take choke between 4mH and 10mH, no more than 0R33 of Rdc, toss it in place of R in your CRC
in case that you're using 33mF caps, already proven, no much thinking
remember that we are dealing with A Class amps here, so no much of rails current modulation
I even remember Pa did use puny I chokes (1mH) in one of bigger PL amps (XA160)
obviously surplus stock of bass chokes needed to go, somewhere
in case that you're using 33mF caps, already proven, no much thinking
remember that we are dealing with A Class amps here, so no much of rails current modulation
I even remember Pa did use puny I chokes (1mH) in one of bigger PL amps (XA160)
obviously surplus stock of bass chokes needed to go, somewhere
I've built three amps, 2xF5 and one Aleph J using the Hammond 159ZJ, 10mH/5A. Its Rdc was very similar to the 4 parallel 0.47 ohm in the previous CRC. After doing the work I compared the results on a scope. Ripple dropped almost 50mV to a few mV and the waveform was not as sharp. Subjectively the clarity and openess in the mids and highs is exceptional. Noise at the 98dB midrange in my towers (F5 actively powers mid and high) is not discernable.
As Zen Mod said, just build it. The part spec I used has been widely shared in the past as the most suitable for a typical power supply used in a diy Firstwatt project.
As Zen Mod said, just build it. The part spec I used has been widely shared in the past as the most suitable for a typical power supply used in a diy Firstwatt project.
hello again 
You know what kids do if you tell them to stop thinking ... they go away and think even more
Sorry if I offended someone or came across a bit too familiar. After reading this forum for some years, I feel like I know some of you a bit.
And me, I had some electronics education (below uni level) and have had this hobby since a teenager. That's when I built my first preamp, simple 2 or 3 bjt stage single ended IIRC. A friend with a good mid end stereo wanted to buy it while it still was on a piece of wood with nails to hold the components and wires in place, looking like a Christmas tree. I have always like siplistic circuits.
Anyway, life happened, I've built some power amps and speakers over the years and now for for different reasons having more time on my hands I want to build some more.
I was collecting parts for M2 ( to be continued or sissified ? ) and Mofo ( to be continued ) for my Alpair 12p's, when I was lucky enough to win Nelson's lottery Vfef, I also have a couple sets of THF51S that need a purpose when time allows.
Now I need to build an amp for the living room. As I don't have space for large Tannoys, horns or open baffels, I need some more 4Ω power. Hence my questions on choke current rating. It will still be 25W classA, but up to 50W in AB.
Anyway, before I leave this, I would like to share my findings regarding these chokes.
Extreme_Boky
My response to you was meant to be; perhaps the sim is good enough for this purpose, and that the parasitics will vary a lot according to the layout of the amp.
If we use the calculator provided by tombo56 we get a resonant frequency between 8.76 and 13.85Hz.
If we were to use 1mH choke, we would need 140mF to get 13.45Hz
If we were to use 2.5mH, we would need 66mF to get 12.39Hz.
The L + C2 is the low pass filter where resonat frequency is calculated.
When we add C1 of the same value as C2 it becomes a pi filter. I suspect that this is when the ringing occurs.
Here is Hammond 159ZL as an example ; 2.5mA 10A DCR=44mΩ
from bottom to top ; cyan red blue green
If we decrease C before L, or increase C after L most of the ringing goes away, more if we increase after C ( 2.5mH needs more C )
The first part of the curve is under startup and the second dip is a stressload of 1-5A for 100ms (0.1 second). Perhaps a bit extreme but needed to expose any ringing under load.
we are talking about; R5 damping resistance.
Here is Hammond 159ZL as an example ; 2.5mA 10A DCR=44mΩ with damping resistor
from bottom to top ; cyan red blue green
I like to see the load curve more or less the same as the CRC. Then I think the ringing is to a minimum. The green here is where it is because it sags less due to more C.
The DCR is the chokes resistance at DC. With this choke it would equal a CRC with R=44mΩ at DC.
I have found that a larger inductance with a smaller DCR will ring more than a larger inductance with a larger DCR.
With the damping resistor we can tune the DCR of the choke.
If we sim the choke 2.5mH/44mΩ +R=100mΩ, together with the same choke with changed DCR to 144mΩ, the resulting curves are identical.
Here is a sim of 159ZJ which seems popular here.
This one looks promising with CLCC.
from bottom to top ; cyan red green
The green one is 33mF + 10mH + 33mF + 33mF.
This is like hiding our best 33mF caps in calm water behind the L, and we can use a more brute of a cap before the L to deal with the hammering ripple from the rectifier.
The square wave you see on the rails is the load square wave.
With music playing we would see the music signal here.
On the green it is smaller due to parallel capacitors; the ESR is halved.
The ESL is also halved. My understanding is that this increases speed of the current through the capacitor.
Music is the current through psu capacitor.
If you read the long but very interesting thread Power Supply Resevoir Size you will find that the music rides on the rails and the music is the current through especially the last psu capacitor modulated by the output transistors / fets. So it it good to have a clean rail voltage.
It is also good to have the last capacitor close to the Amlifier with short tightly twisted wires for low inductance.
This is not only about ripple, but also the fact that the choke soaks up all the high frequency mess from the net and the rectifier.
Here you can see that the output impedance is also improved, with a less hump around 10Hz.
Thanks to ozorfis and tombo56 for helping me set up this spice model.
Perhaps you should consider a CLCC instead of CLCRC.
With CLCC you would reduce the ringing and lower the resonant frequency and output impedance.
I am not an expert in Ltspice, but I find it a good tool to try out new ideas with.
If any of you have been contemplating learning LT spice, this is a simple circuit for you to get started with.
I see that quite a few of you have implemented damping resistor or/and more C after L, so I'm not claiming to have reinvented anything here, just trying to learn and put it into perspective for myself.
I also hope that this can give some of us less experienced builders some inspiration to make choices and start building.
I am including the Ltspice file for you to play with if you want.
In capacitors, don't go any lower than 20nH inductance, as it takes a very good quality capacitor of these sizes to get down to 10mH.
Also, when paralleling capacitors after L, it looks like it is best to use the same capacitors. I have tried different sizes, ESR and ESL in parallel, and with some combinations the square wave can get distorted.
I have also seen that two high quality 10mF in parallel in fron of L can give good results.
I hope I am not far out in the woods picking blueberries with all this. If so, I hope to be corrected.
So what can I say;
try it and measure
try it and listen
but most of all, be sceptical
Now I only have to figure out the current rating of these chokes and then start building
Hasta luego
You know what kids do if you tell them to stop thinking ... they go away and think even more
Sorry if I offended someone or came across a bit too familiar. After reading this forum for some years, I feel like I know some of you a bit.
And me, I had some electronics education (below uni level) and have had this hobby since a teenager. That's when I built my first preamp, simple 2 or 3 bjt stage single ended IIRC. A friend with a good mid end stereo wanted to buy it while it still was on a piece of wood with nails to hold the components and wires in place, looking like a Christmas tree. I have always like siplistic circuits.
Anyway, life happened, I've built some power amps and speakers over the years and now for for different reasons having more time on my hands I want to build some more.
I was collecting parts for M2 ( to be continued or sissified ? ) and Mofo ( to be continued ) for my Alpair 12p's, when I was lucky enough to win Nelson's lottery Vfef, I also have a couple sets of THF51S that need a purpose when time allows.
Now I need to build an amp for the living room. As I don't have space for large Tannoys, horns or open baffels, I need some more 4Ω power. Hence my questions on choke current rating. It will still be 25W classA, but up to 50W in AB.
Anyway, before I leave this, I would like to share my findings regarding these chokes.
Extreme_Boky
My response to you was meant to be; perhaps the sim is good enough for this purpose, and that the parasitics will vary a lot according to the layout of the amp.
thanks
thanks
If we use the calculator provided by tombo56 we get a resonant frequency between 8.76 and 13.85Hz.
If we were to use 1mH choke, we would need 140mF to get 13.45Hz
If we were to use 2.5mH, we would need 66mF to get 12.39Hz.
The L + C2 is the low pass filter where resonat frequency is calculated.
When we add C1 of the same value as C2 it becomes a pi filter. I suspect that this is when the ringing occurs.
Here is Hammond 159ZL as an example ; 2.5mA 10A DCR=44mΩ
from bottom to top ; cyan red blue green
If we decrease C before L, or increase C after L most of the ringing goes away, more if we increase after C ( 2.5mH needs more C )
The first part of the curve is under startup and the second dip is a stressload of 1-5A for 100ms (0.1 second). Perhaps a bit extreme but needed to expose any ringing under load.
we are talking about; R5 damping resistance.
Here is Hammond 159ZL as an example ; 2.5mA 10A DCR=44mΩ with damping resistor
from bottom to top ; cyan red blue green
I like to see the load curve more or less the same as the CRC. Then I think the ringing is to a minimum. The green here is where it is because it sags less due to more C.
The DCR is the chokes resistance at DC. With this choke it would equal a CRC with R=44mΩ at DC.
I have found that a larger inductance with a smaller DCR will ring more than a larger inductance with a larger DCR.
With the damping resistor we can tune the DCR of the choke.
If we sim the choke 2.5mH/44mΩ +R=100mΩ, together with the same choke with changed DCR to 144mΩ, the resulting curves are identical.
Here is a sim of 159ZJ which seems popular here.
This one looks promising with CLCC.
from bottom to top ; cyan red green
The green one is 33mF + 10mH + 33mF + 33mF.
This is like hiding our best 33mF caps in calm water behind the L, and we can use a more brute of a cap before the L to deal with the hammering ripple from the rectifier.
The square wave you see on the rails is the load square wave.
With music playing we would see the music signal here.
On the green it is smaller due to parallel capacitors; the ESR is halved.
The ESL is also halved. My understanding is that this increases speed of the current through the capacitor.
Music is the current through psu capacitor.
If you read the long but very interesting thread Power Supply Resevoir Size you will find that the music rides on the rails and the music is the current through especially the last psu capacitor modulated by the output transistors / fets. So it it good to have a clean rail voltage.
It is also good to have the last capacitor close to the Amlifier with short tightly twisted wires for low inductance.
This is not only about ripple, but also the fact that the choke soaks up all the high frequency mess from the net and the rectifier.
Here you can see that the output impedance is also improved, with a less hump around 10Hz.
Thanks to ozorfis and tombo56 for helping me set up this spice model.
Perhaps you should consider a CLCC instead of CLCRC.
With CLCC you would reduce the ringing and lower the resonant frequency and output impedance.
I am not an expert in Ltspice, but I find it a good tool to try out new ideas with.
If any of you have been contemplating learning LT spice, this is a simple circuit for you to get started with.
I see that quite a few of you have implemented damping resistor or/and more C after L, so I'm not claiming to have reinvented anything here, just trying to learn and put it into perspective for myself.
I also hope that this can give some of us less experienced builders some inspiration to make choices and start building.
I am including the Ltspice file for you to play with if you want.
In capacitors, don't go any lower than 20nH inductance, as it takes a very good quality capacitor of these sizes to get down to 10mH.
Also, when paralleling capacitors after L, it looks like it is best to use the same capacitors. I have tried different sizes, ESR and ESL in parallel, and with some combinations the square wave can get distorted.
I have also seen that two high quality 10mF in parallel in fron of L can give good results.
I hope I am not far out in the woods picking blueberries with all this. If so, I hope to be corrected.
So what can I say;
try it and measure
try it and listen
but most of all, be sceptical
Now I only have to figure out the current rating of these chokes and then start building
Hasta luego
Attachments
Very comprehensive summary of discussed so far and a very good starting point for everyone that is interested to investigate further.
That is a point I agree with. Good power supply is foundation of a good amplifier. That’s why I opted to use super regulators for my amplifier build. CLC is, probably, the best power supply one can get with simple solutions.
That is a point I agree with. Good power supply is foundation of a good amplifier. That’s why I opted to use super regulators for my amplifier build. CLC is, probably, the best power supply one can get with simple solutions.
Thanks a lot both of you
That means a lot coming from the two of you.
Super regulator in power amp, care to share your solution?
I do like the sissy solution though
I have seen a few reports on capacitance multipliers in power amps taking some of the "bloom" away.
Priceless you should get your jester title back
That means a lot coming from the two of you.
Super regulator in power amp,
care to share your solution?I do like the sissy solution though
I have seen a few reports on capacitance multipliers in power amps taking some of the "bloom" away.
Priceless
you should get your jester title back
Super regulator in power amp,
I have seen a few reports on capacitance multipliers in power amps taking some of the "bloom" away.
IMO, capacitance multipliers are failed electronic choke approximations. They are definitely NOT regulators. Cap Mx is excellent at removing ripple, but relies on output capacitor for its output impedance and AC current performance. It can have specific “audio signature”. They will never find a place in my amplifier build.
I want a power supply to be as close as possible like ideal voltage source: zero output impedance, infinite voltage and current slew rate. Not possible of course, but it can behave like one for an audio bandwidth amplifier.
Jung super regulator is enough close to that ideal. I’ve made my version for high current, up to 10 A, and wide voltage range. It works in my LuDEF. However, parts availability is terrible for almost two years now, and doesn’t improve. Most parts for regulator were available a week ago, and some now missing will arrive in few weeks. After initial design posting, parts were available only for a few days, and since then, all parts never could can be found at the same moment or one source. It’s a DOA project. For an academic consideration, here’s the design: R21 PS module
Industrial high power wire wound R with heatsink is the best for PSU with limited space and on budget
that looks better
Yeah...
Yeah...
Re inductors in power supply..... I have mixed feelings.... no doubt the high/mid spectrum is cleaned-up, but the bass region is much more difficult to get right... it requires a full retune of other little spots/things...
Thanks for sharing your R21_Positive & R21_Negative designs. Very nice of you.
Hi, a very nice design with impressive mesurements you have here
PSSR, noise, transient responce and output impedance seem to be in a different league.
A pita with parts shortages.
Did you or anyone else compare this with CLC.
If time allows, would you mind sharing some more details on you findings regarding this.Could this be due to the low frequency ringing we get if we are close to a pi filter ?
Could it be due to the peak in output impedance at low frequency ?
Could it be due to current saturation in the choke ?
What I have found so far is that a choke's current rating should be bias current + max expected peak ac current (please correct me if this is wrong)
Does anyone know why the FW amps do not have local power rails' decoupling... (a capacitors' combo that will provide a low impedance power source, close to MOSFETs... well, and for the JFETs for that matter) I've been trying to answer this question for more than a year now and unfortunately, I always conclude that it was the measure taken to differentiate the FW amps range from the Pass Labs amps product range. Of course, I might be wrong here...
.....but then the question stays unanswered....
I'd love to see what others think; or even if anyone has given this fact a thought (or 2).
.....but then the question stays unanswered....
I'd love to see what others think; or even if anyone has given this fact a thought (or 2).
No, as far as I know. I went straight with super regulators. Let’s not turn this thread to CRC vs CLC vs Super Reg, but simply, there is no competition with super regulators. During loud music reproduction, I can’t measure even a fraction of mV voltage variation. There is no point to describe incredible sound stage, details and definition I perceive. Anyway, those can be primarily attributed to the terrific amplifier design (LuDEF) and loudspeakers. PS role is only to enable power amp to show its abilities in the full glory, not to hold it back.
I hope that someone will manage to put R21 regulators in his existing amplifier, where only single PS is used for both channels. I expect that could turn into the jaw dropping experience.
Could be worthy of a mod as some peolple think that's the most important cap in the amp