Quick question about inverter TX order wrt mains

[tim9000]

Hi, I should know better, but I'm having a mental blank. I was just wondering, if you're doing something using non-linear components on the mains, could be anything, maybe bucking to make a DC supply, whatever.
Is it common practice to use an isolation transformer between mains and the rectifier?
Also, if you do this, does this completely mitigate clipping and the high voltages induced from the snapping off and leakage inductance?
(P.S: That was from memory, I can't actually remember what the issues of using non-linear components are, feel free to remind me)
Thanks

 

[davenn]

non-linear, you mean switch mode PSU ?

A transformer is still used to give isolation
the mains is rectified, smoothed, then high freq oscillated which is applied to the transformer primary

eg ...

http://www.computeraudiophile.com/attachments/f8-general-forum/4738d1363502150-power-quality-audio-systems-power-supplies-better-smps-schematic.png

Dave

 

[tim9000]

non-linear, you mean switch mode PSU ?

A transformer is still used to give isolation
the mains is rectified, smoothed, then high freq oscillated which is applied to the transformer primary

eg ...

http://www.computeraudiophile.com/attachments/f8-general-forum/4738d1363502150-power-quality-audio-systems-power-supplies-better-smps-schematic.png

Dave

Thanks Dave, nice diagram, can you please talk about it a little (operation & control)? So it's a double buck converter? Is there a technique to analysing such a circuit from first site? Or is it just developed intuition?

So the bottom transformer is high frequency, the top is low frequency? What's the center-tap for?
I'm not very au fait with control circuitry, but I am interested, especially in in feed-back.

By non-linear yeah I meant SMPS or even just the rectifier itself.

Specifically what function does C102 do? I imagine C103 just smooths the rectification?

The LF101 is just a smoothing inductor isn't it? It's not actually providing isolation.
Because what my original question was pertaining to, was something like, when the stored charge in a diode rushes out as it becomes reversed biased, the diode looks like a short circuit (from memory) and the leakage inductance of the line causes an induced voltage due to the large di/dt as the diode snaps off. So what would happen to the primary and secondary sides if LF101 was actually an isolation transformer? And is this ever done?

Thank you (I need to develop my circuit intuition)

 

[davenn]

hanks Dave, nice diagram, can you please talk about it a little (operation & control)? So it's a double buck converter? Is there a technique to analysing such a circuit from first site? Or is it just developed intuition?

it's not one I personally work on ... just one of the first I saw on the net and pulled off as an example for you. It doesn't take much time to recognise the main sections and basic functions of the SMPS cct.

So the bottom transformer is high frequency, the top is low frequency? What's the center-tap for?
I'm not very au fait with control circuitry, but I am interested, especially in in feed-back.

both primary sections are high frequency and controlled by the oscillator, the FSQ0565R IC and its surrounding components. The centre tap of the primary isn't used in this particular design

You can see that the output of the 5V rail is monitored, essentially they are measuring variations in the voltage drop across that output inductor L202. This produces an error voltage that is fed back to the controller IC in the primary side via the opto-coupler with provides isolation between the HV and LV stages

By non-linear yeah I meant SMPS or even just the rectifier itself.

I suspected that, and hence I posted a SMPS cct, but you may have been referring to something else, I took the risk

Specifically what function does C102 do? I imagine C103 just smooths the rectification?

The LF101 is just a smoothing inductor isn't it? It's not actually providing isolation.

C102 and LF 101 are a low pass filter ( C101 probably plays a part in it as well) this is there to stop the high frequency signal of the switching oscillator from going back out into the mains supply and yes correct it isn't providing any isolation
Yes, C103 is the smoothing after rectification

Because what my original question was pertaining to, was something like, when the stored charge in a diode rushes out as it becomes reversed biased, the diode looks like a short circuit (from memory) and the leakage inductance of the line causes an induced voltage due to the large di/dt as the diode snaps off.

am not sure that that is quite correct, I'm not familiar with diodes having a stored charge ?? doesn't quite sound right. maybe some one else can chime in and clarify
@jim hardy

there's some thoughts for you

Dave

 

[davenn]

Is it common practice to use an isolation transformer between mains and the rectifier?

I noted that I didn't really fully answer that Q
no, you won't see a isolation transformer between the mains and the rectifier ... I don't think I have ever seen that situation

And just as a Q to you, do you understand why they use high frequency switching in SMPS's ?Dave

 

[jim hardy]

high voltages induced from the snapping off

Snapping off ? From rectifying line frequency? Seems to me dv/dt is so slow at line frequency, the approach to reverse bias is slow, it's more of a slide than a snap. Takes an instant for the charges to move out of the depletion region i suppose, which is oppositeof a snap.
Energy stored in a forward biased diode? New to me.

SMPS 's have swiyching transistors that snap on and off. Energy is stored in leakage inductance of a circuit and it's a nuisance in SMPS 's because it requires snubbing and the fast switching makes hf noise that gets into everything.

I think your memory of trouble from nonlinearity is probably based on harmonic content not di/dt.

 

[CWatters]

Is it common practice to use an isolation transformer between mains and the rectifier?

No.

An isolation transformer between the mains and rectifier would have to operate at the line frequency. One of the advantages of a switch mode power supply is that it can operate at a higher frequency than the line and that allows wound components to be smaller.

https://en.wikipedia.org/wiki/Switched-mode_power_supply#Advantages_and_disadvantages

 

[davenn]

No.

An isolation transformer between the mains and rectifier would have to operate at the line frequency. One of the advantages of a switch mode power supply is that it can operate at a higher frequency than the line and that allows wound components to be smaller.

https://en.wikipedia.org/wiki/Switched-mode_power_supply#Advantages_and_disadvantages

Grrrr I was wanting Tim to come up with some answers to my question to him D

 

[Tom.G]

Snapping off ? From rectifying line frequency? Seems to me dv/dt is so slow at line frequency, the approach to reverse bias is slow, it's more of a slide than a snap. Takes an instant for the charges to move out of the depletion region i suppose, which is oppositeof a snap.

He may have been referring to T_RR, Reverse Recovery Time. Not generally significant at mains frequency, so not usually specified for power diodes.
For small signal and high frequency diodes it can be significant. For instance the 1N4148 has a T_RR of 8nS. See https://www.vishay.com/docs/81857/1n4148.pdf

 

[tim9000]

Thanks for the replies.

Right, so it is a developed intuition, rather than a systematic approach a novice could take.

So why is there only C301 across T1, and not another cap across the 12v part? (are they a Flyback converter?) I'm only just researching...is it 'common mode noise'? (from the HF capacitive coupling between primary and secondary)

You can see that the output of the 5V rail is monitored, essentially they are measuring variations in the voltage drop across that output inductor L202. This produces an error voltage that is fed back to the controller IC in the primary side via the opto-coupler with provides isolation between the HV and LV stages

Ah, I see. So if there was zero volts across the output conductor no light would be emitted from the optocoupler, then what would the IC do? (stop pulsing? Like it only pulses when there is an error voltage, like a hysteresis control?) Also, could you please elaborate on that little circuit, like what that zener diode is for, and the resistors and C205?

C102 and LF 101 are a low pass filter ( C101 probably plays a part in it as well) this is there to stop the high frequency signal of the switching oscillator from going back out into the mains supply and yes correct it isn't providing any isolation
Yes, C103 is the smoothing after rectification

Ah, yeah ok. So are C102 and LF101 specific to be at a resonant frequency LC circuit?

am not sure that that is quite correct, I'm not familiar with diodes having a stored charge ?? doesn't quite sound right. maybe some one else can chime in and clarify
@jim hardy

Snapping off ? From rectifying line frequency? Seems to me dv/dt is so slow at line frequency, the approach to reverse bias is slow, it's more of a slide than a snap. Takes an instant for the charges to move out of the depletion region i suppose, which is oppositeof a snap.
Energy stored in a forward biased diode? New to me.

SMPS 's have swiyching transistors that snap on and off. Energy is stored in leakage inductance of a circuit and it's a nuisance in SMPS 's because it requires snubbing and the fast switching makes hf noise that gets into everything.

I think your memory of trouble from nonlinearity is probably based on harmonic content not di/dt.

Yeah I can't remember how harmonics fit's into the mix, but I would imagine this impulse would contain all sorts of harmonics.
yeah I think what I'm talking about is probably applicable anywhere there is a PN junction. Yes, snubbing does sound familiar actually.
but I am pretty sure it is the leakage of the line wires itself.
I did some quick googling, I haven't had time to read through but I think it is mentioned at the top here:
https://en.wikipedia.org/wiki/Step_recovery_diode

So I don't think it matters that at what frequency the diode is being used at, as soon as it goes reverse bias, the di/dt is the same? (Hence the same leakage induced voltage)

I noted that I didn't really fully answer that Q
no, you won't see a isolation transformer between the mains and the rectifier ... I don't think I have ever seen that situation

And just as a Q to you, do you understand why they use high frequency switching in SMPS's ?

Ah, ok, I was thinking that since an isolation TX would act as a low pass filter that it might be good to use leading into the rectifier. (also maybe containing the leakage inductance in the TX and not the power line or house wiring)
Yes (I think so, I'm a bit rusty), its because the higher the frequency the smaller the inductor you require to generate the same voltage due to the greater di/dt, something along those lines...
I also forgot about the ramifications CWatters made about TX size for line frequency.

Grrrr I was wanting Tim to come up with some answers to my question to him

Well I actually wrote this sort of in sequence, so I didn't read his comment until the end anyway.

Thanks

Edit:

He may have been referring to TRR, Reverse Recovery Time. Not generally significant at mains frequency, so not usually specified for power diodes.
For small signal and high frequency diodes it can be significant. For instance the 1N4148 has a TRR of 8nS. See https://www.vishay.com/docs/81857/1n4148.pdf

Thanks I'll check that out now.

 

[davenn]

but I am pretty sure it is the leakage of the line wires itself.
I did some quick googling, I haven't had time to read through but I think it is mentioned at the top here:
https://en.wikipedia.org/wiki/Step_recovery_diode

Step recover diodes are used in PSU's where there is rectification at high freq needed
eg, you will probably find that the 2 output rectifier diodes for the 12V and 5V rails are probably step recover diodes ... D201 and D202
You will also see them in boost/buck regulator switching supplies
The is because normal rectifier diodes like the 1N4001 to 4007, 1N5404 etc, cannot handle the high switching frequencies
But they are quite happy at mains frequencies of 50 - 60 Hz give or take a bitDave

 

[davenn]

Yes (I think so, I'm a bit rusty), its because the higher the frequency the smaller the inductor you require to generate the same voltage due to the greater di/dt, something along those lines...
.

sort of along those lines ... tho, more correctly, As the frequency increases, the core becomes less lossy
So instead of needing a big and heavy laminated iron core at 50/60 Hz, at 40 - 100 kHz ( common SMPS switching freq range)
the core can be made much smaller and use compressed ferrite powder instead of iron laminationsDave

 

[tim9000]

So does a step recover diode or the 1N4001 (TRR of 8nS) count as a 'soft reverse recovery' diode?

As the frequency increases, the core becomes less lossy
So instead of needing a big and heavy laminated iron core at 50/60 Hz, at 40 - 100 kHz ( common SMPS switching freq range)
the core can be made much smaller and use compressed ferrite powder instead of iron laminations

Ah, yeah less loss, of course.
Thanks in advance for anymore responses regarding my last post, cheers

 

[jim hardy]

So I don't think it matters that at what frequency the diode is being used at, as soon as it goes reverse bias, the di/dt is the same? (Hence the same leakage induced voltage)

??
current through a junction is i= e^qv/kt and that t is temperature not time
i just can't see the relationship you're suggesting.

 

[tim9000]

??
current through a junction is i= e^qv/kt and that t is temperature not time
i just can't see the relationship you're suggesting.

Hey Jim!
(Disclaimer I haven't read it very well) I believe it is explained on page 3:
https://www.google.com.au/url?sa=t&...f1x_OE1L9cFyU8few&sig2=MO1v52REt42YuXX7wU8eJg

Cheers

 

[jim hardy]

Hey Jim!
(Disclaimer I haven't read it very well) I believe it is explained on page 3:
https://www.google.com.au/url?sa=t&...f1x_OE1L9cFyU8few&sig2=MO1v52REt42YuXX7wU8eJg

Ahhh, okay

so the conduction doesn't stop at zero current but continues ( i guess while depletion region is re-established ? ) , and the diode interrupts not zero current which would be di/dt = 0 but some finite backward amount of it ...

 

[tim9000]

Ahhh, okay

so the conduction doesn't stop at zero current but continues ( i guess while depletion region is re-established ? ) , and the diode interrupts not zero current which would be di/dt = 0 but some finite backward amount of it ...

I believe so, hence why I was wondering if an isolation TX would prevent leakage inductance from the mains contributing to EMR if you used a non-soft recovery diode.

I've never had an electronics class that I've taken to, I suppose it's just not my strong suit and I never put the required effort in. I wish there was some sort of electronics Feynman I had of had when I took the subject, who shifted my perception early on (but I was overwhelmed by complexity).
I am actually quite curious as to the control method and understanding general operation (as in my post #10):

So why is there only C301 across T1, and not another cap across the 12v part? (are they a Flyback converter?) I'm only just researching...is it 'common mode noise'? (from the HF capacitive coupling between primary and secondary)

Ah, I see. So if there was zero volts across the output conductor no light would be emitted from the optocoupler, then what would the IC do? (stop pulsing? Like it only pulses when there is an error voltage, like a hysteresis control?) Also, could you please elaborate on that little circuit, like what that zener diode is for, and the resistors and C205?

Ah, yeah ok. So are C102 and LF101 specific to be at a resonant frequency LC circuit?

Thanks

 

[jim hardy]

I believe so, hence why I was wondering if an isolation TX would prevent leakage inductance from the mains contributing to EMR if you used a non-soft recovery diode.

at RF frequencies it'd probably couple as much hf noise back via interwinding capacitance as by transformer action..
A foil shield between the windings is commonly employed
good writeups here
https://www.canadatransformers.com/what-is-electrostatic-shield/
http://www2.schneider-electric.com/...live/FAQS/123000/FA123947/en_US/Isolation.pdf

 

[Tom.G]

Ref. post #10:

So why is there only C301 across T1, and not another cap across the 12v part?

C301 is used only to tie the Load Ground (output of the supply) to Earth Ground from the mains. This is effective for high frequency noise while providing DC and mains frequency isolation of the load. Note that C301 is rated at 1KV, the designer wanted to make sure it was effective for any reasonable circumstances!

 

[tim9000]

C301 is used only to tie the Load Ground (output of the supply) to Earth Ground from the mains. This is effective for high frequency noise while providing DC and mains frequency isolation of the load. Note that C301 is rated at 1KV, the designer wanted to make sure it was effective for any reasonable circumstances!

So nothing to do with common mode noise capacitive coupling between the coils? I could have sworn though, that I'd seen a capacitor on the top of a hf TX, like if in this case C301 was connected to where the dots were...
Also, really silly question, but how does C301 provide isolation at mains frequency?

I was hoping @davenn might get a chance to chime in and address some of:

Ah, I see. So if there was zero volts across the output conductor no light would be emitted from the optocoupler, then what would the IC do? (stop pulsing? Like it only pulses when there is an error voltage, like a hysteresis control?) Also, could you please elaborate on that little circuit, like what that zener diode is for, and the resistors and C205?

Ah, yeah ok. So are C102 and LF101 specific to be at a resonant frequency LC circuit?

Thanks again!

 

[davenn]

Hi Tim

sorry, have had a few busy days and nites

Ah, I see. So if there was zero volts across the output conductor no light would be emitted from the optocoupler, then what would the IC do? (stop pulsing? Like it only pulses when there is an error voltage, like a hysteresis control?)

I haven't gone into the fine details of the IC
Basically for all these types of IC's are a PWM control (Pulse Width Modulation) what that means is that, in the IC, there is an oscillator circuit
that switches a FET transistor on and off very fast. The frequency of the oscillator can be controlled by a PWM circuit ( this is also in that IC)
The error voltage coming from the opto-coupler is fed into pin4 Vfb and is referenced to a voltage across a resistor inside the IC.
As the Vfb voltage increases, it slows down the PWM duty cycle and this in turn lowers the internal oscillator frequency which lowers the switching
speed of the FET meaning less voltage is induced across the Primary winding (labelled pins 1-2-3) to the secondary windings of T1

Also, could you please elaborate on that little circuit, like what that zener diode is for, and the resistors and C205?

OK firstly, those components in the lower right ... R201,202,203,204,205, C205 and IC301
all form part of a shunt regulator circuit for driving the LED in the opto coupler

The zener diode ZD101, R105, 107, C106, 107 and D102 that all feed off a secondary winding of T1 pins 4 and 5
these are just for supplying a DC supply voltage to the IC, note they feed into pin3 Vcc of the ICcheers
Dave

 

[Tom.G]

So nothing to do with common mode noise capacitive coupling between the coils? I could have sworn though, that I'd seen a capacitor on the top of a hf TX, like if in this case C301 was connected to where the dots were...
Also, really silly question, but how does C301 provide isolation at mains frequency?

The dots at the transformer windings indicate the start of the winding, they are used so you know which pin to use to get the polarity you want at the transformer output.

For C301, the key point is that it connects to Earth Ground at the left end and the Load Ground (Load Common) at the right end. There is no connection to the Mains voltage. Its relatively low value of 4.7nF presents a rather high impedance at the Mains frequency but lower impedance at higher frequencies.

 

[tim9000]

Thanks for the replies!

The error voltage coming from the opto-coupler is fed into pin4 Vfb and is referenced to a voltage across a resistor inside the IC.
As the Vfb voltage increases, it slows down the PWM duty cycle and this in turn lowers the internal oscillator frequency which lowers the switching
speed of the FET meaning less voltage is induced across the Primary winding (labelled pins 1-2-3) to the secondary windings of T1

Ah yeah. Is there a comparitor involved in that IC? Reading that, I remembered seeing a triangular pulse train compared to a sine wave to determine switching?...duty cycle?
So when you say less voltage across the transformer primaries, is this all operating within...continuous conduction mode? So less average voltage going into the transformers, shorter time ON less the voltage ramps up?

OK firstly, those components in the lower right ... R201,202,203,204,205, C205 and IC301
all form part of a shunt regulator circuit for driving the LED in the opto coupler

The zener diode ZD101, R105, 107, C106, 107 and D102 that all feed off a secondary winding of T1 pins 4 and 5
these are just for supplying a DC supply voltage to the IC, note they feed into pin3 Vcc of the IC

I'll have to look into how to design regulator circuits in more detail, but What's IC201? some sort of voltage monitoring zener diode? (I've never seen that notation)

I only just noticed, so the top TX gets full rectified line voltage, but I'm a bit unclear on the feeding of the bottom TX (via the voltage division of SYNC?)?
...so the bottom supplied by pins 4 and 5? Isn't that the primary of the bottom of TX? So how much is SYNC? I assume it's in phase with something?
So how does the feedback pin work? What does shorting the pin via the opto or having a voltage across C105 make the IC do?

The dots at the transformer windings indicate the start of the winding, they are used so you know which pin to use to get the polarity you want at the transformer output.

For C301, the key point is that it connects to Earth Ground at the left end and the Load Ground (Load Common) at the right end. There is no connection to the Mains voltage. Its relatively low value of 4.7nF presents a rather high impedance at the Mains frequency but lower impedance at higher frequencies.

Ah, ok I can see that having a common ground on both sides is necessary to prevent current loops or something, and yeah that would be a pretty high impedance at mains frequency. But I swear I saw a video about high frequency TXs needing a capacitor between coils due to high capacitive coupling voltages, and in the video the capacitor was on the dotted side. Yeah I think I remember the dot convention. Oh, the polarity of the TX is going to swing positive and negative anyway, so having a cap up there would still work for that purpose? (I was getting the DC and AC aspect confused, but when the input circuit is shunted the TX polarity will flip the other way even though it is only feed by DC) But it's necessary to have it on the bottom here for grounding purposes. Does that all ring true to you?

Much Obliged. (sorry If I'm a pain with so many questions)

 

[Tom.G]

Well... mostly. Dave will probably be back with a detailed description that will go a long way toward understanding the ckt.

BTW. IC201 is known as a Programmable Zener or as a Programmable Shunt Regulator. When the Sense pin (the one coming out the side) goes above 2.5V the device starts conducting between Cathode and Anode. In the ckt. when IC201 starts conducting it pulls its Cathode towards Ground, turning on the optocoupler IC301 due to the voltage drop across R202.

A datasheet for IC201 is available at https://www.fairchildsemi.com/datasheets/KA/KA431.pdf

 

[tim9000]

Well... mostly. Dave will probably be back with a detailed description that will go a long way toward understanding the ckt.

BTW. IC201 is known as a Programmable Zener or as a Programmable Shunt Regulator. When the Sense pin (the one coming out the side) goes above 2.5V the device starts conducting between Cathode and Anode. In the ckt. when IC201 starts conducting it pulls its Cathode towards Ground, turning on the optocoupler IC301 due to the voltage drop across R202.

A datasheet for IC201 is available at https://www.fairchildsemi.com/datasheets/KA/KA431.pdf

Ah ok cool. Thanks
I assume I didn't say anything glaringly wrong then.

 

[tim9000]

Dave will probably be back with a detailed description that will go a long way toward understanding the ckt.

Here's hoping.

I only just noticed, so the top TX gets full rectified line voltage, but I'm a bit unclear on the feeding of the bottom TX (via the voltage division of SYNC?)?
...so the bottom supplied by pins 4 and 5? Isn't that the primary of the bottom of TX? So how much is SYNC? I assume it's in phase with something?
So how does the feedback pin work? What does shorting the pin via the opto or having a voltage across C105 make the IC do?

I looked up the IC. There's a block diagram for the converter here:
https://www.fairchildsemi.com/datasheets/FS/FSQ0565RQ.pdf
"Sync:
This pin is internally connected to the sync-detect comparator for quasi-resonant switch-
ing. In normal quasi-resonant operation, the threshold of the sync comparator is 1.2V/1."

But I'm still not sure about the operation...

 

[tim9000]

Hi,
I got a quick chance to look back at this thread (one of many on my to-do list) and I only just noticed that the whole bottom section of T1 is a Secondary, so in effect there are 3 secondary coils. So is the bottom left 5V actually used by an external component, or is it just used for feedback purposes?

I'm still confused regarding the IC pins 4 and 5. How does Vfb increase at all? Because doesn't it just short to ground when the opto is ON? And how does c105 make it more stable?
So when there is a voltage on pin 5 (Synch) @davenn you said that it slows the PWM down, did you have an actual waveform to demonstrate approximately what this means for converter operation?

Cheers

 

[davenn]

reposted to save swapping pages all the time

http://www.computeraudiophile.com/attachments/f8-general-forum/4738d1363502150-power-quality-audio-systems-power-supplies-better-smps-schematic.png

Hi,
I got a quick chance to look back at this thread (one of many on my to-do list) and I only just noticed that the whole bottom section of T1 is a Secondary, so in effect there are 3 secondary coils. So is the bottom left 5V actually used by an external component, or is it just used for feedback purposes?

Yes 3 secondaries
The bottom left secondary doesn't have a stated voltage on the diag. all we know is that it is rectified and used as a Vcc supply for the chip
there is also a small voltage taken off and fed to Pin 5, the sync pin

I'm still confused regarding the IC pins 4 and 5. How does Vfb increase at all? Because doesn't it just short to ground when the opto is ON? And how does C105 make it more stable?

If look above at the block diag for the IC, you will see that the voltage is actually supplied internally in the IC at Pin 4. Now you need to understand that the opto-coupler doesn't just switch on and off.
Depending on the brightness of the LED in the couple ( which is controlled by the regulator in the lower right corner) the conduction level of the coupler transistor will vary from full conduction = almost short cct to gnd to zero conduction = open cct.
This controls that internally supplied voltage and that variation is used along with that reference voltage
is fed via 3R to the PWM control stage

So when there is a voltage on pin 5 (Synch) @davenn you said that it slows the PWM down, did you have an actual waveform to demonstrate approximately what this means for converter operation?

The sync pin is picking up the switching freq of the oscillator coming off that lower left winding.
the sync'ing signal is used inside the IC to minimise the output MOSFET's switching noise and losses
I don't know enough to describe any deeper how that all works.
Dave

 

[tim9000]

reposted to save swapping pages all the time

The bottom left secondary doesn't have a stated voltage on the diag.

Damn-it, I meant bottom right. Sorry. I was typing that ad hoc and I should have looked at it again, does the 5V 2A output actually go to anything external? Sorry about the mistake.

the IC, you will see that the voltage is actually supplied internally in the IC at Pin 4.

I don't really remember how electronic current sources work, but I do take your point about the opto having degrees of conduction, that is something I didn't realize.

I'm still a bit confused: So for Vfb to increase (slowing down the PWM cycle) does that mean that the opto is going open circuit, or shorting to ground?
What's going on with the comparitors? Like when the opto is shorting to ground does that mean that the V_ref and V_CC is flowing to ground and so the PWM comparitor is grounded?
So pins 6 and 3 are what determines V_ref and V_CC? What's going on there?

The sync pin is picking up the switching freq of the oscillator coming off that lower left winding.
the sync'ing signal is used inside the IC to minimise the output MOSFET's switching noise and losses
I don't know enough to describe any deeper how that all works.

So what Synch is doing with the voltage from pin 5 is only taking a frequency, then using that frequency to create a pulse train to send into the flip flop and Nand gate? So it wants to use that frequency because any lower won't work and any higher will be inefficient?

The only basic experience I've had in this area is that I remember seeing a diagram years ago of a triangular wave and a reference voltage, and everytime the reference level the MOSFET (or whatever) turns off (or on?) creating a pulse train, I assume this is something like that?

Thanks heaps for your help Dave!

 

[Jony130]

You do not read this post and the links ?
https://www.physicsforums.com/threa...specific-smps-example-is-it-a-flyback.873084/
Or this
http://www.dos4ever.com/flyback/flyback.html

As for C301 capacitor, C301 is a part of a common-mode suppression, and is referred to as Y-capactor. More can be find here
https://eewiki.net/display/Motley/Minimizing+EMI+Problems+in+Noisy+Switching+Converter+Circuits++...and+Bears

Sync pin is used to "detect" damped oscillation form by primary side inductance and parasitic capacitance of the switch (source-drain capacitance of the MOSFET).
Because main IC is a quasi resonant controller current mode PWM controller and this Sync pin "set" MOSFET Turn-ON time.
More here
https://www.fairchildsemi.com/application-notes/AN/AN-4150.pdf (page 2)
And here about current mode PWM controller
http://artofelectronics.net/wp-content/uploads/2016/02/AoE3_chapter9.pdf (page 81 chapter 9.6.9)

The only basic experience I've had in this area is that I remember seeing a diagram years ago of a triangular wave and a reference voltage, and everytime the reference level the MOSFET (or whatever) turns off (or on?) creating a pulse train, I assume this is something like that?

Yes, basic PWM modulator

http://www.planetanalog.com/author.asp?section_id=483&doc_id=559556
http://www.ti.com/lit/ug/slau508/slau508.pdf

 

[davenn]

Damn-it, I meant bottom right. Sorry. I was typing that ad hoc and I should have looked at it again, does the 5V 2A output actually go to anything external? Sorry about the mistake.

Yes it's a 5V output ... the PSU has 5V and 12V outputs

I'm still a bit confused: So for Vfb to increase (slowing down the PWM cycle) does that mean that the opto is going open circuit, or shorting to ground?

it's varying between those to maximum conditions ... probably reaches neither of them

What's going on with the comparitors? Like when the opto is shorting to ground does that mean that the Vref and VCC is flowing to ground and so the PWM comparitor is grounded?
So pins 6 and 3 are what determines Vref and VCC? What's going on there?

see previous comment
and jony130 has answered most of the rest Dave