Hi Mike,
Had you tried any other caps before going to all that effort?
So far I have to say changing the caps has been the biggest improvement.
Was never happy with the highs until now and the bass can really be felt even at lower volumes.
You can also mount your caps in parallel by putting one on either side of the board, might save some filing.
Regards,
Chris
Had you tried any other caps before going to all that effort?
So far I have to say changing the caps has been the biggest improvement.
Was never happy with the highs until now and the bass can really be felt even at lower volumes.
You can also mount your caps in parallel by putting one on either side of the board, might save some filing.
Regards,
Chris
Hey Chris,
Yes, I have been been replacing various caps. So far I have removed the input caps to the AD8620 and replaced the 22ufd reg. caps with 27ufd Panasonic FC's. Also replaced the 470ufd with Panasonic FC's, but didn't care for the sound much. Might try some Elna Cerfines there next. Currently useing Panasonic M series there. I have not tried replacing the bootstrap cap yet. Is this worth the effort? Also my new boards have the CRD's installed. I'm kind of on Rev 3 with this amp. Originaly started with the older ST boards. I thougt about mounting the second cap under the board but had some concerns with RF. I remember Bruno talking about keeping things close to the ground plane. Might be a bit anal maybe? Filing down one of the caps wasn't too hard anyway. All I can say is it was defenitely worth it.
Yes, I have been been replacing various caps. So far I have removed the input caps to the AD8620 and replaced the 22ufd reg. caps with 27ufd Panasonic FC's. Also replaced the 470ufd with Panasonic FC's, but didn't care for the sound much. Might try some Elna Cerfines there next. Currently useing Panasonic M series there. I have not tried replacing the bootstrap cap yet. Is this worth the effort? Also my new boards have the CRD's installed. I'm kind of on Rev 3 with this amp. Originaly started with the older ST boards. I thougt about mounting the second cap under the board but had some concerns with RF. I remember Bruno talking about keeping things close to the ground plane. Might be a bit anal maybe? Filing down one of the caps wasn't too hard anyway. All I can say is it was defenitely worth it.
Hi Mike,
Actually I wondered if you'd tried any other output filter caps first but that's OK, and I'm glad it worked out.
When you replaced the 22uF reg caps with FC caps did you try them on their own or do a few other mods at the same time? Curious to know because they're supposed to be high ESR caps for damping purposes, but I know some have changed them to low ESR and I wonder how well they got away with it. I've left mine stock and am very happy with the sound now for the first time.
The sound of the FC's really sucks if you don't do the output filter caps as well. It leaves it with a muffled/dead sound totally lacking in the higher frequencies. Change the filter cap along with them though and it's a new beast! I believe the filter cap and decoupling caps are chosen in conjuction with one another for the best overall sound.
I much prefer the combination of FC's decoupling and polypropylene filter cap. I do think it lacks a bit of microdetail though, possibly due to the FC's, the sound doesn't seem to wrap around you to the extent it did before, but aside from that it's very musical and I find much more pleasing, it's really made a huge difference, much more so than swapping to an AD8620.
I changed the bootstrap cap to FM at the same time I did the FC's so I can't really comment on the sound of it alone. I didn't think it would do much and just wanted lower ESR, so I didn't bother testing it by itself.
I've no doubt Bruno is correct about keeping it close to the ground plane, but the size of the cap I chose ruled out the possibility of mounting it top side. I also gambled and went with a wider lead pitch than the max recommended of 10mm.
It did increase EMI, as measured by my AM radio a few feet away, on the very lowest frequency, there's only a slight increase in static and you can still hear a very weak station over it. The modules still aren't in a chassis either, so the increase couldn't have been much, and more importantly I can't hear it at all through the amp, which is the main thing.
On your next modules it might not hurt to try one filter cap on either side before all the filing?
I still have the CRD's on the way and expect them this week, I know it can still only get better, but at this point it's really hard to believe it. What an experience.
Regards,
Chris
Actually I wondered if you'd tried any other output filter caps first but that's OK, and I'm glad it worked out.
When you replaced the 22uF reg caps with FC caps did you try them on their own or do a few other mods at the same time? Curious to know because they're supposed to be high ESR caps for damping purposes, but I know some have changed them to low ESR and I wonder how well they got away with it. I've left mine stock and am very happy with the sound now for the first time.
The sound of the FC's really sucks if you don't do the output filter caps as well. It leaves it with a muffled/dead sound totally lacking in the higher frequencies. Change the filter cap along with them though and it's a new beast! I believe the filter cap and decoupling caps are chosen in conjuction with one another for the best overall sound.
I much prefer the combination of FC's decoupling and polypropylene filter cap. I do think it lacks a bit of microdetail though, possibly due to the FC's, the sound doesn't seem to wrap around you to the extent it did before, but aside from that it's very musical and I find much more pleasing, it's really made a huge difference, much more so than swapping to an AD8620.
I changed the bootstrap cap to FM at the same time I did the FC's so I can't really comment on the sound of it alone. I didn't think it would do much and just wanted lower ESR, so I didn't bother testing it by itself.
I've no doubt Bruno is correct about keeping it close to the ground plane, but the size of the cap I chose ruled out the possibility of mounting it top side. I also gambled and went with a wider lead pitch than the max recommended of 10mm.
It did increase EMI, as measured by my AM radio a few feet away, on the very lowest frequency, there's only a slight increase in static and you can still hear a very weak station over it. The modules still aren't in a chassis either, so the increase couldn't have been much, and more importantly I can't hear it at all through the amp, which is the main thing.
On your next modules it might not hurt to try one filter cap on either side before all the filing?
I still have the CRD's on the way and expect them this week, I know it can still only get better, but at this point it's really hard to believe it. What an experience.
Regards,
Chris
Some Help
Hi Chis,
I'm a newbie to DIYAUdio, but a vetran analog and SMPS guy. Some monolithic class D experience too. I'm very excited about the UcD stuff.
I know all my questions are likely answered somewhere, but since you seem to be a regular, I thought I'd just ask you a few questions.
My UcD400 are one the way and I'd like to know a few things:
What is the dead time?
What is the delay?
What is the daughter board on the UCD?
Do the amps come with schematics, and if not where can I get one?
On the outside chance the duaghter board is the op amps and feedback components, do you know what connector they use?
I really want to try my hand at designing a modulator filter that I"ve been pondering.
Lots of questions I know.
Thanks for any help you can muster.
by the way, is there a way to search that puts you to the exact place the search hit happens rather than just a thread of hundreds of entries?
I"ve spent a good 8 hours in the last few days finding all sorts of tid bits, but rarely what I'm looking for!
Thanks again
Mike
Hi Chis,
I'm a newbie to DIYAUdio, but a vetran analog and SMPS guy. Some monolithic class D experience too. I'm very excited about the UcD stuff.
I know all my questions are likely answered somewhere, but since you seem to be a regular, I thought I'd just ask you a few questions.
My UcD400 are one the way and I'd like to know a few things:
What is the dead time?
What is the delay?
What is the daughter board on the UCD?
Do the amps come with schematics, and if not where can I get one?
On the outside chance the duaghter board is the op amps and feedback components, do you know what connector they use?
I really want to try my hand at designing a modulator filter that I"ve been pondering.
Lots of questions I know.
Thanks for any help you can muster.
by the way, is there a way to search that puts you to the exact place the search hit happens rather than just a thread of hundreds of entries?
I"ve spent a good 8 hours in the last few days finding all sorts of tid bits, but rarely what I'm looking for!
Thanks again
Mike
Hi Mike,
Welcome to the forum.
Deadtime, can't tell ya /don't know. You'll find it's adjustable by the pot on your module and comes preset for best possible THD.
You'll find a few mentions of circuit delays in the reference design thread, and I think on the patent as well.
The daughter board is the discrete comparator and is covered in epoxy. You're not going to find any kind of a pinout or anything for it.
The amps don't come with a schematic like you're looking for, and such a thing is not public domain for reasons obvious enough. You might read the "technology" link at the Hypex website.
You might try one of the DIY versions you've likely seen here to run your experiments with.
The search function isn't much help here.
One trick people use is to click on "show printable version" and then "show all posts in this thread", then you can use control+f to scan through it for keywords.
Or google for keyword site:www.diyaudio.com
That sometimes works but often misses alot.
The best you can do is free up alot more than 8 hours and read it all.
Regards,
Chis
Welcome to the forum.
Deadtime, can't tell ya /don't know. You'll find it's adjustable by the pot on your module and comes preset for best possible THD.
You'll find a few mentions of circuit delays in the reference design thread, and I think on the patent as well.
The daughter board is the discrete comparator and is covered in epoxy. You're not going to find any kind of a pinout or anything for it.
The amps don't come with a schematic like you're looking for, and such a thing is not public domain for reasons obvious enough. You might read the "technology" link at the Hypex website.
You might try one of the DIY versions you've likely seen here to run your experiments with.
The search function isn't much help here.
One trick people use is to click on "show printable version" and then "show all posts in this thread", then you can use control+f to scan through it for keywords.
Or google for keyword site:www.diyaudio.com
That sometimes works but often misses alot.
The best you can do is free up alot more than 8 hours and read it all.
Regards,
Chis
Re: Some Help
I check the option "show results as post" which gets you to the actual post with the keywords.
(if that is what you are looking for)
Cheers,
Johan
Portlandmike said:
I check the option "show results as post" which gets you to the actual post with the keywords.
(if that is what you are looking for)
Cheers,
Johan
Thanks
Thanks for the warm welcome and the help.
Sounds like the best mode is to get at the 680nF cap from what I've read so far, and with that going to lower ESR bypass Caps like the FC's.
What are some US availible caps to try besides the wima 330nF's.
Has anyone got an opinion on mica's. I can get 100nF's from digikey, and it would take 6 or 7 to do the job. They are 10 bucks a pop, but mica's seem to be pretty ideal caps.
Also, I'd love to try a Teflon given Bruno's favorable comments on them, but were does a guy find one?
Once I get these units working maybe I'll do my own design since its something I like to do.
Has anyone ever dared to mod the inductor?
I use to work for a semi company and was doing some analysis of class D distortion. My findings were that dead time is pretty much the major root cause of the distortion. Once way to deal with it is to just run the ripple current really high. Kind of analogus to turning up the bias on a A/B amp.
Does anyone know what the inductance value in the UcD400 is?
I'm thinking removing a turn or two might not be too tough.
Has anyone played with regulated linear power supplies?
Best Regards,
Mike
Thanks for the warm welcome and the help.
Sounds like the best mode is to get at the 680nF cap from what I've read so far, and with that going to lower ESR bypass Caps like the FC's.
What are some US availible caps to try besides the wima 330nF's.
Has anyone got an opinion on mica's. I can get 100nF's from digikey, and it would take 6 or 7 to do the job. They are 10 bucks a pop, but mica's seem to be pretty ideal caps.
Also, I'd love to try a Teflon given Bruno's favorable comments on them, but were does a guy find one?
Once I get these units working maybe I'll do my own design since its something I like to do.
Has anyone ever dared to mod the inductor?
I use to work for a semi company and was doing some analysis of class D distortion. My findings were that dead time is pretty much the major root cause of the distortion. Once way to deal with it is to just run the ripple current really high. Kind of analogus to turning up the bias on a A/B amp.
Does anyone know what the inductance value in the UcD400 is?
I'm thinking removing a turn or two might not be too tough.
Has anyone played with regulated linear power supplies?
Best Regards,
Mike
Couple of questions....
I was wondering if anyone had the schematic of the front end of the UcD modules or knew what all the front end was doing? Does it provide the gain as well or is it just a buffer?
I am a little curious if the Borbeley All-JFet "super buffers" would be a good idea for a discrete front end? If gain is needed then maybe the All-JFet line amp might be a better solution.
Has anyone attempted the independent supplies for the front end yet? If so, are you using a super regulator kit, or your own concoction? I don't know what super regulator kits are out there.
This is a great thread and has been a very intriguing read (read pgs 1-32 & 48-69, still need to read the missing portion).
I was wondering if anyone had the schematic of the front end of the UcD modules or knew what all the front end was doing? Does it provide the gain as well or is it just a buffer?
I am a little curious if the Borbeley All-JFet "super buffers" would be a good idea for a discrete front end? If gain is needed then maybe the All-JFet line amp might be a better solution.
Has anyone attempted the independent supplies for the front end yet? If so, are you using a super regulator kit, or your own concoction? I don't know what super regulator kits are out there.
This is a great thread and has been a very intriguing read (read pgs 1-32 & 48-69, still need to read the missing portion).
half the answer
You can visit the hypex.nl site and look at technology.
It will show you the basics of the front end.
Its basically an instrumentation amp front end. It provides gain and thus improves CMRR.
I don't know, but suspect that you really do want some gain there as the input of amp, less the op amps, is a comparator, and comparators speed is a function of overdrive to some degree.
For what you might gain in a buffer, you will take a hit for sure in PSSR, very likely mess up the switching speeds too becuase of the lower overdrive to the comparator. Just my guess.
From what I've read, and I'm new, the biggest bang out there is the output filter cap, and the mains caps.
My amps shipped monday, so I really don't comment from any experience yet.
Best Regards,
Mike
You can visit the hypex.nl site and look at technology.
It will show you the basics of the front end.
Its basically an instrumentation amp front end. It provides gain and thus improves CMRR.
I don't know, but suspect that you really do want some gain there as the input of amp, less the op amps, is a comparator, and comparators speed is a function of overdrive to some degree.
For what you might gain in a buffer, you will take a hit for sure in PSSR, very likely mess up the switching speeds too becuase of the lower overdrive to the comparator. Just my guess.
From what I've read, and I'm new, the biggest bang out there is the output filter cap, and the mains caps.
My amps shipped monday, so I really don't comment from any experience yet.
Best Regards,
Mike
Re: Couple of questions....
Mike,
This thread isn’t that long, were you referring to the UcD400 Q/A thread? If so, yes it is filled with gems of wisdom as is this thread. I feel it would be time well spent to read them both all the way through.
The front end issue is a question of providing buffering for the comparator as its input impedance is rather low @ 1.8Kohm. It provides gain as well as a means of accepting single ended or balanced input while providing good common mode rejection with low distortion, all with very low DC offset and noise. Rather daunting tasks when taken in total but the best of the new opamps do an admirable job. To replace all this with a discreet stage requires a design with all of the attributes of the best opamps including DC offset and drift as well as extremely low noise. I don’t know if the design you are referring to is up to the task as I haven’t analyzed it myself. I would look up the detailed specs of a quality chip like the ad8620 and compare them to what you want to try. It will be a give or take situation and the bottom line is if it is worth the time and money for what you gain in sonics. My guess is probably not. However the power for this stage is an open question and a great place for improvement.
I have found the output filter cap to be the single most sensitive part to change. I have tried several with most being a significant improvement to sonics. With this said I am now looking forward to trying the new Auricap which is now on the way, I should have them tomorrow. Diycable.com should have their stock shortly.
Roger.
JoshK said:
Mike,
This thread isn’t that long, were you referring to the UcD400 Q/A thread? If so, yes it is filled with gems of wisdom as is this thread. I feel it would be time well spent to read them both all the way through.
The front end issue is a question of providing buffering for the comparator as its input impedance is rather low @ 1.8Kohm. It provides gain as well as a means of accepting single ended or balanced input while providing good common mode rejection with low distortion, all with very low DC offset and noise. Rather daunting tasks when taken in total but the best of the new opamps do an admirable job. To replace all this with a discreet stage requires a design with all of the attributes of the best opamps including DC offset and drift as well as extremely low noise. I don’t know if the design you are referring to is up to the task as I haven’t analyzed it myself. I would look up the detailed specs of a quality chip like the ad8620 and compare them to what you want to try. It will be a give or take situation and the bottom line is if it is worth the time and money for what you gain in sonics. My guess is probably not. However the power for this stage is an open question and a great place for improvement.
I have found the output filter cap to be the single most sensitive part to change. I have tried several with most being a significant improvement to sonics. With this said I am now looking forward to trying the new Auricap which is now on the way, I should have them tomorrow. Diycable.com should have their stock shortly.
Roger.
Hi Roger et al,
Is it still a "drop in replacement"?. Can you give us tolerance, temp co, voltage, lead pitch, squashed or plain radial? method of constuction (metallized/flim n foil/ stacked etc) anything ?
Will they be made available factory direct? If so I'm sure a group buy would be an interesting idea for them. Possibly along with coupling caps for those who require it. Projected/estimated cost?
I as well found the output filter one of the biggest upgrades, or most significant, but I think I was in error to have said that, as it stands to reason one would immediatly remove or upgrade the electrolytic AC coupling caps... really you have to look at the system as a whole.
The biggest mica I've seen from digikey is 10nF. I considered using them for the feedback loops but I understand they're under the epoxy so I forgot about trying it.
I agree with Roger, a discrete input stage may be an excellent upgrade, but it would indeed have to be a very good input stage to out do the op amps, and you're likely to be giving up the CMRR.
As far as gain I dont' think the comparator relies on being overdriven. It has alot of gain built in.
As per the regulated input stage supplies, if you look at what Hypex is calling an audiophile auxiliary supply for their 700 module, it seems you don't really need to go all out in that respect.
Regards,
Chris
Is it still a "drop in replacement"?. Can you give us tolerance, temp co, voltage, lead pitch, squashed or plain radial? method of constuction (metallized/flim n foil/ stacked etc) anything ?
Will they be made available factory direct? If so I'm sure a group buy would be an interesting idea for them. Possibly along with coupling caps for those who require it. Projected/estimated cost?
I as well found the output filter one of the biggest upgrades, or most significant, but I think I was in error to have said that, as it stands to reason one would immediatly remove or upgrade the electrolytic AC coupling caps... really you have to look at the system as a whole.
The biggest mica I've seen from digikey is 10nF. I considered using them for the feedback loops but I understand they're under the epoxy so I forgot about trying it.
I agree with Roger, a discrete input stage may be an excellent upgrade, but it would indeed have to be a very good input stage to out do the op amps, and you're likely to be giving up the CMRR.
As far as gain I dont' think the comparator relies on being overdriven. It has alot of gain built in.
As per the regulated input stage supplies, if you look at what Hypex is calling an audiophile auxiliary supply for their 700 module, it seems you don't really need to go all out in that respect.
Regards,
Chris
oops, I bet your right.
I bet your right on the Mica's. I must be off by 10.
Roger, thanks for the Auricap tip. I'll look forward to hearing how it sounds. And for the thread tip. I'll take you up on it.
You know, I did a discrete preamp with two NPN's and like 3 resistors that I haven't bested yet. Its THD was below what I could measure with my AP System one. LIke down 110dB.
That all happened after my guru buddy helped me tweak it by soft biasing the current source.
Once I get my UcD400's,and listen to them a bit, maybe its an experiement to do.
Perhaps then a regulated power supply would be needed.
More than anything right now, I just want a decent amp since my handbuilt toasted on me.
Best Regards,
Mike
I bet your right on the Mica's. I must be off by 10.
Roger, thanks for the Auricap tip. I'll look forward to hearing how it sounds. And for the thread tip. I'll take you up on it.
You know, I did a discrete preamp with two NPN's and like 3 resistors that I haven't bested yet. Its THD was below what I could measure with my AP System one. LIke down 110dB.
That all happened after my guru buddy helped me tweak it by soft biasing the current source.
Once I get my UcD400's,and listen to them a bit, maybe its an experiement to do.
Perhaps then a regulated power supply would be needed.
More than anything right now, I just want a decent amp since my handbuilt toasted on me.
Best Regards,
Mike
Hi Mike,
You have some interesting ideas, I'd like to know what you mean by soft biasing the current source.
I'd like to know how you plan on implementing it/ what parts you'll use etc. I'm more than happy to help you get the best out of it.
The UCD180 thread is also very long but well worth a good read.
Cheers,
Chris
You have some interesting ideas, I'd like to know what you mean by soft biasing the current source.
I'd like to know how you plan on implementing it/ what parts you'll use etc. I'm more than happy to help you get the best out of it.
The UCD180 thread is also very long but well worth a good read.
Cheers,
Chris
soft bias for current sources
Chris,
First, before I get to soft bias, I need to get to good sounding current sources first.
IMO most all designs I see don't do current sources well.
Most designs will use a rather low drop on the Re of the current source transistor.
If you look at it as an amp, your asking for lots of gain. I like to put LOTS of volts accross the emitter resistor. With a MPSA18 for example, I'd split the voltage on the npn and the Re, but my experience has been you need about 10V or more VCE to make MPSA18's work best.
Having the Re large really helps the current source as the frequencies rise. Hey, its just degeneration. That is usually good. For some reason most designs will maximize the voltage on the transistor and minimize the V(Re). Sometimes this makes sense, like if you want big swings, but even then, its my bet that a couple more volts V(Re) would be an improvement almost always.
In my 2-npn preamp it consisted of a MPSA18A follower, loaded by MPSA with a 10V zener biasing the base from the negative rail.
It ran on +/-15~+/-25V rails.
Yeah, I used a RC filter on the zener as I recall too, but that's not the final yet.
It sounded pretty good concidering how simple it was, but not amazing.
I dragged it up to a buddy of mine who lives on the North Shore of Lake Superior. He is a gifted individual with amazing hearing and hearing memory. (The best amp I've heard to date is one he designed without the aid of simulators or even a scope, just by listeing in the circuits.)
Anyways, we listened, and he ask to see the schematic.
He said, "lets pull out that zener and cap and add a resistor".
But.... but, that won't be as good a current source.
Well anyways, we did it, and it was an amazing.
We bypassed the V(CSbase,-V) with a cap, mostly just to show ME that it was a negative result.
The rest of the night was spent tweaking the bias current, if I recall it was like 4mA in the end, and mostly tweaking the simple divider thevinin that biased the base of the current source. It was quite cool. Bottom line, you want a thevinin impedance on the base that is about twice that of Re, but rangese of 2-6X are worth playing with.
When its right it stage and dynamics are world class.
Hope that's not to wordy an answer.
Best Regards,
Mike
Chris,
First, before I get to soft bias, I need to get to good sounding current sources first.
IMO most all designs I see don't do current sources well.
Most designs will use a rather low drop on the Re of the current source transistor.
If you look at it as an amp, your asking for lots of gain. I like to put LOTS of volts accross the emitter resistor. With a MPSA18 for example, I'd split the voltage on the npn and the Re, but my experience has been you need about 10V or more VCE to make MPSA18's work best.
Having the Re large really helps the current source as the frequencies rise. Hey, its just degeneration. That is usually good. For some reason most designs will maximize the voltage on the transistor and minimize the V(Re). Sometimes this makes sense, like if you want big swings, but even then, its my bet that a couple more volts V(Re) would be an improvement almost always.
In my 2-npn preamp it consisted of a MPSA18A follower, loaded by MPSA with a 10V zener biasing the base from the negative rail.
It ran on +/-15~+/-25V rails.
Yeah, I used a RC filter on the zener as I recall too, but that's not the final yet.
It sounded pretty good concidering how simple it was, but not amazing.
I dragged it up to a buddy of mine who lives on the North Shore of Lake Superior. He is a gifted individual with amazing hearing and hearing memory. (The best amp I've heard to date is one he designed without the aid of simulators or even a scope, just by listeing in the circuits.)
Anyways, we listened, and he ask to see the schematic.
He said, "lets pull out that zener and cap and add a resistor".
But.... but, that won't be as good a current source.
Well anyways, we did it, and it was an amazing.
We bypassed the V(CSbase,-V) with a cap, mostly just to show ME that it was a negative result.
The rest of the night was spent tweaking the bias current, if I recall it was like 4mA in the end, and mostly tweaking the simple divider thevinin that biased the base of the current source. It was quite cool. Bottom line, you want a thevinin impedance on the base that is about twice that of Re, but rangese of 2-6X are worth playing with.
When its right it stage and dynamics are world class.
Hope that's not to wordy an answer.
Best Regards,
Mike
Active current source.
Mike,
It sounds like you are talking about an active current source that is modulated by the signal to some degree. This can be useful if you don't approach the rail voltage too closely as it will run out of drive. Putting a filter cap on the base will improve the rail performance but now it is no longer active. You don't see this design too often as power amp current source usage is most always concerned with full power close to the rail issues, thus the small res. to the emitter and the hard zener or forward diode drop reference voltage.
Roger
Mike,
It sounds like you are talking about an active current source that is modulated by the signal to some degree. This can be useful if you don't approach the rail voltage too closely as it will run out of drive. Putting a filter cap on the base will improve the rail performance but now it is no longer active. You don't see this design too often as power amp current source usage is most always concerned with full power close to the rail issues, thus the small res. to the emitter and the hard zener or forward diode drop reference voltage.
Roger
most amplifiers do use current sources
Roger,
Most amps I'd venture do use or could benifit from a current source.
Any diff pair can benifit for example, although you see lots that use a resistor becuase it sounds better than a transistor current source w/o much degeneration.
The same principle applies to active loads. More degeneration is a good thing as it moves the pole out there, just like bruno talked about in his UcD paper. I think the problem is most amps are trying to get the last watt or couple volts of swing even though they almost always don't sound good there, at least in linear stuff.
Putting a filter cap on the base DOES NOT improve performance though. I'm not sure why, but it doesn't. That's barely measurable in my experience but extreemly audible.
I don' t know if it would matter in the UCD comparator but it would be one my list of experiments.
A good friend of mine who is designing very high end speaker systems (100k$) took the hint after hearing my system and redesigned his active crossovers using that very current source. He told me the results were validated in simulation, but I haven't looked at it myself. My best experience in optimization has always been by listening once your in the ball park.
The other fatal sin is to share your bias voltage among several current sources. Lots more intermod stuff results.
By the way, I tried fets too, and they sounded good, more laid back, but lacked dynamics. The bipolar as I recall won the current source postition, and I was in favor of the bipolar follower too, which suprised me. I think it had more to do with output Z. A two stage would be interesting.
Best Regards,
Mike
Roger,
Most amps I'd venture do use or could benifit from a current source.
Any diff pair can benifit for example, although you see lots that use a resistor becuase it sounds better than a transistor current source w/o much degeneration.
The same principle applies to active loads. More degeneration is a good thing as it moves the pole out there, just like bruno talked about in his UcD paper. I think the problem is most amps are trying to get the last watt or couple volts of swing even though they almost always don't sound good there, at least in linear stuff.
Putting a filter cap on the base DOES NOT improve performance though. I'm not sure why, but it doesn't. That's barely measurable in my experience but extreemly audible.
I don' t know if it would matter in the UCD comparator but it would be one my list of experiments.
A good friend of mine who is designing very high end speaker systems (100k$) took the hint after hearing my system and redesigned his active crossovers using that very current source. He told me the results were validated in simulation, but I haven't looked at it myself. My best experience in optimization has always been by listening once your in the ball park.
The other fatal sin is to share your bias voltage among several current sources. Lots more intermod stuff results.
By the way, I tried fets too, and they sounded good, more laid back, but lacked dynamics. The bipolar as I recall won the current source postition, and I was in favor of the bipolar follower too, which suprised me. I think it had more to do with output Z. A two stage would be interesting.
Best Regards,
Mike
Hi,
Putting a cap there could very well depend on the quality and type of cap, the level of noise it has to filter.. etc.
Would it ring with the base at all, or even change the phase adversly? There's always more than first meets the eye.
No your reply wasn't too wordy.
That's not something you'll get to play with on the UCD either, unless you want to remove some epoxy, but then you're likely not have a working amp after.
So, I just soldered on the bread crumb CRD's, what a pain!! The first three went alright... that last one though.... grrrr. Now I'm half blind too.
At first I was going to do them all at once but the second module proved to be much more difficult for some strange reason, so I wound up doing an A/B test.
The difference is noticeable but nothing drastic. It's just a higher quality sound overall, I guess more refined in microdetail, cleaner. Really it helped the sound of the AD op amp alot. Side by side one sounded a little bit flat and lifeless while the other was really clear and dynamic, tighter.
I'm tempted to do the same on my soundcard, should my vision ever return. Damn bread crumbs.
I got some X2 polypropylene caps to bypass the rectifier with... think it's worth it?
Kind of bored now, might start work on an "audiophile" 78XX /79XX based aux supply later this week.
Also while I really like the added bass of the FC caps, I think they're responsible for it being a bit less holographic, the sound really doesn't wrap around you like it did with those brown stock caps. Could this be what others don't like about the FC's? I wonder if there's a cap out there that would let us have it all so to speak.
Regards,
Chris
Putting a cap there could very well depend on the quality and type of cap, the level of noise it has to filter.. etc.
Would it ring with the base at all, or even change the phase adversly? There's always more than first meets the eye.
No your reply wasn't too wordy.
That's not something you'll get to play with on the UCD either, unless you want to remove some epoxy, but then you're likely not have a working amp after.
So, I just soldered on the bread crumb CRD's, what a pain!! The first three went alright... that last one though.... grrrr. Now I'm half blind too.
At first I was going to do them all at once but the second module proved to be much more difficult for some strange reason, so I wound up doing an A/B test.
The difference is noticeable but nothing drastic. It's just a higher quality sound overall, I guess more refined in microdetail, cleaner. Really it helped the sound of the AD op amp alot. Side by side one sounded a little bit flat and lifeless while the other was really clear and dynamic, tighter.
I'm tempted to do the same on my soundcard, should my vision ever return. Damn bread crumbs.
I got some X2 polypropylene caps to bypass the rectifier with... think it's worth it?
Kind of bored now, might start work on an "audiophile" 78XX /79XX based aux supply later this week.
Also while I really like the added bass of the FC caps, I think they're responsible for it being a bit less holographic, the sound really doesn't wrap around you like it did with those brown stock caps. Could this be what others don't like about the FC's? I wonder if there's a cap out there that would let us have it all so to speak.
Regards,
Chris
Sorry if this is obvious, but what are CRD's? I have read this thread in entirety and don't recall seeing what these were.
JoshK said:
I think CRD stands for Current Regulating Diode. In fact it is a mosfet that is used with the gate and source shorted so that it acts as a current source. Hooking up such a current source at the output of the AD8620 opamp apparantly improves the sound quality of that opamp. Likely due to improved power supply rejection and output of the opamp will be Class A biased as well.
The CRD that should be used is the SST511, you can get it at mouser.com, just ordered some a few days ago.
Hope this helps
Gertjan
- Status
- Not open for further replies.