Pro and Con and Info of Low vs High Frequency Transformer

[Mechatrommer]

currently me considering which transformer to be used in PSU design. either the normal 50/60Hz iron cored or high frequency (100KHz-500KHz) ferrite cored transformer. i only know one advantage of hi-freq transformer is its small size and lighter. what about efficiency and other aspects? the PSU planned to be in range somewhere 0-12V, 1-10A, the normal iron cored is heavy at this range, so using hi-freq is tempting. but i'm afraid i will be tricked by another aspect, any thought?

ps: while at it, i will also appreciate any links to high frequency transformer about info, spec, supply, type, pwm control circuit etc, thanks.

[IanB]

The 50/60 Hz transformers are powered from the mains and are used in linear supplies. The high frequency transformers are used in switch mode designs. Totally different circuit topology and not comparable at all.

[Mechatrommer]

i dont mean the difference in topology/circuit. but say, power efficiency between two similar rated transformer.

[alm]

How can you compare the transformers without considering topology? A high frequency transformer will saturate at 50/60 Hz, a 50/60 Hz transformer will be very inefficient at 100 kHz.

[Mechatrommer]

i mean like this. let say i have 2 transformer, A and B.

A is designed for 50/60Hz operation, iron cored and output 40Vpp 50/60Hz AC. further regulation output 12Vdc driving a load at 10A.
B is designed for 100KHz operation, ferrite cored and output 40Vpp 100KHz AC. further regulation output 12Vdc driving a load at 10A.

which one will give me lesser electricity usage if both is operated at same time duration in a month, A or B? well, rule of thumb answer pls.

[IanB]

The question is irrelevant as the two transformers are not comparable and can't be substituted.

The only way to answer your question is to look at a complete power supply design from front to back with specific parts identified and then measure the overall power supply performance.

As a rule of thumb, switching supplies will have a higher operating efficiency when higher output currents are needed at regulated voltages (10 A counts as high in this context). The greater efficiency comes at the cost of greater design complexity and possibly more noise on the output.

[Neilm]

A mains transformer running at 50/60Hz will have very little loss, only in the windings and a small amount of core loss. They are also easy (relevantly) to make / design, and very reliable.

A switch mode supply working in kHz will have not only the losses in the core to consider, but also losses in the switching device and those inherent in the topology it is working in. If there are two FETs switching, then there will be two sets of switching losses. A SPMS will also be most efficient at a specific loading usually about 85% of full (although there are ways of improving things at low loads). A well designed and well laid out supply could be over 90% efficient. The same circuit poorly laid out with only slightly less good components could be only 75% and be prone to going bang.

Yours

Neil

[Mechatrommer]

ok now i got something. that smps is less efficient as low current load, right? but last time i heard smps can be 90% efficient as neil comfirmed it, but for normal 50/60Hz transformer, its only 50% efficient i heard/read. so is this a myth?

[IanB]

Everything is a myth.

You are quoting all sorts of random numbers without a clear specification or definition of what you are trying to measure or even that you have a working design.

It's like asking which goes faster, a car or a train?

[Neilm]

There are many sorts of SMPS chip available. Some of these include techniques for improving the efficiency of the supply. I recently used a dedicated chip to charge a Li-Ion battery. It gets 24V in and charges the battery to 12.6V. When I measured the efficiency, it was over 90% from full load (3A) to when it finished (0.3A). The only problems I had were the FETs that I was originally told to use (we already used it on other products) didn't work at the frequency and got very hot (lowering the efficiency but I never bothered to measure it).

As  I tend to work on hand held / portable stuff, having large heavy mains transformers tends not to be feasable, so I don't really have any experience with 50/60Hz transformers.

Neil

[IanB]

but for normal 50/60Hz transformer, its only 50% efficient i heard/read

What does this even mean? Stop throwing around random factoids and mis-remembered details and start doing science. Produce a specification, consider designs that can meet that specification, and then do a comparative evaluation of them to see which one is most efficient, or cost effective, or other criteria you have to judge them against.

For instance if you take a mains transformer rated at 12 V 10 A RMS on the secondary, you feed that through a bridge rectifier, and use it to drive a resistive load like a lamp or a heater, then the transformer will be very efficient.

If you start saying you want the DC output to be filtered and regulated with a low ripple, then the efficiency of those other parts will start to matter. The transformer itself will also become less efficient because it will now have a more difficult load to drive.

So it all depends. Random factoids won't cut it.

[jahonen]

ok now i got something. that smps is less efficient as low current load, right? but last time i heard smps can be 90% efficient as neil comfirmed it, but for normal 50/60Hz transformer, its only 50% efficient i heard/read. so is this a myth?

Maybe true for small transformers, but bigger ones are over 90%. That of course applies to AC/AC conversion efficiency only (just the transformer efficiency) with linear load.

Regards,
Janne

[Mechatrommer]

Stop throwing around random factoids and mis-remembered details and start doing science.

science needs money. if i have money, i wont ask

[Zero999]

I'd say a high frequency transformer has fewer losses.

All right you can't compare them side by side exactly but just look at how efficient an electronic halogen transformer is compared to a traditional mains frequency transformer, the former is 95% efficient, the latter is typically 85%. Given that most of the losses in the electronic transformer will be in the switching transistors, the actual transformer on its own will be much more efficient, although that's purely theoretical.

[Mechatrommer]

Given that most of the losses in the electronic transformer will be in the switching transistors, the actual transformer on its own will be much more efficient, although that's purely theoretical.

i believe the bolded quotation is my original question. if we take the topology into account, a linear voltage regulator will have an efficiency of Vout/Vin (Vin is unregulated voltage output on transformer secondary, Vout is PSU regulated output), so say we are regulating 1V from 20V, the efficiency is only 5%!

[ejeffrey]

Yeah, it is just too complicated to say.  It only makes sense to compare two specific devices under specific operating conditions, or to compare two approaches under a common optimization metric including cost, efficiency, and so forth.  A 50/60 Hz transformer followed by a linear regulator will essentially always be less efficient than a SMPS, but that is due to many factors, not only the transformer itself, and is as much guided by cost effectiveness than anything else.

The advantages for low frequency transformers are: lower eddy current losses, stray capacitances, and antenna losses.  The advantages for high frequency devices is that magnetizing current is much less.  This means you need less inductance so you can use fewer turns of larger gauge copper, reducing copper losses, and cores of lower susceptibility, which allows lower hysteresis losses and higher resistivity for less eddy current losses even at high frequency.

However, this is dwarfed by the main differences between a linear and switch-mode supply.  A SMPS transfers less energy per cycle due to the higher operating frequency, and therefore can store that energy in an inductor or gapped transformer.  This allows efficient regulation through duty cycle control.  A linear supply instead stores the energy in a capacitor and uses a variable resistor to regulate the voltage, dissipating a lot of power in the process.  This makes SMPSs much more efficient than linear power supplies with line frequency transformers, but isn't really related to the losses in the transformer at all.

Finally, an SMPS must consider switching losses in the MOSFETs and diodes.  A line transformer supply simply doesn't have these losses because the AC waveform comes to you from the power company.  These losses get worse with frequency, so generally high frequency SMPSs have more loss than lower frequency SMPS.  It isn't really cut and dry though, because the higher frequency you go, the smaller and cheaper the transformers, inductors, and capacitors can be, allowing for cheaper 'oversizing' for improved efficiency.  A properly cost-optimized design will have to trade off the cost of the silicon vs. the cost of the passive devices vs. the desired efficiency, plus performance requirements like ripple, noise,  regulation, and emissions to find an optimal design.

If I had to give an answer to your relatively meaningless question, I would probably say that the physics favors  low frequency but economics favors high frequency.  If your really want to set a record for minimum transformer losses and aren't concerned with cost or output power, a low frequency transformer will probably win.  That lets you minimize antenna losses and eddy currents, which are a problem even at zero power.  Signal transformers can be really high efficiency at low audio frequency, but are expensive, bulky, and very low power.  If you have to make something cost effective, a high frequency transformer will win every time.

[amspire]

ok now i got something. that smps is less efficient as low current load, right? but last time i heard smps can be 90% efficient as neil comfirmed it, but for normal 50/60Hz transformer, its only 50% efficient i heard/read. so is this a myth?

You could probably make steel cored transformers with 98% efficiency  or better, but it is limited by economics. To reduce winding resistance, you need thicker wire, but then you need more bobbin space for the wire, so the whole transformer has to be bigger, which means each winding turn is now much longer so you need to make it even thicker ... You can end up turning a 1Kg transformer into a 20Kg transformer very easily. The ultimate quality steel core transformers end up being relegated to metrology measurement tasks like ultra-precision transformer bridges and dividers.

In practice, weight and economics means you usually design a transformer with as much losses as you can manage comfortably.

Switching power supplies can be over 95% efficiency, but again it comes down to economics. Going from 94% efficiency to 95% efficiency could double the price - who is going to pay for that? Most people are happy for anything over 80% efficiency.

Richard

[Psi]

currently me considering which transformer to be used in PSU design. either the normal 50/60Hz iron cored or high frequency (100KHz-500KHz) ferrite cored transformer.

Are you planing on using switchmode for either option?

[IanB]

You could probably make steel cored transformers with 98% efficiency  or better, but it is limited by economics.

I read that the DOE has mandated new efficiency standards since 2010 for power distribution transformers in the USA. New transformers are mandated to be up to 99.5% efficient and never less than 97% efficient depending on size and type. Evidently this increases size, weight and cost, but since it is a government regulation customers have to take the hit.

[Zero999]

Given that most of the losses in the electronic transformer will be in the switching transistors, the actual transformer on its own will be much more efficient, although that's purely theoretical.

i believe the bolded quotation is my original question. if we take the topology into account, a linear voltage regulator will have an efficiency of Vout/Vin (Vin is unregulated voltage output on transformer secondary, Vout is PSU regulated output), so say we are regulating 1V from 20V, the efficiency is only 5%!

What's that got to do with it?

It's perfectly possible to use an old fashioned iron transformer/rectifier with a switching regulator on the secondary to achieve a higher efficiency than a linear regulator.

[Mechatrommer]

Are you planing on using switchmode for either option?

cant decide yet as i have much unknown parameters right now. the latest idea is to use both linear+smps transformer/regulator (due to low voltage requirement pwm chip unable to cope with). maybe my original question (and my mind) is not very clear about this issue. but now i have to agree with ejeffrey's and amspire's reply, its not about transformer efficiency alone. in the end, the final product that matters. but i think ejeffrey has answered the original question spot on

If your really want to set a record for minimum transformer losses and aren't concerned with cost or output power, a low frequency transformer will probably win

but it differ from Hero's answer about theoritical efficiency of smps transformer but well i guess it doesnt matter as everyone agreed, it doesnt matter. but then, the "weight" issue still a dilemma for me. sorry i didnt mention that portability of the psu is one of my main concern, beside the efficiency.

[saturation]

One thing that can save you money and time is simply tear down and examine any wall wart PSU made in the past 10 years; at worst its minimal parts and is easy to study.  At best, its well made and you'll see many features added versus the cheapo one.  These PSU are all over the place and are cheaper than buying parts.

https://www.eevblog.com/forum/index.php?topic=6148.0

There's a lot of SMPS design on the web.  Very recently on eevblog a book has come to light to rival Art of E.  Its huge, but has a lot of expert advice, its been out since 2007, is very well written and its free!

http://www.thetan.plus.com/seekrets/index.html

[Mechatrommer]

One thing that can save you money and time is simply tear down and examine any wall wart PSU

you got my senses remotely, i'm doing it right now, except not 10 years, most of my stocks are recent, including the cheapo's which i'm concentrating since they are simpler to dechipher. including the one similar to mikeelectricstuff's (link) which i already blown apart, sadly i havent been able to dechipher the blocking oscillator circuit. i will avoid complex design right now to avoid dizziness thanks for the book link.

[saturation]

For safety, since these are mains powered devices its helps to use a variac and an isolation transformer.  They are easy to find internationally and you can see how the PSU responds to lower input voltage and isolate yourself from the mains.  I think them a must have for designing and troubleshooting mains powered PSU, linear or isolated. 

There are many copies of the Staco designed benchtop variac and you can plug this into any isolation transformer.

Chinese copy, about $50-70.  I have one of these, I tore it down the transformer is very well made, electrically functional but the finish is crude.  YMMV depending on what 'clone' you end up with.  If you operate it at 50% its capacity, it very likely safe to use.   Its about 4 x 5 inches in size.

 

Staco Original, about $300-500!

One thing that can save you money and time is simply tear down and examine any wall wart PSU

you got my senses remotely, i'm doing it right now, except not 10 years, most of my stocks are recent, including the cheapo's which i'm concentrating since they are simpler to dechipher. including the one similar to mikeelectricstuff's (link) which i already blown apart, sadly i havent been able to dechipher the blocking oscillator circuit. i will avoid complex design right now to avoid dizziness thanks for the book link.

[Zero999]

The advantages for low frequency transformers are: lower eddy current losses, stray capacitances, and antenna losses.  The advantages for high frequency devices is that magnetizing current is much less.  This means you need less inductance so you can use fewer turns of larger gauge copper, reducing copper losses, and cores of lower susceptibility, which allows lower hysteresis losses and higher resistivity for less eddy current losses even at high frequency

The copper losses tend to dominate in a low frequency transformer so all things being equal, a higher frequency transformer will be more efficient and past a certain frequency an air core transformer becomes a more viable solution but then skin effect losses increase.

However, this is dwarfed by the main differences between a linear and switch-mode supply.  A SMPS transfers less energy per cycle due to the higher operating frequency, and therefore can store that energy in an inductor or gapped transformer.  This allows efficient regulation through duty cycle control.  A linear supply instead stores the energy in a capacitor and uses a variable resistor to regulate the voltage, dissipating a lot of power in the process.  This makes SMPSs much more efficient than linear power supplies with line frequency transformers, but isn't really related to the losses in the transformer at all.

Finally, an SMPS must consider switching losses in the MOSFETs and diodes.  A line transformer supply simply doesn't have these losses because the AC waveform comes to you from the power company.  These losses get worse with frequency, so generally high frequency SMPSs have more loss than lower frequency SMPS.  It isn't really cut and dry though, because the higher frequency you go, the smaller and cheaper the transformers, inductors, and capacitors can be, allowing for cheaper 'oversizing' for improved efficiency.  A properly cost-optimized design will have to trade off the cost of the silicon vs. the cost of the passive devices vs. the desired efficiency, plus performance requirements like ripple, noise,  regulation, and emissions to find an optimal design.

The comments regarding the linear regulator aren't relevant to the discussion.

Low frequency transformers are more frequently used without a linear regulator than with one and even high frequency transformers are often used outside of a switching regulator for powering halogen lamps. It's true you can't compare them side by side but most of what you've said applies to linear vs switching regulators than anything else.