Toroidal vs SMPS for Class D Modules

[Doodski]

If you check Yorkville sounds ratings for toroidal amps, you will note they were one of the few manufacturer, to also add the burst ratings. Almost any toroidal amplifier, using the same measuring tactics, currently used to measure smps power, would easily double their current and actual specifications, and this is after the fact, that many of these high powered pro amplifiers used under rated transformers to begin with.

I've handled thousands of SMPS toroidal transformers in high end car amps when I used to repair them in the late 80s and through the 90s. I'm familiar with the high current type of SMPS and they are certainly different than anything I've seen in a home audio application. Big difference.

For one simply add up the costs. SMPS is comprised of nothing but transistors, diodes, smaller value capacitors etc.

I'm not used to adding up expenses for home audio circuitry but now that you mention this I suppose you have a valid point.

When SMPS was invented, they changed the way power amplifiers were rated. Power was now rated as a 22ms burst, despite the fact 500ms burst would be required to mimic a complete audio signal, and 3 seconds full burst would be required for EDM bass. Today most SMPS are measured this way, and you can find WIlliston Audio labs using the AMP DYNO, to measure thousands of amplifiers, this exact same way. AS such I have always said we are using measuring specifications of power , that do not even reproduce an audio signal.

True that.

I figured a solid 500W/ch would require a a more capable transformer.

 

[Johnplayerson]

CHeck out how this Yorkville 1200 watt amp looks like when it is rated like an SMPS.

Amplifier class	H
Continuous Average Power @ 8 Ohms BCD (Watts)	750 (x2)
Continuous Average Power @ 4 Ohms BCD (Watts)	1200 (x2)
Continuous Average Power @ 2 Ohms BCD (Watts)	NA
Continuous Average Power Bridged BCD (Watts)	2400 @ 8 Ohms
Burst Average Power @ 8 Ohms BCD(Watts)	1200 (x2)
Burst Average Power @ 4 Ohms BCD (Watts)	2175 (x2)
Burst Average Power @ 2 Ohms BCD (Watts)	NA
Burst Average Power Bridged BCD (Watts)	4350 @ 8 Ohms
Frequency Response (Hz, +/- 1dB)	20-20,000
Hum and Noise (un / Aweighted -dB)	0.971698113207547
THD -1kHz- 4 OHMS	0.0001
THD - 20Hz-20kHz, 4 Ohms	less than 0.1%
Slew Rate (V/uS)	25
Slew Rate Bridged (V/uS)	50
Damping Factor (30 Hz - 400 Hz @ 8 Ohms)	600
Crosstalk (1kHz / 20Hz-20kHz)	-75 / -60 dB
Input Impedance - Bal/Unbal (Ohms)	20,000/10,000
Input Sensitivity (Vrms) For Full Power Out	1.4 V
Max Voltage Gain (dB)	36
CMRR @ 60Hz (min/typ)	48/56 dB
Stereo / Mono / Bridge )S/M/B)	S/M/B
Protection	DC,Load,Thermal
Limiter	Peak
High Pass Filter	40Hz, 12dB Octave
Cooling	Internal Fan
Cooling Path	Front to Rear
Fan Filter	User Serviceable
Inputs - XLR	2
Inputs - 1/4" Jacks	2
Outputs - Speakon 4-pin	2
Outputs - Binding Post	2
Power Consumption (typ/max)	1130/1800 Watts
Rack Spaces	2
Transformer Type	Toroidal
Exterior Finish	Baked, Black Painted
Dimensions (DWH /D fm ears, inches)	19 x 17.3 x 3.5 x 16.2
Dimensions (DWH /D fm ears, cm)	48.3 x 43.9 x 8.9 x 41
Weight (lbs / kg)	42.3 / 19.2

Note that Yorkville used burst average.... this is even more conservative then burst peak.. used by SMPS ratings.

 

[Johnplayerson]

I've handled thousands of SMPS toroidal transformers in high end car amps when I used to repair them in the late 80s and through the 90s. I'm familiar with the high current type of SMPS and they are certainly different than anything I've seen in a home audio application. Big difference.

I'm not used to adding up expenses for home audio circuitry but now that you mention this I suppose you have a valid point.

True that.

I figured a solid 500W/ch would require a a more capable transformer.

It is way better to have a transformer that will far exceed the power required of it. In the past most expensive Hi fi amplifiers came with OVERBUILT transformers. Also depending on the max voltage for which the transformer is made some are rated in VA at 50hz, A transformer rated at 50hz would have a much higher va rating at 60hz. The higher the frequency, the less size transformer is needed for the same va rating. SMPS changes the frequency to the other end of the scale completely. However , still a tiny transformer is needed to run the SMPS itself.

So in the case of everyday 200 watt into 8 ohm amplifiers like the hypex............ they could easily to made to use a transformer. Weight would not be the end of the world. Its not like the modules need the toroidal transformer of some of my 5000 kilowatt amplifiers.

 

[Doodski]

CHeck out how this Yorkville 1200 watt amp looks like when it is rated like an SMPS.

Amplifier class	H
Continuous Average Power @ 8 Ohms BCD (Watts)	750 (x2)
Continuous Average Power @ 4 Ohms BCD (Watts)	1200 (x2)
Continuous Average Power @ 2 Ohms BCD (Watts)	NA
Continuous Average Power Bridged BCD (Watts)	2400 @ 8 Ohms
Burst Average Power @ 8 Ohms BCD(Watts)	1200 (x2)
Burst Average Power @ 4 Ohms BCD (Watts)	2175 (x2)
Burst Average Power @ 2 Ohms BCD (Watts)	NA
Burst Average Power Bridged BCD (Watts)	4350 @ 8 Ohms
Frequency Response (Hz, +/- 1dB)	20-20,000
Hum and Noise (un / Aweighted -dB)	0.971698113207547
THD -1kHz- 4 OHMS	0.0001
THD - 20Hz-20kHz, 4 Ohms	less than 0.1%
Slew Rate (V/uS)	25
Slew Rate Bridged (V/uS)	50
Damping Factor (30 Hz - 400 Hz @ 8 Ohms)	600
Crosstalk (1kHz / 20Hz-20kHz)	-75 / -60 dB
Input Impedance - Bal/Unbal (Ohms)	20,000/10,000
Input Sensitivity (Vrms) For Full Power Out	1.4 V
Max Voltage Gain (dB)	36
CMRR @ 60Hz (min/typ)	48/56 dB
Stereo / Mono / Bridge )S/M/B)	S/M/B
Protection	DC,Load,Thermal
Limiter	Peak
High Pass Filter	40Hz, 12dB Octave
Cooling	Internal Fan
Cooling Path	Front to Rear
Fan Filter	User Serviceable
Inputs - XLR	2
Inputs - 1/4" Jacks	2
Outputs - Speakon 4-pin	2
Outputs - Binding Post	2
Power Consumption (typ/max)	1130/1800 Watts
Rack Spaces	2
Transformer Type	Toroidal
Exterior Finish	Baked, Black Painted
Dimensions (DWH /D fm ears, inches)	19 x 17.3 x 3.5 x 16.2
Dimensions (DWH /D fm ears, cm)	48.3 x 43.9 x 8.9 x 41
Weight (lbs / kg)	42.3 / 19.2

Cool. I've never seen a comparison of burst test mode versus a standard FTC test on the same amp a class H in this test. That burst power test reallly makes a big difference.

Continuous Average Power @ 8 Ohms BCD (Watts)	750 (x2)
Continuous Average Power @ 4 Ohms BCD (Watts)	1200 (x2)
Continuous Average Power @ 2 Ohms BCD (Watts)	NA
Continuous Average Power Bridged BCD (Watts)	2400 @ 8 Ohms
Burst Average Power @ 8 Ohms BCD(Watts)	1200 (x2)
Burst Average Power @ 4 Ohms BCD (Watts)	2175 (x2)
Burst Average Power @ 2 Ohms BCD (Watts)	NA
Burst Average Power Bridged BCD (Watts)	4350 @ 8 Ohms

 

[Doodski]

The higher the frequency, the less size transformer is needed for the same va rating. SMPS changes the frequency to the other end of the scale completely. However , still in tiny transformer is needed to run the SMPS itself.

Yes, the car amps' SMPS operating frequency was generally around ~50kHz +/-. I'm told now they are up to 400Khz for some lower current applications. I don't understand why but I've also been advised that lower frequency at the SMPS is better for high current output. Any ideas on that?

 

[restorer-john]

For one simply add up the costs. SMPS is comprised of nothing but transistors, diodes, smaller value capacitors etc. As my favorite man Randy Slone, has said, when he worked at Heathkit back in the day, they were not allowed to bend over to pick up any components dropped, while working on the production line. At the end of the day all dropped components would be swept up, and dropped into grab bags to sell to hobby electronic shops. It cost more money to bend over to pick up a dropped component, then it was worth. I have always said SMPS is like replacing a transformer supply, with one of these worthless grab bags. Mr Slone himself said entire amplifiers could be made less then the wholesale oem cost of a torodal transformer.

When SMPS was invented, they changed the way power amplifiers were rated. Power was now rated as a 22ms burst, despite the fact 500ms burst would be required to mimic a complete audio signal, and 3 seconds full burst would be required for EDM bass. Today most SMPS are measured this way, and you can find WIlliston Audio labs using the AMP DYNO, to measure thousands of amplifiers, this exact same way. AS such I have always said we are using measuring specifications of power , that do not even reproduce an audio signal.

If you check Yorkville sounds ratings for toroidal amps, you will note they were one of the few manufacturer, to also add the burst ratings. Almost any toroidal amplifier, using the same measuring tactics, currently used to measure smps power, would easily double their current and actual specifications, and this is after the fact, that many of these high powered pro amplifiers used under rated transformers to begin with.

Transformers are specified out at CONSTANT power. So a 500va transformer should run a 500 watt heater if need be, on a constant full burn at 500 watts indefinitely. The dynamic nature of music, has average power of the transformer like this, way down at just a few hundred watts at most. as such a transformer of 500va should be able to easily handle twice it's VA rating, when used for music. However if one is actually going to consider a 22 ms burst, power worth measuring, It would probably peak out 4 times higher.

Not quite sure why you are focusing in on toroidal transformers as being the 'gold standard'. EI and C/R core transformers also deliver considerably more than their continuous ratings. It also depends greatly on the filter capacitance.

And saying the 'invention' of SMPS supplies changed the way power amplifiers were rated is just plain wrong on every level. SMPSs have been around in power amplifiers since the 1970s and all of them produced rated power in accordance with FTC regulations which is continuous (5 minutes minimum), not "22mS" bursts. Modern Class Ds can produce phenomenal output powers in small packages but sure, their long term power outputs are limited by their ability (or lack thereof) to get rid of heat, both in the power supply and the output devices. But Class AB is no better and arguably a lot worse when it comes to getting rid of heat.

The standard EIA toneburst used for dynamic power was 20mS: 480mS (500mS total) with the 480mS being at -20dB. The EIAJ (Japanese) dynamic power was 1kHz 8 waves at 0dB, 24 waves at zero, so a 1 in 4 ratio. In reality, the two produce much the same results- I use them all the time for testing. The CEA 490-A burst is essentially just the old IHF/EIA test.

 

[kiwifi]

Large, linear power transformers are often audible at mains frequency, humming away at 50/60Hz. SMPS are silent in comparison because they are switching at hundreds of kilohertz. I am referring to mechanical noise here, not anything that you hear at the loudspeaker.

 

[Doodski]

Large, linear power transformers are often audible at mains frequency, humming away at 50/60Hz. SMPS are silent in comparison because they are switching at hundreds of kilohertz. I am referring to mechanical noise here, not anything that you hear at the loudspeaker.

In SMPS the transformers are known to hiss. Loud hissing.

 

[Sokel]

CHeck out how this Yorkville 1200 watt amp looks like when it is rated like an SMPS.

Amplifier class	H
Continuous Average Power @ 8 Ohms BCD (Watts)	750 (x2)
Continuous Average Power @ 4 Ohms BCD (Watts)	1200 (x2)
Continuous Average Power @ 2 Ohms BCD (Watts)	NA
Continuous Average Power Bridged BCD (Watts)	2400 @ 8 Ohms
Burst Average Power @ 8 Ohms BCD(Watts)	1200 (x2)
Burst Average Power @ 4 Ohms BCD (Watts)	2175 (x2)
Burst Average Power @ 2 Ohms BCD (Watts)	NA
Burst Average Power Bridged BCD (Watts)	4350 @ 8 Ohms
Frequency Response (Hz, +/- 1dB)	20-20,000
Hum and Noise (un / Aweighted -dB)	0.971698113207547
THD -1kHz- 4 OHMS	0.0001
THD - 20Hz-20kHz, 4 Ohms	less than 0.1%
Slew Rate (V/uS)	25
Slew Rate Bridged (V/uS)	50
Damping Factor (30 Hz - 400 Hz @ 8 Ohms)	600
Crosstalk (1kHz / 20Hz-20kHz)	-75 / -60 dB
Input Impedance - Bal/Unbal (Ohms)	20,000/10,000
Input Sensitivity (Vrms) For Full Power Out	1.4 V
Max Voltage Gain (dB)	36
CMRR @ 60Hz (min/typ)	48/56 dB
Stereo / Mono / Bridge )S/M/B)	S/M/B
Protection	DC,Load,Thermal
Limiter	Peak
High Pass Filter	40Hz, 12dB Octave
Cooling	Internal Fan
Cooling Path	Front to Rear
Fan Filter	User Serviceable
Inputs - XLR	2
Inputs - 1/4" Jacks	2
Outputs - Speakon 4-pin	2
Outputs - Binding Post	2
Power Consumption (typ/max)	1130/1800 Watts
Rack Spaces	2
Transformer Type	Toroidal
Exterior Finish	Baked, Black Painted
Dimensions (DWH /D fm ears, inches)	19 x 17.3 x 3.5 x 16.2
Dimensions (DWH /D fm ears, cm)	48.3 x 43.9 x 8.9 x 41
Weight (lbs / kg)	42.3 / 19.2

Note that Yorkville used burst average.... this is even more conservative then burst peak.. used by SMPS ratings.

And this is an SMPS amp's ratings (portion of this module is also measured by Amir)

Yep,the 1200 watt rating is misleading under conditions but on the other hands 90 secs is well over the 500ms limit.

 

[restorer-john]

Large, linear power transformers are often audible at mains frequency, humming away at 50/60Hz. SMPS are silent in comparison because they are switching at hundreds of kilohertz. I am referring to mechanical noise here, not anything that you hear at the loudspeaker.

Good transformers are quiet. Go pick up half a dozen random SMPS supplies in your house and have a listen. I'll bet half of them are audibly noisy.

 

[Sokel]

I'll bet half of them are audibly noisy.

Mine had a slight coil noise for about a week,then it decided to stop moaning and accept it's cruel fate.
From then on it's completely silent,other that a tick when it wakes from standby.

 

[restorer-john]

Yep,the 1200 watt rating is misleading under conditions but on the other hands 90 secs is well over the 500ms limit.

90 seconds is also well under the requirements for continuous power output claims for amplifier products sold, or advertised in the US, according to FTC regulations.

 

[restorer-john]

The thing about coil whine is you hear it when the amp is idling or asleep. I wouldn't care one bit if audible coil whine ocurred under extreme loads at high powers, but that's not the case.

 

[Sokel]

90 seconds is also well under the requirements for continuous power output claims for amplifier products sold, or advertised in the US, according to FTC regulations.

That's why I said misleading,it's in the fashion of all modern class D amps.I look at it as an 2x300 amp really only with enough headroom (that's why I have it for lows).

 

[Johnplayerson]

Not quite sure why you are focusing in on toroidal transformers as being the 'gold standard'. EI and C/R core transformers also deliver considerably more than their continuous ratings. It also depends greatly on the filter capacitance.

And saying the 'invention' of SMPS supplies changed the way power amplifiers were rated is just plain wrong on every level. SMPSs have been around in power amplifiers since the 1970s and all of them produced rated power in accordance with FTC regulations which is continuous (5 minutes minimum), not "22mS" bursts. Modern Class Ds can produce phenomenal output powers in small packages but sure, their long term power outputs are limited by their ability (or lack thereof) to get rid of heat, both in the power supply and the output devices. But Class AB is no better and arguably a lot worse when it comes to getting rid of heat.

The standard EIA toneburst used for dynamic power was 20mS: 480mS (500mS total) with the 480mS being at -20dB. The EIAJ (Japanese) dynamic power was 1kHz 8 waves at 0dB, 24 waves at zero, so a 1 in 4 ratio. In reality, the two produce much the same results- I use them all the time for testing. The CEA 490-A burst is essentially just the old IHF/EIA test.

Most modern amplifiers implement toroidal transformers. Further the toroidal can pack more punch in a smaller space then an EI the reason of course they were invented. Would hate to see the size of some pro audio amplifiers If they used an EI, that also requires greater shielding.

The ms burst specifications changes is not wrong at all. They can be found everywhere and the subject matter was part of a presentation at one of the AES conventions. I had read the whole paper at one time. BACK in the day, no one cared to measure bursts. And yes while some smps supplies were around for a long time, they were not around for extremely high powered amplifiers until more recent decades. The company speaker power which arrived some 8 yrs ago or so, use to advertised their SMPS was good for a full 3 seconds burst, vs the others that were rated at just 22ms when determining power output. Use to be there would be a little asterisk, saying 22ms burst, when pro audio amplifiers quoted output ratings. Not surprisingly, you no longer find this anymore in the spec sheets


Nearing 60 yrs old, I am still buying toroidal amplifiers, but because most are phased out.......... one has to go to china to get them. Transformer amps have been expanding rather then withdrawn over there. These ones are a bit too large for me. I like to stick to 100 lbs tops for an amplifier. When one uses them in multiples, it hardly becomes necessary to give any amplifier , anywhere near a full workout.

Seems to this day, anyone producing Class D is still reaching to the past to compare to high quality toroidal amplifiers, Keeping in mind that class D is the amplifier module, and SMPS is the power supply.

 

[Johnplayerson]

And this is an SMPS amp's ratings (portion of this module is also measured by Amir)

View attachment 256227

View attachment 256228

Yep,the 1200 watt rating is misleading under conditions but on the other hands 90 secs is well over the 500ms limit.

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.

 

[egellings]

Not quite sure why you are focusing in on toroidal transformers as being the 'gold standard'. EI and C/R core transformers also deliver considerably more than their continuous ratings. It also depends greatly on the filter capacitance.

And saying the 'invention' of SMPS supplies changed the way power amplifiers were rated is just plain wrong on every level. SMPSs have been around in power amplifiers since the 1970s and all of them produced rated power in accordance with FTC regulations which is continuous (5 minutes minimum), not "22mS" bursts. Modern Class Ds can produce phenomenal output powers in small packages but sure, their long term power outputs are limited by their ability (or lack thereof) to get rid of heat, both in the power supply and the output devices. But Class AB is no better and arguably a lot worse when it comes to getting rid of heat.

The standard EIA toneburst used for dynamic power was 20mS: 480mS (500mS total) with the 480mS being at -20dB. The EIAJ (Japanese) dynamic power was 1kHz 8 waves at 0dB, 24 waves at zero, so a 1 in 4 ratio. In reality, the two produce much the same results- I use them all the time for testing. The CEA 490-A burst is essentially just the old IHF/EIA test.

Mean length of turn is shorter in a toroidal than it is in a comparable E-I type. Less resistance with less wire means less heating loss and a stiffer voltage. The toroidal core makes for a more efficient magnetic circuit as well.

 

[DonH56]

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

 

[sq225917]

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

 

[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

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.