I've been reading various items on this, including the following threads from this esteemed forum: PI-Filter, PS question - CLC vs C filtering, First Pass project, need Aleph-M PSU advice please and proper design of pi-filter for Aleph 30.
There appear to be varying ideas on the usefulness of such filters for the power supply, especially for a class A, balanced amplifier with regulation on the power supply. Nelson Pass' post in the second thread especially seems to indicate this.
And one other thing I've been wondering is, how much of the effect of the filter is caused by the series resistance of the inductor used in the filter, and how much is due to the inductance? Most of the inductors of the 2mH variety would run between 0.2 and 0.5 ohms, which is almost exactly the value of the resistor that you would use if your were going to use a CRC type filter.
So the question, what's more useful in general, CLC, CRC or just plain C?
In my case, looking at a ZV6 with a regulated power supply (like the one described in ZV5), I probably won't bother with the extra space/cost vs performance that the four inductors would give me. And, yes, I know I'm doing a lot of talking without much action, but I'm still buying parts (and I'm a real slow-starter at the best of times - just ask the missus...
).
There appear to be varying ideas on the usefulness of such filters for the power supply, especially for a class A, balanced amplifier with regulation on the power supply. Nelson Pass' post in the second thread especially seems to indicate this.
And one other thing I've been wondering is, how much of the effect of the filter is caused by the series resistance of the inductor used in the filter, and how much is due to the inductance? Most of the inductors of the 2mH variety would run between 0.2 and 0.5 ohms, which is almost exactly the value of the resistor that you would use if your were going to use a CRC type filter.
So the question, what's more useful in general, CLC, CRC or just plain C?
In my case, looking at a ZV6 with a regulated power supply (like the one described in ZV5), I probably won't bother with the extra space/cost vs performance that the four inductors would give me. And, yes, I know I'm doing a lot of talking without much action, but I'm still buying parts (and I'm a real slow-starter at the best of times - just ask the missus...
).
The ripple voltage is of saw teeth shape after the caps.
When the size of PSU caps is getting bigger, the ripple is getting
smaller. And, the additional L or R reduces the ripple further down.
If the L and R have the same ac resistance, they give the
ripple reduction at the same rate. The only difference is that the L
changes the ripple shape to the sine wave while thr R keeps it to
the saw teeth shape. But, I do not know how much these
different shape of ripple voltage will affect the sound, assuming
that the ripple voltage is very low anyhow . . .
When the size of PSU caps is getting bigger, the ripple is getting
smaller. And, the additional L or R reduces the ripple further down.
If the L and R have the same ac resistance, they give the
ripple reduction at the same rate. The only difference is that the L
changes the ripple shape to the sine wave while thr R keeps it to
the saw teeth shape. But, I do not know how much these
different shape of ripple voltage will affect the sound, assuming
that the ripple voltage is very low anyhow . . .
Babowana said:The only difference is that the L changes the ripple shape to the sine wave while thr R keeps it to the saw teeth shape.
. . .
The penny has dropped!
So how do you find these wonderful inductors to do this for you? I guess there is always "roll your own" (now where is that inductor calculator?). As big air-cored inductors seem to cost about $50 each (for a 2mH) in Australia. Maybe Mundorf has a iron-cored one available here for less. And I can't afford to get them sent from the US any more as it's too expensive now
.Thanks, Babowana!
Hi,
you can look at the effect of L and R filters with PSU Designer2 (search).
A CLC filter has a lot more ripple reduction than a CRC one without losing too much voltage and generating too much heat.
Up to 3A of bias you can use a siple 1.8mH, 1.4mm air choke (less than 15€)
For my Aleph-X (7-8A bias) I´m using Torobar coils wich work fine.
William
you can look at the effect of L and R filters with PSU Designer2 (search).
A CLC filter has a lot more ripple reduction than a CRC one without losing too much voltage and generating too much heat.
Up to 3A of bias you can use a siple 1.8mH, 1.4mm air choke (less than 15€)
For my Aleph-X (7-8A bias) I´m using Torobar coils wich work fine.
William
For such a simple subject, this generates far more than its fair share of angst. Never have understood why. Just me being silly, I suppose.
Here's the hierarchy from most best to least best:
1) REGULATED
Pro--excellent isolation from line problems, absolute control over rail voltage
Con--heat, complexity, potential for transient problems (ringing) if not executed properly, potential for dynamic restriction, potential for noise
2) CLC
Pro--good isolation from line problems, very quiet
Con--lack of control over rail voltage, bulky, heavy
3) CRC
Pro--good isolation from line problems, very quiet
Con--lack of control over rail voltage, not as good at eliminating high frequency aberrations as CLC
4) C and Capacitance multiplier
Pro--decent isolation from line problems
Con--lack of control over rail voltage, not as good at eliminating high frequency aberrations as CRC
The list is far from exhaustive. For instance you could have something like a CRCLC-Capacitance multiplier-C. Where would that fit in? Flip a coin. Also, my inclusion of Capacitance multipliers with straight C filters is rather arbitrary. I did so because a capacitance multiplier acts like an arbitrarily large capacitor; in principle, you could duplicate the effect with a really, really big cap. Given that capacitance multipliers are generally used in combination with other things, it gets complicated. If someone wants to put it higher up in the hierarchy, I won't put up much of a fight.
But so-and-so used this filter and what's-his-face used the other filter and they both say their way is the best and I just don't know what to do...
This is one of those human psychology things. If they used it and it worked out okay, to them it's the best thing in the world. Just ask anyone who's ever built a speaker. They always love it, even if it sounds like fingernails on a blackboard. (Beranek's Law)
One general rule of thumb applies: Make it big.
Or, as a fellow I know used to say: If some is good, then more is better, and too much is just enough.
As for inductors vs. resistors, this is covered ad nauseum in every electronics text that's ever been written. Basically, resistors are neutral as far as frequency goes. DC (aka 0Hz) gets attenuated exactly the same as 120Hz, which gets attenuated exactly the same as RF. Inductors attenuate high frequencies more than low frequencies. RF is attenuated more than DC. If a resistor is .5 Ohms and an inductor has a DC resistance of .5 Ohms, then in that limited sense they will behave the same...at DC. But that's not why you use an inductor. If you already had perfect DC, you wouldn't need the extra filtration.
Quit griping and build something. You can always add an L or an R later if you think you need it. Power supplies are easy.
Quote: Nothing is hard, but thinking makes it so. (Meaning if you think it's hard, it will be. Clear your mind and the problem will solve itself.)
Grey
DISCLAIMER: The young 'uns are being particularly demanding. I caught a few typos and oddities, but can't claim that I got them all.
Here's the hierarchy from most best to least best:
1) REGULATED
Pro--excellent isolation from line problems, absolute control over rail voltage
Con--heat, complexity, potential for transient problems (ringing) if not executed properly, potential for dynamic restriction, potential for noise
2) CLC
Pro--good isolation from line problems, very quiet
Con--lack of control over rail voltage, bulky, heavy
3) CRC
Pro--good isolation from line problems, very quiet
Con--lack of control over rail voltage, not as good at eliminating high frequency aberrations as CLC
4) C and Capacitance multiplier
Pro--decent isolation from line problems
Con--lack of control over rail voltage, not as good at eliminating high frequency aberrations as CRC
The list is far from exhaustive. For instance you could have something like a CRCLC-Capacitance multiplier-C. Where would that fit in? Flip a coin. Also, my inclusion of Capacitance multipliers with straight C filters is rather arbitrary. I did so because a capacitance multiplier acts like an arbitrarily large capacitor; in principle, you could duplicate the effect with a really, really big cap. Given that capacitance multipliers are generally used in combination with other things, it gets complicated. If someone wants to put it higher up in the hierarchy, I won't put up much of a fight.
But so-and-so used this filter and what's-his-face used the other filter and they both say their way is the best and I just don't know what to do...
This is one of those human psychology things. If they used it and it worked out okay, to them it's the best thing in the world. Just ask anyone who's ever built a speaker. They always love it, even if it sounds like fingernails on a blackboard. (Beranek's Law)
One general rule of thumb applies: Make it big.
Or, as a fellow I know used to say: If some is good, then more is better, and too much is just enough.
As for inductors vs. resistors, this is covered ad nauseum in every electronics text that's ever been written. Basically, resistors are neutral as far as frequency goes. DC (aka 0Hz) gets attenuated exactly the same as 120Hz, which gets attenuated exactly the same as RF. Inductors attenuate high frequencies more than low frequencies. RF is attenuated more than DC. If a resistor is .5 Ohms and an inductor has a DC resistance of .5 Ohms, then in that limited sense they will behave the same...at DC. But that's not why you use an inductor. If you already had perfect DC, you wouldn't need the extra filtration.
Quit griping and build something. You can always add an L or an R later if you think you need it. Power supplies are easy.
Quote: Nothing is hard, but thinking makes it so. (Meaning if you think it's hard, it will be. Clear your mind and the problem will solve itself.)
Grey
DISCLAIMER: The young 'uns are being particularly demanding. I caught a few typos and oddities, but can't claim that I got them all.
Thanks for your replies everyone. Especially Grey, very clear and erudite. I'll take your advice "Quit griping and build something. You can always add an L or an R later if you think you need it." when the bits arrive. You guys are blessed over there, big heatsinks are rare as hen's teeth over here.
wuffwaff said:
1.4mm is about 15 gauge.
I'm having trouble working out how the DC current handling of an air core inductor can be calculated. Is there any formula to figure this out?
For example, after thinking of CRC from my Mini-A stereo amp project, i've been looking at CLC -- having found 17 gauge inductors available locally. How can I work out how much current these will take?
ssmith said:
First calculate the total length (L) of the wire in cm, and calculate
the cross sectional area (A) of the wire in cm2. As the copper has
the resistivity (RO) of 1.724 (microhm.cm), the DC resistance of
coil is
R = RO*L/A/1000,000 ohms
Or, simply measure resistance using DMM ^^.
With the calculated or measured DC resistance, you could know
the power dissipation in Watt. Your thickness of wire depends on
how much hot (warm) coils in service you want to have. Good luck!
Nelson Pass said:Either a CLC or CRC is still a lot better than C. As an example,
I have here a supply with 30,000 uF after the rectifier, then
.12 ohms, then another 30,000 uF. The difference on either side
of the resistor is about 10 dB worth of ripple, and that's a lot.
thanks a lot for the suggestion Mr Pass.
I was using the following topology CLCLCRCC with R=3.3ohm.
So i'll try to lower this value
I would like to ask you guys, this things:
Since i was suggested that a regulated power supply is, in general if well done, better for a preamplifier
in case, would it be convenient to put the stab stage after clc... section or before?
To me, it lools more convenient t oput it after.
Do you guys have any suggestion on a good topology to implement to stabilyze the rail voltage (without using LM or any chip)?
The rail voltage that i should have is about 60V (the supply for the P1.7).
What ripple value can be considered to be VERY good?
Thanks a lot in advantage.
Best,
Stefano.
ssmith said:
Not sure, but isn't this simply an application of P=I^2*R (current squared times DCR resistance)? So, for me, 8.8amps (squared) times resistance (between .2 and .3) gives somewhere between 15 and 23 Watts.
With the size of the coils I'd be looking at, that would be "not a problem".
Corrections, please...
There is a good explanation of Capacitance Multiplier for those interested:
http://sound.westhost.com/project15.htm
http://sound.westhost.com/project15.htm
Cloth Ears said:
The penny has dropped!
So how do you find these wonderful inductors to do this for you? I guess there is always "roll your own" (now where is that inductor calculator?). As big air-cored inductors seem to cost about $50 each (for a 2mH) in Australia. Maybe Mundorf has a iron-cored one available here for less. And I can't afford to get them sent from the US any more as it's too expensive now
Thanks, Babowana!
The "roll your own" solution is not that much trouble, and hands you a very flexible solution for future projects as well.
Take a look here for a little inspiration.....most of the inductor-stuff you see there can be made on the kitchen table
www.briangt.com/gallery/magura
Magura
Attachments
Considering the sound quality alone, and nothing else; there is nothing like a big CLC supply The second best is the regulated supply! A very good example is to be found here: http://www.passdiy.com/pdf/zen-ver3.pdf Do mind, though that if you make a dual-rail from that PDF, the lower fet must be turned a 180 degrees!! I never thought that I would catch NP in making a mistake..he-he
Steen
The second best is the regulated supply! A very good example is to be found here: http://www.passdiy.com/pdf/zen-ver3.pdf Do mind, though that if you make a dual-rail from that PDF, the lower fet must be turned a 180 degrees!! I never thought that I would catch NP in making a mistake..he-he Steen
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.