win by >> wrapping >> >> a P loop around the plant to make it fast and simple, and then >> putting >> >> an integrating loop around the whole works. The trick works >> great, but >> >> his "pseudo-derivative" nomenclature is claptrap. >> >> >> >> Cheers >> >> >> >> Phil Hobbs
>> http://electrooptical.net > > How come everyone hates Phelan so much? Adding proportional feedback > to a summing node after the integrator is equivalent to > differentiating and summing before the integrator, without actually > building a differentiator, hence the pseudo. Supposedly PDF is more > immune to nose disturbances in the feedback sensors. Wasn't Phelan a > mechanical engineer, so he's going to be doing things that seem odd > to an EE.
I don't hate Phelan at all, but reading his stuff you do get the impression that his head was too big to fit through the door. He was always spouting off about how everybody else's approaches were obsolete and so on, and publishing "universal" controller schematics that look like some plumber's nightmare. Apparently he pissed off everybody in the field back in the day, so they ignored him more or less completely. Which is too bad, because some of his stuff was 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 Wednesday, January 15, 2014 8:12:47 PM UTC-5, Phil Hobbs wrote:
> On 1/15/2014 8:06 PM, bloggs.fredbloggs.fred@gmail.com wrote: > > > On Wednesday, January 15, 2014 5:40:29 PM UTC-5, Phil Hobbs wrote: > > How come everyone hates Phelan so much? Adding proportional feedback > > to a summing node after the integrator is equivalent to > > differentiating and summing before the integrator, without actually >> building a differentiator, hence the pseudo. Supposedly PDF is more > > immune to nose disturbances in the feedback sensors. Wasn't Phelan a > > mechanical engineer, so he's going to be doing things that seem odd > > to an EE. > > > > I don't hate Phelan at all, but reading his stuff you do get the > impression that his head was too big to fit through the door. He was > always spouting off about how everybody else's approaches were obsolet > and so on, and publishing "universal" controller schematics that look > like some plumber's nightmare. Apparently he pissed off everybody in > the field back in the day, so they ignored him more or less completely. > > Which is too bad, because some of his stuff was great. >
Perfect response, thanks Phil. I don't hate Phelan at all. I'm eager to try it. He does come across as a bit of a blow hard. It does give one pause that negative feedback (PDF, whatever you call it) seems to be ignored in the control community. George H.
> > 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