Designing a AC-DC 9 Watt power supply. Input supply: 230 VAC, Output voltage: 48 V, 140 mA output current. Please help me decide the topology: Buck or Flyback? I want to understand the advantages and disadvantages between Buck and flyback so that I can justify my selection of the topology for this particular requirement.

Do you want to be electrocuted or not? The answer to that question tells you all you need to know to choose between a buck and a flyback.

Bob is right that you will have some mighty high voltages in a flyback. Why would you restrict your thoughts to only these two topologies? What is the end use? There are many considerations, for example Topswitch provides very simple solutions for low power devices that include dithering for EMI, but also can have higher ripple and as we said some very high voltages. A half bridge is more complex, but might work quite well as would most quasi-resonants, active clamp forward and a few others.

Steve, I was actually referring to the lack of isolation from the ac mains with a buck converter (assuming he meant a buck converter, and not a forward converter).

Hi Bob,

Thanks for the advice. I meant a buck converter and not a forward converter. Actually, I wanted to know if there are any more differences like in terms of efficiency or BOM cost, or manufacturing complexity, reliability.

A Flyback is more expensive than a buck, as usually a transformer is more expensive than an inductor. Fortunately too, except if you have suicidal tendencies and you like to put your two hands in an electric circuit connected to the power grid (and have also the chance to touch 2 points connected to Line and Neutral), you may even stay alive after designed your buck converter.

The only drawback I can see of a buck is the lack of stability at low output power. Then usually a Zener diode is required to limit the output voltage and this issue is solved.

My question ... why a switcher? ... If you have not done many SMPS, a step down transformer and linear Reg. may get you what you want. ... it would be bigger and less efficient, but you can't argue with simplicity. Depends on how critical the PS performance is to the end product. ... my two cents.

The choice of the topology will depend on your design goals and requirements. The buck converter aims for very high efficiency, typically >90% and low voltage application, thus simplifying BOM cost. On the other hand, flyback or forward topology will be ideal for medium voltage application, which eventually require higher voltage component and bigger magnetic, leading to higher BOM cost. Another factor to consider is the case size of the end product. Again, base on your design requirements, you can perform a topology trade off study to select the topology that will provide the best performance. I hope this help.

Thanks All.
Keil: I am actually performing a topology trade off study in order to select the topology for which I have posted this query. I would appreciate if you could help me do that in a more accurate way.

Shash, you did not provide the most important specification: what is the required isolation, if any, between the ac mains and the load? You say CE certification is needed - but to what standards? I see from your background that you have done street lighting so maybe isolation is not needed...

Also, in your initial posting, you said you were designing a 9 W power supply but 150 mA at 48 Vdc is 72 W. I am guessing the 72 W is correct specification...

Another question, what is the ac mains voltage range? If you are designing for India, China, or similar locations, is the converter required to maintain a proper output over a range of, for example, 140 Vac rms to 300 Vac rms? And must the output remain within specification for mains frequencies from, for example, 40 Hz to 70 Hz?

All of these are important questions that need to be answered before a design can be started.

Also for product design specifications for immunity to power line transients, limits on harmonic and other EM emissions, and operating temperature range (if street light, is the range from -40 degC to 120 degC?) are needed.

Another important question: are you trying to develop a product for production or is this a laboratory/university/hobby/personal project?

Hi Bob,

I am trying to develop this product for production and is intended for indoor environments. The input supply range is 90 - 300 VAC, 50 Hz.
The isolation required is up to 3 KV.
The operating voltage range should be 80 Deg. ambient.

Please let me know for any other information.

Hi,

Based on everyone's inputs and from library research, I have collated the differences between the flyback and buck topologies. Please find it below for review. Please let me know if there is anything incorrect or invalid. I would also appreciate any additions as well.

Buck
--Forward mode
--Magnetic element: Inductor
--Duty Cycle: up to 90%
--Limited to 100 Watt applications
--Exhibits lower output peak to peak ripple voltage. Low AC content, low losses
--Disadvantage: Step down topology only
--Non isolated topology, hence, for safety reasons, the buck topology cannot be used for input voltages greater than 42.5 VDC.
--Lack of stability at low output power levels. Zener diode is required to limit the output voltage and this issue is solved.
--Inexpensive due to inductor
--Good cross regulation with coupled inductors.
--Difficult to operate on Wide mains

Flyback
--Buck-Boost Mode
--Magnetic element: Transformer
--Duty Cycle: up to 50%
--Limited to 150 Watt applications due to high peak currents
--Stable at low output power levels
--Exhibits higher output peak to peak ripple voltage, Large AC content(conduction losses)
--Advantage: Step down, Step up topology
--Isolated topology, hence, safe to operate and handle and can be used for input voltages greater than 42.5 VDC
--Stable at low output power levels
--Expensive due to transformer
--High peak currents
--Leakage inductance difficult to manage
--Operates on Wide mains

One other advantage to the flyback: If you store energy on the output/secondary side you can get good power factor by running in discontinuous mode (which you probably want to do anyway to eliminate diode reverse recovery losses) with constant duty cycle. There are also variations (search on Redl) that use a single switch with a combination of duty cycle and frequency modulation to regulate both the output and shape the input current for high power factor.

A few notes about your notes:
I've built bucks that are over 10kW. They are incredibly useful and by no means limited to 100W or input voltages over 42 V (the 10kW unit was a 450V input to 0-400V output).

Bucks can also be stable at low power levels, you just have to be careful when it slips into discontinuous mode. The easy way around this is find a control IC with extra features like reverse current blocking which means it looks for a negative voltage across low side fet (assumes synchronous buck) and forces the low fet off. Look at the TI part TPS40054 to TPS40060 series chips (note: I would only use ICs like that for low voltage designs).

Bucks require a higher input voltage over the output. So, a changing AC input line doesn't work well with it unless you have a decent size bulk capacitance. This brings other problems like horrible power factor (0.7 or lower) and huge inrush current to charge that capacitance.


The Flyback on the other hand is tougher to get to higher powers; so you're correct in the limit around 150W. I've made them bigger, but the EMC and Radiated emissions are harder to deal with than other designs (fly backs like to couple to ground and heat sinks).

Looking at your design requirements (so far):
1. Isolation: Needed, and up to 3000V. This excludes the buck, unless you us a two stage design which for LEDs is probably not worth it. I would say your choices are fly back or forward converter, but given the power level stick with a fly back.

2. 7.2W requirement, designed for 9.0W. This sounds about right to give yourself a little margin.

3. Efficiency >85%. This may be hard to do with a fly back. Usually they run 80-90% efficient at nominal load. So, you might be getting really close to your limit. A two switch forward might be a better choice if this a "must" requirement, but it's going to cost more.

4. You mentioned CE was the regulating body... At your power level, you probably don't need power factor correction, but it's something you may want to double check as that requirement seems like its being required by any new product these days. Though with a requirement of <10% THD, you probably do need it.
If you need PFC, I would recommend NCL30001 by on-semi (I'm currently using it and it works well), but my design is much larger and requires PFC. They also have a decent reference design for LED lighting with a dimmer for your reference though at about 10x your power level.

An extra note on a PFC fly back. If you require a design of 9W, design your transformer for 18-20W (this is to handle the peak current at high line... and do the math at the lowest possible AC line voltage 80Vac to make sure it will never saturate)

5. What's you sizing/ cost target? And what's your budget/timeline?
The off the shelf still seems like it would meet your criteria, but it could be over your budget... and it probably won't meet your THD requirement.

Honestly, you're building a supply that is less than 10W. Check your THD requirement as it may end up costing you a fortune to hit that when you may not need to. Now is the time to go through that spec because later in the design/build phase will only end up costing you.