Increased peak current of charging through diodes. It sounds more harsh with low ESR and OK with stinking ordinary caps.
😎
And if one were to precede the Low ESR cap/s with a real resistor equivalent to the missing circuit resistance such that the peak current through the charging circuit were the same, would that apples to apples comparison show equivalent sound quality with respect to PSU induced harshness?
Along similar lines, when anyone suggests a CRC supply, I have recommended low quality cap for first C and higher quality cap for last C
As an aside:
I refer to an ordinary Capacitor input filter as an rC and for CRC, I write rCRC.
The significance of that first small value r should not be overlooked.
Along similar lines, when anyone suggests a CRC supply, I have recommended low quality cap for first C and higher quality cap for last C
As an aside:
I refer to an ordinary Capacitor input filter as an rC and for CRC, I write rCRC.
The significance of that first small value r should not be overlooked.
Last edited:
Does not add up to me. The charging current to the caps is dependent on the load on the amp, not the capacitor ESR, or even the capacitance value. The diodes only pass what has been used by the amp. A high ESR cap will produce heat and use more power, so if anything the charging current will be higher to make up the heat lost in the caps.
Low ESR translates to high ripple current capacity. Especially in class A amps you need lots of ripple current capability or you suffer short capacitor life. They heat up and die. One rule of thumb is that you need a ripple rating of at least 5 times the amp load current for a class A.
The continuous ripple current in the first cap of an rCRC filter is very high. Much higher than in subsequent caps.
That demands that sufficient ripple capacity should be provided in the first stage caps.
Otherwise these caps are going to have a much shorter life than the remainder.
This has little to do with esr. Choosing by ripple capacity, is not the same as choosing by esr.
One high ripple capacity 22mF cap rated at 14A ripple may have an esr down around 20milli-ohms. I'm suggesting that the 0r02 is not important. It's the 14A ripple capacity that is important for this duty.
Now pick out a low ripple capacity 22mF cap rated at 8A ripple. This might fail after a much shorter time than the 14A version.
Can we get ripple capacity another way?
Yes.
Choose 10off 2m2F caps rated at 1.8A ripple. The total ripple capacity of the paralleled group is 18A. That is higher than both the single capacitor capabilities. This is likely to result in a good lifetime that might be the best of the three we are using in this hypothetical example.
Now look at the esr of the 2m2F caps. It might typically be around 250 milli-ohms. Did we select these caps by using esr? No. Would we select these caps if had used esr as our criteria? No.
ESR is not what we use when looking for ripple capacity.
That demands that sufficient ripple capacity should be provided in the first stage caps.
Otherwise these caps are going to have a much shorter life than the remainder.
This has little to do with esr. Choosing by ripple capacity, is not the same as choosing by esr.
One high ripple capacity 22mF cap rated at 14A ripple may have an esr down around 20milli-ohms. I'm suggesting that the 0r02 is not important. It's the 14A ripple capacity that is important for this duty.
Now pick out a low ripple capacity 22mF cap rated at 8A ripple. This might fail after a much shorter time than the 14A version.
Can we get ripple capacity another way?
Yes.
Choose 10off 2m2F caps rated at 1.8A ripple. The total ripple capacity of the paralleled group is 18A. That is higher than both the single capacitor capabilities. This is likely to result in a good lifetime that might be the best of the three we are using in this hypothetical example.
Now look at the esr of the 2m2F caps. It might typically be around 250 milli-ohms. Did we select these caps by using esr? No. Would we select these caps if had used esr as our criteria? No.
ESR is not what we use when looking for ripple capacity.
Last edited:
Andrew you are mixing apples and oranges. Ripple current capacity is determined by at least two factors. One is ESR. The lower the ESR the less heat is generated, and the more ripple the capacitor can take. In a given form factor there is a direct relationship between ESR and ripple capability.
However, the second factor is simply heat dissipation. The larger the capacitor diameter they less surface area there is for a given capacitance to dissipate heat. The same with height. A short fat capacitor dissipates heat much more poorly than a tall skinny one.
So when you combine these two factors, yes you may get better cooling and thus more ripple capability when you go to multiple small sized capacitors. But, you are still dissipating more heat than an equivalent low ESR combination. That means the current draw on the bridge is higher, not lower. I don't believe for one second that small variations in current draw have any effect on audio quality, but if it did, then high ESR would result in the higher current draw, not the lower.
The hard facts remain that energy must balance. The current draw at the end of the day is determined by the load, and not even the size of the capacitor.
However, the second factor is simply heat dissipation. The larger the capacitor diameter they less surface area there is for a given capacitance to dissipate heat. The same with height. A short fat capacitor dissipates heat much more poorly than a tall skinny one.
So when you combine these two factors, yes you may get better cooling and thus more ripple capability when you go to multiple small sized capacitors. But, you are still dissipating more heat than an equivalent low ESR combination. That means the current draw on the bridge is higher, not lower. I don't believe for one second that small variations in current draw have any effect on audio quality, but if it did, then high ESR would result in the higher current draw, not the lower.
The hard facts remain that energy must balance. The current draw at the end of the day is determined by the load, and not even the size of the capacitor.
Once again Ron, I cannot agree with your reasoning. You seem to have some misunderstanding of what you are selecting for.
Now if you were a capacitor designer and had a contract to develop a better Ripple Capacity Capacitor, then you will get involved in the internal construction and other fundamental features of the cap. But we are not that.
We choose a cap to meet a duty.
In a rCRC filter, the first cap has high ripple current imposed on it. I still believe the main and sensible selection criteria has to Ripple Capacity, despite your protestations.
You can do all the guessing you like about heat dissipating areas and internal construction, I will rely on the hopefully accurate ripple Capacity specified by the Manufacturer.
Now if you were a capacitor designer and had a contract to develop a better Ripple Capacity Capacitor, then you will get involved in the internal construction and other fundamental features of the cap. But we are not that.
We choose a cap to meet a duty.
In a rCRC filter, the first cap has high ripple current imposed on it. I still believe the main and sensible selection criteria has to Ripple Capacity, despite your protestations.
You can do all the guessing you like about heat dissipating areas and internal construction, I will rely on the hopefully accurate ripple Capacity specified by the Manufacturer.
I'm not at all saying ripple capability is not important. I'm just saying for the same physical size capacitor, I would prefer the one that achieves it with lower ESR. Less heat means longer life. Less heat means less load on the amp. Less heat means longer life for all the other components that share the same box. Capacitor orientation and spacing is also very important. If you cram a whole bunch of small ones together, then your ripple capacity is just theoretical, and actual life will be compromised by poor cooling.
I actually said rCRC.
look back a few posts to:
http://www.diyaudio.com/forums/powe...t-electrolytics-power-supply.html#post2767566
look back a few posts to:
http://www.diyaudio.com/forums/powe...t-electrolytics-power-supply.html#post2767566
I think the best capacitor is the jensen 4 pole capacitor Four Pole Electrolytic Capacitor, Radial, Asymmetrical Pin Config: Reference Audio Mods
I tested many capacitors for PSU and I use a dynamic test with a pulse generator which switch a load to the output for a short time so you can see the transient response on the scoop
I tested many capacitors for PSU and I use a dynamic test with a pulse generator which switch a load to the output for a short time so you can see the transient response on the scoop
Low ESR will mean low heating for a given RMS ripple current. It is not quite as simple as that, because low ESR can also mean slightly higher RMS ripple current for a given DC current draw. The total resistance (transformer+rectifier+ESR) affects the shape of the charging pulse, and hence the ratio between RMS ripple and DC.
If a supply is to be regulated, then it might be worth deliberately adding a little resistance in series with the rectifier, so that the first cap lasts longer due to lower ripple. The voltage drop can be handled by the regulator. You might not be able to do this trick for an unregulated supply as it will degrade intrinsic regulation.
If a supply is to be regulated, then it might be worth deliberately adding a little resistance in series with the rectifier, so that the first cap lasts longer due to lower ripple. The voltage drop can be handled by the regulator. You might not be able to do this trick for an unregulated supply as it will degrade intrinsic regulation.
Jensen have a datasheet about their 4 pole capacitors http://jensencapacitors.com/public/dokumenter/4pole.pdf , at my job I repaired some power 600W AC/DC converters and the jensen capacitors were used in the output
I still do not see any hard specifications on the capacitor. The document is interesting but all it suggests is that it is a low ESR/ESL capacitor. Need to see actual specifications for specific capacitors to determine if it could be any better than other low ESL capacitors in the market.
Also when your intent is to filter and supply power at very high ultrasonic frequencies, it is probably better to just add a metalized film capacitor. And, if you are concerned about distance from the power supply, add it to the PCB or right at the output transistors on the heat sink.
- Status
- Not open for further replies.