I was looking into the idea of using a capacitance multiplier supply based from - http://sound.westhost.com/project15.htm
would this - "go up in capacitance value, you go up in esr and inductance of the cap" still apply since the cap will be a fairly small value but multiplied by the transistor hfe.
also how does snubbing the bridge diodes apply when using fancy diodes i.e schottky hexfets
i've been using this simulation to test out diodes with the snubber network - here
the output plots look like this - here
would this - "go up in capacitance value, you go up in esr and inductance of the cap" still apply since the cap will be a fairly small value but multiplied by the transistor hfe.
also how does snubbing the bridge diodes apply when using fancy diodes i.e schottky hexfets
i've been using this simulation to test out diodes with the snubber network - here
the output plots look like this - here
Rainwulf said:
Some people say that a role of snubber also to eliminate the inductance of big caps.
Carlosfm, a man who created this thread recommends far more other values and his design sounds fine as I know.
Btw, National engineers told me that "the general snubber recommendation is 10 ohms + 0.1uF to set an upper pole at ~160kHz. Using smaller values for the resistor increases the pole location to a higher frequency".
So, with your recommendation that pole shall be much lower I guess?
Soon I'll test power supply made under your spec except snubber values and tell the results.
In appnote 1849 National also recommends to use snubbers to reduce high frequency noise and ringing of the diode.
Thanks and regards.
You cant eliminate inductance by a resistor and capacitor combination.
A resistor and capacitor in series models.. guess, a high esr capacitor!
Yes, use a model of a high esr capacitor to help reduce the ESI of a capacitor.
No. It doesnt work like that. i have even modeled the snubber network on the end of the capacitor chain as shown in various diagrams here. It does nothing at all, apart from short out rf noise caused by the recitiers.
So just put it back where it belongs, on the rectifiers.
Just in case you are wondering, an RC combination has whats called a time constant.
That time time constant is the time it takes for the voltage to rise to a certain percantage of maxiumum voltage.
In other words, it has a rise time.
Guess what else has rise time.
An inductor. Two circuits in parallel dont help each other at all, as they both have a rise and fall time. You want the smallest rise and fall time, IE, low ESR, or a "FAST" psu. That means getting rid of any extra resistance or inductance in a circuit.
You cant cancel inductance, you can just swamp it with more capacitors.
Preferebly smaller ones that have less inductance!
However, please do make both models and see what happens.
I dont have good enough speakers to see what effects different systems of power supplies have, i spent a whole 15 dollars per driver for my surround computer system.
However, i may be mistaken about that snubber values on the diode, but it also depends on the inductance of the transformer you use too.
Go with the appnote values, probably a better idea.
A resistor and capacitor in series models.. guess, a high esr capacitor!
Yes, use a model of a high esr capacitor to help reduce the ESI of a capacitor.
No. It doesnt work like that. i have even modeled the snubber network on the end of the capacitor chain as shown in various diagrams here. It does nothing at all, apart from short out rf noise caused by the recitiers.
So just put it back where it belongs, on the rectifiers.
Just in case you are wondering, an RC combination has whats called a time constant.
That time time constant is the time it takes for the voltage to rise to a certain percantage of maxiumum voltage.
In other words, it has a rise time.
Guess what else has rise time.
An inductor. Two circuits in parallel dont help each other at all, as they both have a rise and fall time. You want the smallest rise and fall time, IE, low ESR, or a "FAST" psu. That means getting rid of any extra resistance or inductance in a circuit.
You cant cancel inductance, you can just swamp it with more capacitors.
Preferebly smaller ones that have less inductance!
However, please do make both models and see what happens.
I dont have good enough speakers to see what effects different systems of power supplies have, i spent a whole 15 dollars per driver for my surround computer system.
However, i may be mistaken about that snubber values on the diode, but it also depends on the inductance of the transformer you use too.
Go with the appnote values, probably a better idea.
I redesigned my amplifier to Rainwolf recommendations and obtained the following results.
Bass reproduction became more light, effortless and tighter, HF oscilation disappear completely, sound bacame more natural (maybe because, using an opportunity, I also replaced my current resistors to Kiwame). However I leave snubbers and 100nF caps on PSU and added snubbers (1.3Ohm+100nF polystyrene and some polyester caps which I had on hand) on diodes.
Likewise I added 10nF MKP cap to 2.2uF MKP cap as input coupling capacitor.
Initially I had small ceramic caps on PSU and IC power supply pins. Now I replaced ceramic caps to polypropylene caps on IC and PSU in snubbers, the others 100nF caps in PSU are polystyrene caps. It seems that with ceramics HF performance of my amplifier was much better than with film caps. HF sound reproduction with ceramics like small bell ringning and similar was so natural, nice and clear until after some hours of work with closed cover of a case of my amplifier I overheated my low quality ceramic caps and had to replace them. Considering mentioned above I inclined to use ceramics with lowest ESR, ESL where it's not in a signal path.
Bass reproduction became more light, effortless and tighter, HF oscilation disappear completely, sound bacame more natural (maybe because, using an opportunity, I also replaced my current resistors to Kiwame). However I leave snubbers and 100nF caps on PSU and added snubbers (1.3Ohm+100nF polystyrene and some polyester caps which I had on hand) on diodes.
Likewise I added 10nF MKP cap to 2.2uF MKP cap as input coupling capacitor.
Initially I had small ceramic caps on PSU and IC power supply pins. Now I replaced ceramic caps to polypropylene caps on IC and PSU in snubbers, the others 100nF caps in PSU are polystyrene caps. It seems that with ceramics HF performance of my amplifier was much better than with film caps. HF sound reproduction with ceramics like small bell ringning and similar was so natural, nice and clear until after some hours of work with closed cover of a case of my amplifier I overheated my low quality ceramic caps and had to replace them. Considering mentioned above I inclined to use ceramics with lowest ESR, ESL where it's not in a signal path.
I apologize for mistake in typing Rainwulf username.
Btw, when I tested my amplifier designed under mentioned specs I obviously noticed that 3.3uF cap between V+ and V- rails make sound worse. It made amplifier to oscillate in HF range. Without this cap music sounds much more rich and natural. It seems that this cap limit a frequency response of the amp.
Regards.
Btw, when I tested my amplifier designed under mentioned specs I obviously noticed that 3.3uF cap between V+ and V- rails make sound worse. It made amplifier to oscillate in HF range. Without this cap music sounds much more rich and natural. It seems that this cap limit a frequency response of the amp.
Regards.
Thank you for testing my design!
I have designed my own amp along the same procedures, however i have very cheap speakers, and i am saving up for some decent ones.
I think the one thing that i am most impressed with is ZERO hum, even from the subwoofer amp (which isnt a gainclone, but its power supply is designed along the same as the one i mentioned above)
I am very glad it worked for you!
I have designed my own amp along the same procedures, however i have very cheap speakers, and i am saving up for some decent ones.
I think the one thing that i am most impressed with is ZERO hum, even from the subwoofer amp (which isnt a gainclone, but its power supply is designed along the same as the one i mentioned above)
I am very glad it worked for you!
has anyone tried these : Suppressors or similar across the bridge diodes ?
They are 100r 100nf snubbers but in one package. Should make things abit easier to retro fit.
They are 100r 100nf snubbers but in one package. Should make things abit easier to retro fit.
AndriyOL: What diodes are you using for your rectifier? I've heard that diode snubbers make the sound worse on soft-recovery diodes, like the MUR860s, which I plan on using in my next design.
Ted205: Those suppressors look cool, I'm also interested in any reports by people who have used them.
Ted205: Those suppressors look cool, I'm also interested in any reports by people who have used them.
Rainwulf said:
Guess I must thank you for your valuable advices in building PSUs.
Almost everyone come to this forum to find such good experience and scientific support. Some come to donate the experience and support. The last ones deserve a gratefulness and respect.
I also don't have superior speakers. I use car speakers with hand made cabinets. I bought some top available from Sony for 75$ for a pair. They are 100W of rated and 420W of max. rms power, 93db/25-33kHz and sound very well as for me.
Redshift187 said:
I use MIC ßU KBPC3506. I don't know who is a manufacturer. Suppose they aren't soft-recovery.
As far as I know MUR860 are old series diodes. Why don't use better ones?
Regards.
I have 400VA 2x25VAC toroid which suppose to have 16A of current. Do you know MUR860 are suitable for my toroid? Power ratings requirements of diodes for PSU is calculated for one diode rating or diodes which forms a bridge, thus - 8A*4=32A? I have been thinking about MUR1560 or smth. similar.
Regards.
Regards.
AndriyOL said:
Calculate it based on the individual diode rating. All the diodes are not conducting all the time, and when 2 diodes are conducting, they are in series so both pass all the current. So MUR860 would be 8A. But that's only with a (big) heatsink! They're only rated at 3A without a heatsink.
mikesnowdon said:
Well, I do see a lot of "because so-and-so used them they must be the best" around here.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.