Larkinisms
in [Design]
|
From: Nunya on 22 Jul 2010 10:36 On Jul 22, 7:23 am, John Larkin <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > On Thu, 22 Jul 2010 06:58:04 -0700 (PDT), Nunya > > > > <jack_sheph...(a)cox.net> wrote: > >On Jul 21, 8:47 pm, John Larkin > ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > >> On Wed, 21 Jul 2010 23:49:36 +0100, "Ian Iveson" > > >> <IanIveson.h...(a)blueyonder.co.uk> wrote: > >> >John Larkin wrote: > > >> > >>> What, George? Is it supposed to be untrue? Explain > >> >how. > > >> >>>> Tim > > >> >>>The magnetic flux density, B, is proportional to i for all > >> >>>inductor/transformer configurations. > > >> >>>B = uH, u is approximately constant. > > >> >>>H = N*i/(2*pi*r) for a toroid. > > >> >>>At 0 current the magnetic flux density is 0. Simple as > >> >>>that. > > >> >> What "I" are you talking about? A transformer has two > >> >> windings. And > >> >> each winding has resistance. > > >> >> Take a typical power transformer. Connect the primary to > >> >> the AC line > >> >> and leave the secondary open. Measure or compute the core > >> >> flux > >> >> density. Now short the secondary. At least before it > >> >> catches fire, > >> >> shorted-secondary flux density will be about half of the > >> >> unloaded > >> >> value. > > >> >> Loading the secondary of a transformer reduces flux > >> >> density... just > >> >> look at the directions of the primary and secondary > >> >> winding currents. > >> >> It's surprising how many people get this wrong, and think > >> >> that a > >> >> transformer "saturates" if you short its output. > > >> >Yes. The flux arising from secondary current is in > >> >opposition to that arising from primary current. The two > >> >currents are related, so the situation is more complicated > >> >than George's Theory appreciates. > > >> An easy way to measure the flux density in a transformer core is with > >> a probe coil, namely any old 3rd winding on the same core. > > >> So get a power transformer with one primary and two secondaries. > >> Connect the primary to the AC line. Measure the voltage on secondary 1 > >> and switch a load on/off secondary 2. If flux density increases with > >> load, the voltage on the sec1 winding will go UP as sec2 is loaded. > > >> Of course, it doesn't anything that stupid. It goes down. > > >> At least AlwaysWrong has a new friend. > > >> John > > >induced voltage on the open winding is not a function of flux density, > >it is a function of the amplitude of the excitation winding. > > The voltage induced is proportional to N * dP/dT, where P is total > flux that the winding encloses, namely the center leg of the core. > Since the number of turns and the frequency and the core dimensions > don't change, the induced voltage tracks flux density. > > You are right if loading sec2 has zero affect on the voltage observed > at sec1, ie, all transformers have perfect cross-regulation. > > Try it. > > John Then the only proper way to get where you want to be is by using a toroidal core. In your scenario, it should not matter where the monitor winding goes, as long as it is in the flux loop path and makes at least one turn around that. Not everything is an E-I core configuration either, as it appears you assume, with your 'center leg' remark. That is also whay I stated that the right way is to use a toroidal core to be certain one is observing ALL of the flux.
From: John Larkin on 22 Jul 2010 11:28 On Thu, 22 Jul 2010 07:36:16 -0700 (PDT), Nunya <jack_shephard(a)cox.net> wrote: >On Jul 22, 7:23�am, John Larkin ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >> On Thu, 22 Jul 2010 06:58:04 -0700 (PDT), Nunya >> >> >> >> <jack_sheph...(a)cox.net> wrote: >> >On Jul 21, 8:47�pm, John Larkin >> ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >> >> On Wed, 21 Jul 2010 23:49:36 +0100, "Ian Iveson" >> >> >> <IanIveson.h...(a)blueyonder.co.uk> wrote: >> >> >John Larkin wrote: >> >> >> > >>> What, George? �Is it supposed to be untrue? �Explain >> >> >how. >> >> >> >>>> Tim >> >> >> >>>The magnetic flux density, B, is proportional to i for all >> >> >>>inductor/transformer configurations. >> >> >> >>>B = uH, u is approximately constant. >> >> >> >>>H = N*i/(2*pi*r) for a toroid. >> >> >> >>>At 0 current the magnetic flux density is 0. Simple as >> >> >>>that. >> >> >> >> What "I" are you talking about? A transformer has two >> >> >> windings. And >> >> >> each winding has resistance. >> >> >> >> Take a typical power transformer. Connect the primary to >> >> >> the AC line >> >> >> and leave the secondary open. Measure or compute the core >> >> >> flux >> >> >> density. Now short the secondary. At least before it >> >> >> catches fire, >> >> >> shorted-secondary flux density will be about half of the >> >> >> unloaded >> >> >> value. >> >> >> >> Loading the secondary of a transformer reduces flux >> >> >> density... just >> >> >> look at the directions of the primary and secondary >> >> >> winding currents. >> >> >> It's surprising how many people get this wrong, and think >> >> >> that a >> >> >> transformer "saturates" if you short its output. >> >> >> >Yes. The flux arising from secondary current is in >> >> >opposition to that arising from primary current. The two >> >> >currents are related, so the situation is more complicated >> >> >than George's Theory appreciates. >> >> >> An easy way to measure the flux density in a transformer core is with >> >> a probe coil, namely any old 3rd winding on the same core. >> >> >> So get a power transformer with one primary and two secondaries. >> >> Connect the primary to the AC line. Measure the voltage on secondary 1 >> >> and switch a load on/off secondary 2. If flux density increases with >> >> load, the voltage on the sec1 winding will go UP as sec2 is loaded. >> >> >> Of course, it doesn't anything that stupid. It goes down. >> >> >> At least AlwaysWrong has a new friend. >> >> >> John >> >> >induced voltage on the open winding is not a function of flux density, >> >it is a function of the amplitude of the excitation winding. >> >> The voltage induced is proportional to N * dP/dT, where P is total >> flux that the winding encloses, namely the center leg of the core. >> Since the number of turns and the frequency and the core dimensions >> don't change, the induced voltage tracks flux density. >> >> You are right if loading sec2 has zero affect on the voltage observed >> at sec1, ie, all transformers have perfect cross-regulation. >> >> Try it. >> >> John > >Then the only proper way to get where you want to be is by >using a toroidal core. What I want? What's that? > > In your scenario, it should not matter where the monitor >winding goes, as long as it is in the flux loop path and >makes at least one turn around that. Not everything is >an E-I core configuration either, as it appears you assume, >with your 'center leg' remark. That is also whay I stated that >the right way is to use a toroidal core to be certain one is >observing ALL of the flux. Most power transformers wrap all the windings around the same core, whatever its geometry. So what the primary induces into the core material, the secondaries see. The flux that matters is enclosed by all the windings [1]. So the voltage induced into an unloaded secondary is a good indicator of the flux density that matters, namely the flux that the windings actually induce and use; this concept has a pretty high f-sub-D [2] factor. The question was whether loading the secondary of a power transformer increases or decreases the core flux density. Which way do you vote? John [1] with minor footnote for leakage fields. Microwave oven transformers, for example, are designed to have lots of leakage. [2] "duh"
From: Nunya on 22 Jul 2010 11:59 On Jul 22, 8:28 am, John Larkin <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > On Thu, 22 Jul 2010 07:36:16 -0700 (PDT), Nunya > > > > <jack_sheph...(a)cox.net> wrote: > >On Jul 22, 7:23 am, John Larkin > ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > >> On Thu, 22 Jul 2010 06:58:04 -0700 (PDT), Nunya > > >> <jack_sheph...(a)cox.net> wrote: > >> >On Jul 21, 8:47 pm, John Larkin > >> ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: > >> >> On Wed, 21 Jul 2010 23:49:36 +0100, "Ian Iveson" > > >> >> <IanIveson.h...(a)blueyonder.co.uk> wrote: > >> >> >John Larkin wrote: > > >> >> > >>> What, George? Is it supposed to be untrue? Explain > >> >> >how. > > >> >> >>>> Tim > > >> >> >>>The magnetic flux density, B, is proportional to i for all > >> >> >>>inductor/transformer configurations. > > >> >> >>>B = uH, u is approximately constant. > > >> >> >>>H = N*i/(2*pi*r) for a toroid. > > >> >> >>>At 0 current the magnetic flux density is 0. Simple as > >> >> >>>that. > > >> >> >> What "I" are you talking about? A transformer has two > >> >> >> windings. And > >> >> >> each winding has resistance. > > >> >> >> Take a typical power transformer. Connect the primary to > >> >> >> the AC line > >> >> >> and leave the secondary open. Measure or compute the core > >> >> >> flux > >> >> >> density. Now short the secondary. At least before it > >> >> >> catches fire, > >> >> >> shorted-secondary flux density will be about half of the > >> >> >> unloaded > >> >> >> value. > > >> >> >> Loading the secondary of a transformer reduces flux > >> >> >> density... just > >> >> >> look at the directions of the primary and secondary > >> >> >> winding currents. > >> >> >> It's surprising how many people get this wrong, and think > >> >> >> that a > >> >> >> transformer "saturates" if you short its output. > > >> >> >Yes. The flux arising from secondary current is in > >> >> >opposition to that arising from primary current. The two > >> >> >currents are related, so the situation is more complicated > >> >> >than George's Theory appreciates. > > >> >> An easy way to measure the flux density in a transformer core is with > >> >> a probe coil, namely any old 3rd winding on the same core. > > >> >> So get a power transformer with one primary and two secondaries. > >> >> Connect the primary to the AC line. Measure the voltage on secondary 1 > >> >> and switch a load on/off secondary 2. If flux density increases with > >> >> load, the voltage on the sec1 winding will go UP as sec2 is loaded. > > >> >> Of course, it doesn't anything that stupid. It goes down. > > >> >> At least AlwaysWrong has a new friend. > > >> >> John > > >> >induced voltage on the open winding is not a function of flux density, > >> >it is a function of the amplitude of the excitation winding. > > >> The voltage induced is proportional to N * dP/dT, where P is total > >> flux that the winding encloses, namely the center leg of the core. > >> Since the number of turns and the frequency and the core dimensions > >> don't change, the induced voltage tracks flux density. > > >> You are right if loading sec2 has zero affect on the voltage observed > >> at sec1, ie, all transformers have perfect cross-regulation. > > >> Try it. > > >> John > > >Then the only proper way to get where you want to be is by > >using a toroidal core. > > What I want? What's that? You seem to think that placing it elsewhere risks missing some of the flux. There is no way to miss ANY of it with a toroid core. > > > > In your scenario, it should not matter where the monitor > >winding goes, as long as it is in the flux loop path and > >makes at least one turn around that. Not everything is > >an E-I core configuration either, as it appears you assume, > >with your 'center leg' remark. That is also whay I stated that > >the right way is to use a toroidal core to be certain one is > >observing ALL of the flux. > > Most power transformers wrap all the windings around the same core, > whatever its geometry. No, John. ALL transformers MUST have any winding it wishes to 'transform' voltage to on the same core. > So what the primary induces into the core > material, the secondaries see. There is ONE secondary. An additional winding would be a tertiary winding. > The flux that matters is enclosed by > all the windings [1]. No. ALL of the flux is 'enclosed' by the core. > So the voltage induced into an unloaded > secondary is a good indicator of the flux density that matters, namely > the flux that the windings actually induce and use; this concept has a > pretty high f-sub-D [2] factor. Yes, voltage rises as flux density increases in the open feedback winding. They are no longer open once they are being "read", however. It is not linear, and has to be modeled on a per transformer design basis. > > The question was whether loading the secondary of a power transformer > increases or decreases the core flux density. Which way do you vote? It decreases, but is tied to the envelope of the hysteresis loop it follows. Of course one wants to design for operation behind the permeability knee on the B-H curve..
From: John Larkin on 22 Jul 2010 12:25 On Thu, 22 Jul 2010 08:59:18 -0700 (PDT), Nunya <jack_shephard(a)cox.net> wrote: >On Jul 22, 8:28�am, John Larkin ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >> On Thu, 22 Jul 2010 07:36:16 -0700 (PDT), Nunya >> >> >> >> <jack_sheph...(a)cox.net> wrote: >> >On Jul 22, 7:23�am, John Larkin >> ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >> >> On Thu, 22 Jul 2010 06:58:04 -0700 (PDT), Nunya >> >> >> <jack_sheph...(a)cox.net> wrote: >> >> >On Jul 21, 8:47�pm, John Larkin >> >> ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote: >> >> >> On Wed, 21 Jul 2010 23:49:36 +0100, "Ian Iveson" >> >> >> >> <IanIveson.h...(a)blueyonder.co.uk> wrote: >> >> >> >John Larkin wrote: >> >> >> >> > >>> What, George? �Is it supposed to be untrue? �Explain >> >> >> >how. >> >> >> >> >>>> Tim >> >> >> >> >>>The magnetic flux density, B, is proportional to i for all >> >> >> >>>inductor/transformer configurations. >> >> >> >> >>>B = uH, u is approximately constant. >> >> >> >> >>>H = N*i/(2*pi*r) for a toroid. >> >> >> >> >>>At 0 current the magnetic flux density is 0. Simple as >> >> >> >>>that. >> >> >> >> >> What "I" are you talking about? A transformer has two >> >> >> >> windings. And >> >> >> >> each winding has resistance. >> >> >> >> >> Take a typical power transformer. Connect the primary to >> >> >> >> the AC line >> >> >> >> and leave the secondary open. Measure or compute the core >> >> >> >> flux >> >> >> >> density. Now short the secondary. At least before it >> >> >> >> catches fire, >> >> >> >> shorted-secondary flux density will be about half of the >> >> >> >> unloaded >> >> >> >> value. >> >> >> >> >> Loading the secondary of a transformer reduces flux >> >> >> >> density... just >> >> >> >> look at the directions of the primary and secondary >> >> >> >> winding currents. >> >> >> >> It's surprising how many people get this wrong, and think >> >> >> >> that a >> >> >> >> transformer "saturates" if you short its output. >> >> >> >> >Yes. The flux arising from secondary current is in >> >> >> >opposition to that arising from primary current. The two >> >> >> >currents are related, so the situation is more complicated >> >> >> >than George's Theory appreciates. >> >> >> >> An easy way to measure the flux density in a transformer core is with >> >> >> a probe coil, namely any old 3rd winding on the same core. >> >> >> >> So get a power transformer with one primary and two secondaries. >> >> >> Connect the primary to the AC line. Measure the voltage on secondary 1 >> >> >> and switch a load on/off secondary 2. If flux density increases with >> >> >> load, the voltage on the sec1 winding will go UP as sec2 is loaded. >> >> >> >> Of course, it doesn't anything that stupid. It goes down. >> >> >> >> At least AlwaysWrong has a new friend. >> >> >> >> John >> >> >> >induced voltage on the open winding is not a function of flux density, >> >> >it is a function of the amplitude of the excitation winding. >> >> >> The voltage induced is proportional to N * dP/dT, where P is total >> >> flux that the winding encloses, namely the center leg of the core. >> >> Since the number of turns and the frequency and the core dimensions >> >> don't change, the induced voltage tracks flux density. >> >> >> You are right if loading sec2 has zero affect on the voltage observed >> >> at sec1, ie, all transformers have perfect cross-regulation. >> >> >> Try it. >> >> >> John >> >> >Then the only proper way to get where you want to be is by >> >using a toroidal core. >> >> What I want? What's that? > > You seem to think that placing it elsewhere risks missing some of >the flux. >There is no way to miss ANY of it with a toroid core. Don't be silly. Even toroids have leakage flux. But that's entirely off the main point. >> >> >> > �In your scenario, it should not matter where the monitor >> >winding goes, as long as it is in the flux loop path and >> >makes at least one turn around that. �Not everything is >> >an E-I core configuration either, as it appears you assume, >> >with your 'center leg' remark. �That is also whay I stated that >> >the right way is to use a toroidal core to be certain one is >> >observing ALL of the flux. >> >> Most power transformers wrap all the windings around the same core, >> whatever its geometry. > > No, John. ALL transformers MUST have any winding it wishes to >'transform' voltage to on the same core. > >> So what the primary induces into the core >> material, the secondaries see. > > There is ONE secondary. An additional winding would be a >tertiary winding. > >> The flux that matters is enclosed by >> all the windings [1]. > > No. ALL of the flux is 'enclosed' by the core. > >> So the voltage induced into an unloaded >> secondary is a good indicator of the flux density that matters, namely >> the flux that the windings actually induce and use; this concept has a >> pretty high f-sub-D [2] factor. > > Yes, voltage rises as flux density increases in the open feedback >winding. >They are no longer open once they are being "read", however. It is not >linear, and has to be modeled on a per transformer design basis. I don't think my Fluke loads down a power transformer secondary many PPMs. As much as it might, it's still linear, so the voltage reading still tells me whether the flux increases or decreases as I do things. >> >> The question was whether loading the secondary of a power transformer >> increases or decreases the core flux density. Which way do you vote? > > It decreases, but is tied to the envelope of the hysteresis loop it >follows. Good. We have one vote for "decreases." Two more ways to address the question: 1. Load the secondary with various resistor values and graph the load curve. Does it look like a linear impedance being loaded, or does it have the sharp nonlinearities associated with saturation? 2. Load the secondary ditto and look at current waveforms. A little playing with a real transformer, a variac, some resistors, and a scope can be educational. John
From: The Phantom on 22 Jul 2010 16:03
On Thu, 22 Jul 2010 08:59:18 -0700 (PDT), Nunya <jack_shephard(a)cox.net> wrote: <snip> >> So what the primary induces into the core >> material, the secondaries see. > > There is ONE secondary. An additional winding would be a >tertiary winding. > Standard terminology is to refer to all the power-delivering windings as secondaries. This terminology is used by the major transformer manufacturers. See "multiple secondaries": http://www.stancor.com/jsp/products.jsp http://www.triadmagnetics.com/catalog_template.php?productCategoryId=63 |