Toroidal vs SMPS for Class D Modules

[egellings]

More VA per cm squared in a toroid, rather than efficiency per se.

Efficiency could just mean same performance in a smaller size.

 

[DonH56]

A gap in a toroidal transformer or EI construction allows a bit of DC in the core without magnetizing it too much. If there is absolutely no DC present, lose the gap and go toroidal.

Yah, I guess I have never seen a toroid with a gap. Again, no dog in this hunt, and far (far!) from my day job (though we use SMPS pretty much everywhere).

 

[Sokel]

So how do you account the huge discrepancy between 220 and 110vac. Both voltages are well within the parameter of being able to supply the power at the outlet. This in itself leads me to wonder about the entire measurement process here. 220vac is always a superior outlet for power amplifiers, but this should not be a superior outlet until one reaches the max one would get out of a 110 outlet. Those measurements show a drop from 90 to 15 under a 110 outlet.

Looks to me like a good one to send to Williston audio labs, and have him put it on the amp dyno.

Thankfully I'm on the 230V side (not that I ever managed to stress it thought) .
In another conversation reduced time explained due to PFC but I'm always skeptical,I think they should have made it the same regardless the low price.

Maybe you can get hints in Amir's measurements of an amp containing the module (not as is though,they use a buffer also).
Note that there is a "2000 watt" version too.

 

[egellings]

Yah, I guess I have never seen a toroid with a gap. Again, no dog in this hunt, and far (far!) from my day job (though we use SMPS pretty much everywhere).

I've not seen a gapped toroidal, either. Maybe the single-ended tube amp people may have those, I dunno.

 

[Kijanki]

Obvious advantages of SMPs are small size, line/load regulation, universal voltage and AC/DC operation. Less obvious is lower electrical noise. SPMS got bad rap from crude computer applications, but resonant mode SMPS can be extremely quiet switching at zero current / zero voltage (LPS switches at max voltage). That's why all Rowland products contain SMPS - even preamplifiers, where power losses are not important. Same goes for Benchmark, that lowered output noise by 10dB switching from linear supply in DAC1 to SMPS in DAC2, DAC3.

As for the size - tiny ferrite core can carry as much power as huge LPS transformer because of higher frequency. I understand that music demands very little average power (unless somebody listens to sine waves), but still transformer has to be oversized (being tested with sine waves). It is because LPS rectifier current is drawn in short spikes of very high amplitude. Such spikes have high RMS to average ratio resulting in higher copper losses. In addition high frequency content produces higher core losses (hysteresis, eddy currents). Transformer in LPS will get much hotter than same transformer with resistive load.

Why, in spite of SMPS advantages, we still have LPS? One reason could be SMPS complexity. Designing simple SMPS can be easy, but designing high power, high frequency resonant mode SMPS is not a trivial job (like Rowland resonant mode 2.5kW SMPS). Another reason is demand. Many people still believe that in order for something to be strong and robust it has to be big and heavy. Designers are simply addressing this demand.

 

[Doodski]

Obvious advantages of SMPs are small size, line/load regulation, universal voltage and AC/DC operation.

You might have forgot to add DC-DC operation.

 

[sarumbear]

I'm confused a bit. Toroids are used in both linear and PWM power supplies.

That is because the question does not make sense. Toroid is a type of transformer, SMPS is a type of AC to DC power supply. You can use a toroidal transformer in a SMPS PSU as well. I think what the OP means, is a mains frequency linear PSU or a high frequency SMPS better for a D-Class amplifier?

 

[Kijanki]

You might have forgot to add DC-DC operation.

I should've said "operation from AC/DC" instead of "AC/DC operation"

 

[Mnyb]

Yes the DC problem with toroidal transformers are very real in rural areas. You can easily have DC components on your AC mains .

Btw this whole tread makes little sense ? It’s about implementation build your ps in whichever style you want it’s the result that counts .

I speculate that this is about making a feature out of problem unregulated power supplies with no regulation usually sags under load , like a typical linear power supply.

But it holds it’s voltage for a short while and has a bunch of stored energy in its caps , so you then measure ( market ) this as “dynamic power” “ burst power” , “power with a paperclip shorting the amp for a nanosecond”
And the temporary overload you can have on iron and copper components is quite large so a transformer supply can be abused over its spec in short periods.

If you use an SMPS just make it big enough to provide all voltage and amps the class-D amp board can handle and that’s it , you now have moved the bottleneck to the amp and power supply is no longer an issue.

 

[Johnplayerson]

I thought lack of a gap made toroidal transformers less efficient than EI? Don't really care, curious. They are more sensitive to saturation from DC (etc.) without a gap. - Don

Toroidal transformers are actually 95 to 99 percent efficient in performing power. Because most never differentiate between the the power supply and the amplifier module, most assume the transformer has a large degree of blame to be inefficient. It is the amplifier module that determines efficiency. When there is a marriage with the modified ab style amplifier module, and Class H. we see very high rates of efficiency in toroidal amplifiers.

While efficiency is great, when one finally converts actual power use, and looks at the 10 times the power to double the sound rule, some extra efficiency leaves almost nothing to benefit as far as extra audio output is concerned. To me the whole idea of efficiency , just plays into the marketing of Smps and class d.

 

[Johnplayerson]

Yes the DC problem with toroidal transformers are very real in rural areas. You can easily have DC components on your AC mains .

Btw this whole tread makes little sense ? It’s about implementation build your ps in whichever style you want it’s the result that counts .

I speculate that this is about making a feature out of problem unregulated power supplies with no regulation usually sags under load , like a typical linear power supply.

But it holds it’s voltage for a short while and has a bunch of stored energy in its caps , so you then measure ( market ) this as “dynamic power” “ burst power” , “power with a paperclip shorting the amp for a nanosecond”
And the temporary overload you can have on iron and copper components is quite large so a transformer supply can be abused over its spec in short periods.

If you use an SMPS just make it big enough to provide all voltage and amps the class-D amp board can handle and that’s it , you now have moved the bottleneck to the amp and power supply is no longer an issue.

A transformer being a bottleneck? Now i heard just about everything!! But Yes, the biggest bottleneck is that its use is far too expensive compared to a grab bag of IGBT's. resistors and capacitors, and really puts a damper on PROFITS.

 

[egellings]

I thought lack of a gap made toroidal transformers less efficient than EI? Don't really care, curious. They are more sensitive to saturation from DC (etc.) without a gap. - Don

That is true; toroidal transformers are more sensitive to DC current saturating the core than E-I types are. I have not seen a gapped toroidal transformer, though.

 

[Johnplayerson]

That is true; toroidal transformers are more sensitive to DC current saturating the core than E-I types are. I have not seen a gapped toroidal transformer, though.

I would be curious how DC current gets into a transformer, when it is plugged into an AC line. The conversion to DC is accomplished by the bridge rectifiers which are AFTER the transformer, of which the DC current is then smoothed by the capacitors.

Toroidal High power Class D amplifiers by Zepher in India.

 

[egellings]

I would be curious how DC current gets into a transformer, when it is plugged into an AC line. The conversion to DC is accomplished by the bridge rectifiers which are AFTER the transformer, of which the DC current is then smoothed by the capacitors.

Toroidal High power Class D amplifiers by Zepher in India.

Asymmetry of the positive and negative halves of the AC power sine wave that can result from asymmetrical loading of them can cause a net DC offset.

 

[DonH56]

The problem is a lot of AC lines are not completely AC. DC at the wall outlet is not all that uncommon, and as @egellings said usually from asymmetry caused by other things in the house injecting noise that has an apparent DC component.

With respect to @Mnyb 's post, the power supply is often a bottle neck (not the transformer itself, though that happens as well, but he was careful in his wording) in achieving maximum power. Voltage sag in the transformer from simple wire (IR) drop happens and reduces the voltage rails, then there is drop through the rectifiers, insufficient charge storage (capacitance), and so forth. All decisions made during the product design. Conventional supplies for power amps tend to be unregulated whilst SMPS designs are regulated by design.

Many of us have dealt with buzzing, humming transformers caused by DC (or DC-equivalent) on the incoming AC line, and similarly had to live with the horrible stench of a transformer that decided to self-destruct. Not to say that exploding SMPS caps with bits of foil and insulator everywhere, or cracked switching transistors and burned PCB traces, are any better...

This whole thread could really have two branches based upon the title: conventional "linear" supplies vs. SMPS, and toroidal vs. EI (etc. ) transformers.

 

[Mnyb]

When I lived in a rural area I injected the DC myself , with my old car heater fan at half power
Half power was half the rectification shoddy design, but of little consequence if the network impedance is very low.
But out where I lived then it gave a mighty transformer hum in my amplifiers .
Some old hair dryers can do this to with their speed/power settings.

 

[Johnplayerson]

So it would seem the supposed greatest country in America, and supposedly the world, has such a crappy electrical grid then, so to speak. Yes i would tend to agree, that it is quite easy for capacitors and switching transistors to fail. The problem however has always been the manufacturers refusal to use parts that would avoid this very thing. While improvements happen over time, you still have a new up an comer like hypex , making modules with chinese caps, then finally switching to higher quality name brand caps, but even then still using standard 2000hr at 105c rated capacitors. Like how many times does the cycle have to repeat itself , until SMPS manufacturers, get the MEMO. ??? To these guys we are talking about dimes and nickles per capacitor savings over using a part that would significantly raise the life and reliabiliy of SMPS.

You can easily count how many dimes and nickels it would take to improve the power supply pictured , that i have seen sold for 300 USD. And also notice the transformer, is still required.

I never see my favorite caps installed in SMPS power supplies, that being the 10000hr UCC KZN and KYB, that are designed with power supplies in mind, yet these capacitors run into being sold out all the time at mouser and digikey. I really wonder what products these are being installed in.

Transformer HUM from DC,? I Have yet to experience that one. Most people no matter what amplifier they have seem to run into hum simply because the product is not built right. This would not exclude SMPS amplifiers. It would also seem transformers do a good job of removing hum for people , when used on pre amp inputs and outputs.

 

[Sokel]

I never see my favorite caps installed in SMPS power supplies, that being the 10000hr UCC KZN and KYB, that are designed with power supplies in mind, yet these capacitors run into being sold out all the time at mouser and digikey. I really wonder what products these are being installed in.

Will these do?

 

[BR52]

The 1000, 2000 or 5000 Hrs are design points for a given temp., ripple current, voltage and so on. If you stay below, you can reach dramatically higher live times. This is the real engineering task and has to be done in coincidence with the expected lifetime of the product. This datasheet values can only interpreted with the other parameters known.

 

[restorer-john]

The 1000, 2000 or 5000 Hrs are design points for a given temp., ripple current, voltage and so on. If you stay below, you can reach dramatically higher live times. This is the real engineering task and has to be done in coincidence with the expected lifetime of the product. This datasheet values can only interpreted with the other parameters known.

Estimated Life Calculator for Capacitors | NICHICON CORPORATION

You will find information about Nichicon's estimated life calculator for capacitors here.

www.nichicon.co.jp