Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Why is the Offline, Isolated, >100W, Single stage Flyback so badly villified?

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,980
Helped
586
Reputation
1,175
Reaction score
566
Trophy points
1,393
Location
cambridge
Activity points
78,635
Hi,
My consultancy is investigating a 24V, 5A Single FET Offline Flyback for a lighting application which will get used for literally only 4 hours per year. (it’s a LED light which lights the pathway to a Marquee used for Private School Pupils’ parents’ end of year Bash.)

Input voltage is the 390VDC output of a PFC stage.
Operation in BCM
Basic Schematic will be as attached (LTspice sim also attached)

Page 3 of the below (SLUP078) document states that single transistor flybacks need an extra winding so as to achieve non dissipative clamping of the leakage inductance spike. Surely this is not correct?
Also, this document, on page 3, also states that if an RCD clamp is used (with a single transistor flyback) then it will dissipate 15-20% of the output power. Surely this is wildly incorrect?

The attached is a simulation of a 24V, 5A flyback running off Vin=390VDC. The transformer coupling factor is 0.994. The total energy dissipated in the RCDZ clamp is just 2.4W. Is the simulation badly inaccurate about this?
If I reduce the transformer coupling to 0.99, and re-run the simulation I get 3.6W of dissipation in the RCDZ clamp components.
These are vey manageable levels of dissipation.

Also, Some years ago I did a PFC’d single stage , offline, isolated one FET flyback (with a PQ32/20 core) and at 240VAC and 100W input power I was getting significantly less than 740mW of dissipation in the RCD clamp resistor. Also, the efficiency for 100W input power was 90%.......slightly higher than the efficiency I was getting at 60W input power.

So why is the >100W Single FET offline, isolated flyback so badly vilified? Is the attached simulation badly inaccurate in relation to primary clamp losses?

150W Flyback Article (SLUP078)
 

Attachments

  • Flyback 24V 5A _SCHEM.pdf
    176.4 KB · Views: 2
  • Flyback 24V 5A _210506_LEAKAGEDO.zip
    2.8 KB · Views: 1

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,980
Helped
586
Reputation
1,175
Reaction score
566
Trophy points
1,393
Location
cambridge
Activity points
78,635
The writers of the above article in top post are villifying the single switch offline flyback...they wrongly accuse it of having a primary clamp which dissipates 15-20% of the output power.
Also, there are absolutely no application notes, across the world, from any semico, of an application getting a 120-200W Offline isolated Flyback applied to it.....when a 120-200W Flyback is often a great solution.

There are also loads of application notes which start off by saying, above 100W, an offline isolated converter cannot use a single switch flyback because of blah blah blah.........and its all nonsense...just further villification.

In this way, it suffers discrimination , and is villified.

I did once go to a big company who had had a 100W offline single stage, single switch flyback LED driver designed as a prototype. They had made the stupid mistake of using a primary clamp consisting of nothing more than an ultrafast diode into an SMA TVS (SMA!!).
….Some years later, I was describing this to a Leading Consultant SMPS designer over a beer at a meeting……expecting him to baulk at the SMA….but all he said was “what on earth were they using a Flyback for 100W for?”


100-200W Isolated offline Flybacks are actually very good…..usually we are all told (by application notes) that we must use a two Transistor Forward Converter or LLC for this power range....
Taking the 2TFC….its transformer may be smaller than a Flyback, but when you add in the output choke, the double diode output, the double fet primary…..not to mention that all those 4 semiconductors need heatsinking !....then on top of that lot you need the upper gate drive for the top fet in the primary.
…you can see that your PCB space used is actually less with a flyback…..and your BOM is actually much smaller and cheaper with a flyback
 
Last edited:

scopeprobe

Junior Member level 3
Joined
Sep 17, 2020
Messages
31
Helped
14
Reputation
28
Reaction score
19
Trophy points
8
Location
Uk
Activity points
372
The writers of the above article in top post are villifying the single switch offline flyback...they wrongly accuse it of having a primary clamp which dissipates 15-20% of the output power.
Also, there are absolutely no application notes, across the world, from any semico, of an application getting a 100-200W Offline isolated Flyback applied to it.....when a 100-200W Flyback is often a great solution.

There are also loads of application notes which start off by saying, above 100W, an offline isolated converter cannot use a single switch flyback because of blah blah blah.........and its all nonsense...just further villification.

In this way, it suffers discrimination , and is villified.

I did once go to a big company who had had a 100W offline single stage, single switch flyback LED driver designed as a prototype. They had made the stupid mistake of using a primary clamp consisting of nothing more than an ultrafast diode into an SMA TVS (SMA!!).
….Some years later, I was describing this to a Leading Consultant SMPS designer over a beer at a meeting……expecting him to baulk at the SMA….but all he said was “what on earth were they using a Flyback for 100W for?”

I wouldn't say its discrimination, In general at higher power levels other topologies just become more attractive due to the fact the currents in a single stage flyback in DCM, TM or BCM become more difficult to efficiently manage both Efficiency and EMI, think switch ratings, transformer quality, circuit damping, EMI filtering etc. To manage these issues costs money and as costs start to rise other slightly more expensive topologies with better modes of control (don't generate the same level current or ripple) become more viable. This is why manufacturer have IC's aimed to specific power levels as they are best practise. Thats not to say you can't use a flyback but in reality there becomes a point a flyback isn't viable as a commercial solution.
 
Last edited:

mtwieg

Advanced Member level 5
Joined
Jan 20, 2011
Messages
3,564
Helped
1,262
Reputation
2,530
Reaction score
1,253
Trophy points
1,393
Activity points
26,795
They're not "vilifying" it. They're just making a case for why the two switch design is better for higher power levels.
1. EMC restrictions are the same regardless of the power level. A 500W supply will produce much higher EMC (10~20dB) higher EMC than an equivalent 5W supply. So at some point as power increases it's worthwhile to switch to another topology with lower EMC.
2. For very low power levels, the extra losses in a single switch flyback don't significantly impact the overall BOM or size of the design. As power level goes up, then you'll find that moving to a two-switch design can save cost and size. That threshold is, allegedly, around ~100W
 

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,980
Helped
586
Reputation
1,175
Reaction score
566
Trophy points
1,393
Location
cambridge
Activity points
78,635
As power level goes up, then you'll find that moving to a two-switch design can save cost and size. That threshold is, allegedly, around ~100W
Thanks, yes, but i believe the alleged level is closer to 200W...as this 200W offline SMPS flyback shows...

I believe cost is the main driver, and the reduction in parts count that a flyback gives you means its nearer 200W for offline , unless output current is really high
 

scopeprobe

Junior Member level 3
Joined
Sep 17, 2020
Messages
31
Helped
14
Reputation
28
Reaction score
19
Trophy points
8
Location
Uk
Activity points
372
Thanks, yes, but i believe the alleged level is closer to 200W...as this 200W offline SMPS flyback shows...

I believe cost is the main driver, and the reduction in parts count that a flyback gives you means its nearer 200W for offline , unless output current is really high
I think you’ll find it will depend on the target market and application. Areas which demand low noise such as aerospace, medical and military may typically choose a low noise topology from the start as it will give an easier life further down the line. Theres no precise line in the sand. As the design engineer its your prerogative to chose the appropriate topology for your application based on your experience (good and bad). Unfortunately you learn from failures and that will set your boundaries.
 

Easy peasy

Advanced Member level 5
Joined
Aug 15, 2015
Messages
3,043
Helped
1,116
Reputation
2,232
Reaction score
1,155
Trophy points
113
Location
Melbourne
Activity points
16,831
There is a horses for courses argument, even PI appear to prefer forward converter above 150 watts, mainly due to rms currents in the pri switch.

Using just an RFI /EMI argument though - it is hard to beat the single or 2 switch flyback up to 250 watts in QR - valley switched - mode, or course this pre-supposes and engineer who can design a transformer with low enough lakage and will meet mains safety regulations.

A 2 phase or even 3 phase flyback system ( QR ) has the inherent advantages of simple control, low RFI, and reduced input current ripple, reduced Vripple on o/p, and only 2 or 3 major magnetics which are all the same.
 

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,980
Helped
586
Reputation
1,175
Reaction score
566
Trophy points
1,393
Location
cambridge
Activity points
78,635
Using just an RFI /EMI argument though - it is hard to beat the single or 2 switch flyback up to 250 watts in QR - valley switched - mode
...Thanks, yes, and its cheaper. My colleague runs a consultancy, and he says the Semico's are totally out of touch with reality.....they are peddling there 150W 2TFC's and 150W LLC's when real companies want "cheap as chips"....and thus Flyback.......not efficiency , and pleasing Greta Thunberg is way way down their priority list.
They dont even mind dirty great heatsinks.......dirty great heatsinks are cheap to design...and often cheap to buy from eg China...what they can't stand Is the multiple heatsinks that you get with a 2TFC or LLC....and all the cost it means in assembling them to the FETs/Diodes.

EMC is up there though, but as discussed, here, the QR flyback is good on EMC.

this pre-supposes and engineer who can design a transformer with low enough lakage and will meet mains safety regulations.
...Thanks, do you mean sometimes triple interleaving is needed when above 100W?...to really cut down the leakage inductance?

EG Pri sec Pri Sec Pri ?
 

Easy peasy

Advanced Member level 5
Joined
Aug 15, 2015
Messages
3,043
Helped
1,116
Reputation
2,232
Reaction score
1,155
Trophy points
113
Location
Melbourne
Activity points
16,831
Do as much interleaving as you can, use LV aux power circuits as shields, add shield wdgs as necessary, try and keep it neat and easy for the winder.
 

scopeprobe

Junior Member level 3
Joined
Sep 17, 2020
Messages
31
Helped
14
Reputation
28
Reaction score
19
Trophy points
8
Location
Uk
Activity points
372
Thanks, yes, but i believe the alleged level is closer to 200W...as this 200W offline SMPS flyback shows...

I believe cost is the main driver, and the reduction in parts count that a flyback gives you means its nearer 200W for offline , unless output current is really high
If you can get away with direct offline, then i agree flybacks are good however meeting harmonic distortion (sub 5% THD) requirements will be a challenge without high turns ratio of the transformer which can result in high secondary currents. If you have to go down the route of a boost preregulator the differences between a flyback and forward converter becomes much less so this is also a consideration. It is possible to shape the voltage reference of a PFC controller to give better harmonic distortion with a lower turns ratio transformer but it can be a bit of a pain to setup.
 
Last edited:

Easy peasy

Advanced Member level 5
Joined
Aug 15, 2015
Messages
3,043
Helped
1,116
Reputation
2,232
Reaction score
1,155
Trophy points
113
Location
Melbourne
Activity points
16,831
you don't need to change the turns ratio from optimal for a flyback - just modulate the gate drive to achieve in input current closer to a sine wave - the Vout of the flyback Tx will always be the Vout on the o/p cap ... if you want higher power factor.
 

scopeprobe

Junior Member level 3
Joined
Sep 17, 2020
Messages
31
Helped
14
Reputation
28
Reaction score
19
Trophy points
8
Location
Uk
Activity points
372
you don't need to change the turns ratio from optimal for a flyback - just modulate the gate drive to achieve in input current closer to a sine wave - the Vout of the flyback Tx will always be the Vout on the o/p cap ... if you want higher power factor.
I agree that this only achieves good power factor but at 200W but won't harmonics be a requirement? I work in Aerospace where harmonics content comes in above 30W. A flyback will likely fail power quality tests on THD if they are a consideration. If he needs good harmonics and good power factor you have to push more of the energy transfer to the peak of the havesine by either using a higher turns ratio or shaping the line voltage reference to modulate less at low line than would normally be necessary. The shaping via modulation can only be done with IC's that uses the line voltage reference.
 

scopeprobe

Junior Member level 3
Joined
Sep 17, 2020
Messages
31
Helped
14
Reputation
28
Reaction score
19
Trophy points
8
Location
Uk
Activity points
372
all your assumptions are non sequitur, also 0.999 PF = very low THD ...
Ok, To clarify i think what i'm stating is you won't get a PF of 0.999 with a standard offline flyback without doing one of the following:

1. Artificially Modifying the Havesine voltage reference waveform to reduce the curent transfer at the low portion of the haversine and pass more energy at the peak voltage
2. Keeping the flyback voltage above the peak line voltage (high turns ratio)

Is this not correct?

A quick search online found this which kind of explains what i'm saying

https://patentimages.storage.googleapis.com/06/60/bd/58cffb1d1cf9d0/US8471488.pdf

I'm happy to be corrected if thats not the case but please explain how.
 

Easy peasy

Advanced Member level 5
Joined
Aug 15, 2015
Messages
3,043
Helped
1,116
Reputation
2,232
Reaction score
1,155
Trophy points
113
Location
Melbourne
Activity points
16,831
@scopeprobe, actually - if you operate a flyback in DCM, then, with a slow feedback volt loop ( i.e. almost fixed pwm ), you will get very good power factor - without any other fancy control.
 

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,980
Helped
586
Reputation
1,175
Reaction score
566
Trophy points
1,393
Location
cambridge
Activity points
78,635
if you operate a flyback in DCM, then, with a slow feedback volt loop ( i.e. almost fixed pwm ), you will get very good power factor - without any other fancy control.if you operate a flyback in DCM, then, with a slow feedback volt loop ( i.e. almost fixed pwm ), you will get very good power factor - without any other fancy control.
Thanks thats very interesting, and consider without even the divider from the half-sine DC bus to "guide" it...would that give good enough harmonics performance to get through EN61000-3-2 etc?
 

LaTeX Commands Quick-Menu:

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top