What should be the transformer rating in respect to wattage of a power amplifier,
As we all know nobody will be feeding a sinewave signal to the amplifier,nor would anybody use it at maximum power , hence what should be the wattage of a transformer for a poweramp of say 250 watts for optimum performance ,keeping in mind the efficency of the amp and the value of filter capacitors.
As we all know nobody will be feeding a sinewave signal to the amplifier,nor would anybody use it at maximum power , hence what should be the wattage of a transformer for a poweramp of say 250 watts for optimum performance ,keeping in mind the efficency of the amp and the value of filter capacitors.
People building 'gain clone' type amplifiers tend to use very small filter caps and 'make up' for this by using larger than normal transformers.
Cheap commercial designs tend to use smaller transformers and try and 'make up' for this by using larger than normal capacitors.
I prefer a more balanced method.
For a capacitor input filter I find bigger 'sounds' better up to a point, the point being 2VA of transformer for every 1VA the ampifier is rated to produce. Beyond this point the improvement in sound vs cost and size seem to say 'not worth it'.
I have built with 1VA of transformer for 1VA of rated power, the amplifier worked well, but sounded a bit 'thin'.
During testing with the smaller transformer (driving hard into clipping on FM radio) the transformer burned up.
For a 250W amplifier I would use a 400VA~500VA transformer.
For 50hz and a single 4 ohm load I would use a pair of 8,000µF filter caps with around 47µF as a bypass cap.
Cheap commercial designs tend to use smaller transformers and try and 'make up' for this by using larger than normal capacitors.
I prefer a more balanced method.
For a capacitor input filter I find bigger 'sounds' better up to a point, the point being 2VA of transformer for every 1VA the ampifier is rated to produce. Beyond this point the improvement in sound vs cost and size seem to say 'not worth it'.
I have built with 1VA of transformer for 1VA of rated power, the amplifier worked well, but sounded a bit 'thin'.
During testing with the smaller transformer (driving hard into clipping on FM radio) the transformer burned up.
For a 250W amplifier I would use a 400VA~500VA transformer.
For 50hz and a single 4 ohm load I would use a pair of 8,000µF filter caps with around 47µF as a bypass cap.
Wouldn't this depend on the amplifier design quite a bit. As an extreme example for a given output power a class A SE amp would need a much larger transformer than a class AB push pull design. Also differing amounts of small signal circuitry could be an issue if there were a lot of signal processing etc. One would need to know the quiescent current requirements first wouldn't they?
I guess your rule of thumb is based on a basic class AB push pull right?
I guess your rule of thumb is based on a basic class AB push pull right?
Regardless of class, as said earlier if you can calculate the total maximum dissipation of the amp, I would go conservative with 2X the VA per watt dissipated......
So for a 200 watt per channel amp (2 channels) I'd go preferably with a 800VA (minimum of 600VA ). This will give you good headroom if you went into 4 ohms. Also an occasional 2 ohm load will not kill the trafo instantly as it would if you went cheap. E.g. The ubiquitas Adcom 555 is a 200wpc amp and uses a 675VA toroid.
With larger Trafo's you'll see that the voltage drop under load will be less, meaning you can use lower DC rails in your amp thereby avoiding the typical glare high rail amps have and also make the OP stage more reliable.
For example, I have a Kenwood M2a power amp rated at 220 wpc @8e (247w at clip) that uses small trafos... dual 250VA E+I types (500va total) but high rails +/- 91.5 vdc. This set up is less desirable (in the purist sense) than maybe a Nakamichi PA-7II which is also rated at 220wpc (250w at clip) but uses only +/- 70 volt rails to achieve the same. But that uses a 700VA Toroid. It is more reliable and stable at lower loads. (Lets not get into the design of the two amps but speaking strictly from a PS perspective).
I then proceeded to increase the Trafo size on my Kenwood to experiment, I used a 800VA trafo that gave me rails of only +/- 86vdc (about 5.5 volts per rail lower than kenwood's own trafos).
The sound was fuller, less constricted, had less glare and was warmer.... oh and by the way, the output at clipping increased from 247watts per channel to 372 watts per channel...
The point I am making is that by using a larger Trafo, you can increase power and sometimes make the amp better sounding and more reliable.... that is if u can afford a larger trafo.
On the capacitance, I tend to go with a minimum of 10,000uF per 100 watt. So for a 100 watt (2 channels) amp I'd uses 10,000uF per rail for a total of 20,000uF. This is the absolute minimum I'd use...
So for a 200 watt per channel amp (2 channels) I'd go preferably with a 800VA (minimum of 600VA ). This will give you good headroom if you went into 4 ohms. Also an occasional 2 ohm load will not kill the trafo instantly as it would if you went cheap. E.g. The ubiquitas Adcom 555 is a 200wpc amp and uses a 675VA toroid.
With larger Trafo's you'll see that the voltage drop under load will be less, meaning you can use lower DC rails in your amp thereby avoiding the typical glare high rail amps have and also make the OP stage more reliable.
For example, I have a Kenwood M2a power amp rated at 220 wpc @8e (247w at clip) that uses small trafos... dual 250VA E+I types (500va total) but high rails +/- 91.5 vdc. This set up is less desirable (in the purist sense) than maybe a Nakamichi PA-7II which is also rated at 220wpc (250w at clip) but uses only +/- 70 volt rails to achieve the same. But that uses a 700VA Toroid. It is more reliable and stable at lower loads. (Lets not get into the design of the two amps but speaking strictly from a PS perspective).
I then proceeded to increase the Trafo size on my Kenwood to experiment, I used a 800VA trafo that gave me rails of only +/- 86vdc (about 5.5 volts per rail lower than kenwood's own trafos).
The sound was fuller, less constricted, had less glare and was warmer.... oh and by the way, the output at clipping increased from 247watts per channel to 372 watts per channel...
The point I am making is that by using a larger Trafo, you can increase power and sometimes make the amp better sounding and more reliable.... that is if u can afford a larger trafo.
On the capacitance, I tend to go with a minimum of 10,000uF per 100 watt. So for a 100 watt (2 channels) amp I'd uses 10,000uF per rail for a total of 20,000uF. This is the absolute minimum I'd use...
Transformer Spec
1VA per Watt is fine if you are reproducing sine waves. However, you are not. You are reproducing music. I find that a VA of about 2/3 of the amp's RMS power rating is fine for a Class AB amp. With Class D amplifiers, the VA of the transformer can be 55% of the RMS power rating.
Just be awaye that if you run the amp at full power with a sine wave, you will eventually trip out the thermal breaker. However, even the safety agencies do not do this test, At most, they use pink noise at 1/3 power to simulate music though most of their long term tests are pink noise at 1/10 power.
Myself, to simulate music, I run a tone burst with the peak 20db over the average level, 50mS wide repeated every 500mS. The peak is set so the amplifier is in clipping. I find this does best to simulate the power demand of rock music, simulating a 120 beat per minute song with a heavy beat. With this test, VA demand through the transformer never exceeds 1/3 of the RMS power rating.
Don't waste money sizing the transformer to play sine waves at full power for hours at a time. You're not building a power inverter. We don't listen to sine waves.
1VA per Watt is fine if you are reproducing sine waves. However, you are not. You are reproducing music. I find that a VA of about 2/3 of the amp's RMS power rating is fine for a Class AB amp. With Class D amplifiers, the VA of the transformer can be 55% of the RMS power rating.
Just be awaye that if you run the amp at full power with a sine wave, you will eventually trip out the thermal breaker. However, even the safety agencies do not do this test, At most, they use pink noise at 1/3 power to simulate music though most of their long term tests are pink noise at 1/10 power.
Myself, to simulate music, I run a tone burst with the peak 20db over the average level, 50mS wide repeated every 500mS. The peak is set so the amplifier is in clipping. I find this does best to simulate the power demand of rock music, simulating a 120 beat per minute song with a heavy beat. With this test, VA demand through the transformer never exceeds 1/3 of the RMS power rating.
Don't waste money sizing the transformer to play sine waves at full power for hours at a time. You're not building a power inverter. We don't listen to sine waves.
Absolutely true for commercial amps, no need to oversize the transformer and for speakers that behave well, with nice impedance curves and good sensitivity.
When it comes to DIY I see no reason not to oversize the transformer (2x rating is ok).
I find this article good for DIY power supply issues: http://www.tnt-audio.com/clinica/ssps1_e.html
PS. We do not listen to sine waves but don't try weak amps on speakers like let's say an older Infinity Kappa 9 (86dB sensitivity and more like a short in the lower bass region, this speaker will happily do 29dB -2dB with a beefy amp, otherwise a lesser amp is likely to go presto kaboom or rush into protection mode/limiting if it has one)
Kappa 9 impedance curve: http://www.rageaudio.com.au/kappa9.jpg
When it comes to DIY I see no reason not to oversize the transformer (2x rating is ok).
I find this article good for DIY power supply issues: http://www.tnt-audio.com/clinica/ssps1_e.html
PS. We do not listen to sine waves but don't try weak amps on speakers like let's say an older Infinity Kappa 9 (86dB sensitivity and more like a short in the lower bass region, this speaker will happily do 29dB -2dB with a beefy amp, otherwise a lesser amp is likely to go presto kaboom or rush into protection mode/limiting if it has one)
Kappa 9 impedance curve: http://www.rageaudio.com.au/kappa9.jpg
XFMR Rating
Actually, commercial use is harder on the amplifier than home or DIY use. We play highly compressed music on our systems for hoursat a time with a low peak to average ratio. Our amps are consistently running at the clipping level or into heavy limiting. The amount of VA really needed in a home amp is seldom over 1/3 of the RMS rating of the amplifier.
In our commercial amplifiers we operate speakers that are like 100db 1W/1M efficient and routinely put out SPLs like 136db at 1 meter in order to get 105db at the audience. If you look at how UL test our products compared to a home amplifier, comercial amplifiers are tested at both 1/10 and 1/3 rated power while hi-fi amps are just tested at 1/10 power.
The safety agencies recognize that amplifiers reproduce music, not sine waves and they find transformers rated as less than 1/2 the RMS power rating of the amplifier to be acceptable.
That said, damping factor in an amplifier depends on their being a very low source impedance. This not only includes the speaker wiring and the amplifier itself but also includes the transformer, bridge reftifier, wiring and filter capacitors. We have determined that all of these affect the tightness of the bass but the biggest efect is from the filter capacitors and the wiring from the capacitors to the power amplifier.
This is because the capacitors are what supply the power to run the amplifier. The transformer and bridge merely recharge the capacitors. However, this recharge does happen faster if the transformer is over size. The effect of a smaller transformer is that the current pulse that recharges the capacitors is wider, meaning that at high volumes there is more ripple on the supply. On the other hand, less electromagnetic interference or buzz is created.
The point is, that an oversize transformer won't hurt anything and there may be the same sort of subtle effects that occur with the use of things like Monster Cable at work here. However, to make any effect, the use of an oversized transformer has to be accompanied by all the other audiophile tricks, including great speakers, to make the difference between a good amplifier and a true audiophile product.
Just realize that going past 1/2 the RMS rating of the amplifier for a transformer VA rating will do very little good in measurable results and is not required from a safety standpoint. You are entering on the difficult to quantify, often very subtle but very real differences you get in the audiophile realm.
Actually, commercial use is harder on the amplifier than home or DIY use. We play highly compressed music on our systems for hoursat a time with a low peak to average ratio. Our amps are consistently running at the clipping level or into heavy limiting. The amount of VA really needed in a home amp is seldom over 1/3 of the RMS rating of the amplifier.
In our commercial amplifiers we operate speakers that are like 100db 1W/1M efficient and routinely put out SPLs like 136db at 1 meter in order to get 105db at the audience. If you look at how UL test our products compared to a home amplifier, comercial amplifiers are tested at both 1/10 and 1/3 rated power while hi-fi amps are just tested at 1/10 power.
The safety agencies recognize that amplifiers reproduce music, not sine waves and they find transformers rated as less than 1/2 the RMS power rating of the amplifier to be acceptable.
That said, damping factor in an amplifier depends on their being a very low source impedance. This not only includes the speaker wiring and the amplifier itself but also includes the transformer, bridge reftifier, wiring and filter capacitors. We have determined that all of these affect the tightness of the bass but the biggest efect is from the filter capacitors and the wiring from the capacitors to the power amplifier.
This is because the capacitors are what supply the power to run the amplifier. The transformer and bridge merely recharge the capacitors. However, this recharge does happen faster if the transformer is over size. The effect of a smaller transformer is that the current pulse that recharges the capacitors is wider, meaning that at high volumes there is more ripple on the supply. On the other hand, less electromagnetic interference or buzz is created.
The point is, that an oversize transformer won't hurt anything and there may be the same sort of subtle effects that occur with the use of things like Monster Cable at work here. However, to make any effect, the use of an oversized transformer has to be accompanied by all the other audiophile tricks, including great speakers, to make the difference between a good amplifier and a true audiophile product.
Just realize that going past 1/2 the RMS rating of the amplifier for a transformer VA rating will do very little good in measurable results and is not required from a safety standpoint. You are entering on the difficult to quantify, often very subtle but very real differences you get in the audiophile realm.
Yes
A 300W amp will run just fine with a transformer with a core rated at 150VA continuous when you reproduce music. It will overheat if you reproduce sine waves at full power but music is a far less strenuous load than sine waves. That is why sine wave testing alone does not give a true picture on how loud an amplifier will play or how it will perform.
For my 1600W commercial amp the transformer is rated at 925VA. Our 250W amp uses a 225VA transformer and the safety authorities are quite haapy with it. We never, ever have any transformer failures. Never one in manufacturing thousands of amplifiers. No matter how bad the customers abuse the amps, we never even see any heat damage let alone any burning out.
Just remember that music is not a sine wave but a complex interaction of many frequencies and levels. That 90% of the time an amplifier is running at less than 10% power, more like 1% in the case of larger amplifiers, even when being played very loud and that most of the power to reproduce dynamics comes from the main capacitors and not from the transformer.
To measure power effectively and to get measurements that give a better picture of sound quality required much more sophictacat instruments than most people need. I'm a bit spoiled having three Audio Precisions in my lab but I find that nothing less gives anything but the most gross correlation between measured specifications and sound quality.
However, the bottom line is that having loads of filter capacitors makes a far bigger difference to sound quality than the transformer. However, the bridge rectifier has to be sized for the amount of capacitor you are using. I suggest 25A bridges as a minimum because the more capacitance you have, the greater the surge current when the transformer is charging the capacitors.
A 300W amp will run just fine with a transformer with a core rated at 150VA continuous when you reproduce music. It will overheat if you reproduce sine waves at full power but music is a far less strenuous load than sine waves. That is why sine wave testing alone does not give a true picture on how loud an amplifier will play or how it will perform.
For my 1600W commercial amp the transformer is rated at 925VA. Our 250W amp uses a 225VA transformer and the safety authorities are quite haapy with it. We never, ever have any transformer failures. Never one in manufacturing thousands of amplifiers. No matter how bad the customers abuse the amps, we never even see any heat damage let alone any burning out.
Just remember that music is not a sine wave but a complex interaction of many frequencies and levels. That 90% of the time an amplifier is running at less than 10% power, more like 1% in the case of larger amplifiers, even when being played very loud and that most of the power to reproduce dynamics comes from the main capacitors and not from the transformer.
To measure power effectively and to get measurements that give a better picture of sound quality required much more sophictacat instruments than most people need. I'm a bit spoiled having three Audio Precisions in my lab but I find that nothing less gives anything but the most gross correlation between measured specifications and sound quality.
However, the bottom line is that having loads of filter capacitors makes a far bigger difference to sound quality than the transformer. However, the bridge rectifier has to be sized for the amount of capacitor you are using. I suggest 25A bridges as a minimum because the more capacitance you have, the greater the surge current when the transformer is charging the capacitors.
Dan,
Good post, clear, I 100% agree. I sometimes shock people by explaining to them that their expensive over-the-top xformer is effectively disconnected from their amp some 80 to 90% of the time... Indeed, because the xformer only has small time slots to re-fill those caps, your bridge needs to be quite robust as you say.
Jan Didden
Good post, clear, I 100% agree. I sometimes shock people by explaining to them that their expensive over-the-top xformer is effectively disconnected from their amp some 80 to 90% of the time... Indeed, because the xformer only has small time slots to re-fill those caps, your bridge needs to be quite robust as you say.
Jan Didden
Did you know that they have car audio contests to see who
plays the loudest ? dB drag for example.
They design their audio system to only play a single tone
continously, therefore if you want to be a high powered
car audio amplfiier designer you need to think about this
market where people do listen to sine waves -->
/hehe
plays the loudest ? dB drag for example.
They design their audio system to only play a single tone
continously, therefore if you want to be a high powered
car audio amplfiier designer you need to think about this
market where people do listen to sine waves -->
/hehe
Yes, ocassionally the winner of that car audio contest drives past my house and causes my window panes to rattle violently.
There are two things being discussed here. One is what size xformer will work safely. The second is does size matter to sound.
I must admit I've never heard of a transformer bursting into flames in an audio amp. I agree that the average power demanded by the amp & speaker during unclipped music play will be less than that of a constant sinewave. So you need to estimate this average power and make sure the transformer is rated at least this high and is fused appropriately.
Now, there are other considerations. In a rectifier/capacitor supply the xformer has to recharge the caps very fast. This causes very high peak current demand. So you need to consider whether the transformer will cope with this peak demand. You might calculate you need 5A average current but 20A or more peak during worst-case charging. This may cause you to choose a higher rated xformer than is required based on average power delivery.
There are two things being discussed here. One is what size xformer will work safely. The second is does size matter to sound.
I must admit I've never heard of a transformer bursting into flames in an audio amp. I agree that the average power demanded by the amp & speaker during unclipped music play will be less than that of a constant sinewave. So you need to estimate this average power and make sure the transformer is rated at least this high and is fused appropriately.
Now, there are other considerations. In a rectifier/capacitor supply the xformer has to recharge the caps very fast. This causes very high peak current demand. So you need to consider whether the transformer will cope with this peak demand. You might calculate you need 5A average current but 20A or more peak during worst-case charging. This may cause you to choose a higher rated xformer than is required based on average power delivery.
Car Audio
In SPL contests they do indeed play sine waves but just long enough to get an SPL reading. They are indeed overloading the cores of the DC-DC inverter transformers and if left on too long, these will thermal off. The rest of the time, even with heavy rap music, the duty cycle is seldom over 25%.
Please understand that when a transformer is rated at 925VA and you are using it on a 1600W amp that the 925W rating is not an absolute maximum. It is merely the amount of power I can pull through it indefinitely without the transformer over heating. The transformer will quite happily pass 2000VA of power through it for short periods of time. Same with SPL testing.
LIke a car starter. A car starter will burn out if cranked for more than a couple minutes. Before Kettering, electric starters were seen to be impractical because a motor to crank over a car engine needed to be as big as a car engine if it was to last. However, Kettering realized that you did not need to crank the car engine for a long period of time. He realized that if a car did not start in say 30 seconds, it was not going to start at all. SO he wound a small motor to draw a huge amount of current, overloaded the crap out of it and the electric starter was born.
Same here. In an SPL test with sine waves, they overload the crap out of the transformer because we know that the published VA on a transformer is NOT an absolute limit. It is only the rating for indefinite use without overheating.
The need to supply over 100% of the full rated power of the amp with sine waves, for hours on end does not exist with music. If you want to spend money on an oversized transformer because you think you need it to make an amp work, you don't. If you want to spend the money because you think it makes an unmeasurable audiophile difference, in the same class as using monster cable and special CD player feet, then you may have a point.
In SPL contests they do indeed play sine waves but just long enough to get an SPL reading. They are indeed overloading the cores of the DC-DC inverter transformers and if left on too long, these will thermal off. The rest of the time, even with heavy rap music, the duty cycle is seldom over 25%.
Please understand that when a transformer is rated at 925VA and you are using it on a 1600W amp that the 925W rating is not an absolute maximum. It is merely the amount of power I can pull through it indefinitely without the transformer over heating. The transformer will quite happily pass 2000VA of power through it for short periods of time. Same with SPL testing.
LIke a car starter. A car starter will burn out if cranked for more than a couple minutes. Before Kettering, electric starters were seen to be impractical because a motor to crank over a car engine needed to be as big as a car engine if it was to last. However, Kettering realized that you did not need to crank the car engine for a long period of time. He realized that if a car did not start in say 30 seconds, it was not going to start at all. SO he wound a small motor to draw a huge amount of current, overloaded the crap out of it and the electric starter was born.
Same here. In an SPL test with sine waves, they overload the crap out of the transformer because we know that the published VA on a transformer is NOT an absolute limit. It is only the rating for indefinite use without overheating.
The need to supply over 100% of the full rated power of the amp with sine waves, for hours on end does not exist with music. If you want to spend money on an oversized transformer because you think you need it to make an amp work, you don't. If you want to spend the money because you think it makes an unmeasurable audiophile difference, in the same class as using monster cable and special CD player feet, then you may have a point.
Re: Car Audio
I've yet to see a car audio amplifier that will actually "overload" the transformer core in the SMPS. The limiting factor is not the core, or windings, but the switching devices. The thermal shutdown is watching the switching fets, not the core.
As for big transformers, I like my amps to weigh a ton! Sort of like when AT&T was putting steel plates in their phones, so buyers thought they were getting a 'beefy' phone. When you pick one of my amps, it feels like a 'good amp'
-Dan
dmfraser said:
I've yet to see a car audio amplifier that will actually "overload" the transformer core in the SMPS. The limiting factor is not the core, or windings, but the switching devices. The thermal shutdown is watching the switching fets, not the core.
As for big transformers, I like my amps to weigh a ton! Sort of like when AT&T was putting steel plates in their phones, so buyers thought they were getting a 'beefy' phone. When you pick one of my amps, it feels like a 'good amp'
-Dan
In SPL contests they do indeed play sine waves but just long enough to get an SPL reading.
/true
But the owner of the vehicle will still be playing his tones all day
long to aggrevate people with this sound system and perhaps
blowing up his woofers and amplifiers that are not worthly. /heh
Car audio is an abusive environment and amplfiers that are
not designed to be robust will get a bad reputation, so it's
in their best interest to make it beefy.
/true
But the owner of the vehicle will still be playing his tones all day
long to aggrevate people with this sound system and perhaps
blowing up his woofers and amplifiers that are not worthly. /heh
Car audio is an abusive environment and amplfiers that are
not designed to be robust will get a bad reputation, so it's
in their best interest to make it beefy.
a very interesting thread......
within an arbitrary 'line' of transformers with the same construction, one generally finds that the % Regulation drops with increasing VA ratings. So one would typically find that using a smaller VA rating transformer (all things being somewhat equal) that rail sag under load would increase under with smaller transformers.
With respect to that regulation %, what are the implications, both technical and subjective, resulting from the use of a smaller transformer?
within an arbitrary 'line' of transformers with the same construction, one generally finds that the % Regulation drops with increasing VA ratings. So one would typically find that using a smaller VA rating transformer (all things being somewhat equal) that rail sag under load would increase under with smaller transformers.
With respect to that regulation %, what are the implications, both technical and subjective, resulting from the use of a smaller transformer?
Dan makes some interesting points when it comes to thinking from the economics perspective. I would also tend to agree on the trafo not burning out so quickly...
However, looking at higher-end (non-commercial) home amps, they usually overrate their trafos by a large amount. Some of it is for marketing reasons but some of it also has to do with sonics and other considerations like low impedance drive etc.
I have done experiments where the larger trafo almost always sounded better, I got more output and was able to lessen the voltage rails thereby making the design more reliable. That being said, I may be digressing from the true spirit of the thread.
However, looking at higher-end (non-commercial) home amps, they usually overrate their trafos by a large amount. Some of it is for marketing reasons but some of it also has to do with sonics and other considerations like low impedance drive etc.
I have done experiments where the larger trafo almost always sounded better, I got more output and was able to lessen the voltage rails thereby making the design more reliable. That being said, I may be digressing from the true spirit of the thread.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.