I'm looking to build another power amp and would like to consider using either a CLC (Pi) or CRC filter. It's for +35V - 0V - -35V 100W amp. Three questions:
1. Is it possible/sensible (legal even) to use inductors designed for crossovers?
2. How do I calculate the required inductance value
3. Is it sensible to have a high value cap first, then the inductor then a lower value cap?
Also (sorry if I'm asking too much) how do I calculate R value and power in a CRC filter.
Thanks
1. Is it possible/sensible (legal even) to use inductors designed for crossovers?
2. How do I calculate the required inductance value
3. Is it sensible to have a high value cap first, then the inductor then a lower value cap?
Also (sorry if I'm asking too much) how do I calculate R value and power in a CRC filter.
Thanks
Let's see your first amp!
Why not squeeze in a CRC or CLC in there first?
1. Yes!
2. Bigger is better
3. Yes, or the same
Why not squeeze in a CRC or CLC in there first?
1. Yes!
2. Bigger is better
3. Yes, or the same
https://www.diyaudio.com/forums/pass-labs/101993-pi-filter-clc-vs-crc-vs.html#post1211990
Just read Grey's (GRollins) advice.
Just read Grey's (GRollins) advice.
Thanks
AS you can (hopefully) see there is no room for extra, well anything really. But a really helpful response. Just wondering what is big in inductance terms - familiar with capacitance but less so inductors when it comes to power supplies.
AS you can (hopefully) see there is no room for extra, well anything really. But a really helpful response. Just wondering what is big in inductance terms - familiar with capacitance but less so inductors when it comes to power supplies.
Attachments
Last edited:
I did see GRollins post before posting mine actually. I'm surprised there is not more in this thread
Inductors in 120Hz power supply filters will have large values, in the order of Henries.
These are much larger values than found in speaker crossovers, which are in the order of mH.
These are much larger values than found in speaker crossovers, which are in the order of mH.
I was under the impression that the inductors were there to reduce high frequency noise on the power lines. Wouldn't a few mH do that?
Don't get sucked down that psud hole.
All of the inductors I ever used in CLC were xo inductors in the mH range. One even has a steel core.
You just want to be sure the inductor doesn't get too hot from the heat (P=V^2/R) of its internal DC resistance, and that you account for the voltage drop in the rails.
All of the inductors I ever used in CLC were xo inductors in the mH range. One even has a steel core.
You just want to be sure the inductor doesn't get too hot from the heat (P=V^2/R) of its internal DC resistance, and that you account for the voltage drop in the rails.
Say you want a low pass filter with the corner at 12Hz to significantly attenuate 120Hz ripple.
Then 12Hz = 1 / (2Pi x sqrt(L x C) )
If C = 20uF, after rearranging we have:
L = 1 / (20uF x (2Pi x 12)^2) = 8.8 Henries
Here's the filter inductor in the Dynaco Stereo 70 amplifier, value 1.75 Henries.
C-354 CHOKE - Dynakit Parts
Then 12Hz = 1 / (2Pi x sqrt(L x C) )
If C = 20uF, after rearranging we have:
L = 1 / (20uF x (2Pi x 12)^2) = 8.8 Henries
Here's the filter inductor in the Dynaco Stereo 70 amplifier, value 1.75 Henries.
C-354 CHOKE - Dynakit Parts
Last edited:
Yes, that's for a solid state amplifier, not tubes as in the example.
Tube circuits use capacitors around 1000 times smaller in capacity.
Very few solid state amplifiers use inductors in their power supply.
Tube circuits use capacitors around 1000 times smaller in capacity.
Very few solid state amplifiers use inductors in their power supply.
Last edited:
As a simple example a 300ohm + 470uF as a RC filter gets to 1Hz as a low pass with a suit case of issues - the current for capacitor, the current through the resistor (at inrush) and the voltage drop 300ohms gives plenty to think about. Also 470uF wouldn’t go well with valve regulators.
Just been adjusting my design to limit the current values at inrush and during operation.
Still not entirely happy with it yet but getting there.
Just been adjusting my design to limit the current values at inrush and during operation.
Still not entirely happy with it yet but getting there.
Power filter frequency is lower, so you want more H.
Power filter impedance is higher, so you want more H. (400V 100mA is like 4,00 ohms, rather than 4r-16r.)
You want about 100 times the inductance. You can do this by taking the wire off, slicing it to 1/10th cross-section (1/3rd diameter) (buy new wire!), and re-winding.
If space is tight you sure want modern e-caps not iron cores.
R-C filter: figure how much voltage drop you can tolerate. More drop is better filtering but less audio output. Figure out how much ripple filtering you *really* need (no blue-sky numbers). Do math.
I think chokes could be used more in linear power supplies if quality is an important factor. I know they cost a lot and take a lot of space, but sometimes it seems DC chokes are almost totally forgotten in linear applications (switched power supplies are another matter).
Hammond still produce a line of DC chokes also for larger current, up to 10A. D.C. Reactors or filter chokes (153-159 Series) - Hammond Mfg.. These are designed for DC with an air gap (without air gap inductance would drop fast due to core saturation).
I'm planning on using the 159ZJ (10 mH 5A) in a 5A 30V supply together with 10mF + 10mF (low ESR). This will reduce ripple with a great order of magnitude and the voltage drop is small due to 0.16 ohm resistance. The cutoff frequency is 11 Hz. (This circuit will also have a regulator after, but that's another story). Simulated also in PSUD.
What got me wondering is some articles warning about inductance "ringing" or self resonance during transients. I will try to test this. I have seen some circuits implementing a "snubber" on the choke.
Here is someone confirming there could be ringing:
CLC vs. CRC
In this case it seems to help with a series resistor, if resistance of the choke is too low (though a bit wasteful solution).
Just an idea. Could a parallel RC filter be used and the Quasimodo circuit be used to determine snubber values, if there indeed were a ringing problem?
Hammond still produce a line of DC chokes also for larger current, up to 10A. D.C. Reactors or filter chokes (153-159 Series) - Hammond Mfg.. These are designed for DC with an air gap (without air gap inductance would drop fast due to core saturation).
I'm planning on using the 159ZJ (10 mH 5A) in a 5A 30V supply together with 10mF + 10mF (low ESR). This will reduce ripple with a great order of magnitude and the voltage drop is small due to 0.16 ohm resistance. The cutoff frequency is 11 Hz. (This circuit will also have a regulator after, but that's another story). Simulated also in PSUD.
What got me wondering is some articles warning about inductance "ringing" or self resonance during transients. I will try to test this. I have seen some circuits implementing a "snubber" on the choke.
Here is someone confirming there could be ringing:
CLC vs. CRC
In this case it seems to help with a series resistor, if resistance of the choke is too low (though a bit wasteful solution).
Just an idea. Could a parallel RC filter be used and the Quasimodo circuit be used to determine snubber values, if there indeed were a ringing problem?