"As good as BG caps are, I still think that there are better film caps out there."
for signal coupling, yes, certainly, but BGs are pretty good for the price, and the "tube-like" softness
does partially disappear after 4 to 6 weeks 
"The big size of those caps is an inconvenience in that type of design,"
yes, definitely
"I very much like Siemens MKV caps."
me too, but also MIT RTX
"it seems like Mundorf Supreme Silver are the ones to go with I never tried them though."
Many German audiophiles like Mundorf, and many loudspeaker manufacturers here use Mundorfs in the crossovers of their top-price loudspeakers. That said, I was not impressed by Mundorf MCap polypropylene, but Mundorf MCap ZN is different species, and well worth pursuing--agree fully with "coredump" on sonics. Because of large size, I have never used MCap Supreme (silver or normal)--the L-cancelling version. But I have sometimes had subjective problems with "artificial smoothness" of highs with some silver foil caps.
"Geheimtipp": Mundorf also make excellent air-core foil inductors, with immaculate foil stock
"What I plan on improving is revised layout, getting rid of ferrite beads, using better caps for battery bapass, better input coupling caps and better overall bypass scheme."
The eternal experimenter! I am surprised that double-mono TDA 1543s is not on your list of experiments: your layout can accomodate another TDA 1543 on the bottom of the CS8412 with the pins still matching up. Are you using batteries again? Please keep us posted!
"I also want to try those bigger Caddocks resistors (what value are you using here?)"
I followed data-sheet very closely on selection of I/V and Vref resistor values, also paying attention to Doede's results with FFT tests.
I used for Caddock TF020, MK-132, Vishay, and Holcro:
I/V = 2k21
Vref = 1k40
for Riken .5watt I used:
I/V = 2k2
Vref = 1k5
This combination cannot drive certain end stages, but should work with your gainclone. Higher I/V values yielded better dynamics at the price of poorer distortion spectrum, and worse conversion linearity--both apparent to me on FFT analysis and through subjective listening.
I never had a chance to use TF020 before in my projects because of the long pitch, but here they fit into my layout perfectly. I was positively suprised by their performance after a two-week break-in. The TF020s have the naturalness and transparency of the Vishays, but not their mellowness/blandness. (I would say that the TF020s are only mellow when the music is mellow.) They also have the spectacular spatial presentation of the MK-132s, but very little of the latter's muddiness. With TF020 detail is also better than Vishay--with Vishay, I felt I had to listen very hard to hear the details, now the detail is made more obvious. But with both Vishay and TF020, I felt that NOS TDA 1543 "grows up," becomes more adult and sophisticated...
"Did you use input buffer chip?"
No. A transformer from Scientific Conversions.
"I assume both of you are running your DACs at 8V?"
Yes, my TDA1543 is run at 8volts: the manufacturing tolerances of the TDA1543 make selection an issue when running at this voltage: I take the best out of four. Supplier: reichelt.de
Best regards
for signal coupling, yes, certainly, but BGs are pretty good for the price, and the "tube-like" softness
does partially disappear after 4 to 6 weeks "The big size of those caps is an inconvenience in that type of design,"
yes, definitely
"I very much like Siemens MKV caps."
me too, but also MIT RTX
"it seems like Mundorf Supreme Silver are the ones to go with I never tried them though."
Many German audiophiles like Mundorf, and many loudspeaker manufacturers here use Mundorfs in the crossovers of their top-price loudspeakers. That said, I was not impressed by Mundorf MCap polypropylene, but Mundorf MCap ZN is different species, and well worth pursuing--agree fully with "coredump" on sonics. Because of large size, I have never used MCap Supreme (silver or normal)--the L-cancelling version. But I have sometimes had subjective problems with "artificial smoothness" of highs with some silver foil caps.
"Geheimtipp": Mundorf also make excellent air-core foil inductors, with immaculate foil stock "What I plan on improving is revised layout, getting rid of ferrite beads, using better caps for battery bapass, better input coupling caps and better overall bypass scheme."
The eternal experimenter! I am surprised that double-mono TDA 1543s is not on your list of experiments: your layout can accomodate another TDA 1543 on the bottom of the CS8412 with the pins still matching up. Are you using batteries again? Please keep us posted!
"I also want to try those bigger Caddocks resistors (what value are you using here?)"
I followed data-sheet very closely on selection of I/V and Vref resistor values, also paying attention to Doede's results with FFT tests.
I used for Caddock TF020, MK-132, Vishay, and Holcro:
I/V = 2k21
Vref = 1k40
for Riken .5watt I used:
I/V = 2k2
Vref = 1k5
This combination cannot drive certain end stages, but should work with your gainclone. Higher I/V values yielded better dynamics at the price of poorer distortion spectrum, and worse conversion linearity--both apparent to me on FFT analysis and through subjective listening.
I never had a chance to use TF020 before in my projects because of the long pitch, but here they fit into my layout perfectly. I was positively suprised by their performance after a two-week break-in. The TF020s have the naturalness and transparency of the Vishays, but not their mellowness/blandness. (I would say that the TF020s are only mellow when the music is mellow.) They also have the spectacular spatial presentation of the MK-132s, but very little of the latter's muddiness. With TF020 detail is also better than Vishay--with Vishay, I felt I had to listen very hard to hear the details, now the detail is made more obvious. But with both Vishay and TF020, I felt that NOS TDA 1543 "grows up," becomes more adult and sophisticated...
"Did you use input buffer chip?"
No. A transformer from Scientific Conversions.
"I assume both of you are running your DACs at 8V?"
Yes, my TDA1543 is run at 8volts: the manufacturing tolerances of the TDA1543 make selection an issue when running at this voltage: I take the best out of four. Supplier: reichelt.de
Best regards
"I am surprised that double-mono TDA 1543s is not on your list of experiments: your layout can accomodate another TDA 1543 on the bottom of the CS8412 with the pins still matching up. Are you using batteries again?"
Yes I'm using batteries. Did you noticed any improvement with separate batteries for receiver and chip? I'm using 2 of them now, and it seems to be better (although initially it seemed worse).
As to dual mon, how exactly is it implemented?
I noticed you didn't have AN8008 available to try. If you want I can send you one, as I'm curious how you'd like them.
Ultimately I'd like to move DAC into the transport with I2S direct. Although by brief comparison the setup with separate DAC and CD-Pro seemed to sound better than one piece Marantz CD-94 (with Tent clock) run from a battery. I was a bit puzzled by that.
Yes I'm using batteries. Did you noticed any improvement with separate batteries for receiver and chip? I'm using 2 of them now, and it seems to be better (although initially it seemed worse).
As to dual mon, how exactly is it implemented?
I noticed you didn't have AN8008 available to try. If you want I can send you one, as I'm curious how you'd like them.
Ultimately I'd like to move DAC into the transport with I2S direct. Although by brief comparison the setup with separate DAC and CD-Pro seemed to sound better than one piece Marantz CD-94 (with Tent clock) run from a battery. I was a bit puzzled by that.
I have MCap Supreme Silver in Oil on my 8x1543 NOSDAC. The Mundorfs are very liquid, detailed (not analytical) and very musical. It's like having the best of Auricaps and Jensen PIOs in one cap.
They have been burning for 28hrs. In the meantime, I am finalizing my 16x1543 & I will transfer the MCaps and the chips from the 8x1543 DAC there when done.
Peter,
You mentioned that you'd be ditching the ferrite beads? Any particular reason, sonics-wise? I am placing these on the Vins of all the chips of my revised DAC. I'd be interested to know why you decided not to use them anymore.
TIA,
Grant
They have been burning for 28hrs. In the meantime, I am finalizing my 16x1543 & I will transfer the MCaps and the chips from the 8x1543 DAC there when done.
Peter,
You mentioned that you'd be ditching the ferrite beads? Any particular reason, sonics-wise? I am placing these on the Vins of all the chips of my revised DAC. I'd be interested to know why you decided not to use them anymore.
TIA,
Grant
Well, I was reading this post by Charles Hansen:
http://www.diyaudio.com/forums/showthread.php?postid=382332#post382332
http://www.diyaudio.com/forums/showthread.php?postid=382332#post382332
"What process do you use to select the best TDA1543 out of the four?"
socket the IC, use a spectrum analyzer at digital silence and also a signal (say, sine at -6dB). Deviation between each chip can be quite large, and, strangely, departure from linearity and distortion will almost always be higher on left channel (roughly corresponds with "measurements" section of Stereophile review of Shigaraki DAC), and DC offset is almost always higher for right channel (see also thread here "Flying higher and higher with the Doede DAC"). If you are stacking TDA 1543s a la DDDAC, you will have to throw out a lot of silicon.
"Yes I'm using batteries. Did you noticed any improvement with separate batteries for receiver and chip?"
yes. Separate supplies eliminates a lot of crosstalk between the receiver and the DAC chip. Make sure each batter negative pole is fed to the chip supplied and ground from there--i.e. do not tie negatives at the battery itself. But while battery PS is best in terms of ripple, it is not panacea to cure all evils: batteries themselves have their own noise (TNT site has some nice plots of battery noise), noise will still be injected from regs, and you still have to bypass locally.
"I'm using 2 of them now, and it seems to be better (although initially it seemed worse)."
For the full monty, you can use 4, one for the SN 74179, one for the TDA 1543, two for the CS 8412.
"As to dual mon, how exactly is it implemented?"
Ha ha. I thought you would tell me. 47 Labs has a dual mono NOS TDA 1434, called the "Gemini." Recent review said that there are two separate pcbs, each with its own TDA 1543T and CS 8412 (smd version) fed from same EAIJ signal. Simple, but will work, it seems.
"I noticed you didn't have AN8008 available to try. If you want I can send you one, as I'm curious how you'd like them."
Thanks. I can compare with the Motorolas I am using. I was waiting until I needed more parts from Digi-Key before I ordered some--expensive to ship small order here, esp. with all that bio-friendly protective packaging!
"Ultimately I'd like to move DAC into the transport with I2S direct."
Good idea! why use a receiver when you do not have to? I am planning the same inside a Z-Systems digital preamp, positioning TDA 1543 directly above a CS8402 transmitter--the pins match perfectly!
"Although by brief comparison the setup with separate DAC and CD-Pro seemed to sound better than one piece Marantz CD-94 (with Tent clock) run from a battery. I was a bit puzzled by that."
Can't tell. Bit puzzling esp. if you are running at consumer level between the CD-Pro and the NOS TDA 1543 with a long interconnect. Normally, running CS8412 at pro level, fully diffential, allows it to perform better when boxed separately.
Kuei outlined a scheme to replace SN74179 (which unfortunately is supply sensitive) using a fast dual op amp (I think he recommended LM6182) set up to boost S/PDIF signal (i.e. .5V) to 10V P-P TTL
Kittaylor: What SMD resistors were you using for passive I/V, and why did you choose them in the end after trying out all of those other resistors you tried (AN Tantals, Welwyn, etc.)? I never wanted to try out SMD resistors for I/V in 3D layout because I needed at least 5mm pitch to breach the distance between the TDA output pins and the ground plane, and working with through-hole resistors made experiments with placement there much easier. I also never wanted to try out active I/V with TDA 1543, because the unique feature of this chip is that it is the one dac chip where passive I/V is not a big problem. Kusonoki is explicit on this, although he sometimes implies that the choice of TDA1543 was by accident at first. Thus I decided very early in design process to exploit this possibility of TDA1543 and take a vacation from active I/V. And all of the resistors I tested in passive I/V were all "above average" in dynamics--combined with multiple PSUs, the DAC even has "too much slam" for some people who have heard it here. Please don't misinterpret, this is just meant to explain my reasoning--I have lived too long in Germany to sound polite in English.
Best regards.
socket the IC, use a spectrum analyzer at digital silence and also a signal (say, sine at -6dB). Deviation between each chip can be quite large, and, strangely, departure from linearity and distortion will almost always be higher on left channel (roughly corresponds with "measurements" section of Stereophile review of Shigaraki DAC), and DC offset is almost always higher for right channel (see also thread here "Flying higher and higher with the Doede DAC"). If you are stacking TDA 1543s a la DDDAC, you will have to throw out a lot of silicon.
"Yes I'm using batteries. Did you noticed any improvement with separate batteries for receiver and chip?"
yes. Separate supplies eliminates a lot of crosstalk between the receiver and the DAC chip. Make sure each batter negative pole is fed to the chip supplied and ground from there--i.e. do not tie negatives at the battery itself. But while battery PS is best in terms of ripple, it is not panacea to cure all evils: batteries themselves have their own noise (TNT site has some nice plots of battery noise), noise will still be injected from regs, and you still have to bypass locally.
"I'm using 2 of them now, and it seems to be better (although initially it seemed worse)."
For the full monty, you can use 4, one for the SN 74179, one for the TDA 1543, two for the CS 8412.
"As to dual mon, how exactly is it implemented?"
Ha ha. I thought you would tell me.
47 Labs has a dual mono NOS TDA 1434, called the "Gemini." Recent review said that there are two separate pcbs, each with its own TDA 1543T and CS 8412 (smd version) fed from same EAIJ signal. Simple, but will work, it seems."I noticed you didn't have AN8008 available to try. If you want I can send you one, as I'm curious how you'd like them."
Thanks. I can compare with the Motorolas I am using. I was waiting until I needed more parts from Digi-Key before I ordered some--expensive to ship small order here, esp. with all that bio-friendly protective packaging!
"Ultimately I'd like to move DAC into the transport with I2S direct."
Good idea! why use a receiver when you do not have to? I am planning the same inside a Z-Systems digital preamp, positioning TDA 1543 directly above a CS8402 transmitter--the pins match perfectly!
"Although by brief comparison the setup with separate DAC and CD-Pro seemed to sound better than one piece Marantz CD-94 (with Tent clock) run from a battery. I was a bit puzzled by that."
Can't tell. Bit puzzling esp. if you are running at consumer level between the CD-Pro and the NOS TDA 1543 with a long interconnect. Normally, running CS8412 at pro level, fully diffential, allows it to perform better when boxed separately.
Kuei outlined a scheme to replace SN74179 (which unfortunately is supply sensitive) using a fast dual op amp (I think he recommended LM6182) set up to boost S/PDIF signal (i.e. .5V) to 10V P-P TTL
Kittaylor: What SMD resistors were you using for passive I/V, and why did you choose them in the end after trying out all of those other resistors you tried (AN Tantals, Welwyn, etc.)? I never wanted to try out SMD resistors for I/V in 3D layout because I needed at least 5mm pitch to breach the distance between the TDA output pins and the ground plane, and working with through-hole resistors made experiments with placement there much easier. I also never wanted to try out active I/V with TDA 1543, because the unique feature of this chip is that it is the one dac chip where passive I/V is not a big problem. Kusonoki is explicit on this, although he sometimes implies that the choice of TDA1543 was by accident at first. Thus I decided very early in design process to exploit this possibility of TDA1543 and take a vacation from active I/V. And all of the resistors I tested in passive I/V were all "above average" in dynamics--combined with multiple PSUs, the DAC even has "too much slam" for some people who have heard it here. Please don't misinterpret, this is just meant to explain my reasoning--I have lived too long in Germany to sound polite in English.
Best regards.
I wouldn't be using receiver chip when I2S direct connecting. It just presently, everything is separate. Keeping things separate allows more experimentation. If the DAC is inside player, it's not always convenient to take things apart.
BTW, I noticed that the DAC is still sensitive to enclosure and for now it sounds the best when not boxed.
Since I don't have spectrum analyser, I listen to few chips and chose the one sounding best. BTW, there is a quite a difference between chips with a white stripe and without it. Which ones measured better?
BTW, I noticed that the DAC is still sensitive to enclosure and for now it sounds the best when not boxed.
Since I don't have spectrum analyser, I listen to few chips and chose the one sounding best. BTW, there is a quite a difference between chips with a white stripe and without it. Which ones measured better?
Balanced Output
> If balanced outputs be used, it somewhat complicates 3-D, p2p layout.
This is the main reason why I am not using this at the moment.
We have built a DDDAC which I believe is similar circuit, with reclocking, standard parts (dale 0.1%, MKP & FC caps). Compared to my Meridian 508 (20 bit, box standard, no mods), it was at most on par with the phono outputs from the 508 (somewhat different sound, some music is better on one and some on the other). I still clearly prefer the 508 balanced.
Which is also why I'm going for the Sony SCD mods, at least for now. But I am open to hear from your experiences.
Patrick
> If balanced outputs be used, it somewhat complicates 3-D, p2p layout.
This is the main reason why I am not using this at the moment.
We have built a DDDAC which I believe is similar circuit, with reclocking, standard parts (dale 0.1%, MKP & FC caps). Compared to my Meridian 508 (20 bit, box standard, no mods), it was at most on par with the phono outputs from the 508 (somewhat different sound, some music is better on one and some on the other). I still clearly prefer the 508 balanced.
Which is also why I'm going for the Sony SCD mods, at least for now. But I am open to hear from your experiences.
Patrick
FWIW, I preferred 'normal' metal film resistors (1%) very much to Vishay-Dales with my DDDAC, particularly in the signal path (I/V stage). Vishays seem to change the whole tonality towards 'bright and thin'. I'd like to borrow a 'scope one day and try to measure this effect, as I also had it with my GC.
Best regards,
Oliver
Best regards,
Oliver
Slawney
No problem with your English.
For passive I/V I used Multicomp 1% 0.125W 1206 SMD resistors. These can be soldered directly between the TDA1543 and the copper tape ground plane, a little tricky (use a chisel shaped tip) but the shortest path.
You can also shield the TDA1543 with copper tape and soldered the resistors directly on to the top of the chip, which is easier to do and allows quick comparisons.
The main reason for trying them was a "less is more" ideological preference for small things. I think they have the least sound of their own out of all the resistors I've tried, leaving me to concentrate on tweaking the DAC elsewhere.
To be honest, active I/V with AD817 doesn't sound all that different to passive resistor I/V. You might say the balance of colorations was subjectively preferable. I was just something I tried for the sake of it.
No problem with your English.
For passive I/V I used Multicomp 1% 0.125W 1206 SMD resistors. These can be soldered directly between the TDA1543 and the copper tape ground plane, a little tricky (use a chisel shaped tip) but the shortest path.
You can also shield the TDA1543 with copper tape and soldered the resistors directly on to the top of the chip, which is easier to do and allows quick comparisons.
The main reason for trying them was a "less is more" ideological preference for small things. I think they have the least sound of their own out of all the resistors I've tried, leaving me to concentrate on tweaking the DAC elsewhere.
To be honest, active I/V with AD817 doesn't sound all that different to passive resistor I/V. You might say the balance of colorations was subjectively preferable. I was just something I tried for the sake of it.
Kit:
thanks for the tips on positioning and soldering SMD I/V and Vref resistors. Many audiophiles slag SMD resistors for magnetism and sonics, but your experiences seem to indicate that they can be equally as good as boutique, and satisfying from a design perspective that emphasizes simplicity.
I like to ground these resistors as close to one another as possible on the foil plane, which seems easier to do with axial leads.
"To be honest, active I/V with AD817 doesn't sound all that different to passive resistor I/V. You might say the balance of colorations was subjectively preferable. I was just something I tried for the sake of it."
So I guess the SMD resistors was "less is more"
and the AD817 was "more is the same"
Has anyone tried an LCR filter before the regs? What values?
I have only ever measured the TDA1543s with the white stripe.
Best regards,
James
thanks for the tips on positioning and soldering SMD I/V and Vref resistors. Many audiophiles slag SMD resistors for magnetism and sonics, but your experiences seem to indicate that they can be equally as good as boutique, and satisfying from a design perspective that emphasizes simplicity.
I like to ground these resistors as close to one another as possible on the foil plane, which seems easier to do with axial leads.
"To be honest, active I/V with AD817 doesn't sound all that different to passive resistor I/V. You might say the balance of colorations was subjectively preferable. I was just something I tried for the sake of it."
So I guess the SMD resistors was "less is more"
and the AD817 was "more is the same"
Has anyone tried an LCR filter before the regs? What values?
I have only ever measured the TDA1543s with the white stripe.
Best regards,
James
I'm using LCRCRC filtering in AC powered version of a DAC, but untill I get proper PCB, I won't be able to comment much on the effect of particular elements in this setup. It was mostly influenced by Kusunoki earlier supplies. The caps are 3,300u and resistors 10-20 ohm. The inductor is on order of 1mH, IIRC.
Nevertherless, AC powered supply wasn't any worse than batteries. The strength of batteries is in very good midrange, AC supply is better in high frequency extention, when so called air seems to be more present.
Nevertherless, AC powered supply wasn't any worse than batteries. The strength of batteries is in very good midrange, AC supply is better in high frequency extention, when so called air seems to be more present.
Everytime I have tried SM resistors they have been awful.
I'm not sure if its the brand or just being sm that makes them
that way. Very compressed and grey and thin. I mod a lot
of sonys for my own personal use and they are the first thing to
go and always a big improvement. Holcos, Dales or Resistas
are my preferred with tants thrown in to taste.
I'm not sure if its the brand or just being sm that makes them
that way. Very compressed and grey and thin. I mod a lot
of sonys for my own personal use and they are the first thing to
go and always a big improvement. Holcos, Dales or Resistas
are my preferred with tants thrown in to taste.
regulator entry filter
My idea as far as supply filtering is to do it in three/four stages:
1. A discrete or integrated (i.e. Schaffner) line filter before the transformer
2.) LCR filtering after the diodes--like Kusonoki
3.) A "regulator entry filter" mounted directly on the regulator pins
4.) Local decoupling with 1206 ceramics on the supply pins of CS8412, etc.
Does anyone have suggestions for "stage 3" in the above scheme?
I am looking for a minimum number of components (say, one SMD resistor, two caps, one fixed inductor) to be mounted directly on the "Vin" and "ground" leads of a TO-220-or-smaller linear regulator: something like a low-pass filter that attenuates mid-range and higher-frequency RFI, with mild damping, minimum resonance, and which does not lead to an oscillating, unstable regulator as result.
My idea as far as supply filtering is to do it in three/four stages:
1. A discrete or integrated (i.e. Schaffner) line filter before the transformer
2.) LCR filtering after the diodes--like Kusonoki
3.) A "regulator entry filter" mounted directly on the regulator pins
4.) Local decoupling with 1206 ceramics on the supply pins of CS8412, etc.
Does anyone have suggestions for "stage 3" in the above scheme?
I am looking for a minimum number of components (say, one SMD resistor, two caps, one fixed inductor) to be mounted directly on the "Vin" and "ground" leads of a TO-220-or-smaller linear regulator: something like a low-pass filter that attenuates mid-range and higher-frequency RFI, with mild damping, minimum resonance, and which does not lead to an oscillating, unstable regulator as result.
Re: regulator entry filter
http://www.diyaudio.com/forums/showthread.php?s=&threadid=7265&highlight=
Please seeslawney said:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=7265&highlight=
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.