As always a very nice video. The air gap of the flyback transformer should only be made on the middle leg, for example by shortening it with sandpaper. Otherwise, the transformer will emit a significant amount of magnetic field, which will negatively affect the EMC. I also recommend adding additional shielding winding on the primary side and FLUX BAND before varnishing the transformer.
Ferrites are very hard, and "sandpaper" isn't going to work well at all. Silicon carbide is about the softest material with much hope of working. Manually gapping the centre leg is a tedious task. Of course in large production having the centre leg properly ground to spec isn't an issue (tell your supplier what inductance you want, NOT what gap to use). If you are going to use a shorting band around the outside of the transformer gapping all legs with a spacer is generally acceptable. I've used cores with spacers in flyback converters that passed radiated EMC requirements.
Did anyone else hear Diode Gone Wild in your head nonstop saying "☠☠SUPER DODGY!!!☠☠"? Steve's work here isn't dodgy at all, but that's just my natural reaction when I see one of these small transformers.
Please note that the drawings specify Triple Insulated Wire for the secondaries. This is important to comply with safety standards such as IEC 62368. The transformer is too small to be able to provide the margins that are needed for the creepage distance requirements, so TIW must be used instead.
Triple insulated wire is a great boon for small transformers. It's expensive but real problem solver. If you needed to meet some of the older VDE creepage and clearance requirements with ordinary magnet wire you lost 4 mm at each end of your bobbin. You'd probably have to select your transformer core based on that as much as electrical/magnetic performance.
Steve -- This mains-powered isolated SMP video series is highly valuable, following as it does PI's "cookbook", and really joining the dots. It would be great if you could dig a little deeper into the key transformer topics of 1. arranging the air gap for minimum EMC, 2. an insulation layer between primary and secondary in aid of achieving some target of pri-sec withstand voltage, and 3. a shield layer between primary and secondary to reduce transmission of switching noise voltage. And complementarily, how to measure these for a hand-built transformer or for a commercially-built one. Also 4, Resulting "touch current" and "touch voltage" (if the secondary is not grounded to mains ground). Other commenters have contributed some notes on some of these items, it would be really satisfying and useful to see them demonstrated in the series!
I was looking at getting a motorised one ages ago, but I think the windings would just be messier, faster 😢
Рік тому+3
Very interesting video. I tried to replace a similar SMPSU in audio equipment a few years ago. But I was not able to find a replacement for the Flyback transformer. I sent it to a shop in China that told me they would be able to make another one, but It didn't work and fried the board. What will be the best way to reverse-engineer the board to find the required winding?
It is cool ehh. Everyone always unwinds them haha. I plan to try my hand at a very low wattage one ad well. I need an lcr meter first. It's the only thing I'm really lacking.
Transformer winding is the most painful part of building power supplies. I do like your winder tough, I usually just secure a hand-crack "americaner" drill into a vice. It's only lacking a counter, and it was super cheap to get.
I've designed a lot of switchers. Transformer winding has *_never_* been the most painful part. It is a tedious undertaking for production, which is why it is best farmed out to specialist contract manufacturer.
@hardstyle905 Winding magnetics is tedious grunt work. Winding one or two by hand is one thing. Winding hundreds or thousands is a a job for a specialist company.
@hardstyle905 I'll match your "lol" with a "duh!" Hobby dweebs don't need large quantities. Not everyone who designs a transformer is a hobbyist. Apparently there are people slow on the uptake and unaware of this.
@hardstyle905 No, not doing any of those things. I don't even live in your [rude words] country. Do keep up making an [rude word] of yourself. Why is it that juvenile dweebs feel compelled to write "lol" all the time? Don't bother answering, you go straight to trash now.
This is what I can find, maybe a safety issue "It is also dangerous to pull a vacuum on varnish (flash point). Usually vacuum impregnate is used on air dry epoxy or oil filled transformers."
The application note mentions purposely increasing the interwinding capacitance as a means to counteract the effect of the leakage inductance. The idea to remove the need for a primary clamp. I'm not sure if vacuum impregnation could affect it.
@@sdgelectronics Good show! All other things being equal, I can't see how the varnish replacing the air can decrease the inter-winding capacitance. (The solidified varnish will have a higher dielectric constant, than the air it replaces.) Pulling a vacuum on varnish causes the VOCs (solvents) to rapidly boil out, and makes quite a mess. I think varnishing each layer, as it is wound, reduces the available winding area, and may decrease the inter-winding capacitance. (The layers will be farther apart. ) Anyway, I don't tend to see manufacturers impregnating these small SMPS transformers with varnish... there doesn't seem to be a good way to do it? ... and... proper application of transformer tape holds the windings tightly in place.
Varnish is readily available from vendors that supply materials to motor rewinding shops. The problem is that you probably won't be able to get it from such a supplier in small quantities. Most manufacturers don't put it up in volumes less than about 4 litres/1 US gallon.
In production it is common to use magnet wire that can be heat stripped. The wires are wrapped around the pins two or three turns then all are soldered simultaneously in a solder pot. You can do it manually pin by pin with an iron, but you need to be careful. Transformer tape is almost always polyester (Mylar is a well-known brand name). The tape itself is usually clear and colorless - the color is in the adhesive. Various adhesives are available. Some are heat curable. I used to buy the tape from a local vendor that sold supplies for motor rewinding shops. Now some of the electronics supply vendors stock it. For production you would typically go to a specialist supplier called a "converter" that would slit and spool the tape to your requirements (tape normally comes from the manufacturer in wide rolls, sometimes called "logs", half a metre or more in length. Occasionally a woven glass cloth tape is used for special purposes, but not between windings like the polyester tape.
There isn't a lot of merit in a split primary for a small flyback "transformer" since it isn't actually used as a transformer in terms of power delivery. It does reduce the leakage inductance somewhat which means the snubber on the primary has less power to cope with. [edit] I'd forgotten how he did the primary winding with "excessive" layers because of the difficulty in winding a minimal number of neat, uniform layers, Every layer added increases proximity effect losses which can be important. With the very fine wire being used it probably doesn't make much difference at the switching frequency used.
@@d614gakadoug9 I was more thinking about the capacitance of the windings. Or was it that having the split windings provide better magnetic coupling... I can't remember.
@@xenoxaos1 Yes, split windings can definitely make a worthwhile improvement in magnetic coupling. It can also reduce proximity effect losses which are sort of similar to skin effect losses when you have multi-layer windings. It can be rather a pain to do because you have to end one part, put on the secondary, then start the second part. My preference is to take each end to its own pin on the bobbin if there are enough pins. Some designs, such as push-pull, use a centre-tapped primary, so splitting it sort of "natural." Secondaries in push-pull, full-bridge and half-bridge converters are often centre-tapped, too, so you can match primary halves to the the secondary halves for best coupling and lowest proximity effect losses. Sometimes electrostatic shields are useful between primary and secondary for EMI/RFI management. With split windings that means two shields. When I was making transformers by hand for prototypes I'd use thin copper foil for shields. Commercial winding houses may prefer to lay down a single layer of closely-spaced fine wire for a shield because it is quicker with automated equipment. (Only one end of such a winding is connected, otherwise it just short-circuits everything.) In a flyback "transformer" the thing you want to do is minimize the "leakage inductance." This is modeled as inductance in series with the primary winding but with no magnetic coupling to the secondary winding. Split windings can help reduce leakage inductance. There is nothing in the circuit that provide an inherent desirable path through which the energy stored in the leakage inductance can be discharged, so it typically has to be dissipated in a the "snubber" which typically is made with a diode and capacitor in series from the switch drain (or collector) to the positive supply rail. A resistor across the cap discharges the energy it "absorbs" from the leakage inductance spike. A good design will get the leakage inductance down to around 1% to 2% of the primary inductance If you use a discrete power MOSFET for the switch you might be able to use the avalanche characteristics of the body diode of the FET to take care of the leakage inductance energy. You can't do that with these Power Integrations devices. I often say that in switchmode power supply design everything is in conflict with everything else. A lot of factors weigh into how you approach optimizing things.
Start on pin 7? Are you using counter clockwise numbering from the bottom view instead of usual top view or am I counting wrong here? ua-cam.com/video/fQmyrBEPsek/v-deo.html Same goes for pins 5 and 6 too. On pins 1&2 you seemed to calculate normal way from the top vie. That would make the phase of the secundary inverted to what was intended.
@@sdgelectronics You can always bodge the connections, just swap to wires with bit more to invert the secondary. That happens so easily.Been there, done that. Hope I saved you from some debugging and damaged test boards. I was so waiting to see you measure if you checked the polarity too or not. I often do, especially with very complex high turn transformers that have really bit me a couple times before.
As always a very nice video. The air gap of the flyback transformer should only be made on the middle leg, for example by shortening it with sandpaper. Otherwise, the transformer will emit a significant amount of magnetic field, which will negatively affect the EMC. I also recommend adding additional shielding winding on the primary side and FLUX BAND before varnishing the transformer.
Ferrites are very hard, and "sandpaper" isn't going to work well at all. Silicon carbide is about the softest material with much hope of working. Manually gapping the centre leg is a tedious task. Of course in large production having the centre leg properly ground to spec isn't an issue (tell your supplier what inductance you want, NOT what gap to use).
If you are going to use a shorting band around the outside of the transformer gapping all legs with a spacer is generally acceptable. I've used cores with spacers in flyback converters that passed radiated EMC requirements.
Did anyone else hear Diode Gone Wild in your head nonstop saying "☠☠SUPER DODGY!!!☠☠"? Steve's work here isn't dodgy at all, but that's just my natural reaction when I see one of these small transformers.
😂😂😂
Please note that the drawings specify Triple Insulated Wire for the secondaries. This is important to comply with safety standards such as IEC 62368. The transformer is too small to be able to provide the margins that are needed for the creepage distance requirements, so TIW must be used instead.
Triple insulated wire is a great boon for small transformers. It's expensive but real problem solver.
If you needed to meet some of the older VDE creepage and clearance requirements with ordinary magnet wire you lost 4 mm at each end of your bobbin. You'd probably have to select your transformer core based on that as much as electrical/magnetic performance.
Steve -- This mains-powered isolated SMP video series is highly valuable, following as it does PI's "cookbook", and really joining the dots. It would be great if you could dig a little deeper into the key transformer topics of 1. arranging the air gap for minimum EMC, 2. an insulation layer between primary and secondary in aid of achieving some target of pri-sec withstand voltage, and 3. a shield layer between primary and secondary to reduce transmission of switching noise voltage. And complementarily, how to measure these for a hand-built transformer or for a commercially-built one. Also 4, Resulting "touch current" and "touch voltage" (if the secondary is not grounded to mains ground). Other commenters have contributed some notes on some of these items, it would be really satisfying and useful to see them demonstrated in the series!
Thanks, I plan to make a higher power converter soon, so I will cover the topics there!
@@sdgelectronics Excellent!
Same coil winder as mine......the delux version with the didgital readout. I used to have the original analogue dial gauge one which had two speed!
I was looking at getting a motorised one ages ago, but I think the windings would just be messier, faster 😢
Very interesting video. I tried to replace a similar SMPSU in audio equipment a few years ago. But I was not able to find a replacement for the Flyback transformer. I sent it to a shop in China that told me they would be able to make another one, but It didn't work and fried the board. What will be the best way to reverse-engineer the board to find the required winding?
I'd love to see comparison video showing practical difference in inductance and capacitance of various parts at low and high test frequencies.
Thanks for the video.
Really cool video! Thanks!
It will never work - the tape needs to be yellow! LOL
Great video, Steve. Nice to see the process from start to finish. Looks a fantastic job. Thanks for sharing.
It is cool ehh. Everyone always unwinds them haha. I plan to try my hand at a very low wattage one ad well. I need an lcr meter first. It's the only thing I'm really lacking.
Transformer winding is the most painful part of building power supplies.
I do like your winder tough, I usually just secure a hand-crack "americaner" drill into a vice. It's only lacking a counter, and it was super cheap to get.
I've designed a lot of switchers. Transformer winding has *_never_* been the most painful part. It is a tedious undertaking for production, which is why it is best farmed out to specialist contract manufacturer.
@hardstyle905
Winding magnetics is tedious grunt work. Winding one or two by hand is one thing. Winding hundreds or thousands is a a job for a specialist company.
@hardstyle905
I'll match your "lol" with a "duh!"
Hobby dweebs don't need large quantities. Not everyone who designs a transformer is a hobbyist. Apparently there are people slow on the uptake and unaware of this.
@hardstyle905
No, not doing any of those things.
I don't even live in your [rude words] country.
Do keep up making an [rude word] of yourself.
Why is it that juvenile dweebs feel compelled to write "lol" all the time? Don't bother answering, you go straight to trash now.
I wonder if the stuff the core is made from is porous (like graphite) and the varnish will impregnate the core altering it's properties.
This is what I can find, maybe a safety issue "It is also dangerous to pull a vacuum on varnish (flash point). Usually vacuum impregnate is used on air dry epoxy or oil filled transformers."
Great video! Thank you!
Excellent video explanation, & helped clarify many concepts, Thank you.
The varnish we use to conformal coat our PCBs and transformers used to take my skin off, no mater what i cleaned it off with.
The transformer varnish I have contains Xylene.....harmful.
Now I'm curious to know why they said not to use a vacuum pump to get all the air bubbles out?!!
The application note mentions purposely increasing the interwinding capacitance as a means to counteract the effect of the leakage inductance. The idea to remove the need for a primary clamp. I'm not sure if vacuum impregnation could affect it.
@@sdgelectronics Good show! All other things being equal, I can't see how the varnish replacing the air can decrease the inter-winding capacitance. (The solidified varnish will have a higher dielectric constant, than the air it replaces.) Pulling a vacuum on varnish causes the VOCs (solvents) to rapidly boil out, and makes quite a mess. I think varnishing each layer, as it is wound, reduces the available winding area, and may decrease the inter-winding capacitance. (The layers will be farther apart. ) Anyway, I don't tend to see manufacturers impregnating these small SMPS transformers with varnish... there doesn't seem to be a good way to do it? ... and... proper application of transformer tape holds the windings tightly in place.
Nice work, let's see how the PSU works when assembled.
Very interesting. What did you use as transformer varnish? Regular varnish ?
Look for electrical varnish
It was sold specifically as transformer varnish and has that traditional smell of warm transformer. I can't find where I bought it from though
@@sdgelectronics Ok, thanks. I'll look out for it from the usual places.
Google "glpt varnish"
Varnish is readily available from vendors that supply materials to motor rewinding shops. The problem is that you probably won't be able to get it from such a supplier in small quantities. Most manufacturers don't put it up in volumes less than about 4 litres/1 US gallon.
Thank you
Great video, how about making another one and VAC Potting it and compare the two, just for a laugh ?...cheers
Should the secondary side use a TIW type cable?
What would diodegonewild have to say about your winding?
verrry dodgy
at the end did u solder the wires or did u leave them like that ? also what type of tape is that ?
In production it is common to use magnet wire that can be heat stripped. The wires are wrapped around the pins two or three turns then all are soldered simultaneously in a solder pot. You can do it manually pin by pin with an iron, but you need to be careful.
Transformer tape is almost always polyester (Mylar is a well-known brand name). The tape itself is usually clear and colorless - the color is in the adhesive. Various adhesives are available. Some are heat curable.
I used to buy the tape from a local vendor that sold supplies for motor rewinding shops. Now some of the electronics supply vendors stock it. For production you would typically go to a specialist supplier called a "converter" that would slit and spool the tape to your requirements (tape normally comes from the manufacturer in wide rolls, sometimes called "logs", half a metre or more in length.
Occasionally a woven glass cloth tape is used for special purposes, but not between windings like the polyester tape.
I would probably have done a split primary. Half the turns under and half over the secondary.
There isn't a lot of merit in a split primary for a small flyback "transformer" since it isn't actually used as a transformer in terms of power delivery. It does reduce the leakage inductance somewhat which means the snubber on the primary has less power to cope with.
[edit] I'd forgotten how he did the primary winding with "excessive" layers because of the difficulty in winding a minimal number of neat, uniform layers, Every layer added increases proximity effect losses which can be important. With the very fine wire being used it probably doesn't make much difference at the switching frequency used.
@@d614gakadoug9 I was more thinking about the capacitance of the windings. Or was it that having the split windings provide better magnetic coupling... I can't remember.
@@xenoxaos1
Yes, split windings can definitely make a worthwhile improvement in magnetic coupling. It can also reduce proximity effect losses which are sort of similar to skin effect losses when you have multi-layer windings.
It can be rather a pain to do because you have to end one part, put on the secondary, then start the second part. My preference is to take each end to its own pin on the bobbin if there are enough pins.
Some designs, such as push-pull, use a centre-tapped primary, so splitting it sort of "natural." Secondaries in push-pull, full-bridge and half-bridge converters are often centre-tapped, too, so you can match primary halves to the the secondary halves for best coupling and lowest proximity effect losses.
Sometimes electrostatic shields are useful between primary and secondary for EMI/RFI management. With split windings that means two shields. When I was making transformers by hand for prototypes I'd use thin copper foil for shields. Commercial winding houses may prefer to lay down a single layer of closely-spaced fine wire for a shield because it is quicker with automated equipment. (Only one end of such a winding is connected, otherwise it just short-circuits everything.)
In a flyback "transformer" the thing you want to do is minimize the "leakage inductance." This is modeled as inductance in series with the primary winding but with no magnetic coupling to the secondary winding. Split windings can help reduce leakage inductance. There is nothing in the circuit that provide an inherent desirable path through which the energy stored in the leakage inductance can be discharged, so it typically has to be dissipated in a the "snubber" which typically is made with a diode and capacitor in series from the switch drain (or collector) to the positive supply rail. A resistor across the cap discharges the energy it "absorbs" from the leakage inductance spike. A good design will get the leakage inductance down to around 1% to 2% of the primary inductance
If you use a discrete power MOSFET for the switch you might be able to use the avalanche characteristics of the body diode of the FET to take care of the leakage inductance energy. You can't do that with these Power Integrations devices.
I often say that in switchmode power supply design everything is in conflict with everything else. A lot of factors weigh into how you approach optimizing things.
Should use a sewing machine for winding the small stuff
Start on pin 7? Are you using counter clockwise numbering from the bottom view instead of usual top view or am I counting wrong here? ua-cam.com/video/fQmyrBEPsek/v-deo.html Same goes for pins 5 and 6 too.
On pins 1&2 you seemed to calculate normal way from the top vie. That would make the phase of the secundary inverted to what was intended.
I've messed up the secondary winding 😂 time to rewind
@@sdgelectronics You can always bodge the connections, just swap to wires with bit more to invert the secondary.
That happens so easily.Been there, done that. Hope I saved you from some debugging and damaged test boards.
I was so waiting to see you measure if you checked the polarity too or not. I often do, especially with very complex high turn transformers that have really bit me a couple times before.