BTW, I am also concerned with transformer efficiency, as it reflects directly downstream.
For example, I am considering building in a power line input filter, to clean up the junk before it even gets to the transformers. Years of experience helps here, and I already have the critical parts, so it's easy to send them off to Wayne, it's a small package, the only potential problem being that they are fairly large and need some real estate inside the case.
It can easily be proved thatwith the filter, most electronics will exhibit a slightly lowered power consumption, anywhere from 2 to 10%. Also, less heat, nothing spectacular, but still, it's there. Small benefits tend to add up.
My reasoning is simple - if a Krell or a Levinson benefit from an external unit (and they do, on more than one occasion), then it's reasonable to expect this baby will benefit as well.
For example, I am considering building in a power line input filter, to clean up the junk before it even gets to the transformers. Years of experience helps here, and I already have the critical parts, so it's easy to send them off to Wayne, it's a small package, the only potential problem being that they are fairly large and need some real estate inside the case.
It can easily be proved thatwith the filter, most electronics will exhibit a slightly lowered power consumption, anywhere from 2 to 10%. Also, less heat, nothing spectacular, but still, it's there. Small benefits tend to add up.
My reasoning is simple - if a Krell or a Levinson benefit from an external unit (and they do, on more than one occasion), then it's reasonable to expect this baby will benefit as well.
Last edited:
Just my way of thinking - but I tend to put caps into the 'PSU' part and the PSRR is only a characteristic of the 'amp proper' (meaning the semiconductors part).
As for the location - next to the output devices - not sure, can't quite see the benefit myself. When this clearly is of value is when the amp's bridged and the current loop (excluding the speaker) is tight. So right up to the pins of my TDA8566 chip amp it makes sense to me. But when the OPS is discrete the loops look so huge that I can't quite work out why its done. Decoupling has to have the effect of reducing the current loop area in my estimation and I can't see how it does in many instances where (say) 100nF is added to the collector of an OP tranny. I'm open to persuasion though
As for the location - next to the output devices - not sure, can't quite see the benefit myself. When this clearly is of value is when the amp's bridged and the current loop (excluding the speaker) is tight. So right up to the pins of my TDA8566 chip amp it makes sense to me. But when the OPS is discrete the loops look so huge that I can't quite work out why its done. Decoupling has to have the effect of reducing the current loop area in my estimation and I can't see how it does in many instances where (say) 100nF is added to the collector of an OP tranny. I'm open to persuasion though
My experience has shown there are no easy answers - sometimes what intuitively feels good, because it's a big, meaty addition, shows up on a sim as being almost completely pointless, and other times a somewhat trivial seeming alteration can make a world of difference, when added to a model. I'm a great believer in sim's - I generally refine them to as close to reality as I think reasonable, then put the bits together - and it works, first up ... touch wood!! ,
Everything depends upon everything - I like to hit the model with the absolutely worst scenario I can think of, and see what it does. If the behaviour is still clean then there is a good chance that the real world device will reasonably behave itself ...
, Everything depends upon everything - I like to hit the model with the absolutely worst scenario I can think of, and see what it does. If the behaviour is still clean then there is a good chance that the real world device will reasonably behave itself ...
For the sake of science Man , good god ....!!!!
Yes, ala PS Audio, I'm with you on that .......
Just my way of thinking - but I tend to put caps into the 'PSU' part and the PSRR is only a characteristic of the 'amp proper' (meaning the semiconductors part).
As for the location - next to the output devices - not sure, can't quite see the benefit myself. When this clearly is of value is when the amp's bridged and the current loop (excluding the speaker) is tight. So right up to the pins of my TDA8566 chip amp it makes sense to me. But when the OPS is discrete the loops look so huge that I can't quite work out why its done. Decoupling has to have the effect of reducing the current loop area in my estimation and I can't see how it does in many instances where (say) 100nF is added to the collector of an OP tranny. I'm open to persuasion though
I'm in agreement here .....
My experience has shown there are no easy answers - sometimes what intuitively feels good, because it's a big, meaty addition, shows up on a sim as being almost completely pointless, and other times a somewhat trivial seeming alteration can make a world of difference, when added to a model. I'm a great believer in sim's - I generally refine them to as close to reality as I think reasonable, then put the bits together - and it works, first up ... touch wood!!
Everything depends upon everything - I like to hit the model with the absolutely worst scenario I can think of, and see what it does. If the behaviour is still clean then there is a good chance that the real world device will reasonably behave itself ...
Sims are nothing more than Genesis Frank, when you plug it in the real work begins...
Interestingly, a.wayne, that hasn't been the way for me - perhaps through luck, or the gods being on my side, the actual circuit as designed by sim has been the final. Which is not to say that lots of fiddling in the physical areas has not gone on - in the real world one has to contend with the fact that parts have physical behaviours which impact on the electrical.
Since changing anything will always change the sound I never chase finding a 'certain sound' - I pursue nulling, attenuating audible distortion of a certain type, Richard's 'noise modulation'. This is the hardest to 100% kill - but, the most effective way of achieving extremely satisafying sound
Since changing anything will always change the sound I never chase finding a 'certain sound' - I pursue nulling, attenuating audible distortion of a certain type, Richard's 'noise modulation'. This is the hardest to 100% kill - but, the most effective way of achieving extremely satisafying sound
For the sake of science Man , good god ....!!!!
My inherent laziness constrains me to take the path of least resistance. When a sim becomes unwieldy - needs lots of setting up and is likely to take ages to run I prefer to run it in reality rather than inside a machine. That's mainly because I use sims not so much to deliver results more to deliver insights. For insights the sim needs to be really interactive - if the loop of mod-sim-think_of_next_mod takes too long my attention's bound to wander. So I've started out making measurements of PSRR of some of my designs. I've also at Frank's encouragement made recordings of PSU noise.
Toroids have two significant advantages over classic trafos, all other things being the same: they can deliver over and above nominal quite a bit more without saturation of the core than classic trafos, and they have their center field of emissions turned not towards the outside, but towards the inside center.
If you can have the ring made of sintered material, then they are in a league of their own, but unfortunately, so is their price. But in peaks, they would be able to deliver over twice their nominal power, something no classic E or I core trafo could ever do. True powerhouses, but all too rare.
I think what people are referring to with regard to torods is their ability to pass line hash from primary to secondary and hence to the amplifier. However a line conditioner at the mains input and choke somewhere in the power supply can solve this issue - and then all you are left with is torods advantages over EI's.
Shoog
Shoog
Quite so, and worse. I read somewhere years ago that toroids actually amplify line hash by 1-2 dB. Not nice.
On the other hand, I make power line filters, so I am not worried. Some years ago, I suggested to Wayne that in view of his extreme power requirements, he should think hard about filtering the mains.
Easy does it, senator, the power line filter is a very simple add-on for the entire system. If you're gonna do it, do it right and do it across the board, and put a full stop to that.
- Status
- Not open for further replies.