Hi Rob, I suspect adding a snubber circuit could benefit my circumstances, though what is gained on the one hand is lost, to some degree, by what is given away on the other. Reducing tank resonance will eliminate some unwanted AC components, though the resistors will add their own noise and will attenuate the relative gain offered by teflon vs electrolytic dielectric. I think I haven't the patience to work out a snubber for my various applications, especially given my general viewpoint above.
Speaking to that viewpoint, and for what it's worth, I've owned some very good equipment in my many years chasing this hobby --- Edge, Viva, cj, Emm Labs, Manger, Elgar etc --- and I have seen over and over the hard limit parts quality and parts count place on the performance of any given circuit design. The compromising effect of parts quality and quantity imply each other: every part distorts, and the more parts in a given circuit, the more overall distortion, all other things being equal. The larger question lurking in all this is whether improved parts quality offers greater gains than improved circuit design. My experience, regardless the answer to this question, is that best-quality parts offer a degree of improvement I'm unwilling to do without. I am now so finely attuned to the sound of electrolytics --- a dry, dull, uninvolving hazy sludge that rides upon audio signals --- that I refuse to listen to electrolytics-only equipment. My Emm Labs DAC is the latest example of this, and I can't wait to have a qualified modder help me improve its insides.
Speaking to that viewpoint, and for what it's worth, I've owned some very good equipment in my many years chasing this hobby --- Edge, Viva, cj, Emm Labs, Manger, Elgar etc --- and I have seen over and over the hard limit parts quality and parts count place on the performance of any given circuit design. The compromising effect of parts quality and quantity imply each other: every part distorts, and the more parts in a given circuit, the more overall distortion, all other things being equal. The larger question lurking in all this is whether improved parts quality offers greater gains than improved circuit design. My experience, regardless the answer to this question, is that best-quality parts offer a degree of improvement I'm unwilling to do without. I am now so finely attuned to the sound of electrolytics --- a dry, dull, uninvolving hazy sludge that rides upon audio signals --- that I refuse to listen to electrolytics-only equipment. My Emm Labs DAC is the latest example of this, and I can't wait to have a qualified modder help me improve its insides.
ALW said:
Andy-yes the value of resistor for a snubber tends to fall in the range of a fraction of one ohm up to a few ohms at most in my experience. And one could always buy vishays if worried about resistor noise!
Tom, re your comment about mains noise, I think one of the effects a power supply resonance MAY have is to increase an amplifiers susceptibility to mains (or power supply) noise, particularly as the resonance tends to be in a frequency region where the amp has little or no power supply rejection. Then there is the possibility of causing instability in amps with overall negative feedback...
Rob
Mains noise is one form of noise I consider. Another is noise generated by capacitors and other passive components. That kind of noise, especially that arising from electrolytics, is to me quite noticeable. I suspect bypass capacitors of sufficiently low dielectric constant are about the only thing that can appreciably reduce it.
Re: Bypass caps- No
As a long time lurker here, please allow me to say thanks to all who share their expertise, and voice a question here with respect to power supply design. I have seen where some suggest a very small value resistor series'd between the paralleled filter caps. I assume this would lower the Q of the resonant circuit and damp the resonance, and provide some amount of isolation, but of course would limit peak current as well. What about a small value choke in series, perhaps even a ferrite bead with a few turns through it? Would this potentially clean up some of the AR/RF trash on the rails, or simply sum with the ESL and change the frequency of resonance? Just thinking out loud here…
C
Robert F said:
As a long time lurker here, please allow me to say thanks to all who share their expertise, and voice a question here with respect to power supply design. I have seen where some suggest a very small value resistor series'd between the paralleled filter caps. I assume this would lower the Q of the resonant circuit and damp the resonance, and provide some amount of isolation, but of course would limit peak current as well. What about a small value choke in series, perhaps even a ferrite bead with a few turns through it? Would this potentially clean up some of the AR/RF trash on the rails, or simply sum with the ESL and change the frequency of resonance? Just thinking out loud here…
C
Re: Re: Bypass caps- No
Yes, I think it should work, actually want to try it. I want to buy a couple of ferrite cores and wind the GND, - and + wire through the same core in the same direction. Want to experiment a bit with the inductance value so probably try a few different cores to play with. In my case, I want to use it to provide extra HF isolation between an SMPS and an additional capacitor bank.
Best regards
Gertjan
cc00541 said:
Yes, I think it should work, actually want to try it. I want to buy a couple of ferrite cores and wind the GND, - and + wire through the same core in the same direction. Want to experiment a bit with the inductance value so probably try a few different cores to play with. In my case, I want to use it to provide extra HF isolation between an SMPS and an additional capacitor bank.
Best regards
Gertjan
ok, back on track!
ive put together my list of the rest of the parts i need. it may make some people cringe, grab for the keyboard and begin typing feverishly, but oh well.
im just gonna go cheap, with the idea of upgrading later. i just have NO cash right now. ive even scrapped doing stereo amps, im just doing two monos. ill build them so i can just slap in the extra module when i get the cash.
if anyone has any obvious suggestions, im open to it. but if its going to cost significantly more, im probably not too interested.
first off, i plan on modding the module. for one reason, its cheap. for both modules, my bill looks to be around $30. i can handle that.
i plan on replacing the PSU caps with Panasonic FC, 680uf 100v caps. people seem to like these, and they are cheap. then, the boostrap cap, the black gate standard 220uf 16v is only $3 each from percyaudio. i will remove the input caps, i dont need them. i will then replace the input decoupling caps with black gate PK, 22uf, 50V. they are a whopping $0.95 each. so, the total for those little mods is around $30 for BOTH modules.
next, the PSU. here is where im going nice and cheap. for each mono, i will keep my 1.7KVA transformers. since they are wired for dual secondaries, ill stick with dual bridge rectifiers. im going to stick with generics for now. i can get 400V 25A bridges for $1.50 each. and i think i have tons of them in my parts bin.
i was playing around with my preamp project thats giving me headaches and i put the PSU on the scope. it was pretty clean. so,i decided to play around with different configurations to see what noisy looks like and what clean looks like. long story short, i found out that if i threw some .33uf wima caps on the rectifiers, the noise went almost completely away. it was pretty cool to see. so, for a couple of bucks, why not add some small value caps across the bridges. since bruno suggests 47nf, im going to go with some wima MKP2 0.047uf's. i need 16 and they are only $0.42 each. another cheap mod.
for the main PSU caps im going to go with the cornell dublier's from apexjr. they are 12,000uf 80v and only $5 each. two of those per channel will do me. i thought about bypassing them, etc. but, bruno and others say that bypassing doesnt do a whole lot. other people say it does. people say that you need to start with good caps to make bypassing worthwhile. i say, screw it. ill do it when i can afford a good supply.
so, thats what ive come up with. i get my cases this week (maybe tomorrow), and thats when im really going to move forward. if you have suggestions on some small and cheap things i could do to improve the sound, id be interested.
ive put together my list of the rest of the parts i need. it may make some people cringe, grab for the keyboard and begin typing feverishly, but oh well.
im just gonna go cheap, with the idea of upgrading later. i just have NO cash right now. ive even scrapped doing stereo amps, im just doing two monos. ill build them so i can just slap in the extra module when i get the cash.
if anyone has any obvious suggestions, im open to it. but if its going to cost significantly more, im probably not too interested.
first off, i plan on modding the module. for one reason, its cheap. for both modules, my bill looks to be around $30. i can handle that.
i plan on replacing the PSU caps with Panasonic FC, 680uf 100v caps. people seem to like these, and they are cheap. then, the boostrap cap, the black gate standard 220uf 16v is only $3 each from percyaudio. i will remove the input caps, i dont need them. i will then replace the input decoupling caps with black gate PK, 22uf, 50V. they are a whopping $0.95 each. so, the total for those little mods is around $30 for BOTH modules.
next, the PSU. here is where im going nice and cheap. for each mono, i will keep my 1.7KVA transformers. since they are wired for dual secondaries, ill stick with dual bridge rectifiers. im going to stick with generics for now. i can get 400V 25A bridges for $1.50 each. and i think i have tons of them in my parts bin.
i was playing around with my preamp project thats giving me headaches and i put the PSU on the scope. it was pretty clean. so,i decided to play around with different configurations to see what noisy looks like and what clean looks like. long story short, i found out that if i threw some .33uf wima caps on the rectifiers, the noise went almost completely away. it was pretty cool to see. so, for a couple of bucks, why not add some small value caps across the bridges. since bruno suggests 47nf, im going to go with some wima MKP2 0.047uf's. i need 16 and they are only $0.42 each. another cheap mod.
for the main PSU caps im going to go with the cornell dublier's from apexjr. they are 12,000uf 80v and only $5 each. two of those per channel will do me. i thought about bypassing them, etc. but, bruno and others say that bypassing doesnt do a whole lot. other people say it does. people say that you need to start with good caps to make bypassing worthwhile. i say, screw it. ill do it when i can afford a good supply.
so, thats what ive come up with. i get my cases this week (maybe tomorrow), and thats when im really going to move forward. if you have suggestions on some small and cheap things i could do to improve the sound, id be interested.
Maybe you don't want my 2 cents worth but....
1. (in no order of importance)
"snubbing" the supply is a good thing, always. With good caps I've found it may not be required, I imagine propely snubbing it will improve things further, and just slapping on whatever value like wouldn't.
Given my setup, I'm happy with it unsnubbed.
With your 5$ industrial caps, snubbing may be far more critical to get good sound out of them, and it's cheap enough as well.
2. Seriously, reconsider those 50V black gates. Those caps should not be replaced.
3. Don't throw a black gate in there just to say you have a black gate in there, the mod for the bootstrap cap is an OSCON for the lowest ESR possible.
Regards,
Chris
1. (in no order of importance)
"snubbing" the supply is a good thing, always. With good caps I've found it may not be required, I imagine propely snubbing it will improve things further, and just slapping on whatever value like wouldn't.
Given my setup, I'm happy with it unsnubbed.
With your 5$ industrial caps, snubbing may be far more critical to get good sound out of them, and it's cheap enough as well.
2. Seriously, reconsider those 50V black gates. Those caps should not be replaced.
3. Don't throw a black gate in there just to say you have a black gate in there, the mod for the bootstrap cap is an OSCON for the lowest ESR possible.
Regards,
Chris
classd4sure said:
1. ok, so im good with the snubber caps. and 47nf is supposedly the correct value. so, im good there.
2. ok, so those decoupling caps shouldnt be replaced? fair enough. i just thought if it was cheap, why not.
3. ok, so i need an oscon for that place. any ideas where i can get them? i checked the usuals and couldnt find where to get them.
so beyond that, is there anything else i should consider?
cowanrg said:
1. Yeah for the rectifiers. I was talking about the snubbers used after the reservoir caps, something like 100nF//R+C, in order to compensate for inductance. That will change the sound of your caps immensly. So try it without, and if you find bandwidth or frequency response is a little weird, give that a shot.
2. Not those ones, they're part of the damping network and their high ESR is what they're there for.
3. There's a 180uF 20V oscon available at partsconnexion. It's not cheap for what it is. Then you have to ask yourself, is it better to have the extra capacitance (normally 220uF) or the lower ESR? Unless you can find a 220uF elsewhere. That's one I'll probably leave stock myself.
Maybe also consider the output filter cap, that's cheap too and should be worthwhile.
Regards,
Chris
classd4sure said:
i was avoiding the output filter cap because i was just completely unsure what to put there. it seems a lot of people are adding big caps to the backside, etc. i dont mind modding the module, but i just didnt see anything where people had concluded what was good and what wasnt... i dont want to be a test subject.
as far as the oscon, would 180uf be that bad? its not THAT much lower than the stock part...
and what you are saying about the caps on the reftifiers, you are saying that adding a RC networks to the supply caps needs to be tried and can alter the sound, or are you referring to the rectifiers? i didnt think adding them to the rectifiers had as much to do with sound, it was more about lowering the noise caused by the diodes...
im still learning all this stuff. i know how to use tools, not design circuits
Hi
I find info available on the hotrod thread covers the output filter cap rather well.
I'll just be getting the vishay cap. Nothing exotic. Can't think of the # off hand.
About the bootstrap cap, that oscon would probably be fine. I dont' know if you'd notice any change in sound at all though.
You're dead on with the rectifiers. I think we confused ourselves there, snub them like you were going to.
This is what I was refering to in particular:
"for the main PSU caps im going to go with the cornell dublier's from apexjr. they are 12,000uf 80v and only $5 each. two of those per channel will do me. i thought about bypassing them, etc. but, bruno and others say that bypassing doesnt do a whole lot. other people say it does. people say that you need to start with good caps to make bypassing worthwhile. i say, screw it. ill do it when i can afford a good supply."
I think it will be more critical you do snub them with industrial type caps, but yeah I'd try it without first too. Snubbing them isn't just a parallel cap though, it would resonate. Ideally it would be designed for the supply as it's a compensation network, which is why I'm not just throwing whatever values in there for mine. I think it would improve it further though if I did. Standard snubbing values are something like a 100nF poly or equivalent in parallel with 0.5R + 22nF?
So yeah that little network can change the sound of your caps entirely.
We're all learning. Did you make a soft start yet? I'm still using the ol 100W..... lazy!
Regards,
Chris
I find info available on the hotrod thread covers the output filter cap rather well.
I'll just be getting the vishay cap. Nothing exotic. Can't think of the # off hand.
About the bootstrap cap, that oscon would probably be fine. I dont' know if you'd notice any change in sound at all though.
You're dead on with the rectifiers. I think we confused ourselves there, snub them like you were going to.
This is what I was refering to in particular:
"for the main PSU caps im going to go with the cornell dublier's from apexjr. they are 12,000uf 80v and only $5 each. two of those per channel will do me. i thought about bypassing them, etc. but, bruno and others say that bypassing doesnt do a whole lot. other people say it does. people say that you need to start with good caps to make bypassing worthwhile. i say, screw it. ill do it when i can afford a good supply."
I think it will be more critical you do snub them with industrial type caps, but yeah I'd try it without first too. Snubbing them isn't just a parallel cap though, it would resonate. Ideally it would be designed for the supply as it's a compensation network, which is why I'm not just throwing whatever values in there for mine. I think it would improve it further though if I did. Standard snubbing values are something like a 100nF poly or equivalent in parallel with 0.5R + 22nF?
So yeah that little network can change the sound of your caps entirely.
We're all learning. Did you make a soft start yet? I'm still using the ol 100W..... lazy!
Regards,
Chris
classd4sure said:
haha, soft start? i use a lightbulb in series, turn on the amp. then turn off the amp. quickly, take out the light bulb, swap in a normal power cord, turn it back on. no tripped breaker
thats the softstart for now. im debating whether i want to spend money for a softstart module. these things will NEVER go off. i will only use the mute circuit. ill eventually do SOMETHING, i just dont want to pay a lot for it...
yeah, im re-reading through the hotrodding thread, but its SO much. see my post above and see what you think.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.