On Tue, 17 May 2016 13:49:43 -0400, Phil Hobbs wrote:> On 05/17/2016 01:30 PM, Tim Wescott wrote: >> On Tue, 17 May 2016 12:02:12 -0500, John S wrote: >> >>> On 5/17/2016 11:57 AM, Tim Wescott wrote: >>>> On Tue, 17 May 2016 11:52:12 -0500, John S wrote: >>>> >>>>> On 5/17/2016 10:54 AM, Tim Wescott wrote: >>>>>> On Tue, 17 May 2016 08:15:42 -0500, John S wrote: >>>>>> >>>>>>> On 5/17/2016 1:41 AM, Tim Wescott wrote: >>>>>>>> I've just spent over two weeks getting ready to do my next video. >>>>>>>> It was a combination of one of those vast underestimations one >>>>>>>> occasionally makes, combined with falling into a bit of an >>>>>>>> obsession. >>>>>>>> >>>>>>>> I am, at this point, not only wondering if it was worth it, but >>>>>>>> questioning my sanity in carrying on even when the going went >>>>>>>> beyond tough to just plain crazy. >>>>>>>> >>>>>>>> At any rate, a good video needs a visual aid, and I decided that >>>>>>>> my video needed to demonstrate stability with a pendulum. >>>>>>>> Moreover, it needed a pendulum that could be worked >>>>>>>> electronically. So, I've >>>>>>>> >>>>>>>> * Disassembled a hard drive for it's head positioner. This took >>>>>>>> a day or two. >>>>>>>> >>>>>>>> * Decided that wasn't good enough and wound my own custom coil >>>>>>>> (220 feet of #40 wire, woo hoo!). This took a false start (18 >>>>>>>> feet of #34 wire) and several days. >>>>>>>> >>>>>>>> * Mounted the coil into a custom pendulum, running on Real Ball >>>>>>>> Bearings. Several more days, and if you touch it wrong the Q >>>>>>>> goes down from about 80 to about 10, then you have to fiddle with >>>>>>>> it for several minutes so the moving parts don't rub. >>>>>>>> >>>>>>>> * Built an oscillator that uses the pendulum as its resonator >>>>>>>> (this is where stability comes in -- is an oscillator stable? >>>>>>>> How is it stable? >>>>>>>> What if it's showing chaotic behavior?). This was astonishingly >>>>>>>> frustrating, and didn't finally work until I carefully modeled >>>>>>>> the pendulum as a resonator AND took the coil inductance into >>>>>>>> account in the circuit. This part too about a week. >>>>>>>> >>>>>>>> And for all that, I now have the time base for an exceptionally >>>>>>>> inaccurate electro-mechanical clock! Check out the picture. >>>>>>>> That's one cycle of the pendulum, running off of a "tick-toc" >>>>>>>> circuit that (A) >>>>>>>> minimizes the load on the pendulum (to give a high loaded Q, >>>>>>>> essential for wringing as much accuracy as possible out of a >>>>>>>> pendulum, never mind that it's made of wood, masking tape, and >>>>>>>> car parts that I picked up off the floor), and (B) has to be >>>>>>>> started by hand (I wanted to demonstrate a hard limit cycle). >>>>>>>> >>>>>>>> http://wescottdesign.com/movies/stability_teaser.gif >>>>>>>> >>>>>>>> More on all of this when I post the video. >>>>>>>> >>>>>>>> >>>>>>> For your next demo, use an electromagnet to lift a metal ball and >>>>>>> hold it suspended. Sense the height with a light sensor. Use PID >>>>>>> to achieve stability. >>>>>>> >>>>>>> I saw an article that did this 30 or so years ago. They used a >>>>>>> hollow steel ball with a map of the earth painted on. Can't >>>>>>> remember the diameter of the ball, but maybe 1". >>>>>> >>>>>> I've done that. You need a honkin' big electromagnet to make it >>>>>> work with a plain steel load. >>>>>> >>>>>> The executive desk-toys with the floating globes use big (30mm dia >>>>>> x 10mm) rare-earth magnets, and float the ball a little bit below >>>>>> the neutral point. I believe that they use hall effect sensors to >>>>>> detect the magnet proximity. >>>>>> >>>>>> >>>>> Oops! Sorry! I didn't know you were already doing it. >>>> >>>> Well, the project is on long-term hold, but I did get as far as >>>> getting one of those desk toys and starting to take it apart. >>>> >>>> >>> Would it be inappropriate to demonstrate PID? Just curious. >> >> I suspect that it would be a good way to demonstrate a number of >> principles in control systems, PID controllers included. >> >> A lot of people seem to separate "PID" from other controllers -- >> somewhere in this thread someone made a comment about "oh, that >> wouldn't work with a PID controller". Yet most advanced controllers >> really boil down to a PID controller with: >> >> * A different way of arriving at the gains; >> * various linear and nonlinear decorations; >> * and a fancy name. >> >> But -- that's a rant for another day. >> >> > Well, there are various lead-lag type tricks that require basically > lowpass filtering the D term. PLLs and diffusion-dominated temperature > controls don't like terms with huge noise gain. > > When I was a post-doc, in about 1988, I did a motion controller for a > piezo bimorph used in a scanning force microscope. It had a notch > filter for the lowest resonance, two integrators (with lead-lag to make > the loop stable) and two more poles to get rid of wideband crap that > would excite the higher resonances. Worked great--when tweaked for best > settling time, the loop BW was about 30% of the resonant frequency, vs. > 3% for the previous version. (The resonance had a Q of about 30, so > with just one integrator and no notch, the BW had to be backed off a > really long way to prevent oscillation.) > > Would you include that in "decorated PID"?By the time you've got that much stuff stuck on the loop then I may just insist that there's a PID controller (or at least PI) at the core. You're making me realize that the line is blurry -- but I've certainly made loops involving a band-limited derivative (which is just another way of making a lead-lag filter) and a notch, and still considered it a "PID".> I generally do it that way, i.e. use frequency compensation ideas, > concentrate on phase margin, and then watch out for the transient > response, because when you do something fancy like the above, the > settling behaviour is liable to be a bit strange. Of course when one of > my loops misbehaves a bit during testing, no machinery is ruined and > nobody gets hurt.When I need to seriously tune for performance that's how I do it, too.> Windup and nonlinear slewing are sometimes issues too, of course. > > Overall it seems like the distinction between control guys and other EEs > is a bit like the difference between civil and mechanical--a lot of the > same concepts but completely different emphasis. Is that fair?Hmm. That's an interesting idea. One certainly can't sit in the middle of the "EE" room and refuse to budge out of it -- you have to understand the thing you're controlling, which means that you at least need to be able to talk to people of whatever discipline will help you to understand the plant dynamics. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!
Stability and Insanity
Started by ●May 17, 2016
Reply by ●May 17, 20162016-05-17
Reply by ●May 17, 20162016-05-17
On Tue, 17 May 2016 11:37:11 -0700, John Larkin wrote:> On Tue, 17 May 2016 01:41:51 -0500, Tim Wescott > <seemywebsite@myfooter.really> wrote: > >>I've just spent over two weeks getting ready to do my next video. It >>was a combination of one of those vast underestimations one occasionally >>makes, combined with falling into a bit of an obsession. >> >>I am, at this point, not only wondering if it was worth it, but >>questioning my sanity in carrying on even when the going went beyond >>tough to just plain crazy. >> >>At any rate, a good video needs a visual aid, and I decided that my >>video needed to demonstrate stability with a pendulum. Moreover, it >>needed a pendulum that could be worked electronically. So, I've >> >>* Disassembled a hard drive for it's head positioner. This took a day >>or two. >> >>* Decided that wasn't good enough and wound my own custom coil (220 feet >>of #40 wire, woo hoo!). This took a false start (18 feet of #34 wire) >>and several days. >> >>* Mounted the coil into a custom pendulum, running on Real Ball >>Bearings. Several more days, and if you touch it wrong the Q goes down >>from about 80 to about 10, then you have to fiddle with it for several >>minutes so the moving parts don't rub. >> >>* Built an oscillator that uses the pendulum as its resonator (this is >>where stability comes in -- is an oscillator stable? How is it stable? >>What if it's showing chaotic behavior?). This was astonishingly >>frustrating, and didn't finally work until I carefully modeled the >>pendulum as a resonator AND took the coil inductance into account in the >>circuit. This part too about a week. >> >>And for all that, I now have the time base for an exceptionally >>inaccurate electro-mechanical clock! Check out the picture. That's one >>cycle of the pendulum, running off of a "tick-toc" circuit that (A) >>minimizes the load on the pendulum (to give a high loaded Q, essential >>for wringing as much accuracy as possible out of a pendulum, never mind >>that it's made of wood, masking tape, and car parts that I picked up off >>the floor), and (B) has to be started by hand (I wanted to demonstrate a >>hard limit cycle). >> >>http://wescottdesign.com/movies/stability_teaser.gif >> >>More on all of this when I post the video. > > We seldom encounter chaotic behavior in electronic circuits... on > purpose, I guess. > > Superregenerative receivers are kind of chaotic, but mostly just noisy. > > A laser driving some fiber with lots of reflections can be chaotic, > but the only nonlinear element is the laser itself.There's not a lot of utility in a chaotic circuit. Chaotic systems dynamics were popular for about a decade starting in the early 1980's, but to my knowledge there wasn't much practical use that came out of the systems themselves, but there was certainly more understanding of the phenomenon. The circuit I'll use (Chua's circuit) bears a lot of resemblance to a plain-old power oscillator that's squegging as a consequence of too much ambition on the designer's part. My intuition is that a squegging oscillator circuit is, in fact, chaotic. If I had the time I'd collect some and see. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!
Reply by ●May 17, 20162016-05-17
On 18/05/16 01:54, Tim Wescott wrote:> On Tue, 17 May 2016 08:15:42 -0500, John S wrote: >> For your next demo, use an electromagnet to lift a metal ball and hold >> it suspended. Sense the height with a light sensor. Use PID to achieve >> stability. > I've done that. You need a honkin' big electromagnet to make it work > with a plain steel load.Could you use a smaller electromagnet below, partly canceling the pull of a rare-earth magnet from above?
Reply by ●May 17, 20162016-05-17
On Tue, 17 May 2016 14:33:50 -0400, Phil Hobbs wrote:> On 05/17/2016 01:02 PM, John S wrote: >> On 5/17/2016 11:57 AM, Tim Wescott wrote: >>> On Tue, 17 May 2016 11:52:12 -0500, John S wrote: >>> >>>> On 5/17/2016 10:54 AM, Tim Wescott wrote: >>>>> On Tue, 17 May 2016 08:15:42 -0500, John S wrote: >>>>> >>>>>> On 5/17/2016 1:41 AM, Tim Wescott wrote: >>>>>>> I've just spent over two weeks getting ready to do my next video. >>>>>>> It was a combination of one of those vast underestimations one >>>>>>> occasionally makes, combined with falling into a bit of an >>>>>>> obsession. >>>>>>> >>>>>>> I am, at this point, not only wondering if it was worth it, but >>>>>>> questioning my sanity in carrying on even when the going went >>>>>>> beyond tough to just plain crazy. >>>>>>> >>>>>>> At any rate, a good video needs a visual aid, and I decided that >>>>>>> my video needed to demonstrate stability with a pendulum. >>>>>>> Moreover, it needed a pendulum that could be worked >>>>>>> electronically. So, I've >>>>>>> >>>>>>> * Disassembled a hard drive for it's head positioner. This took a >>>>>>> day or two. >>>>>>> >>>>>>> * Decided that wasn't good enough and wound my own custom coil >>>>>>> (220 feet of #40 wire, woo hoo!). This took a false start (18 >>>>>>> feet of #34 wire) and several days. >>>>>>> >>>>>>> * Mounted the coil into a custom pendulum, running on Real Ball >>>>>>> Bearings. Several more days, and if you touch it wrong the Q goes >>>>>>> down from about 80 to about 10, then you have to fiddle with it >>>>>>> for several minutes so the moving parts don't rub. >>>>>>> >>>>>>> * Built an oscillator that uses the pendulum as its resonator >>>>>>> (this is where stability comes in -- is an oscillator stable? How >>>>>>> is it stable? >>>>>>> What if it's showing chaotic behavior?). This was astonishingly >>>>>>> frustrating, and didn't finally work until I carefully modeled the >>>>>>> pendulum as a resonator AND took the coil inductance into account >>>>>>> in the circuit. This part too about a week. >>>>>>> >>>>>>> And for all that, I now have the time base for an exceptionally >>>>>>> inaccurate electro-mechanical clock! Check out the picture. >>>>>>> That's one cycle of the pendulum, running off of a "tick-toc" >>>>>>> circuit that (A) >>>>>>> minimizes the load on the pendulum (to give a high loaded Q, >>>>>>> essential for wringing as much accuracy as possible out of a >>>>>>> pendulum, never mind that it's made of wood, masking tape, and car >>>>>>> parts that I picked up off the floor), and (B) has to be started >>>>>>> by hand (I wanted to demonstrate a hard limit cycle). >>>>>>> >>>>>>> http://wescottdesign.com/movies/stability_teaser.gif >>>>>>> >>>>>>> More on all of this when I post the video. >>>>>>> >>>>>>> >>>>>> For your next demo, use an electromagnet to lift a metal ball and >>>>>> hold it suspended. Sense the height with a light sensor. Use PID to >>>>>> achieve stability. >>>>>> >>>>>> I saw an article that did this 30 or so years ago. They used a >>>>>> hollow steel ball with a map of the earth painted on. Can't >>>>>> remember the diameter of the ball, but maybe 1". >>>>> >>>>> I've done that. You need a honkin' big electromagnet to make it >>>>> work with a plain steel load. >>>>> >>>>> The executive desk-toys with the floating globes use big (30mm dia x >>>>> 10mm) rare-earth magnets, and float the ball a little bit below the >>>>> neutral point. I believe that they use hall effect sensors to >>>>> detect the magnet proximity. >>>>> >>>>> >>>> Oops! Sorry! I didn't know you were already doing it. >>> >>> Well, the project is on long-term hold, but I did get as far as >>> getting one of those desk toys and starting to take it apart. >>> >>> >> Would it be inappropriate to demonstrate PID? Just curious. > > A levitated ball is a nearly undamped, nonlinear second order system, > which seems like a good test case. Tim's pendulum with a Q of 80 is > good too, and more linear for small deflections. > > Cheers > > Phil HobbsA nearly undamped, nonlinear _and unstable_ second-order system. So it's a really good example of how -- through the magic of closed-loop control -- you can stabilize an unstable system. IIRC at any point the linearized transfer function is a H(s) = ----------- s (s - b) where a and b are constants, and b is positive (the sign of a depends on your sign conventions, but that doesn't affect stability until you try wrapping control around the thing). -- Tim Wescott Control systems, embedded software and circuit design I'm looking for work! See my website if you're interested http://www.wescottdesign.com
Reply by ●May 17, 20162016-05-17
On 18/05/2016 10:44, Tim Wescott wrote:> On Tue, 17 May 2016 14:33:50 -0400, Phil Hobbs wrote: > >> On 05/17/2016 01:02 PM, John S wrote: >>> On 5/17/2016 11:57 AM, Tim Wescott wrote: >>>> On Tue, 17 May 2016 11:52:12 -0500, John S wrote: >>>> >>>>> On 5/17/2016 10:54 AM, Tim Wescott wrote: >>>>>> On Tue, 17 May 2016 08:15:42 -0500, John S wrote: >>>>>> >>>>>>> On 5/17/2016 1:41 AM, Tim Wescott wrote: >>>>>>>> I've just spent over two weeks getting ready to do my next video. >>>>>>>> It was a combination of one of those vast underestimations one >>>>>>>> occasionally makes, combined with falling into a bit of an >>>>>>>> obsession. >>>>>>>> >>>>>>>> I am, at this point, not only wondering if it was worth it, but >>>>>>>> questioning my sanity in carrying on even when the going went >>>>>>>> beyond tough to just plain crazy. >>>>>>>> >>>>>>>> At any rate, a good video needs a visual aid, and I decided that >>>>>>>> my video needed to demonstrate stability with a pendulum. >>>>>>>> Moreover, it needed a pendulum that could be worked >>>>>>>> electronically. So, I've >>>>>>>> >>>>>>>> * Disassembled a hard drive for it's head positioner. This took a >>>>>>>> day or two. >>>>>>>> >>>>>>>> * Decided that wasn't good enough and wound my own custom coil >>>>>>>> (220 feet of #40 wire, woo hoo!). This took a false start (18 >>>>>>>> feet of #34 wire) and several days. >>>>>>>> >>>>>>>> * Mounted the coil into a custom pendulum, running on Real Ball >>>>>>>> Bearings. Several more days, and if you touch it wrong the Q goes >>>>>>>> down from about 80 to about 10, then you have to fiddle with it >>>>>>>> for several minutes so the moving parts don't rub. >>>>>>>> >>>>>>>> * Built an oscillator that uses the pendulum as its resonator >>>>>>>> (this is where stability comes in -- is an oscillator stable? How >>>>>>>> is it stable? >>>>>>>> What if it's showing chaotic behavior?). This was astonishingly >>>>>>>> frustrating, and didn't finally work until I carefully modeled the >>>>>>>> pendulum as a resonator AND took the coil inductance into account >>>>>>>> in the circuit. This part too about a week. >>>>>>>> >>>>>>>> And for all that, I now have the time base for an exceptionally >>>>>>>> inaccurate electro-mechanical clock! Check out the picture. >>>>>>>> That's one cycle of the pendulum, running off of a "tick-toc" >>>>>>>> circuit that (A) >>>>>>>> minimizes the load on the pendulum (to give a high loaded Q, >>>>>>>> essential for wringing as much accuracy as possible out of a >>>>>>>> pendulum, never mind that it's made of wood, masking tape, and car >>>>>>>> parts that I picked up off the floor), and (B) has to be started >>>>>>>> by hand (I wanted to demonstrate a hard limit cycle). >>>>>>>> >>>>>>>> http://wescottdesign.com/movies/stability_teaser.gif >>>>>>>> >>>>>>>> More on all of this when I post the video. >>>>>>>> >>>>>>>> >>>>>>> For your next demo, use an electromagnet to lift a metal ball and >>>>>>> hold it suspended. Sense the height with a light sensor. Use PID to >>>>>>> achieve stability. >>>>>>> >>>>>>> I saw an article that did this 30 or so years ago. They used a >>>>>>> hollow steel ball with a map of the earth painted on. Can't >>>>>>> remember the diameter of the ball, but maybe 1". >>>>>> >>>>>> I've done that. You need a honkin' big electromagnet to make it >>>>>> work with a plain steel load. >>>>>> >>>>>> The executive desk-toys with the floating globes use big (30mm dia x >>>>>> 10mm) rare-earth magnets, and float the ball a little bit below the >>>>>> neutral point. I believe that they use hall effect sensors to >>>>>> detect the magnet proximity. >>>>>> >>>>>> >>>>> Oops! Sorry! I didn't know you were already doing it. >>>> >>>> Well, the project is on long-term hold, but I did get as far as >>>> getting one of those desk toys and starting to take it apart. >>>> >>>> >>> Would it be inappropriate to demonstrate PID? Just curious. >> >> A levitated ball is a nearly undamped, nonlinear second order system, >> which seems like a good test case. Tim's pendulum with a Q of 80 is >> good too, and more linear for small deflections. >> >> Cheers >> >> Phil Hobbs > > A nearly undamped, nonlinear _and unstable_ second-order system. So it's > a really good example of how -- through the magic of closed-loop control > -- you can stabilize an unstable system. > > IIRC at any point the linearized transfer function is > > a > H(s) = ----------- > s (s - b) > > where a and b are constants, and b is positive (the sign of a depends on > your sign conventions, but that doesn't affect stability until you try > wrapping control around the thing). >https://www.youtube.com/watch?v=D9NyYYVcspk
Reply by ●May 18, 20162016-05-18
On Tuesday, May 17, 2016 at 5:05:57 AM UTC-7, Evgeny Filatov wrote: (snip)> "In Drosophila and many other animals, including humans, the heart of > the circadian clock is a delayed negative feedback loop based on > transcription regulators: accumulation of certain gene products switches > off the transcription of their own genes, but with a delay, so that the > cell oscillates between a state in which the products are present and > transcription is switched off, and one in which the products are absent > and transcription is switched on."I have heard about genes with really long introns, which slows down transcription, and the suggestion that the reason is timing, but no details on what is being timed. That might be the one, though. -- glen
Reply by ●May 18, 20162016-05-18
On Tuesday, May 17, 2016 at 10:30:09 AM UTC-7, Tim Wescott wrote:> On Tue, 17 May 2016 12:02:12 -0500, John S wrote:(snip)> > Would it be inappropriate to demonstrate PID? Just curious.> I suspect that it would be a good way to demonstrate a number of > principles in control systems, PID controllers included.> A lot of people seem to separate "PID" from other controllers -- > somewhere in this thread someone made a comment about "oh, that wouldn't > work with a PID controller". Yet most advanced controllers really boil > down to a PID controller with:Seems like there are some where PID is more than is needed. Household ovens that I know use on/off control, which is close enough for most food, especially adding the thermal mass of the rest of the oven. A proportional controller might be a little better, maybe with zero cross switching SCRs. Are there cases where more than PID, such as second derivative, are useful? -- glen
Reply by ●May 18, 20162016-05-18
On Wednesday, May 18, 2016 at 3:33:34 PM UTC+12, herrman...@gmail.com wrote:> On Tuesday, May 17, 2016 at 10:30:09 AM UTC-7, Tim Wescott wrote: > > On Tue, 17 May 2016 12:02:12 -0500, John S wrote: > > (snip) > > > Would it be inappropriate to demonstrate PID? Just curious. > > > I suspect that it would be a good way to demonstrate a number of > > principles in control systems, PID controllers included. > > > A lot of people seem to separate "PID" from other controllers -- > > somewhere in this thread someone made a comment about "oh, that wouldn't > > work with a PID controller". Yet most advanced controllers really boil > > down to a PID controller with: > > Seems like there are some where PID is more than is needed. > > Household ovens that I know use on/off control, which is close enough for most food, > especially adding the thermal mass of the rest of the oven. A proportional controller > might be a little better, maybe with zero cross switching SCRs. > > Are there cases where more than PID, such as second derivative, are useful? > > -- glenMultiple phase-advances are common in high performance control systems. Multiple integrators too. You need to get the gain as high as possible at low frequencies, go through unity gain at around -20dB/decade and a steep roll-off at high frequencies. Sometimes you can squeeze through flat (asymptotically) through the unity gain crossing. if you put too many phase advances you can end up with three unity gain crossings, which is interesting. Basically PID is ok for basic control, low bandwidth applications.
Reply by ●May 18, 20162016-05-18
On Tue, 17 May 2016 20:33:31 -0700, herrmannsfeldt wrote:> On Tuesday, May 17, 2016 at 10:30:09 AM UTC-7, Tim Wescott wrote: >> On Tue, 17 May 2016 12:02:12 -0500, John S wrote: > > (snip) >> > Would it be inappropriate to demonstrate PID? Just curious. > >> I suspect that it would be a good way to demonstrate a number of >> principles in control systems, PID controllers included. > >> A lot of people seem to separate "PID" from other controllers -- >> somewhere in this thread someone made a comment about "oh, that >> wouldn't work with a PID controller". Yet most advanced controllers >> really boil down to a PID controller with: > > Seems like there are some where PID is more than is needed. > > Household ovens that I know use on/off control, which is close enough > for most food, especially adding the thermal mass of the rest of the > oven. A proportional controller might be a little better, maybe with > zero cross switching SCRs. > > Are there cases where more than PID, such as second derivative, are > useful?Agreed, and yes. "Control" doesn't necessarily equal "PID" -- sometimes just a simple ON/OFF is not only sufficient, but is actually better than PID. I have an excellent book on adaptive control ("Adaptive Control" by Astrom and Wittenmark). One chapter of the book is devoted to alternatives. One really good alternative to adaptive control, sometimes, is some sort of relay control (like the on/off control in your oven). You have to accept some chattering around your target point, but in the right system, done the right way, it can be very robust. It's actually been used quite successfully in unmanned aircraft that have to work over a wide speed range, but according to the authors when it was tried in manned aircraft it worked well but freaked out the pilots. There are also times -- quite common in industrial control, AFAIK, where you don't want to mess around with a derivative term or any other leading term. Derivatives amplify noise and make nasty high-speed plant behaviors come out of the woodwork. In such cases a PI controller may be the best thing. If a plant pretty much acts like a memoryless element, then just integral control may be best (EE's use this when they wrap a diode with an op-amp to get log or exponential circuits). And finally, in addition to times when less than PID is useful, there are cases where more than PID is useful. I've implemented a controller that was proportional-integral-double integral, that vastly improved the performance of the thing that it was used on. Since the plant itself was also integrating, the thing would oscillate either if you turned the master gain up too high, or down too low. That's why I said that "most" controllers look an awful lot like PID. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!
Reply by ●May 18, 20162016-05-18
On Wed, 18 May 2016 10:05:16 +1000, Clifford Heath wrote:> On 18/05/16 01:54, Tim Wescott wrote: >> On Tue, 17 May 2016 08:15:42 -0500, John S wrote: >>> For your next demo, use an electromagnet to lift a metal ball and hold >>> it suspended. Sense the height with a light sensor. Use PID to achieve >>> stability. >> I've done that. You need a honkin' big electromagnet to make it work >> with a plain steel load. > > Could you use a smaller electromagnet below, partly canceling the pull > of a rare-earth magnet from above?Yes. IMHO, a thingie that's hanging suspended from a point looks more impressive than a thingie that's hanging suspended between two points. A thingie that's floating _above_ a point is more impressive yet, but if you do that then all of a sudden you have to control it in three dimensions instead of one (or perhaps six if you can't make it inherently stable in rotation). You can float an aluminum pan above an electromagnet and have it be stable, but you can't do the same thing with plain old magnets. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!
