Further to the dc protection....
Regarding your friends speakers catching fire....
With the CE approvals (and indeed UL) I think it makes sense for all amplifiers which can deliver enough dc to be hazardous to incorporate some sort of dc protection but I also think that loudspeaker manufacturers knowing that this can be a common fault condition should also ensure that loudspeakers can not physically catch fire. I think many loudspeaker makers think that because their products contain no active electronics they are somehow exempt from having to meet any safety requirements, but they too should test for this...just my opinion i am afraid - i am not sure how many do test
Robin
Regarding your friends speakers catching fire....
With the CE approvals (and indeed UL) I think it makes sense for all amplifiers which can deliver enough dc to be hazardous to incorporate some sort of dc protection but I also think that loudspeaker manufacturers knowing that this can be a common fault condition should also ensure that loudspeakers can not physically catch fire. I think many loudspeaker makers think that because their products contain no active electronics they are somehow exempt from having to meet any safety requirements, but they too should test for this...just my opinion i am afraid - i am not sure how many do test
Robin
Jaka Racman said:
You'll have to be patient with me because I am largely ignorant of official (=academically sanctioned) terminogy in any of the fields I'm active in. I find it easier and even quicker to "reinvent the wheel" than to try studying from paper how the wheel is supposed to work. After that I understand what the text books are saying, but not before. The advantage is that while most of the time I'm redoing something that already exists, sometimes I get something which didn't pre-exist.
So far this approach has worked for me in class D, AD/DA, deltasigma modulators and linear (opamp) amps. Now I'm just turning to SMPS.
The clamp is intended to recover energy from the leakage inductance, store it in the cap and deliver it in forward mode. Adding a diode would defeat that purpose.
So how do you call such a topology?
Robin:
>There is always the problem of how you tackle it without putting something nasty in the audio path (typically a relay at the amplifier output).<
I consider this to be a questionable approach, and not only because of the effect on the sonic path. When a relay in the output pops open, this will frequently be accompanied by a transient spike which can spell instant death to the tweeters - sometime more.
>I have seen other schemes that crowbar the supply fuses or trip out a relay in the mains feed that forms part of the soft start.<
I find methods like this to be preferable (or cut out both rails in the case of transformers with dual secondaries), as the low impedance of the speaker load will pull down the power supply to the output stage fairly quickly, without any transients.
>Using an SMPS actually simplifies this as it is easy to shut down the supply without having to put extra components in the audio path.<
Agreed.
regards, jonathan carr
>There is always the problem of how you tackle it without putting something nasty in the audio path (typically a relay at the amplifier output).<
I consider this to be a questionable approach, and not only because of the effect on the sonic path. When a relay in the output pops open, this will frequently be accompanied by a transient spike which can spell instant death to the tweeters - sometime more.
>I have seen other schemes that crowbar the supply fuses or trip out a relay in the mains feed that forms part of the soft start.<
I find methods like this to be preferable (or cut out both rails in the case of transformers with dual secondaries), as the low impedance of the speaker load will pull down the power supply to the output stage fairly quickly, without any transients.
>Using an SMPS actually simplifies this as it is easy to shut down the supply without having to put extra components in the audio path.<
Agreed.
regards, jonathan carr
OT: DC protection
DC protection is sometimes omitted in low-power systems where amps and speakers are co-packaged and inseparable, and where safety testing has demonstrated that there is no fire hazard.
I am certain that no UL approval is granted to separate amplifiers without dc protection. If a house burns down due to a failed amplifier, and the amp has no UL listing, the manufacturer is required to foot the bill all by himself.
(note to non-US readers: UL means Underwriters Laboratory and an underwriter is an insurance company. No UL approval means no insurance).
It is legal to sell non-UL approved equipment (at your own risk!) in most of the US except California. Similar safety requirements hold in Europe, and are squarely mandatory (no no personal risk taking is allowed).
I would never consider putting an amplifier in use (apart from experimental testing when you're always present) that doesn't have DC protection on board.Robin said:
DC protection is sometimes omitted in low-power systems where amps and speakers are co-packaged and inseparable, and where safety testing has demonstrated that there is no fire hazard.
I am certain that no UL approval is granted to separate amplifiers without dc protection. If a house burns down due to a failed amplifier, and the amp has no UL listing, the manufacturer is required to foot the bill all by himself.
(note to non-US readers: UL means Underwriters Laboratory and an underwriter is an insurance company. No UL approval means no insurance).
It is legal to sell non-UL approved equipment (at your own risk!) in most of the US except California. Similar safety requirements hold in Europe, and are squarely mandatory (no no personal risk taking is allowed).
Hi Bruno,
I am not trying to criticisize your your design, but understand it and maybe give some helpful advice, if it is necessary at all (you seem to have things under control at all times).
From your description I think your design is close to this one. I think that you transform coupled inductor into separate transformer and series inductor on the secondary side and connect clamping capacitor to ground.
Regarding series diode you are correct, it is not required. I reverse oriented clamp transistor in my drawing.
Here is a short list of alternative single switch PFC topologies. I found that reading is sometimes advantageous to prevent reinventing the wheel, especially when you find that the wheel has already been patented. The last and most frustrating experience I had in this area was UcD patent, but there were others, like coupled inductor in forward converters or Peavey ampliverter. On the other hand, too much reading makes you difficult to think out of the box.
Best regards,
Jaka Racman
BTW, if the first link is really similar to your design, it was the first hit while googling for "active clamp flyback single stage pfc".
I am not trying to criticisize your your design, but understand it and maybe give some helpful advice, if it is necessary at all (you seem to have things under control at all times).
From your description I think your design is close to this one. I think that you transform coupled inductor into separate transformer and series inductor on the secondary side and connect clamping capacitor to ground.
Regarding series diode you are correct, it is not required. I reverse oriented clamp transistor in my drawing.
Here is a short list of alternative single switch PFC topologies. I found that reading is sometimes advantageous to prevent reinventing the wheel, especially when you find that the wheel has already been patented. The last and most frustrating experience I had in this area was UcD patent, but there were others, like coupled inductor in forward converters or Peavey ampliverter. On the other hand, too much reading makes you difficult to think out of the box.
Best regards,
Jaka Racman
BTW, if the first link is really similar to your design, it was the first hit while googling for "active clamp flyback single stage pfc".
Hi Jaka,
I was't seeing your comment as criticism, only explaining why I can't answer to seemingly simple questions like "are you using a clamped forward converter"...
It looks like the converter in this paper doesn't have the zero-load problem that my circuit does.
I was't seeing your comment as criticism, only explaining why I can't answer to seemingly simple questions like "are you using a clamped forward converter"...
Sounds correct. I do have a full wave rectifier on the output.Jaka Racman said:
It looks like the converter in this paper doesn't have the zero-load problem that my circuit does.
DC Protection
Hi Bruno,
I know this is slightly off thread and should probably be on the UcD thread but it has been mentioned on here so I wanted to ask.
I agree with your comments on dc protection and wonder what is incorporated on the Hypex UcD modules, and assume it is probably left up to the user. I assume that shutting down the amplifier enable is insufficient in the case of output mosfet damage.
Robin
Hi Bruno,
I know this is slightly off thread and should probably be on the UcD thread but it has been mentioned on here so I wanted to ask.
I agree with your comments on dc protection and wonder what is incorporated on the Hypex UcD modules, and assume it is probably left up to the user. I assume that shutting down the amplifier enable is insufficient in the case of output mosfet damage.
Robin
The hypex modules do not incorporate DC protection because usually you have just one per box, instead of 1 per channel. So it's left to the user to make such a centralised protection. Maybe it would be a good idea to incorporate a detection on the module (status pin).
I hear that JP is making a psu tailored for the modules that incorporates a dc protection.
Simply turning off the amplifier will no longer do when it's blown. Molten silicon is hard to turn off
I hear that JP is making a psu tailored for the modules that incorporates a dc protection.
Simply turning off the amplifier will no longer do when it's blown. Molten silicon is hard to turn off
Don't wish to sound anti-climatic, but maybe a better solution is just to use a choke after the mains bridge rectifier. It creates a simple PI-section filter. It does not give a high PFC, but I think it is good for audio. Here is an example where the choke value is chosen for 20 amperes peak to charge the primary capacitor from full wave-rectified 220vac. I have been using this method on my projects already. The file below shows the simple calculations for the example above.
Not using switching devices (switches) before the filter capacitor prevents dead spots in the response where the rectified voltage has dropped below the useful range of the switches when they are under a heavy load. At that time efficiency has fallen very low as well. There you have my simplified take on things.
Bruno, I feel that way also about keeping my mind from being distracted by too many outside ideas. I like to think of a circuit, try it, then troubleshoot it. I prefer to study how individual parts behave and then use the ones needed for the circuit.
Not using switching devices (switches) before the filter capacitor prevents dead spots in the response where the rectified voltage has dropped below the useful range of the switches when they are under a heavy load. At that time efficiency has fallen very low as well. There you have my simplified take on things.
Bruno, I feel that way also about keeping my mind from being distracted by too many outside ideas. I like to think of a circuit, try it, then troubleshoot it. I prefer to study how individual parts behave and then use the ones needed for the circuit.
Attachments
Agree completely.Jaka Racman said:
All readers interested in switching power circuits (power supplies and class d amplifiers) should ponder this truth and its immutable consequences. The several functions performed by a switching power supply with power factor correction all will extract their costs in terms of total volume of magnetics and total sum of switch V*I ratings, regardless of whether compacted into one stage of power processing (such as a flyback with output energy storage) or spread out among several stages (such as a boost inductor followed by a push-pull transformer dumping into an output energy store).
One can also see this when comparing a full bridge to a half bridge in class d output topologies with equal power handling capabilities. For the same total mosfet die area, the full bridge has four devices of half the voltage rating and half the on resistance. Yet, total losses and die power densities remain the same. Design choices then become driven by second order topological differences such as the number of gate driver stages or issues such as power supply pumping.
PFC really starts to make sense when the d-amp's peak power capability reaches to the tens of kilowatts. Then, the amp's rail voltage is high enough to where its capacitors store energy efficiently (at low voltages 'lytic cap size goes as CV rather than as CV2). With enough energy storage at the dc link between the PFC and d-amp, it can isolate the line and the PFC front end from the audio power peaks. The PFC keeps the wall breaker from tripping prematurely so that it can deliver the maximum possible power (up to twice the limit for a conventional rectifier/capacitor power supply). Thus, from 2kW of average line power, sustained music power peaks of 10kW to 20kW are theoretically possible when a suitable PFC switcher is mated to the right class d amplifier.Jaka Racman said:
Professional audio is largely exempt from harmonic and flicker requirements, but it has strict standards regarding leakage currents. Leakage limits (.7ma?) must be met at high line and with an open neutral (unless double insulation is used and there is no earth connection). Can't have the sound techs getting zapped at gig just from man handling gear strung out on well used ac extension cords.Jaka Racman said:
It seems reasonable that an amp big enough to put out more than a hundred volts and tens of amps should be protected against fire with fuse or breaker as is a similarly sized ac branch circuit. 
Regards -- analog(spiceman)
Jaka & analog,
Home truth there. The reason why I want to contract both functions is because of board real estate, not to save silicon area or ferrite volume. SMPS are already substantially lighter than 50Hz linear supplies but they are physically much bigger (and their efficiency is lower but that is another matter).
Home truth there. The reason why I want to contract both functions is because of board real estate, not to save silicon area or ferrite volume. SMPS are already substantially lighter than 50Hz linear supplies but they are physically much bigger (and their efficiency is lower but that is another matter).
single stage vs 2 stage pfc
Hi all, allow me to jump into this interesting discussion.
Bruno, the active clamp single stage isolated pfc approach you are considering is worth trying IMHO. A flyback implementation shouldn't be excessively hard to develop if you follow the detailed design rules in Robert Watson's thesis (first link provided by Jaka in post#26). Moreover, the 500W experimental design he describes in some detail has a 48V DC output, so it would only need halving the secondary wire section and duplicating the secondary to get a working prototype with two 48V rails and more than enough power for the UcD180!
From there, if you find the rail voltage ripple at double the mains freq (100hz or 120Hz) due to single stage is a problem (which it shouldn't be if output storage caps are big enough which they have to be for audio ripple anyway as has been said before), you can always get rid of it (admittedly at the cost of power factor correction by adding storage to the mains rectifier bridge!
Bruno Putzeys said:
Hi all, allow me to jump into this interesting discussion.
Bruno, the active clamp single stage isolated pfc approach you are considering is worth trying IMHO. A flyback implementation shouldn't be excessively hard to develop if you follow the detailed design rules in Robert Watson's thesis (first link provided by Jaka in post#26). Moreover, the 500W experimental design he describes in some detail has a 48V DC output, so it would only need halving the secondary wire section and duplicating the secondary to get a working prototype with two 48V rails and more than enough power for the UcD180!
From there, if you find the rail voltage ripple at double the mains freq (100hz or 120Hz) due to single stage is a problem (which it shouldn't be if output storage caps are big enough which they have to be for audio ripple anyway as has been said before), you can always get rid of it (admittedly at the cost of power factor correction
by adding storage to the mains rectifier bridge!Hi,
i am trying to establish design constraints for what would be considered a good power supply for classD amplifier. Bruno has already put forth some requirements, which I will try to consider. Questions are:
Input voltage range? Probably the only answer is worldwide input 90-265V AC, 50-60Hz.
Insulation Class? I can not answer that, but i suppose ClassII is a requirement. That then boils down to 4kV test voltage and 8mm creepage distance betveen primary and secondary circuits. Small low leakage transformer design is difficult with this constraint, so probably use of triple insulated wire would be a necessity.
Earth leakage current or enclosure leakage current requirement? My usual limit would be 0.5mA which would limit total Y capacitance of to approx 4.7nF. Bruno said he is not fond of Y caps at all, so even less would be desired.
Conducted EMI requirements? Is there any limit for 9kHz-150kHz range? If there is none, then it is probably a good idea to have switching frequency below 150kHz.
From the system point of wiev, should switching frequency be constant or variable? Variable switching frequency lowers requirements for EMI filter attenuation, but may interfere with ClassD amplifier operation.
Again, from the system point of wiev, what would be optimal switching frequency for classD amplifiers switching on the typical range of 350kHz to 450kHz?
Output requirements? Probably symmetrical output is required for all self oscillating amplifiers. Single output would be sufficient only for open loop modulated full bridge designs.
What is the maximum allowed supply voltage for full bridge designs? I have read that full bridges can't be UL listed as proffessional amplifiers because of the DC component at the output, yet we have some commercial designs using single supplies (Panasonic XR45). But they are limited to 100W per output. Problem can of course be solved by using split supply also for full bridges.
Typical output voltage requirement? Probably somewhere in the 50V-100V range per rail?
Output voltage regulation and limiting? Bruno said voltage sagging would be desirable for his designs. This is possible up to a certain output current, then you have to protect the switches. SMPS usually employ three modes of current limiting:
-foldback current limit where output current is proportional to output voltage, so short circuit current is usually low. Amplifier driven with full input and no startup muting would probably have problems starting.
-constant current output limit where power supply changes from voltage source to constant current source after some maximum load. This type of limiting is somewhat problematic in fixed frequency forward designs.
-hiccup current limit. Here power supply shuts down in presence of the short circuit and then periodically retries soft start until short circuit is removed. A variant of this is latched shutdown, where short circuit has to be removed, and then mains switch recycled.
So, is sagging really required and what is the best type of current limiting?
Minimum load condition? I suppose classD idling current would be somewhere in the 100mA-500mA range, depending on the number of channels being driven.
Can this load be considered as present in all conditions, or there is necessity for bleeder resistors in case where only primary regulation is used?
Required protection functions? Primary fuse is obvious, but secondary protection not. Is overvolage protection on the secondary mandatory? It should probably shut the power supply down during the overvoltage condition and then restart. Also overvoltage protection should probably be independent for each rail. Loss or undervoltage on one rail? Is any action required?
Thermal design? Peak to average power ratio? Maximum sustained peak output power time? Compared to 50Hz transformers SMPS have lower mass so their capacity to sustain long (several minutes) peak powers is not that great. Regarding short peaks, output capacitors are probably not enough, so magnetics should be designed for high peak output powers and relatively small average power. This is not problem for transformers, but is problem for inductors. Small inductors with large energy storage require large air gaps. Result is troublesome stray field, which has detrimential impact on EMI filter, since it induces voltage in it. Low mu relative toroids are no good for this although they would provide maximum energy storage per volume. Probably Bruno's idea about pot core without center post is a good solution. Low inductor volume necessitates use of high frequency switching, which on the other hand requires larger heatsinks. For optimal design it is therefore mandatory to know exact load requirements.
Other mandatory agency approval requirements, present and considered in the future? While I understand PFC is not required for the time beeing, is it even considered? Even the first version of IEC61000-3-2 (harmonics) has proposed metod for audio amplifier testing.
This is all I can think of it now, but I am sure there a some other questions to be answered.
Best regards,
Jaka Racman
i am trying to establish design constraints for what would be considered a good power supply for classD amplifier. Bruno has already put forth some requirements, which I will try to consider. Questions are:
Input voltage range? Probably the only answer is worldwide input 90-265V AC, 50-60Hz.
Insulation Class? I can not answer that, but i suppose ClassII is a requirement. That then boils down to 4kV test voltage and 8mm creepage distance betveen primary and secondary circuits. Small low leakage transformer design is difficult with this constraint, so probably use of triple insulated wire would be a necessity.
Earth leakage current or enclosure leakage current requirement? My usual limit would be 0.5mA which would limit total Y capacitance of to approx 4.7nF. Bruno said he is not fond of Y caps at all, so even less would be desired.
Conducted EMI requirements? Is there any limit for 9kHz-150kHz range? If there is none, then it is probably a good idea to have switching frequency below 150kHz.
From the system point of wiev, should switching frequency be constant or variable? Variable switching frequency lowers requirements for EMI filter attenuation, but may interfere with ClassD amplifier operation.
Again, from the system point of wiev, what would be optimal switching frequency for classD amplifiers switching on the typical range of 350kHz to 450kHz?
Output requirements? Probably symmetrical output is required for all self oscillating amplifiers. Single output would be sufficient only for open loop modulated full bridge designs.
What is the maximum allowed supply voltage for full bridge designs? I have read that full bridges can't be UL listed as proffessional amplifiers because of the DC component at the output, yet we have some commercial designs using single supplies (Panasonic XR45). But they are limited to 100W per output. Problem can of course be solved by using split supply also for full bridges.
Typical output voltage requirement? Probably somewhere in the 50V-100V range per rail?
Output voltage regulation and limiting? Bruno said voltage sagging would be desirable for his designs. This is possible up to a certain output current, then you have to protect the switches. SMPS usually employ three modes of current limiting:
-foldback current limit where output current is proportional to output voltage, so short circuit current is usually low. Amplifier driven with full input and no startup muting would probably have problems starting.
-constant current output limit where power supply changes from voltage source to constant current source after some maximum load. This type of limiting is somewhat problematic in fixed frequency forward designs.
-hiccup current limit. Here power supply shuts down in presence of the short circuit and then periodically retries soft start until short circuit is removed. A variant of this is latched shutdown, where short circuit has to be removed, and then mains switch recycled.
So, is sagging really required and what is the best type of current limiting?
Minimum load condition? I suppose classD idling current would be somewhere in the 100mA-500mA range, depending on the number of channels being driven.
Can this load be considered as present in all conditions, or there is necessity for bleeder resistors in case where only primary regulation is used?
Required protection functions? Primary fuse is obvious, but secondary protection not. Is overvolage protection on the secondary mandatory? It should probably shut the power supply down during the overvoltage condition and then restart. Also overvoltage protection should probably be independent for each rail. Loss or undervoltage on one rail? Is any action required?
Thermal design? Peak to average power ratio? Maximum sustained peak output power time? Compared to 50Hz transformers SMPS have lower mass so their capacity to sustain long (several minutes) peak powers is not that great. Regarding short peaks, output capacitors are probably not enough, so magnetics should be designed for high peak output powers and relatively small average power. This is not problem for transformers, but is problem for inductors. Small inductors with large energy storage require large air gaps. Result is troublesome stray field, which has detrimential impact on EMI filter, since it induces voltage in it. Low mu relative toroids are no good for this although they would provide maximum energy storage per volume. Probably Bruno's idea about pot core without center post is a good solution. Low inductor volume necessitates use of high frequency switching, which on the other hand requires larger heatsinks. For optimal design it is therefore mandatory to know exact load requirements.
Other mandatory agency approval requirements, present and considered in the future? While I understand PFC is not required for the time beeing, is it even considered? Even the first version of IEC61000-3-2 (harmonics) has proposed metod for audio amplifier testing.
This is all I can think of it now, but I am sure there a some other questions to be answered.
Best regards,
Jaka Racman
Maybe, but I wouldn't want to burden a high power supply with auto ranging, because it adds too much extra cost. A 115/230 voltage selector switch (or internal jumpers for the diy-ers) would still allow single model inventory and sales. This is not a product to be carried around the world like a laptop PC. Once set, very few end users would ever need to change the voltage setting.
Agency rated triple insulation wire (i.e. http://www.rubaduewire.com/html/news.html) is useful in small transformers where the winding window is not much wider than the required creepage distances. For transformers large enough to handle more than a couple of hundred watts or so, the safety setbacks (keepout zones at the ends of the winding window) don't subtract from the winding volume enough to justify the very much higher cost of the specialty wire. Rubadue type wire probably would make sense in the power supply for a UcD180, but not for a 1000 watt model. But, however it's done, as a product designer, I'd wouldn't want to cut any corners on safety by skimping on spacings, even for those market niches that allowed it.
Doing a quick calculation assuming a 0.5mA basic limit with 240VAC/50Hz at high line and with an open neutral yields a maximum total leakage capacitance of 6nF. Take out a few hundred pF for strays and cut the remainder back by 20 percent for component tolerances and you're down to 4.7nF. That budget might be spent on a few 1nF film caps on the power board and a couple of 330pF ceramics right by the ac power inlet.
Keeping the SMPS fundamental below 150kHz is a good idea and smearing the frequency also helps meet the conducted EMI requirements (which, unlike with radiated, allow averaging). Since the class d amp may very well be variable frequency itself, synchronizing the SMPS to it may be problematic. The SMPS absolutely must not interfere with the class d amp in any case, so, if it proves useful, why not go with a variable frequency design there as well?
Here I must disagree.
Self oscillation and full bridge outputs are completely compatible. In fact, once the power level rises to the point where parallel switches would be needed anyway, it's foolish not to consider a inherently pumping-free bridge.There is no reason that a bridge cannot be free of dc (just like when two UcD180 are bridged). The output power stage supply only need have its center be pseudo grounded as that "connection" passes no real current. Am I missing something?
Current day mosfet technology can easily accommodate output voltages three times as high as with bipolar designs (600V verses 200V devices). This might save considerable cabling costs in large venues with long runs between speakers (such as sports arenas). Even standard amps of high power might conceivably require rail voltage near 200 volts.
Unless the capacitive energy storage is able to provide all the low frequency peak audio power, the power supply will need a sustained short term peak current limit several time higher than its average limit. In this case, severe foldback or hiccup limiting would be unacceptable. On the other hand, paying for enough audio energy storage would allow a less expensive power supply design and keep the house lights from flickering to the beat of the music.
Just use a bridged design.
The inductor stresses in the power supply can't be any worse than those in the class d amplifier itself since it must pass the audio power peaks. The same design approach should work as well in both places.
What's needed is a toroidal core with a distributed air gap that is made like one of those chocolate oranges that breaks into wedges. The core would be built up of alternating wedges of magnetic and inert material. If the winding are symmetrically and evenly distributed there would be very little external field. With the windings on the outside, cooling would be excellent.
Thanks for the though provoking exchange. -- analog(spiceman)-
NCP1651 single stage pfc controller
Googling for single stage pfc controllers I came across the onsemi NCP1651, they claimed it was the industry's first of its kind when they released it about a year ago, there may be others now. Here is the product page:
http://www.onsemi.com/site/products/summary/0,4450,NCP1651,00.html
Extract: "The NCP1651 is an active power factor correction controller that is designed for operation over the universal input range (85 VAC- 265 VAC)in 50/60 Hz power systems. It provides a low cost, low component count solution for isolated ac- dc converters with mid- high output voltage requirements and eases the task of meeting the IEC1000-3-2 harmonic requirements for converters in the range of 50W - 250W."
Seems fairly well documented, with app notes, a 90W 48V ref board (couldnt find the pcb layout though) and an excel design spreadsheet.
Googling for single stage pfc controllers I came across the onsemi NCP1651, they claimed it was the industry's first of its kind when they released it about a year ago, there may be others now. Here is the product page:
http://www.onsemi.com/site/products/summary/0,4450,NCP1651,00.html
Extract: "The NCP1651 is an active power factor correction controller that is designed for operation over the universal input range (85 VAC- 265 VAC)in 50/60 Hz power systems. It provides a low cost, low component count solution for isolated ac- dc converters with mid- high output voltage requirements and eases the task of meeting the IEC1000-3-2 harmonic requirements for converters in the range of 50W - 250W."
Seems fairly well documented, with app notes, a 90W 48V ref board (couldnt find the pcb layout though) and an excel design spreadsheet.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.