Ok, call it reclocking, so... "I wouldn't use a PLL in the ASRC, neither in the reclocking circuit"
not sure it adds to the discussion but I answer:
1-) I've got 3 or 4 TDA 1541 (rest of old broken philips or marantz and think also like most of us it is still a good dac chip )
2-) please read again all what I wrote above, I think you will find a good enough answer of what i believe and what I don't believe.
Dear fellow members,
keep in mind the CD-77 is a system, not a DAC, there is a big difference.
If ever this workgroup will start, my suggestion is to think to a system rather than a DAC (that's the reason I did talk of DAC, clock and source as a system).
Obviously, it will take a long long time to develop such a system.
keep in mind the CD-77 is a system, not a DAC, there is a big difference.
If ever this workgroup will start, my suggestion is to think to a system rather than a DAC (that's the reason I did talk of DAC, clock and source as a system).
Obviously, it will take a long long time to develop such a system.
Thank you, It's usefull,
use simple 7812, 7912 pre-regulators for +5v, -5v : why not use better LDO ? SOIC MIcrel seem to be appreciate by many designers i.e. it's better because small grounf loop and can be soldered nearer of the device (short traces ?) ? maybe are they not better, is that you want to mean ?
if you have the energy, go for dem-reclocking (not very easy) I have to say it is the part i had difficulty to understand in the ECDesigns thread that I read twice ! And more after the V7 !
use passive IV. Ok here Rhopoint Z foil ! Better in the smd design or better the bulk one (Through hole) ? 200 ohms ? less ?
use a simple tube output stage (Audio Note Design or Mr. Fikus' design) which design ? Mu follower e.g. Something near the AMR CD777 ? : o,e first triode à la Telefunken and a 6 something after ? Problem here for output : Lundhal or over transformer are expensive; Teflon caps as well; A good enough MKP : which one is used to be good here ?
Last edited:
Yes maybe a very good one but is it possible to have something good enough (once again : better than most of the crap KITs is already a good enough goal) whitout spend a fortune or more than 6 months for a first throw ?
So, you want someone else to spend many thousands of dollars in order for you to be able to use those components you have?
I recognize the shematic of Lampizator... 
I don't.
Joshua agree with us that such a kit don't exist yet with its own goals or not enough easy for assembly.
This is a try to do better. That's why I thought he was in the ideal position as OP to drive the discussions. He changed his mind but not forbidd and then agree for such a beginning here.
I think it is a place to learn as wrote an american people before thinking I, and maybe you, were discouraging ! I think many are but not me, he missunderstood by lake of experience than a reference device is very usefull.
My limit is the time I can share and my level of english as I already wrote. Both are missing. But we have first fellows goodwilled like Andrea, Presapian, you, etc.
If the OP change his mind to think this time it's not the good place, I think he will say it to us.
I don't think this my project but a try to answer to the hope of the OP which is not alone to want it but more beamed : "Good enough" and DIY, the go no-go because a limit of time. I call this a beginning of participation.
But unhapilly, i find people loose themselves in rethoric instead action or lake of will.
Here once again is a to much long answer of what I see around me (not only at DIYAUDIO) : people spend more time to find excuse because both tey don't really want and catch the vertigo on a chair !
I would like to give an impulse... It's not easy ! Maybe it's impossible... In fact I think it is without a project manager to drive the discussion and fellows.
To be continued !
I don't.
Joshua agree with us that such a kit don't exist yet with its own goals or not enough easy for assembly.
This is a try to do better. That's why I thought he was in the ideal position as OP to drive the discussions. He changed his mind but not forbidd and then agree for such a beginning here.
I think it is a place to learn as wrote an american people before thinking I, and maybe you, were discouraging ! I think many are but not me, he missunderstood by lake of experience than a reference device is very usefull.
My limit is the time I can share and my level of english as I already wrote. Both are missing. But we have first fellows goodwilled like Andrea, Presapian, you, etc.
If the OP change his mind to think this time it's not the good place, I think he will say it to us.
I don't think this my project but a try to answer to the hope of the OP which is not alone to want it but more beamed : "Good enough" and DIY, the go no-go because a limit of time. I call this a beginning of participation.
But unhapilly, i find people loose themselves in rethoric instead action or lake of will.
Here once again is a to much long answer of what I see around me (not only at DIYAUDIO) : people spend more time to find excuse because both tey don't really want and catch the vertigo on a chair !
I would like to give an impulse... It's not easy ! Maybe it's impossible... In fact I think it is without a project manager to drive the discussion and fellows.
To be continued !
The 1541A loves to see a virtual ground at its output, not 47 or 200 ohm that exceeds the voltage swing limit (+/- 25mV) producing some distorsion.
That's the reason I'm experimenting on a IV tube stage where the dac see 1 ohm only, pratically a virtual ground.
Ok ... where are the SM digital bypass caps?
... not exactly. Those are R and L:
View attachment BOTTEM.PSC.pdf
I'm guessing in accordance with Guido's supply decoupling overview:
View attachment Supply_decoupling.pdf
But that was 1999 and the lead inductance of a through hole component means it can't decouple at very high freqs. With so much EMI around these days, SM chips and SM caps are almost a requirement... I imagine Guido selected low interference chips and then designed specific decoupling with the TH caps, SM resistors and SM inductors ... but wouldn't that make is hard to DIY? You'd need the same components he used?
... not exactly. Those are R and L:
View attachment BOTTEM.PSC.pdf
I'm guessing in accordance with Guido's supply decoupling overview:
View attachment Supply_decoupling.pdf
But that was 1999 and the lead inductance of a through hole component means it can't decouple at very high freqs. With so much EMI around these days, SM chips and SM caps are almost a requirement... I imagine Guido selected low interference chips and then designed specific decoupling with the TH caps, SM resistors and SM inductors ... but wouldn't that make is hard to DIY? You'd need the same components he used?
Last edited:
If you're even a bit serious about building a design you could easily study what successful others have already offered with respect to this in the DIY community.
If you respect Mr.T's designs, then accept his advice about the output (DC) voltage compliance not violating distortion spectra above a certain value - the value of which he has made public.
Indeed, he has shared enough info over the years to build. Start with his older designs, do some study and see what he may have changed over the years. His description of power supplies, shunt regs etc are all there.
Cross reference with what other notable manufacturers are doing.. Pedja uses separate windings for each supply. So, if you specify a transformer, just do that. He also mentions CM RC filtering after the rectifiers.. you can see the resistor values from the internal pics of his DAC.
WRT decoupling caps.. AMR have made it public that they use SMD Panasonic PPS capacitors??.. the same that you can see in the internal pics of Pedja's tda1541A DAC. So.. just use those. What values?.. read around, have a look at some of T's older schematics.
You have now Power supplies, decoupling caps, I/V resistor values - build an output stage. Use T's 6922 to start. What cap to use?.. search the archives. Otherwise use what AMR were using 2.2uF Mundorf ZN.
I2S attenuation. Did you want to use I2S or PCM mode (simultaneous mode) to the tda1541A?. What are the advantages etc?. Where to find a decent I2S to PCM converter?. Its all in the archives.
The biggest problem is going to be PCB layout. You could copy something, but it would need to be known as 'good'. You might find someone (Pedja?) who has discontinued a version of tda1541A commercial product and offer something in exchange. Charity contribution of his choosing per item, whatever it is. If it conflicts with current commercial offerings I doubt you'd get far, so look for a discontinued model that was well regarded. There should be more than one.
Just build the tda1541 with associated supplies and output provision. Put a header at the BCK/WS/DATA L/DATA R inputs so that people can build it in a modular form. They can use what they want, just get the core board 'right'.
if you have the energy, go for dem-reclocking (not very easy)
I would suggest reading more into DEM clocking, as opposed to reclocking. As in, synching it to an already reclocked (or not) clock source. This doesn't need to be on the core board, either. Leave it as an option. Just put two thru hole points at pins 16, 17 and use a simple capacitor to start with.
There is a data extract from Philips that mentions if "fDEM is 4x signal frequency then absolute reduction for DEM filtering is needed." When you read about DEM through the archives you see all about Grundig 9009 where DEM is synched to WS. People tried it and liked it, even with NOS. Yet, 9009 runs 4x OS, so WS is 176.4kHz. Just happens to line up exactly with what the data extract says if you feed the chip with 44.1kHz signal. Problem with NOS is that in this case fDEM would be 44.1kHz if you synched to WS. Ideally, you'd want a very clean divider to get 4x actual signal frequency. So, 176kHz for 44.1kHz, or 768kHz for 192kHz (simultaneous mode allows for). Then work back to find what clock source you're going to use..
In all honesty with a bit of application, if you get the PCB layout for a small core board sorted, the rest can be SOTA for minimal cost. Have a look at Ians FIFO>ISO>re-CLOCK>I2S to PCM offerings. He runs a group buy, the guys already on the list could surely use the numbers.. .. .. And you know, even if you can get just obtain the gerbers from a known to be good discontinued model, for a worthwhile contribution per unit made, the FIFO guy might be open to running off the PCB's for the 1541A board too.
Again, you just gotta find an amicable way to get a decent PCB. Difficult?. I don't know. Worth a try?, definitely. As we all know, theres no shortage of worthwhile charities out there which would surely benefit. And for those funds to come from an endeavour started by enthusiasts who are in pursuit of a better way to appreciate one Gods greatest gifts, then all the better.
What price to put on a PCB when the proceeds to charity?… it very quickly stops being about the individual and opens up into something better..
If you respect Mr.T's designs, then accept his advice about the output (DC) voltage compliance not violating distortion spectra above a certain value - the value of which he has made public.
Indeed, he has shared enough info over the years to build. Start with his older designs, do some study and see what he may have changed over the years. His description of power supplies, shunt regs etc are all there.
Cross reference with what other notable manufacturers are doing.. Pedja uses separate windings for each supply. So, if you specify a transformer, just do that. He also mentions CM RC filtering after the rectifiers.. you can see the resistor values from the internal pics of his DAC.
WRT decoupling caps.. AMR have made it public that they use SMD Panasonic PPS capacitors??.. the same that you can see in the internal pics of Pedja's tda1541A DAC. So.. just use those. What values?.. read around, have a look at some of T's older schematics.
You have now Power supplies, decoupling caps, I/V resistor values - build an output stage. Use T's 6922 to start. What cap to use?.. search the archives. Otherwise use what AMR were using 2.2uF Mundorf ZN.
I2S attenuation. Did you want to use I2S or PCM mode (simultaneous mode) to the tda1541A?. What are the advantages etc?. Where to find a decent I2S to PCM converter?. Its all in the archives.
The biggest problem is going to be PCB layout. You could copy something, but it would need to be known as 'good'. You might find someone (Pedja?) who has discontinued a version of tda1541A commercial product and offer something in exchange. Charity contribution of his choosing per item, whatever it is. If it conflicts with current commercial offerings I doubt you'd get far, so look for a discontinued model that was well regarded. There should be more than one.
Just build the tda1541 with associated supplies and output provision. Put a header at the BCK/WS/DATA L/DATA R inputs so that people can build it in a modular form. They can use what they want, just get the core board 'right'.
if you have the energy, go for dem-reclocking (not very easy)
I would suggest reading more into DEM clocking, as opposed to reclocking. As in, synching it to an already reclocked (or not) clock source. This doesn't need to be on the core board, either. Leave it as an option. Just put two thru hole points at pins 16, 17 and use a simple capacitor to start with.
There is a data extract from Philips that mentions if "fDEM is 4x signal frequency then absolute reduction for DEM filtering is needed." When you read about DEM through the archives you see all about Grundig 9009 where DEM is synched to WS. People tried it and liked it, even with NOS. Yet, 9009 runs 4x OS, so WS is 176.4kHz. Just happens to line up exactly with what the data extract says if you feed the chip with 44.1kHz signal. Problem with NOS is that in this case fDEM would be 44.1kHz if you synched to WS. Ideally, you'd want a very clean divider to get 4x actual signal frequency. So, 176kHz for 44.1kHz, or 768kHz for 192kHz (simultaneous mode allows for). Then work back to find what clock source you're going to use..
In all honesty with a bit of application, if you get the PCB layout for a small core board sorted, the rest can be SOTA for minimal cost. Have a look at Ians FIFO>ISO>re-CLOCK>I2S to PCM offerings. He runs a group buy, the guys already on the list could surely use the numbers.. .. .. And you know, even if you can get just obtain the gerbers from a known to be good discontinued model, for a worthwhile contribution per unit made, the FIFO guy might be open to running off the PCB's for the 1541A board too.
Again, you just gotta find an amicable way to get a decent PCB. Difficult?. I don't know. Worth a try?, definitely. As we all know, theres no shortage of worthwhile charities out there which would surely benefit. And for those funds to come from an endeavour started by enthusiasts who are in pursuit of a better way to appreciate one Gods greatest gifts, then all the better.
What price to put on a PCB when the proceeds to charity?… it very quickly stops being about the individual and opens up into something better..
Last edited:
Hi AudiLapDance.
Sorry for the delay... I have soo little time for diy these days...
I was wrong :-(
In the original design there are smd L's and R's at the Bottom-Layer, but no smd capacitors.
PSU caps are located on the Top-layer, (through-hole ceramic caps).
I bought this dac 14 years ago => back to my old notes (as I found no through-hole caps on my assembly):
My unit has been heavily modified, this is the list of mods acc to my (10 year old) notes:
- All HF Decoupling caps were changed to 1210 1uF 16V
- All Power feed to the chips were replaced by Enhanced TL431 Shunt Reg, instead of the original Lm317.
- All digital IC's were packed in self-adhesive, grounded, copper shields.
- Input transformer removed.
- nos implemented
- Analog low-pass-filter removed
- Analog Tube Stage by-passed
- A very high quality 75 ohm RF cable and PCB socket was used instead of the original SPDIF input socket.
Last edited:
Very sound advice...
Exactly what i have tried to get across (unsucessfully).
Many clever and passionate people here build low freq valve amps etc.
Dacs are different as they are HF Mixed Signal Implementations.
Low freq design practices do not apply to higher frequencies without modifications.
Last edited:
Hi there,
As far as I know, I don't owe you anything.
Furthermore, I distastes you attempts to coerce me. The more you attempt, the more you put me off.
I'm out of this project, I explained above why, I'll repeat it.
1. Such a project calls for investment of many thousands of dollars, which I don't have.
2. Such a project will require thousands of hours, which I don't have at present. I'm busy in another project of a documentary movie.
3. My aim at DIY is to obtain a better sound quality than I have already, at an affordable cost. This is unattainable.
3a. I'm not into DIY neither for learning nor for the joy of doing it myself.
3b. The best sounding CDP is the AMR CD-77, which I cannot afford. Embarking a project of developing a DAC may well cost me more than the AMR CD-77.
3c. The second best sounding CDP/DAC is the AMR CD-777, which I own. There is no point for me investing even 1 hour and $1 in something that may not sound better.
3d. The older Tent DAC is of about the same sound quality as the AMR CD-777. The newer Tent DAC is of a lesser sound quality.
3e. Should I embark a project of developing a DAC, the best I may possibly attain, after thousands of hours and thousands of dollars, is the sound quality of Tent DAC.
3f. It will be stupid of me to spend so much time and so much money (should I somehow gather the money) for something of a lesser sound quality than I already have.
End of story.
Not entirely true.
The output compliance of TDA1541a, mentioned in the datasheet, is quite conservative.
I performed many measurements in the late 90:s, concluding that THD is not negatively effected
as long as the IV resistor value is kept to less than 50 ohms (with a twist).
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.