Power factor correction
in [Design]
|
From: Jan Panteltje on 4 Jan 2010 07:29 On a sunny day (Mon, 04 Jan 2010 07:53:30 +0200) it happened Paul Keinanen <keinanen(a)sci.fi> wrote in <q4v2k5hqkr88dmmja3fssrc7o7cdrnpipb(a)4ax.com>: >On Tue, 29 Dec 2009 13:44:36 GMT, Jan Panteltje ><pNaonStpealmtje(a)yahoo.com> wrote: > >>On a sunny day (Tue, 29 Dec 2009 07:17:31 -0600) it happened "Tim Williams" >><tmoranwms(a)charter.net> wrote in <Yhn_m.18015$wC3.5829(a)newsfe07.iad>: >> >>>PFC is a flyback thing. But flybacks suck over 100W, where forward looks >>>better. But forward sucks over wide duty cycle ranges. What do they do, >>>just bigger flybacks? Nothing at all? >>> >>>Tim >>> >>>-- >>>Deep Friar: a very philosophical monk. >>>Website: http://webpages.charter.net/dawill/tmoranwms >> >>For much higher power 3 phase rectifiers are used. >>That leaves only a small ripple. >>I am not sure if you need PFC in such a case. > >The ordinary 3 phase 6 pulse rectifier is a quite nasty polluter of >the mains. > >To reduce the harmonics, 12 pulse rectifiers have been used, with one >set of rectifiers connected to the wye windings and the other set from >separate delta windings. For high power systems, you still are going >to need a medium/low voltage transformer on site, so it is not a big >deal having separate wye and delta secondary windings on that >transformer. > >If you want to get away with the heavy 50 Hz transformer, there seems >to be various more or less patented 3 phase systems using flyback etc. >systems to reduce the PFC, but these seems to be hugely complex and >have special floating requirements for the DC side. > >If floating output or voltages other than those obtainable by simple >rectifiers are required, it might be easier to simply get three single >phase SMPS with PFC at the input and connect the DC outputs in >parallel. > >If the input voltage ratings permit, instead of wye, connect the >inputs into delta to avoid any mains neutral polluting and also allow >higher voltage and hence higher power (1.7x) for a specific >semiconductor amperage. Yes, the ones I have dealt with also had a huge heavy inductor beteen the rectifiers and the filtercaps. L ========= -------------------------^^^^^^^^^---------------- + | | | | --- --- --- | / \ / \ / \ | --- --- --- | | | | | --- | | | | | | | === --- ) ------| | | | | --- C --- ) ----- ) ----- | | | | | | --- --- --- | / \ / \ / \ | --- --- --- | | | | | -------------------------------------------------- - The inductor smoothes the main current, and reduces capacitor ripple current. For something like 4kVA and up.
From: MooseFET on 4 Jan 2010 09:26 On Jan 4, 4:29 am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: > On a sunny day (Mon, 04 Jan 2010 07:53:30 +0200) it happened Paul Keinanen > <keina...(a)sci.fi> wrote in <q4v2k5hqkr88dmmja3fssrc7o7cdrnp...(a)4ax.com>: > > > > >On Tue, 29 Dec 2009 13:44:36 GMT, Jan Panteltje > ><pNaonStpealm...(a)yahoo.com> wrote: > > >>On a sunny day (Tue, 29 Dec 2009 07:17:31 -0600) it happened "Tim Williams" > >><tmoran...(a)charter.net> wrote in <Yhn_m.18015$wC3.5...(a)newsfe07.iad>: > > >>>PFC is a flyback thing. But flybacks suck over 100W, where forward looks > >>>better. But forward sucks over wide duty cycle ranges. What do they do, > >>>just bigger flybacks? Nothing at all? > > >>>Tim > > >>>-- > >>>Deep Friar: a very philosophical monk. > >>>Website:http://webpages.charter.net/dawill/tmoranwms > > >>For much higher power 3 phase rectifiers are used. > >>That leaves only a small ripple. > >>I am not sure if you need PFC in such a case. > > >The ordinary 3 phase 6 pulse rectifier is a quite nasty polluter of > >the mains. > > >To reduce the harmonics, 12 pulse rectifiers have been used, with one > >set of rectifiers connected to the wye windings and the other set from > >separate delta windings. For high power systems, you still are going > >to need a medium/low voltage transformer on site, so it is not a big > >deal having separate wye and delta secondary windings on that > >transformer. > > >If you want to get away with the heavy 50 Hz transformer, there seems > >to be various more or less patented 3 phase systems using flyback etc. > >systems to reduce the PFC, but these seems to be hugely complex and > >have special floating requirements for the DC side. > > >If floating output or voltages other than those obtainable by simple > >rectifiers are required, it might be easier to simply get three single > >phase SMPS with PFC at the input and connect the DC outputs in > >parallel. > > >If the input voltage ratings permit, instead of wye, connect the > >inputs into delta to avoid any mains neutral polluting and also allow > >higher voltage and hence higher power (1.7x) for a specific > >semiconductor amperage. > > Yes, the ones I have dealt with also had a huge heavy inductor beteen the rectifiers and the filtercaps. > L > ========= > -------------------------^^^^^^^^^---------------- + > | | | | > --- --- --- | > / \ / \ / \ | > --- --- --- | > | | | | > --- | | | | > | | | === > --- ) ------| | > | | | --- C > --- ) ----- ) ----- | | > | | | | > --- --- --- | > / \ / \ / \ | > --- --- --- | > | | | | > -------------------------------------------------- - > The inductor smoothes the main current, and reduces capacitor ripple current. > For something like 4kVA and up. With the coil, the current is near constant. This makes the harmonic content less but still significant. Adding a moderate sized transformer at each phase and doubling the number of diodes will lower the harmonics quite a bit. Basically it is 3 of these ------------------ To bridge ( A---------- ( ------------------ To bridge B----------- ) ) ) C----------- Since the voltage between B and C is at 90 degrees to the voltage from A to ground, the voltage on the secondary only needs to be the tan() of the angle you are shifting the phase by. tan(15)=0.27
From: Jan Panteltje on 4 Jan 2010 10:35 On a sunny day (Mon, 4 Jan 2010 06:26:08 -0800 (PST)) it happened MooseFET <kensmith(a)rahul.net> wrote in <2562a20f-e214-4701-8f38-50366424ad9a(a)a15g2000yqm.googlegroups.com>: >On Jan 4, 4:29�am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: >> On a sunny day (Mon, 04 Jan 2010 07:53:30 +0200) it happened Paul Keinane= >n >> <keina...(a)sci.fi> wrote in <q4v2k5hqkr88dmmja3fssrc7o7cdrnp...(a)4ax.com>: >> >> >> >> >On Tue, 29 Dec 2009 13:44:36 GMT, Jan Panteltje >> ><pNaonStpealm...(a)yahoo.com> wrote: >> >> >>On a sunny day (Tue, 29 Dec 2009 07:17:31 -0600) it happened "Tim Willi= >ams" >> >><tmoran...(a)charter.net> wrote in <Yhn_m.18015$wC3.5...(a)newsfe07.iad>: >> >> >>>PFC is a flyback thing. �But flybacks suck over 100W, where forward = >looks >> >>>better. �But forward sucks over wide duty cycle ranges. �What do t= >hey do, >> >>>just bigger flybacks? �Nothing at all? >> >> >>>Tim >> >> >>>-- >> >>>Deep Friar: a very philosophical monk. >> >>>Website:http://webpages.charter.net/dawill/tmoranwms >> >> >>For much higher power 3 phase rectifiers are used. >> >>That leaves only a small ripple. >> >>I am not sure if you need PFC in such a case. >> >> >The ordinary 3 phase 6 pulse rectifier is a quite nasty polluter of >> >the mains. >> >> >To reduce the harmonics, 12 pulse rectifiers have been used, with one >> >set of rectifiers connected to the wye windings and the other set from >> >separate delta windings. For high power systems, you still are going >> >to need a medium/low voltage transformer on site, so it is not a big >> >deal having separate wye and delta secondary windings on that >> >transformer. >> >> >If you want to get away with the heavy 50 Hz transformer, there seems >> >to be various more or less patented 3 phase systems using flyback etc. >> >systems to reduce the PFC, but these seems to be hugely complex and >> >have special floating requirements for the DC side. >> >> >If floating output or voltages other than those obtainable by simple >> >rectifiers are required, it might be easier to simply get three single >> >phase SMPS with PFC at the input and connect the DC outputs in >> >parallel. >> >> >If the input voltage ratings permit, instead of wye, connect the >> >inputs into delta to avoid any mains neutral polluting and also allow >> >higher voltage and hence higher power (1.7x) for a specific >> >semiconductor amperage. >> >> Yes, the ones I have dealt with also had a huge heavy inductor beteen the= > rectifiers and the filtercaps. >> � � � � � � � � � � � � � � � � �L >> � � � � � � � � � � � � � � � ===== >===== >> � � �-------------------------^^^^^^^^^---------------- + >> � � | � � � | � � � | � � � � � � � � �= > � � � � � | >> � �--- � � --- � � --- � � � � � � � � � = >� � � � �| >> � �/ \ � � / \ � � / \ � � � � � � � � � = >� � � � �| >> � �--- � � --- � � --- � � � � � � � � � = >� � � � �| >> � � | � � � | � � � | � � � � � � � � �= > � � � � � | >> --- | � � � | � � � | � � � � � � � � � �= > � � � � | >> � � | � � � | � � � | � � � � � � � � �= > � � � � �=== >> --- ) ------| � � � | >> � � | � � � | � � � | � � � � � � � � �= > � � � � �--- C >> --- ) ----- ) ----- | � � � � � � � � � � � � �= > � | >> � � | � � � | � � � | � � � � � � � � �= > � � � � � | >> � �--- � � --- � � --- � � � � � � � � � = >� � � � �| >> � �/ \ � � / \ � � / \ � � � � � � � � � = >� � � � �| >> � �--- � � --- � � --- � � � � � � � � � = >� � � � �| >> � � | � � � | � � � | � � � � � � � � �= > � � � � � | >> � � �-------------------------------------------------- - >> The inductor smoothes the main current, and reduces capacitor ripple curr= >ent. >> For something like 4kVA and up. > >With the coil, the current is near constant. This makes the harmonic >content less >but still significant. Adding a moderate sized transformer at each >phase and doubling >the number of diodes will lower the harmonics quite a bit. > >Basically it is 3 of these > > ------------------ To bridge > ( >A---------- > ( > ------------------ To bridge > >B----------- > ) > ) > ) >C----------- > > >Since the voltage between B and C is at 90 degrees to the voltage from >A >to ground, the voltage on the secondary only needs to be the tan() of >the >angle you are shifting the phase by. > >tan(15)=0.27 An interesting solution, never seen that before, maybe I am too old:-)
From: Tim Williams on 4 Jan 2010 11:16 "MooseFET" <kensmith(a)rahul.net> wrote in message news:2562a20f-e214-4701-8f38-50366424ad9a(a)a15g2000yqm.googlegroups.com... > With the coil, the current is near constant. This makes the harmonic > content less > but still significant. Adding a moderate sized transformer at each > phase and doubling > the number of diodes will lower the harmonics quite a bit. > > Basically it is 3 of these > ... And of course, since you have two phases on two wires (let the other be common), you have a linearly independent basis (not orthonormal, but so what) from which you can span the entire vector space of voltage and phase. So you could, for example, produce a 5-phase system with 10 pulses, or 120 if you wanted. You spend a lot on transformers, though. What does that do for current, anyway? Current through each phase is essentially a parabolic pulse. All those added up, in the ratios from which they are generated, should distribute to a fairly constant current, shouldn't they? Well, the sum of (three phase) currents is always zero, but the sum of magnitudes isn't: that must oscillate at the 6th harmonic. So it should be that, in the same way as a choke input filter causes relatively more current draw on the flanks of the sine wave than overall, this arrangement also causes more current draw on the flanks, resulting in an inversely phased 6th harmonic. The curious part is, this nonlinear conclusion was based on the linear construction of vectors: of course the diodes, being nonlinear elements, are the reason, but the voltages don't seem like they should do that. Tim -- Deep Friar: a very philosophical monk. Website: http://webpages.charter.net/dawill/tmoranwms
From: MooseFET on 4 Jan 2010 21:19
On Jan 4, 7:35 am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: > On a sunny day (Mon, 4 Jan 2010 06:26:08 -0800 (PST)) it happened MooseFET > <kensm...(a)rahul.net> wrote in > <2562a20f-e214-4701-8f38-50366424a...(a)a15g2000yqm.googlegroups.com>: > > > > >On Jan 4, 4:29 am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote: > >> On a sunny day (Mon, 04 Jan 2010 07:53:30 +0200) it happened Paul Keinane= > >n > >> <keina...(a)sci.fi> wrote in <q4v2k5hqkr88dmmja3fssrc7o7cdrnp...(a)4ax.com>: > > >> >On Tue, 29 Dec 2009 13:44:36 GMT, Jan Panteltje > >> ><pNaonStpealm...(a)yahoo.com> wrote: > > >> >>On a sunny day (Tue, 29 Dec 2009 07:17:31 -0600) it happened "Tim Willi= > >ams" > >> >><tmoran...(a)charter.net> wrote in <Yhn_m.18015$wC3.5...(a)newsfe07.iad>: > > >> >>>PFC is a flyback thing. But flybacks suck over 100W, where forward = > >looks > >> >>>better. But forward sucks over wide duty cycle ranges. What do t= > >hey do, > >> >>>just bigger flybacks? Nothing at all? > > >> >>>Tim > > >> >>>-- > >> >>>Deep Friar: a very philosophical monk. > >> >>>Website:http://webpages.charter.net/dawill/tmoranwms > > >> >>For much higher power 3 phase rectifiers are used. > >> >>That leaves only a small ripple. > >> >>I am not sure if you need PFC in such a case. > > >> >The ordinary 3 phase 6 pulse rectifier is a quite nasty polluter of > >> >the mains. > > >> >To reduce the harmonics, 12 pulse rectifiers have been used, with one > >> >set of rectifiers connected to the wye windings and the other set from > >> >separate delta windings. For high power systems, you still are going > >> >to need a medium/low voltage transformer on site, so it is not a big > >> >deal having separate wye and delta secondary windings on that > >> >transformer. > > >> >If you want to get away with the heavy 50 Hz transformer, there seems > >> >to be various more or less patented 3 phase systems using flyback etc.. > >> >systems to reduce the PFC, but these seems to be hugely complex and > >> >have special floating requirements for the DC side. > > >> >If floating output or voltages other than those obtainable by simple > >> >rectifiers are required, it might be easier to simply get three single > >> >phase SMPS with PFC at the input and connect the DC outputs in > >> >parallel. > > >> >If the input voltage ratings permit, instead of wye, connect the > >> >inputs into delta to avoid any mains neutral polluting and also allow > >> >higher voltage and hence higher power (1.7x) for a specific > >> >semiconductor amperage. > > >> Yes, the ones I have dealt with also had a huge heavy inductor beteen the= > > rectifiers and the filtercaps. > >> L > >> ===== > >===== > >> -------------------------^^^^^^^^^---------------- + > >> | | | = > > | > >> --- --- --- = > > | > >> / \ / \ / \ = > > | > >> --- --- --- = > > | > >> | | | = > > | > >> --- | | | = > > | > >> | | | = > > === > >> --- ) ------| | > >> | | | = > > --- C > >> --- ) ----- ) ----- | = > > | > >> | | | = > > | > >> --- --- --- = > > | > >> / \ / \ / \ = > > | > >> --- --- --- = > > | > >> | | | = > > | > >> -------------------------------------------------- - > >> The inductor smoothes the main current, and reduces capacitor ripple curr= > >ent. > >> For something like 4kVA and up. > > >With the coil, the current is near constant. This makes the harmonic > >content less > >but still significant. Adding a moderate sized transformer at each > >phase and doubling > >the number of diodes will lower the harmonics quite a bit. > > >Basically it is 3 of these > > > ------------------ To bridge > > ( > >A---------- > > ( > > ------------------ To bridge > > >B----------- > > ) > > ) > > ) > >C----------- > > >Since the voltage between B and C is at 90 degrees to the voltage from > >A > >to ground, the voltage on the secondary only needs to be the tan() of > >the > >angle you are shifting the phase by. > > >tan(15)=0.27 > > An interesting solution, never seen that before, maybe I am too old:-) .... Or maybe too young. Back when rectifiers couldn't handle the power, you needed more diodes and a way to share the current so doing this didn't add parts. |