It sure is a huge beast and cooler transformer is nice. But if you choosed to use bigger caps AND a bigger transformer it migth be of advantage to reduce the formfaktor to get lower harmonic content. You can easely improve your amp do this by using a appropiate ac-choke at the input to the tranny.
but the isue here was the transformer. how can you say a smal transformer is cufficeint for a large class A amp?
and the inrushcurrent.
Trust me, in a class A bigger is not better, lol
I wish it where that easy ;0)
In class B things are different because good dc regulation is important, thats the only reason why I put up with huge caps and huge transformer DESPITE the many disadvantages that come with it.
I wish it where that easy ;0)
In class B things are different because good dc regulation is important, thats the only reason why I put up with huge caps and huge transformer DESPITE the many disadvantages that come with it.
because only about 1.8 times AC rating is required to get the dc power you need.
In class B a ratio of 1.8 would be JUST enough, but if you want a stiff dc you need a much higher ratio. You need much more caps and a much bigger transformer. If the transformer I designed would be for class A it would handle easely 1,2kW dc. But since it is used in a class B amplifier it is loade with only about 60% of that.
In class B a ratio of 1.8 would be JUST enough, but if you want a stiff dc you need a much higher ratio. You need much more caps and a much bigger transformer. If the transformer I designed would be for class A it would handle easely 1,2kW dc. But since it is used in a class B amplifier it is loade with only about 60% of that.
AudioSan, hopefull this explanation makes more sense to you:
Inrush current is the maximum if we have remanence in the core at switch on. Under this worst case situation the limiting factors are:
dc resistans and leakage inductans in the primary circuit.
We can forget about the secondary and the caps as long as the core is saturated.
What is good for class A is not necessarely good for class B.
And the taken decisions influence how a PSU behave at switch on and what steps are necessary to deal with possible inrushcurrent.
In class B we have large current variations. To keep the resulting voltage variations down we have to try to get the lowest possible charging circuit impedans.
This translates into low winding resistance, low transformer temperature, low leakage inductance , low diode forward voltage drop variation, highish rectifier temperature at current maximum, BIG caps with low ESR and ESL and we need to watch out to get a minimum inductance charging current loop.
Transformer related issues have to be solved with more copper.
More copper means bigger core. I found that a big toroid (with more than usual copper on it) is the most rewarding option and can giv below 2,5% regulation at max temp rise
even at 20% lower flux than normally used.
Despite the fluxreduction, all of this has many disadvantages that had to be traded in
to get the good dc regulation needed for class B operation. The fluxreduction I choosed was merely to reduce strays and the ferrostrictive induced noise and was not nearly enough to get rid of possible core saturation under worstcase switch on conditions.
Goes without saying that further filtering is out of the question in a stiff class B supply.
Class A is a lot easier because A means there is almost no current variation the voltage stays stable also with quite lousy dc-regulation.
A powersupply for a class A amp has mainly thermal issues and a (comparable) small transformer with highish regulation will do. No toroid required and normal EI or similar may be even of advantage because the easy of additional shielding. They have by nature a little airgap wich increases leakage inductance, limits inrush current and increases chargingtime and decreases peak currents. The caps can be comparable
small because ripplereduction is better done by adding additional RC or LC filters.
The smaller input-cap has many benefits compared to a overly big cap, like:
lower currents in each and every part of the PSU (before the filterchain).
current related issues like losses, harmonic content, strayfield a.s.o are less troublesome and much easier dealt with and this all translates into better efficience, less component stress and longer life expectation (or higher allowed temperature).
So yes, a PSU can be designed for a quite big class A amplifier without necessarely needing a inrushcurrent limiting circuit or a shortcutted NTC.
But 350W dc is not a big A amplifier it is a monster.
Still, a 630VA transformer with 5% regulation would be more than enough if the caps
are choosen to provide a formfaktor up to 1.8
With 5% regulation and the only currentlimiting factor being the cold dc-resistance of the primary the 10mS halfsine peak at 230Vac +10% would have a theoretical peak of 170A. Offcourse this wont ever happen and the peak will always be considerable lower.
A normal 10A diazed houshold fuse can handle a 10ms halfsine of 250A in comparison.
The fuse would not blow under this conditions.
But offcourse for your monsteramp with those huge caps and a 1000VA with propable below 4%regulation you need a current limiter no question about it.
I hope things make now sense to you
Inrush current is the maximum if we have remanence in the core at switch on. Under this worst case situation the limiting factors are:
dc resistans and leakage inductans in the primary circuit.
We can forget about the secondary and the caps as long as the core is saturated.
What is good for class A is not necessarely good for class B.
And the taken decisions influence how a PSU behave at switch on and what steps are necessary to deal with possible inrushcurrent.
In class B we have large current variations. To keep the resulting voltage variations down we have to try to get the lowest possible charging circuit impedans.
This translates into low winding resistance, low transformer temperature, low leakage inductance , low diode forward voltage drop variation, highish rectifier temperature at current maximum, BIG caps with low ESR and ESL and we need to watch out to get a minimum inductance charging current loop.
Transformer related issues have to be solved with more copper.
More copper means bigger core. I found that a big toroid (with more than usual copper on it) is the most rewarding option and can giv below 2,5% regulation at max temp rise
even at 20% lower flux than normally used.
Despite the fluxreduction, all of this has many disadvantages that had to be traded in
to get the good dc regulation needed for class B operation. The fluxreduction I choosed was merely to reduce strays and the ferrostrictive induced noise and was not nearly enough to get rid of possible core saturation under worstcase switch on conditions.
Goes without saying that further filtering is out of the question in a stiff class B supply.
Class A is a lot easier because A means there is almost no current variation the voltage stays stable also with quite lousy dc-regulation.
A powersupply for a class A amp has mainly thermal issues and a (comparable) small transformer with highish regulation will do. No toroid required and normal EI or similar may be even of advantage because the easy of additional shielding. They have by nature a little airgap wich increases leakage inductance, limits inrush current and increases chargingtime and decreases peak currents. The caps can be comparable
small because ripplereduction is better done by adding additional RC or LC filters.
The smaller input-cap has many benefits compared to a overly big cap, like:
lower currents in each and every part of the PSU (before the filterchain).
current related issues like losses, harmonic content, strayfield a.s.o are less troublesome and much easier dealt with and this all translates into better efficience, less component stress and longer life expectation (or higher allowed temperature).
So yes, a PSU can be designed for a quite big class A amplifier without necessarely needing a inrushcurrent limiting circuit or a shortcutted NTC.
But 350W dc is not a big A amplifier it is a monster.
Still, a 630VA transformer with 5% regulation would be more than enough if the caps
are choosen to provide a formfaktor up to 1.8
With 5% regulation and the only currentlimiting factor being the cold dc-resistance of the primary the 10mS halfsine peak at 230Vac +10% would have a theoretical peak of 170A. Offcourse this wont ever happen and the peak will always be considerable lower.
A normal 10A diazed houshold fuse can handle a 10ms halfsine of 250A in comparison.
The fuse would not blow under this conditions.
But offcourse for your monsteramp with those huge caps and a 1000VA with propable below 4%regulation you need a current limiter no question about it.
I hope things make now sense to you
your theory sounds nice and all that. but theory and real life are not the same.
if you build a class A amp with minimum transformer and minimum capbank, the life spand for theamp will be minimum as well. again, we are back to the killer temp. if you use a 300VA transformer to draw 300W countunius. you will kill it. the transformer will have a temp of 125-135celsius at all times. and will not behave stable. smal capbank will give you spank on your *** the first time you start your washing machine or your refrigerator turns on, and the grid gets a dip. you don't longer have the current bank to back it up.
you talkt earlier about how grid dip and peaks was so dangerous for hifi.
you just created it!
if you build a class A amp with minimum transformer and minimum capbank, the life spand for theamp will be minimum as well. again, we are back to the killer temp. if you use a 300VA transformer to draw 300W countunius. you will kill it. the transformer will have a temp of 125-135celsius at all times. and will not behave stable. smal capbank will give you spank on your *** the first time you start your washing machine or your refrigerator turns on, and the grid gets a dip. you don't longer have the current bank to back it up.
you talkt earlier about how grid dip and peaks was so dangerous for hifi.
you just created it!
- Status
- Not open for further replies.