Allright you yanks:
"Bob's your uncle" is a way of saying "you're all set" or "you've got it made." It's a catch phrase dating back to 1887, when British Prime Minister Robert Cecil (a.k.a. Lord Salisbury) decided to appoint a certain Arthur Balfour to the prestigious and sensitive post of Chief Secretary for Ireland. Not lost on the British public was the fact that Lord Salisbury just happened to be better known to Arthur Balfour as "Uncle Bob." In the resulting furor over what was seen as an act of blatant nepotism, "Bob's your uncle" became a popular sarcastic comment applied to any situation where the outcome was preordained by favoritism. As the scandal faded in public memory, the phrase lost its edge and became just a synonym for "no problem."
God Bless America
"Bob's your uncle" is a way of saying "you're all set" or "you've got it made." It's a catch phrase dating back to 1887, when British Prime Minister Robert Cecil (a.k.a. Lord Salisbury) decided to appoint a certain Arthur Balfour to the prestigious and sensitive post of Chief Secretary for Ireland. Not lost on the British public was the fact that Lord Salisbury just happened to be better known to Arthur Balfour as "Uncle Bob." In the resulting furor over what was seen as an act of blatant nepotism, "Bob's your uncle" became a popular sarcastic comment applied to any situation where the outcome was preordained by favoritism. As the scandal faded in public memory, the phrase lost its edge and became just a synonym for "no problem."
God Bless America
I hate to change the subject...
(actually, that was good. I didn't know about the origins of one of my favorite sayings, and Karen's words were better than anything I got via Google, though I did get the facts)
....but I'm keeping my eyes open for the parts I'll need for an Aleph-X when I get there, and was wondering about supply capacitance.
I know that some of the boys are talking of hundreds of thousands of uF for each rail, and I just wonder about that relative to Nelson's comments, above:
"First and foremost, when there is adequate capacitance across the supply at the circuit after the series pass element, there is little or no difference sonically. By adequate, I mean like 25,000 uF per rail. Or more"
I tried to find out where / how they (the DIYers) concluded that so much was needed in the Aleph-X thread, but the searching facilities just aren't up to the task for that one.
Is Nelson saying that for an amp of the power most are considering (~200VA draw / channel) that 25,000 or so should be enough? I know of course that more is generally merrier, but we don't need anything approaching 10x that, right?
(actually, that was good. I didn't know about the origins of one of my favorite sayings, and Karen's words were better than anything I got via Google, though I did get the facts)
....but I'm keeping my eyes open for the parts I'll need for an Aleph-X when I get there, and was wondering about supply capacitance.
I know that some of the boys are talking of hundreds of thousands of uF for each rail, and I just wonder about that relative to Nelson's comments, above:
"First and foremost, when there is adequate capacitance across the supply at the circuit after the series pass element, there is little or no difference sonically. By adequate, I mean like 25,000 uF per rail. Or more"
I tried to find out where / how they (the DIYers) concluded that so much was needed in the Aleph-X thread, but the searching facilities just aren't up to the task for that one.
Is Nelson saying that for an amp of the power most are considering (~200VA draw / channel) that 25,000 or so should be enough? I know of course that more is generally merrier, but we don't need anything approaching 10x that, right?
As far as I'm concerned, the more the merrier. In this instance, however, I was referring to my own limited ability to distinguish between CLC, CRC, and a Mosfet Capacitor Multiplier as a regulating element, and how much capacitance after the filter was enough to make them about the same sonically from my subjective standpoint.
This was not a blind test.
This was not a blind test.
Are you guys actually using the value of your chokes?
I note that most of you are considering putting a lot of capacitance before the choke.
Short:
If you have a smallish transformer, you want to have small or no capacitance preceding the choke. If you have a very large transformer, you can afford to put more capacitance at that point but you sway away from having choke effect of the choke (sic!).
Long:
I don't know what current levels you are at etc, but my guess is that having this much capacitance before the choke puts your transformer in a "Capacitance mode" instead of a "choke mode". Thus, you get less utilization of your choke and transformer and still high pulse loading of your transformer comparable to not having a choke in the circuit. (Remember, the purpose of the choke is to lengthen the charge pulse drawn from the transformer. As you add more and more capacitance right after rectifiers, you will end up with the opposite effect, regardless of what comes after. The purpose of the capacitor banks are thus to shorten the charge pulse which is the opposite of our goal of using inductors to reduce the magnitude of the ripple
)
I am not saying this is either good or bad .... that would be a function of the transformer and inductors capacity for handling pulses and other factors -- all of which I do not pretend to know.
I played with PSUD2 (free software for designing power supplies) extensively and was a beta tester for it a few years back. It is great, and I would encourage you all to consider playing with the capacitance levels preceding the inductance, perhaps (very small C of very high quality if any preceding LCLC. I believe you will find it enlightening. Having no capacitance directly after the transformer causes large voltage transients at that point (and required PIV of rectifier), but extends the charge pulse the most, and thus improves the transformers ability to deliver power beyond it's rating (sic!) and reduces magnitude of current
pulses drawn from the grid. In the LCLC scenario above, you might want do to "do a Thorsten (paraphrased, sorry Thorsten)" and have "50% of current delivered by final capacitor and 50% by final inductor", or use other algorithms such as placing most of your capacitance at the end (in order to have maximum current ripple over each inductor and minimize output
impedance) and so maximize the effect of having an inductor in the path.
If you are already passing DC, the effect of the inductor on "magnitude of traditional ripple" is limited at best
Another but unrelated trick which has been already been discussed in another group I am member os is to add small series chokes to the inductor to reduce it's parasitic capacitance seen from either end (and thus improve it's parameters, particularly series capacitance).
Petter
I note that most of you are considering putting a lot of capacitance before the choke.
Short:
If you have a smallish transformer, you want to have small or no capacitance preceding the choke. If you have a very large transformer, you can afford to put more capacitance at that point but you sway away from having choke effect of the choke (sic!).
Long:
I don't know what current levels you are at etc, but my guess is that having this much capacitance before the choke puts your transformer in a "Capacitance mode" instead of a "choke mode". Thus, you get less utilization of your choke and transformer and still high pulse loading of your transformer comparable to not having a choke in the circuit. (Remember, the purpose of the choke is to lengthen the charge pulse drawn from the transformer. As you add more and more capacitance right after rectifiers, you will end up with the opposite effect, regardless of what comes after. The purpose of the capacitor banks are thus to shorten the charge pulse which is the opposite of our goal of using inductors to reduce the magnitude of the ripple
)I am not saying this is either good or bad .... that would be a function of the transformer and inductors capacity for handling pulses and other factors -- all of which I do not pretend to know.
I played with PSUD2 (free software for designing power supplies) extensively and was a beta tester for it a few years back. It is great, and I would encourage you all to consider playing with the capacitance levels preceding the inductance, perhaps (very small C of very high quality if any preceding
LCLC. I believe you will find it enlightening. Having no capacitance directly after the transformer causes large voltage transients at that point (and required PIV of rectifier), but extends the charge pulse the most, and thus improves the transformers ability to deliver power beyond it's rating (sic!) and reduces magnitude of currentpulses drawn from the grid. In the LCLC scenario above, you might want do to "do a Thorsten (paraphrased, sorry Thorsten)" and have "50% of current delivered by final capacitor and 50% by final inductor", or use other algorithms such as placing most of your capacitance at the end (in order to have maximum current ripple over each inductor and minimize output
impedance) and so maximize the effect of having an inductor in the path.
If you are already passing DC, the effect of the inductor on "magnitude of traditional ripple" is limited at best
Another but unrelated trick which has been already been discussed in another group I am member os is to add small series chokes to the inductor to reduce it's parasitic capacitance seen from either end (and thus improve it's parameters, particularly series capacitance).
Petter
Petter,
after reading this thread I´ve stopped my order for four Torobar chokes until it`s clear if the saturation current they mention is at dc or at 1kHz. I´ve sent an e-mail to intertechnik so let´s see if they answer.
I´ve also played a lot with PSU Designer and I still think a choke is a very effective way to reduce ripple without losing too much voltage/wasting energy (as you would with an CRC) filter.
Even a 1mH choke (air core) makes sence as it reduces the ripple voltage from over 500mV to about 60 mV in my configuration (with 10A bias). So if the torobars won´t work I´ll get some air core chokes with low impedance (about 0,18 Ohms) that will fit inside my amps.
william
after reading this thread I´ve stopped my order for four Torobar chokes until it`s clear if the saturation current they mention is at dc or at 1kHz. I´ve sent an e-mail to intertechnik so let´s see if they answer.
I´ve also played a lot with PSU Designer and I still think a choke is a very effective way to reduce ripple without losing too much voltage/wasting energy (as you would with an CRC) filter.
Even a 1mH choke (air core) makes sence as it reduces the ripple voltage from over 500mV to about 60 mV in my configuration (with 10A bias). So if the torobars won´t work I´ll get some air core chokes with low impedance (about 0,18 Ohms) that will fit inside my amps.
william
Re: Are you guys actually using the value of your chokes?
You need to have at least enough that the caps will be reliable at the ripple current, which they alone will see and not the caps after the series pass filter element. Take whatever the draw of the circuit is, and multiply it by 4 or so to get a minimum sort of number for the ripple rating (per rail) of the capacitors, regardless of their capacitance value.
Crime Stopper's Textbook:Petter said:
You need to have at least enough that the caps will be reliable at the ripple current, which they alone will see and not the caps after the series pass filter element. Take whatever the draw of the circuit is, and multiply it by 4 or so to get a minimum sort of number for the ripple rating (per rail) of the capacitors, regardless of their capacitance value.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.