Shield?
Chris,
We are dealing with an 8 ohm load here connected directly to the output. Stuff feeding in at this point would need to be very high frequency and/or very large to have an effect on the actual audio. We need to concentrate on this coil radiating into near by sensitive stuff not its being a receiving antenna. Point is, I think the shielding aspect would prove to be far more beneficial than any disadvantages. Ah, the art of compromise!
As for parasitic capacitance it would be in parallel with the output filter cap and would be of no consequence. It would have a much smaller HF component to couple into nearby circuits it were the output side.
As far as shielding goes it would be pretty well shielded for electrostatics but something entirely different would need to be done for electromagnetic EMI. UcD’s have a ferrite shield overall that seems to be quite effective.
Roger
Chris,
We are dealing with an 8 ohm load here connected directly to the output. Stuff feeding in at this point would need to be very high frequency and/or very large to have an effect on the actual audio. We need to concentrate on this coil radiating into near by sensitive stuff not its being a receiving antenna. Point is, I think the shielding aspect would prove to be far more beneficial than any disadvantages. Ah, the art of compromise!
As for parasitic capacitance it would be in parallel with the output filter cap and would be of no consequence. It would have a much smaller HF component to couple into nearby circuits it were the output side.
As far as shielding goes it would be pretty well shielded for electrostatics but something entirely different would need to be done for electromagnetic EMI. UcD’s have a ferrite shield overall that seems to be quite effective.
Roger
Hi Roger,
That statement seems to assume an ideal output filter coil with no parasitic components of its own either.
I can't help but feel EMI is the main concern. All the same it's a worthy theory and worth trying, that's what two modules and two coils are for 🙂
That statement seems to assume an ideal output filter coil with no parasitic components of its own either.
I can't help but feel EMI is the main concern. All the same it's a worthy theory and worth trying, that's what two modules and two coils are for 🙂
Are you aware of the fact that this is a digital volume control, which means that you lose bits as you turn the volume down ?
peufeu said:
I'm aware that at low volume sound is not like it woud be. The amp sound at its best only with volume at 3 o'clock. I fixed the amp with lower gain so that neighbors don't kill me.
Have you some hints for me?
Ciao
Thomas
Volume control?
It might be possible to scale the DAC's reference voltage to scale the output without loosing bits. There are compromises to this but should be less severe than loosing bits of resolution. You need to look into the DAC’s data sheets.
Roger
It might be possible to scale the DAC's reference voltage to scale the output without loosing bits. There are compromises to this but should be less severe than loosing bits of resolution. You need to look into the DAC’s data sheets.
Roger
Re: Volume control?
Hi Roger,
I have a nice tube Pre, but I don't use it for tests because then I hear to it instead of Tripath or UCD. What about a simple stand alone volume control with a switch? (maybe somebody has a link to a good project 😉 )
I'm waiting for your OK to mount the coils. I think EMI is the problem. Do you think an aluminium shield would help?
Ciao
Thomas
sx881663 said:
Hi Roger,
I have a nice tube Pre, but I don't use it for tests because then I hear to it instead of Tripath or UCD. What about a simple stand alone volume control with a switch? (maybe somebody has a link to a good project 😉 )
I'm waiting for your OK to mount the coils. I think EMI is the problem. Do you think an aluminium shield would help?
Ciao
Thomas
Dear Klaus,
Glad you mentioned the Super-E caps 🙂
I've praised them before and got acid critics from an audio expert, (for whom I have big respect as his contributions to the forums are solid) as being impossible. No one wants to test if the claims of the inventor is true or false. (beware)
I have had very good results on a DAC PS (now I'm trying AlWSR).
By coincidence, I have been playing around with UCD400's low voltage PS for opamps, and found that this is a most critical site for sound improvement. I changed the 22uF electros, near T2/T3? for Panasonic FC and found the sound very clean but lacking punch and bass. Paralel 100nf film caps made it cleaner and more elegant. Then I swaped them for some Rubycon ZL, 270uF, bigger caps and they totally changed the amps sound, with more up-front, energetic presentation and deeep bass 😱 HF suffered (got a problem with my NOS dac ) a bit and I thought about // again film caps, when I remembered that I have a bunch of BG-NX 100nf and some super-E ready made to try them 😎
Dear Thomas,
Is it possible that the present configuration of your dedicated supply for UCD's input section is sub-optimal?
It's only a guess. Maybe you have to take off a couple of caps and the transistors (sorry if this was already discoused ). Could you take pictures of both PCB surfaces? ...for a wider look.
Cheers,
M
Glad you mentioned the Super-E caps 🙂
I've praised them before and got acid critics from an audio expert, (for whom I have big respect as his contributions to the forums are solid) as being impossible. No one wants to test if the claims of the inventor is true or false. (beware)
I have had very good results on a DAC PS (now I'm trying AlWSR).
By coincidence, I have been playing around with UCD400's low voltage PS for opamps, and found that this is a most critical site for sound improvement. I changed the 22uF electros, near T2/T3? for Panasonic FC and found the sound very clean but lacking punch and bass. Paralel 100nf film caps made it cleaner and more elegant. Then I swaped them for some Rubycon ZL, 270uF, bigger caps and they totally changed the amps sound, with more up-front, energetic presentation and deeep bass 😱 HF suffered (got a problem with my NOS dac ) a bit and I thought about // again film caps, when I remembered that I have a bunch of BG-NX 100nf and some super-E ready made to try them 😎
Dear Thomas,
Is it possible that the present configuration of your dedicated supply for UCD's input section is sub-optimal?
It's only a guess. Maybe you have to take off a couple of caps and the transistors (sorry if this was already discoused ). Could you take pictures of both PCB surfaces? ...for a wider look.
Cheers,
M
Ok, has anyone compared the UCD and Tripath with one of the Equibit equiped Panasonic receivers?
I've personally compared the Tripath to the Panasonic. And the Tripath has better detail. The Panasonic isnt bad.
Now I keep going back and forth between my Class-A SS amps(Threshold/Pass) and the Pansonic. And I kind of heard this same thing going on with Tripath....
All the digital amps sound more close to me. But I dont think it's due to increased high frequency. The soundstage seems a little less deep, but that is made up for by razor sharp details.
Still, the detail is there with the Class-A SS amps. But it's presented and digested in a different way. It's as if the detail is there but it's coming from farther away and it's less dissected.
Has anyone categorized the good digital amps like that versus SS?
Is the UCD more likely to be tolerant of extremently varied speaker loads?
I've noticed that in going back and forth between various amps that the entire tonal balance of my speakers gets changed. With Class-A SS, it's like the effect the old "loudness" button has... kind of a bumped up bass presence. I dont notice that with digital amps. They seem perfectly flat across the whole frequency range.
Are SS amps more effected by speaker loads that digital?
Would the UCD be better than the Equibit in Panasonic receiver form?
I've personally compared the Tripath to the Panasonic. And the Tripath has better detail. The Panasonic isnt bad.
Now I keep going back and forth between my Class-A SS amps(Threshold/Pass) and the Pansonic. And I kind of heard this same thing going on with Tripath....
All the digital amps sound more close to me. But I dont think it's due to increased high frequency. The soundstage seems a little less deep, but that is made up for by razor sharp details.
Still, the detail is there with the Class-A SS amps. But it's presented and digested in a different way. It's as if the detail is there but it's coming from farther away and it's less dissected.
Has anyone categorized the good digital amps like that versus SS?
Is the UCD more likely to be tolerant of extremently varied speaker loads?
I've noticed that in going back and forth between various amps that the entire tonal balance of my speakers gets changed. With Class-A SS, it's like the effect the old "loudness" button has... kind of a bumped up bass presence. I dont notice that with digital amps. They seem perfectly flat across the whole frequency range.
Are SS amps more effected by speaker loads that digital?
Would the UCD be better than the Equibit in Panasonic receiver form?
Can't answer all your questions for ya, but I'll hit on a few of them.
"Is the UCD more tolerant".... hell, it's immune. It's tonal characteristics remain virtually unchanged across very different speakers of all shapes and sizes, provided it's got the current capability to drive it of course, and they're good to about 1 ohm anyway.
As far as soundstage depth alot of that is directly related to your grounding/wiring/PSU... it won't be the module that limits this but your own ability to wire it properly. The sound wraps right around you and you can fall right in.
Should watch for commercial class D amps, commercial guys always like to undersize the power supply, probably even more so with a class D amp because they really can get away with it, but then you lack the depth and such, so it's truly worth spending some money in that area and comming up with a really good stiff supply.
Start with a cheap stereo setup, play around with different supplies and grounding schemes/wiring techniques.... found out what works and go from there... it's really hard to go wrong.
"Is the UCD more tolerant".... hell, it's immune. It's tonal characteristics remain virtually unchanged across very different speakers of all shapes and sizes, provided it's got the current capability to drive it of course, and they're good to about 1 ohm anyway.
As far as soundstage depth alot of that is directly related to your grounding/wiring/PSU... it won't be the module that limits this but your own ability to wire it properly. The sound wraps right around you and you can fall right in.
Should watch for commercial class D amps, commercial guys always like to undersize the power supply, probably even more so with a class D amp because they really can get away with it, but then you lack the depth and such, so it's truly worth spending some money in that area and comming up with a really good stiff supply.
Start with a cheap stereo setup, play around with different supplies and grounding schemes/wiring techniques.... found out what works and go from there... it's really hard to go wrong.
Daveis said:
That would be very interesting. You are the first I see to have done a comparison between Tripath and Class A SS amps. What speakers and source do you use?
Maxlorenz,
look at page #1 and #4 of this thread for pictures. My UCD has no input caps and the opamps have dedicated PS from a small transformer secondary trough SupplyHG, both from Hypex.
Ciao
Thomas
Have you matched the two amplifier levels ? A difference in levels can be perceived as a sound quality difference.
About volume controls :
Since your DAC probably has a buffered output, and your amplifiers have a decent input impedance, I suggest you try evaluating a simple potentiometer (try a cheap stereo 10K log carbon pot) against the digital volume control. Maybe you're in for a surprise...
Also, you don't need toslink for isolation, most DACs have a pulse transformer at the SPDIF input (you should check, though)
About volume controls :
Since your DAC probably has a buffered output, and your amplifiers have a decent input impedance, I suggest you try evaluating a simple potentiometer (try a cheap stereo 10K log carbon pot) against the digital volume control. Maybe you're in for a surprise...
Also, you don't need toslink for isolation, most DACs have a pulse transformer at the SPDIF input (you should check, though)
Daveis said:
Hi Daveis
i own both Panasonic saxr45 ( the only one with regulated smps ) and UCD180
in stock form or modded , the Panasonic is very transparent but a bit artificial in the mid-treble range , like the equibit-based Tact 2150
UCD is more neutral and natural , with nice soundstage
Alain
peufeu said:
Many thanks, I'll try. I posted something about my dac and sound card in your thread about ethernet dac.
Ciao
Thomas
Hi Thomas.
I tried all kind of Pots. Digital, Alps-Blue, Plastic, Resistor-Ladders. You name it.
Once you've heard a transformer Pot like the TX102 of Stevens&
Billington. You'll be convinced, that this is the way to go. If you want to spend more money you take the silver version.
I promise you'll not go back.
I am running the I/V stage of the 1543 DAC at 90R, which works just great. The Tripath2020 counts 20kR on the input. It works perfect with the TX102. It also works on more extrem impedance
situations just loook it up on the S&B homepage.
Some people claim that the TX does not have the extremest level
of resolution you could probably achieve with just a very few
high-end pre-amps around. The pity even -- in a phantastic Mos-Fet pre you need a Pot! And you need clean power etc.
Some people also claim the dynamic and transient response of
the TX102 is not good enough compared to the best in class pres. I am running it in autotransformer mode, which is a special
way of running the pot. That config exactly addresses these subjects. It improves above areas. Still. Even in normal configI was happy with it for quite some time now.
Of course You can also run it fully differential!! If you have a differantial DAC-out and AMP-in. You can even skip the coupling Cs! (The pity - my 2020 does not have a differential Input.
That's the only thing which is bothering me in my chain.)
You can not arrange the signal path with less material at this performance. You'll hear it 😀
Bottom line.
I stay transformer-passive - That's for sure!
\Klaus
I tried all kind of Pots. Digital, Alps-Blue, Plastic, Resistor-Ladders. You name it.
Once you've heard a transformer Pot like the TX102 of Stevens&
Billington. You'll be convinced, that this is the way to go. If you want to spend more money you take the silver version.
I promise you'll not go back.
I am running the I/V stage of the 1543 DAC at 90R, which works just great. The Tripath2020 counts 20kR on the input. It works perfect with the TX102. It also works on more extrem impedance
situations just loook it up on the S&B homepage.
Some people claim that the TX does not have the extremest level
of resolution you could probably achieve with just a very few
high-end pre-amps around. The pity even -- in a phantastic Mos-Fet pre you need a Pot! And you need clean power etc.
Some people also claim the dynamic and transient response of
the TX102 is not good enough compared to the best in class pres. I am running it in autotransformer mode, which is a special
way of running the pot. That config exactly addresses these subjects. It improves above areas. Still. Even in normal configI was happy with it for quite some time now.
Of course You can also run it fully differential!! If you have a differantial DAC-out and AMP-in. You can even skip the coupling Cs! (The pity - my 2020 does not have a differential Input.
That's the only thing which is bothering me in my chain.)
You can not arrange the signal path with less material at this performance. You'll hear it 😀
Bottom line.
I stay transformer-passive - That's for sure!
\Klaus
Dear Thomas,
I saw your UCD photo. The thing is, IF i remember well, hypex recommends taking off T2-T3 to connect aux supply.
The caps I mentioned are not signal input caps but those cheap electros that are decoupling the low V supply...
(don't remember nº)
I'm interested because eventually I plan to make a dedicated supply for the input section: R-core TX; soft recovery bridges; rubycon ZL caps; possibly BG. I don't want to let elements on PCB which could degrade the sound 🙁
Klaus,
Cheers,
M
I saw your UCD photo. The thing is, IF i remember well, hypex recommends taking off T2-T3 to connect aux supply.
The caps I mentioned are not signal input caps but those cheap electros that are decoupling the low V supply...
(don't remember nº)I'm interested because eventually I plan to make a dedicated supply for the input section: R-core TX; soft recovery bridges; rubycon ZL caps; possibly BG. I don't want to let elements on PCB which could degrade the sound 🙁
Klaus,
Totally agree, again 😉
Cheers,
M
Hi Max
here again the picture of page #4:
the transistors to eliminate are T2 and T3: wich are the caps to eliminate? C14, C15? There is no mention of them in Hypex instructions.
Chris,
what Peufeu meant is that is different the volume command from PatchMix (you too have an Emu card I know) and an external chip that digitally acts as a pot. As far as I can understand it, the first is a software manipulation of data made by DSP, the second manipulates the electric signal. I'd like to learn more about this.
Ciao
Thomas
here again the picture of page #4:
the transistors to eliminate are T2 and T3: wich are the caps to eliminate? C14, C15? There is no mention of them in Hypex instructions.
Chris,
what Peufeu meant is that is different the volume command from PatchMix (you too have an Emu card I know) and an external chip that digitally acts as a pot. As far as I can understand it, the first is a software manipulation of data made by DSP, the second manipulates the electric signal. I'd like to learn more about this.
Ciao
Thomas
Hi ClassD4sure.
What are you refering to, when talking about "Lossless"
I had a chance to try a Burr Brown PGA 2310 solution, when I talk about digital pot. I wonder if you call that lossless.
I still prefer the TX102.
\Klaus
What are you refering to, when talking about "Lossless"
I had a chance to try a Burr Brown PGA 2310 solution, when I talk about digital pot. I wonder if you call that lossless.
I still prefer the TX102.
\Klaus
Hi Thomas,
I can't provide good answers...only good (or bad) questions! 😉
The problem is that everything resonates in this universe as dear Chris allways is remembering me 😀
See this link as a light read before sleep:
http://www.diyhifi.org/forums/viewtopic.php?t=523
The problem is if the combination of C/L/R used sounds good to you or not 🙁
I can't predict how good or bad will some caps arrangement sound. I don't have scope nor know the maths and could never probably use the software that EE's uses so joyously
All I have is courage (read recklessness) and a happy iron...😀
My UCD400ST allways have sounded bad for me on the highs. I have tried many things. Now I run a sort of Pi filter for the main PS wich made a good improvement in overall sound. Recently I tried mods on the +/-15VDC on board supply: did not like the 22uF decoupling electrolytics "near" the darlingtons? (now I realize that your UCD180 lacks these, sorry for the confusion)
I found that //ing 100nF film caps to the electros was worthy, soundwise: cleaner sound.
Swaping those cheap electros for Panasonic FC 47uF//Wima 100nF resulted in very clean sound but without punch and bass.
Then I put there the smallest Rubycon that I had on hand, instead of Pana//Wima: rubycon ZL 270uF!! Not a good fit... but , again, the whole sound presentation changed dramatically: very deep bass, more dinamics and punch. But also harsher highs; back to where I started
After that, encouraged by Klaus' experience, I decided that the culprits were those little decoupling ceramic caps (C6-C7) near OPA2134 power pins. I took them off (my first time with SMD ) and put a pair of Black Gate NX 100nF in "antiparallel configuration" (sorry George). I had to drill the PCB for that 
(kids, don't do that at home 😀 )
Now, the BG are well known for being very slow to burn-in. Untill know I hear absence of the previous HF noise (the Rubycon's did not liked the ceramics, and I don't blame them). The sound is lacking HF extension and "air" but is very, and I mean very, robust. Only time will tell if this is a winner combo...
What's the conclusion of this boring post? I'm affraid that every part on the power and the signal paths can affect adverselly the sound 🙁
The other conclusion is that I don't like "the sound" of ceramics at all!
Any opinions are welcome!
I will try to get some 47uF Rubycon ZA to try instead of the 270uF's.
ClassD tried before //ing tantalums to the ceramics and was happy with the result, if I recall well.
Best regards,
Mauricio
I can't provide good answers...only good (or bad) questions! 😉
The problem is that everything resonates in this universe as dear Chris allways is remembering me 😀
See this link as a light read before sleep:
http://www.diyhifi.org/forums/viewtopic.php?t=523
The problem is if the combination of C/L/R used sounds good to you or not 🙁
I can't predict how good or bad will some caps arrangement sound. I don't have scope nor know the maths and could never probably use the software that EE's uses so joyously
All I have is courage (read recklessness) and a happy iron...😀
My UCD400ST allways have sounded bad for me on the highs. I have tried many things. Now I run a sort of Pi filter for the main PS wich made a good improvement in overall sound. Recently I tried mods on the +/-15VDC on board supply: did not like the 22uF decoupling electrolytics "near" the darlingtons? (now I realize that your UCD180 lacks these, sorry for the confusion)
I found that //ing 100nF film caps to the electros was worthy, soundwise: cleaner sound.
Swaping those cheap electros for Panasonic FC 47uF//Wima 100nF resulted in very clean sound but without punch and bass.
Then I put there the smallest Rubycon that I had on hand, instead of Pana//Wima: rubycon ZL 270uF!! Not a good fit... but , again, the whole sound presentation changed dramatically: very deep bass, more dinamics and punch. But also harsher highs; back to where I started
After that, encouraged by Klaus' experience, I decided that the culprits were those little decoupling ceramic caps (C6-C7) near OPA2134 power pins. I took them off (my first time with SMD
) and put a pair of Black Gate NX 100nF in "antiparallel configuration" (sorry George). I had to drill the PCB for that (kids, don't do that at home 😀 )Now, the BG are well known for being very slow to burn-in. Untill know I hear absence of the previous HF noise (the Rubycon's did not liked the ceramics, and I don't blame them). The sound is lacking HF extension and "air" but is very, and I mean very, robust. Only time will tell if this is a winner combo...
What's the conclusion of this boring post? I'm affraid that every part on the power and the signal paths can affect adverselly the sound 🙁
The other conclusion is that I don't like "the sound" of ceramics at all!
Any opinions are welcome!
I will try to get some 47uF Rubycon ZA to try instead of the 270uF's.
ClassD tried before //ing tantalums to the ceramics and was happy with the result, if I recall well.
Best regards,
Mauricio
- Status
- Not open for further replies.