So I built this board for my fire prevention equipment customer. (An excellent outfit BTW: Argus Fire Control.) It needed two isolated 5-V supplies, one for the analogue stuff and one for RS-485. A few months ago we were discussing those supplies in this very boutique, in the "Coupled Inductors--how coupled is coupled?" thread. I wound up with a flybuck design using an LM3103 sync buck chip. Using a simple fixed duty-cycle model with realistic resistances for the switches and inductors, it simulated very nicely. However this switcher chip doesn't work very well at all in the application, apparently on account of its DCM mode, which makes it stutter unless there's a serious load on the non-isolated side (as in a normal buck). When it stutters, the feedback voltage stays stable, but the isolated outputs collapse, of course--I need the output FETS to keep on switching, and they don't. There's no apparent way to turn this feature off. Looks like I'll have to use something simpler. 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
Flybuck converter funnies
Started by ●April 26, 2016
Reply by ●April 26, 20162016-04-26
On Tue, 26 Apr 2016 15:11:01 -0400, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:>So I built this board for my fire prevention equipment customer. (An >excellent outfit BTW: Argus Fire Control.) It needed two isolated 5-V >supplies, one for the analogue stuff and one for RS-485. > >A few months ago we were discussing those supplies in this very >boutique, in the "Coupled Inductors--how coupled is coupled?" thread. I >wound up with a flybuck design using an LM3103 sync buck chip. > >Using a simple fixed duty-cycle model with realistic resistances for the >switches and inductors, it simulated very nicely. However this switcher >chip doesn't work very well at all in the application, apparently on >account of its DCM mode, which makes it stutter unless there's a serious >load on the non-isolated side (as in a normal buck). > >When it stutters, the feedback voltage stays stable, but the isolated >outputs collapse, of course--I need the output FETS to keep on >switching, and they don't. There's no apparent way to turn this feature >off. > >Looks like I'll have to use something simpler. > >Cheers > >Phil HobbsYeah, the block diagrams of modern switcher controller chips look like the aerial view of Baltimore (with the burning bits too!) -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by ●April 26, 20162016-04-26
On 04/26/2016 03:37 PM, John Larkin wrote:> On Tue, 26 Apr 2016 15:11:01 -0400, Phil Hobbs > <pcdhSpamMeSenseless@electrooptical.net> wrote: > >> So I built this board for my fire prevention equipment customer. (An >> excellent outfit BTW: Argus Fire Control.) It needed two isolated 5-V >> supplies, one for the analogue stuff and one for RS-485. >> >> A few months ago we were discussing those supplies in this very >> boutique, in the "Coupled Inductors--how coupled is coupled?" thread. I >> wound up with a flybuck design using an LM3103 sync buck chip. >> >> Using a simple fixed duty-cycle model with realistic resistances for the >> switches and inductors, it simulated very nicely. However this switcher >> chip doesn't work very well at all in the application, apparently on >> account of its DCM mode, which makes it stutter unless there's a serious >> load on the non-isolated side (as in a normal buck). >> >> When it stutters, the feedback voltage stays stable, but the isolated >> outputs collapse, of course--I need the output FETS to keep on >> switching, and they don't. There's no apparent way to turn this feature >> off. >> >> Looks like I'll have to use something simpler. >> >> Cheers >> >> Phil Hobbs > > Yeah, the block diagrams of modern switcher controller chips look like > the aerial view of Baltimore (with the burning bits too!)The leading contenders so far are the Richtek RT7272 and RT2862A. 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
Reply by ●April 26, 20162016-04-26
On Tue, 26 Apr 2016 15:11:01 -0400, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> Gave us:>So I built this board for my fire prevention equipment customer. (An >excellent outfit BTW: Argus Fire Control.) It needed two isolated 5-V >supplies, one for the analogue stuff and one for RS-485. > >A few months ago we were discussing those supplies in this very >boutique, in the "Coupled Inductors--how coupled is coupled?" thread. I >wound up with a flybuck design using an LM3103 sync buck chip. > >Using a simple fixed duty-cycle model with realistic resistances for the >switches and inductors, it simulated very nicely. However this switcher >chip doesn't work very well at all in the application, apparently on >account of its DCM mode, which makes it stutter unless there's a serious >load on the non-isolated side (as in a normal buck). > >When it stutters, the feedback voltage stays stable, but the isolated >outputs collapse, of course--I need the output FETS to keep on >switching, and they don't. There's no apparent way to turn this feature >off. > >Looks like I'll have to use something simpler. > >Cheers > >Phil HobbsInstead of StarBuck Flying... It could be FlyBuck Smoking!
Reply by ●April 27, 20162016-04-27
>Instead of StarBuck Flying...>It could be FlyBuck Smoking!Yes, it's more of a Skybuck converter at the moment. (The poor guy has had a lot of broken hardware over the years.) Cheers Phil Hobbs
Reply by ●April 27, 20162016-04-27
On Wed, 27 Apr 2016 04:41:44 -0700 (PDT), Phil Hobbs <pcdhobbs@gmail.com> Gave us:>>Instead of StarBuck Flying... > >>It could be FlyBuck Smoking! > >Yes, it's more of a Skybuck converter at the moment. (The poor guy has had a lot of broken hardware over the years.) > >Cheers > >Phil HobbsBetween his concept of ESD and his obsession with taking apart live gear, it is no wonder. Shame he hasn't touched the AC feed terminals what with their double the voltage of here thing. One in each hand. Just a shame. :-)
Reply by ●May 4, 20162016-05-04
On 04/26/2016 03:54 PM, Phil Hobbs wrote:> On 04/26/2016 03:37 PM, John Larkin wrote: >> On Tue, 26 Apr 2016 15:11:01 -0400, Phil Hobbs >> <pcdhSpamMeSenseless@electrooptical.net> wrote: >> >>> So I built this board for my fire prevention equipment customer. (An >>> excellent outfit BTW: Argus Fire Control.) It needed two isolated 5-V >>> supplies, one for the analogue stuff and one for RS-485. >>> >>> A few months ago we were discussing those supplies in this very >>> boutique, in the "Coupled Inductors--how coupled is coupled?" thread. I >>> wound up with a flybuck design using an LM3103 sync buck chip. >>> >>> Using a simple fixed duty-cycle model with realistic resistances for the >>> switches and inductors, it simulated very nicely. However this switcher >>> chip doesn't work very well at all in the application, apparently on >>> account of its DCM mode, which makes it stutter unless there's a serious >>> load on the non-isolated side (as in a normal buck). >>> >>> When it stutters, the feedback voltage stays stable, but the isolated >>> outputs collapse, of course--I need the output FETS to keep on >>> switching, and they don't. There's no apparent way to turn this feature >>> off. >>> >>> Looks like I'll have to use something simpler. >>> >>> Cheers >>> >>> Phil Hobbs >> >> Yeah, the block diagrams of modern switcher controller chips look like >> the aerial view of Baltimore (with the burning bits too!) > > The leading contenders so far are the Richtek RT7272 and RT2862A.Well, I tried, and both stutter at light loads too. The RT2862 datasheet makes no such claim, but does it anyway. The plus side is that when I load down the primary, they produce heaps of isolated power, so the magnetics and so forth are working fine. Anybody got a favourite sync buck that'll handle at least 36V input and 400 mA, and keeps on PWMing at light loads? 'Cuz otherwise I'm going to have to do it the fully-manual way, with a bridge driver and so on (yuck). &*$%%!! 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
Reply by ●May 4, 20162016-05-04
"Phil Hobbs" <pcdhSpamMeSenseless@electrooptical.net> wrote in message news:Nv2dnSAO7JJ4k7fKnZ2dnUU7-bHNnZ2d@supernews.com...> Anybody got a favourite sync buck that'll handle at least 36V input and > 400 mA, and keeps on PWMing at light loads?Was about to say TPS54233 or something like that, but I immediately realized those miss exactly those specs (30V and pulse skipping)... LTC3810 is the only controller I know offhand, but that's rather a lot of brain (and additional circuit) for present levels. I don't know if there's a smaller, cheaper, integrated version out there, but hopefully it's a hint at least. Suppose I might suggest dropping the sync requirement, since the max voltage is high enough not to matter much. But you'd know better than I.. Tim -- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
Reply by ●May 4, 20162016-05-04
On Wed, 4 May 2016 10:52:09 -0500, "Tim Williams" <tiwill@seventransistorlabs.com> wrote:>"Phil Hobbs" <pcdhSpamMeSenseless@electrooptical.net> wrote in message >news:Nv2dnSAO7JJ4k7fKnZ2dnUU7-bHNnZ2d@supernews.com... >> Anybody got a favourite sync buck that'll handle at least 36V input and >> 400 mA, and keeps on PWMing at light loads? > >Was about to say TPS54233 or something like that, but I immediately realized >those miss exactly those specs (30V and pulse skipping)... > >LTC3810 is the only controller I know offhand, but that's rather a lot of >brain (and additional circuit) for present levels. I don't know if there's >a smaller, cheaper, integrated version out there, but hopefully it's a hint >at least. > >Suppose I might suggest dropping the sync requirement, since the max voltage >is high enough not to matter much. But you'd know better than I.. > >TimThe flybuck architecture (which I understand to be a buck with a second winding, to make a negative voltage) needs a pushpull drive to work over the full load range. I've used a diode-capacitor voltage doubler off the switcher node to make a negative supply, similar idea. A catch-diode type buck switcher won't make the negative voltage if the positive supply is unloaded, or if the switcher chip decides to stutter. These days, I mostly buy cheap SIP or DIP converters to make a negative or bipolar power supply. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by ●May 4, 20162016-05-04
"John Larkin" <jjlarkin@highlandtechnology.com> wrote in message news:1u6kib9i52tcvl3skum47ocdnct9tnkqbm@4ax.com...> The flybuck architecture (which I understand to be a buck with a > second winding, to make a negative voltage) needs a pushpull drive to > work over the full load range.Not strictly, but you do need more load on the normal output. Otherwise, the "flyback" doesn't fly all the way back because it's being clamped by the negative output's load. Tim -- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
