Awsome ! Thank you Andrea.
Great, in my case it will be just the I2S to PCM board with no FIFO as it exists some USB to I2S devices with isolator chip and isolated layout...
But does the WS and Blck with the I2S to PCM but without FIFO are reclocked as its MCLK is slaved ? That point is not clear for me. My understanding : it is not without the FIFO and if you use a third party USB to I2S : the fs should stay 32 or 64 fs (manual of I2S to PCM) and the result would be great only if this first board of the chain is also slaved by your MCLK ? Right ? Because ifnthere is a drift of the WS and Data in relation to the MCLK which is stable : we have always a jitter problem (ws & data timing given by the source board !)
Does your board allow to slave 3 devices : dac chip + USB to PCM + I2S to PCM with 3 uf.l connectors ? dac chip is slaved with the two pins (stacking), I see one uf.l, but is it possible to add one just near the first (no drift with such conf : more thanvone uf.l ?). On J1 : I can't understand if the empty uf.l trace give the same frequency than the populated uf.l on your photograph ?
Well if the xo is 6.4 max, the max sampling allowed by the I2S to PCM would be 48 hz ! (128 fs minimum for this device)
Good news given by Ben Mah
If the DAC chip works at MCLK (BCK=MCLK) I'm not sure you need the I2S to PCM board, I cannot see any real reason to not use simply the I2S signal, natural way for the TDA1541A.
This was discussed in Ian's thread some months ago, you can find different opinion about I2S and simultaneous mode.
On the other side I believe that Ian's FIFO should be a must, because it isolates the DAC from any source you are using.
I'm not sure that's possible to sync the I2S to PCM board from the MCLK at 5.6448 MHz (or 6.144 MHz).
I think the best is to ask Ian in the specific thread.
BTW, with my oscillator board you can use any frequency from 5 to 25-30 MHz.
That board does not provide any dividers, but if there will be interest I could develope a separate board to use in sandwich with the oscillator board (I prefer to avoid dividers if possible).
John, Pedja, Thorsten (?) seem to have simultanous mode on their flag ships....
The TDA allow both natural way (that's my understanding about the floating pin 27 to VDD1 pin 26 : any work about the quality of the VDD1 voltage rail on the the 27 pin ? a ferrite bead more than a strad to join them ? is the quality of the voltage change the quality of the sound with simultanous mode ? I read nothing about that... because maybe it's a bad question !).
Totaly agree with dividers after read both John and Marce fellows. The slower the better. So no multiplicators on your board : you need to choose the Crystal you need ? Right ? That's great ! I'm in, let try to low the cost of the Laptech by a GB. Anyway, on DIYHINK you can find also at 8 USD genuine NDK xo sourced from Japan.
Shane tried both and believe the simultanous seems better with him own listening tests.
The TDA allow both natural way (that's my understanding about the floating pin 27 to VDD1 pin 26 : any work about the quality of the VDD1 voltage rail on the the 27 pin ? a ferrite bead more than a strad to join them ? is the quality of the voltage change the quality of the sound with simultanous mode ? I read nothing about that... because maybe it's a bad question !).
Totaly agree with dividers after read both John and Marce fellows. The slower the better. So no multiplicators on your board : you need to choose the Crystal you need ? Right ? That's great ! I'm in, let try to low the cost of the Laptech by a GB. Anyway, on DIYHINK you can find also at 8 USD genuine NDK xo sourced from Japan.
Shane tried both and believe the simultanous seems better with him own listening tests.
Pedja yes, John and Thorsten I don't believe, but maybe I'm wrong.
To compare both I think there is a different way to follow.
I don't know how Shane has compared both, but the right way it's not to add the I2S to PCM board to the FIFO using the clock of the FIFO board as the Master clock, so the logic of the DAC works at a different frequency.
To do the right comparison, IMHO, the DAC should work at MCLK frequency (BCK=MCLK) in case of using I2S.
Does brand new John's reactor design does not allow it ? Hummm my understanding is very poor I have to say. And with all the FGPAs Mr T. use I really don't know, he can't say all since he works for AMR !
As a poor technician I'm agree with you about the MCLK of a great quality near the dac chip. I asked though in the digital question (see above) about a first device which can be slaved. That's right anyway if the I2S to PCM is used than you have to respect both 32 or 64 fs feeding for I2S and minimum 128 fs for the Mclk... so here a 6.x xo can be used but at least here the Bclk can not be equal to the Mclk.
Maybe ears can solve the problem after listening test... In my case I'm for a cheaper and simpliest way...
As a poor technician I'm agree with you about the MCLK of a great quality near the dac chip. I asked though in the digital question (see above) about a first device which can be slaved. That's right anyway if the I2S to PCM is used than you have to respect both 32 or 64 fs feeding for I2S and minimum 128 fs for the Mclk... so here a 6.x xo can be used but at least here the Bclk can not be equal to the Mclk.
Maybe ears can solve the problem after listening test... In my case I'm for a cheaper and simpliest way...
Last edited:
Hi Shane, Andrea,
What about a tube as the Bendix 6900 in NOS ? (not existing anymore ?)
or a modern 6SN7 (or military 5692) ? Sovtek 6H30PI ? What NOS tubes T. uses in him CD-77 ? Notice I know absolutly nothing about tubes ! (Shane you have a bad influence on me about tubes
(your photographs... true DIY !)... but Pedja input also?! (its always the war between the two technologies...)
Andrea, I have a look at your work about shunt PS... impressive. Did you have look at Peufeu measurement about low noise PS in the S-S Thread (I linked two month ago elsewhere in this thread in the beginning when we talked about shunt with Alexiis fellow ?)
Why more a shuntt PS in relation to more modern chip like : TPS7A4700 e.g. : 4.17uV Ultralow noise DAC power supply regulator +-12/15V 1A - DIYINHK for a - 15 V.
More subjective transcient and punch with a shunt ?
(I ask here not for the outputstages but for the 3 Voltage rails of the TDA1541 and eventually for the XO board and inputs stages...). I saw also Abraxalito has a simple Shunt PS shematic with low noise on his blog but seems more involved in the same time by working to reduce dramaticly the noise before the regs and go with ultra low ESR ( around 0.5 Mohms !!! iirc) with caps towers (this is art !).
A question comes in mind about ESL and caps for PS : is ist better to have a strong ESL factor to filter stronger the noise harmonics from the wall for our dacs ?
kind regards
Eldam
What about a tube as the Bendix 6900 in NOS ? (not existing anymore ?)
or a modern 6SN7 (or military 5692) ? Sovtek 6H30PI ? What NOS tubes T. uses in him CD-77 ? Notice I know absolutly nothing about tubes ! (Shane you have a bad influence on me about tubes
(your photographs... true DIY !)... but Pedja input also?! (its always the war between the two technologies...)Andrea, I have a look at your work about shunt PS... impressive. Did you have look at Peufeu measurement about low noise PS in the S-S Thread (I linked two month ago elsewhere in this thread in the beginning when we talked about shunt with Alexiis fellow ?)
Why more a shuntt PS in relation to more modern chip like : TPS7A4700 e.g. : 4.17uV Ultralow noise DAC power supply regulator +-12/15V 1A - DIYINHK for a - 15 V.
More subjective transcient and punch with a shunt ?
(I ask here not for the outputstages but for the 3 Voltage rails of the TDA1541 and eventually for the XO board and inputs stages...). I saw also Abraxalito has a simple Shunt PS shematic with low noise on his blog but seems more involved in the same time by working to reduce dramaticly the noise before the regs and go with ultra low ESR ( around 0.5 Mohms !!! iirc) with caps towers (this is art !).
A question comes in mind about ESL and caps for PS : is ist better to have a strong ESL factor to filter stronger the noise harmonics from the wall for our dacs ?
kind regards
Eldam
Last edited:
Hi Damien,
Thats an expensive tube.. similar to 5687 and 7044, E182CC etc. Voltage gain is just a bit less than 20, it has Rp about 2k and sets up anywhere between 5 and 20mA.
With the CD77, I'm not sure. Gomes/z or something unusual with 12At7 and 5687?.. dunno. I do know that Mr.T uses a CCS loaded 6072 in his personal DAC with the minimal bias thing, though.. and when asked he did comment that he thought D3A to be about the best tube for 1541A. I put two and two together and called it a day, not worrying about it anymore.
Regards,
Shane
Thats an expensive tube.. similar to 5687 and 7044, E182CC etc. Voltage gain is just a bit less than 20, it has Rp about 2k and sets up anywhere between 5 and 20mA.
With the CD77, I'm not sure. Gomes/z or something unusual with 12At7 and 5687?.. dunno. I do know that Mr.T uses a CCS loaded 6072 in his personal DAC with the minimal bias thing, though.. and when asked he did comment that he thought D3A to be about the best tube for 1541A. I put two and two together and called it a day, not worrying about it anymore.
Regards,
Shane
A D3A then a NOS tungsol or Rayton 5687 at 40 buks should be enough for most of us but not with a caps as DC like a CD77 but with a fine choosed trafo for (Bartolucci, Lundhal...) much more better result here is my understanding with my poor technic level... despite of the output stage is not all in a CD77 of course !
But my though is : why a tubestage for prople like me who have all S-S between the DAC and the speakers.... it's a waste. But I like the idea and will learn more about tubes : we have excelent litterature here in France where tubes are more famous (Hiraga and others) than Solid states devices !
btw my best digital device is a DVD-cd Philips player with a lundhal trafo, AK dac chip and hardly tweaked separated analogic PS and carrefully choosed caps for digitals : very very natural, even the good and more modern Subbu is less natural despite a slighty better precision. We can have subjective analogic feeling with S-S ihmo, but sure : heard so many times than a fully tubes system with nos tubes (not the modern ones) with less microphonics and better specs than todays are the best way (if the speakers is not a never finished Crompression drivers system ! )
I surmise the P; Rodgic flagship with S-S but output trafo maid just for it to be a killer also (Cheung is happy with it !)
like very much the pictures of your blog : because it's true diy, not just light tweaking as I do... but have to say I think I have a good experience with caps mixing... spending hours like a fool with its ! Like food : can be very subtle and make miracles... as people said to me ! I maid better than a famous man of my country who tweaks many cd players ! I'm persidtent !
cheers !
But my though is : why a tubestage for prople like me who have all S-S between the DAC and the speakers.... it's a waste. But I like the idea and will learn more about tubes : we have excelent litterature here in France where tubes are more famous (Hiraga and others) than Solid states devices !
btw my best digital device is a DVD-cd Philips player with a lundhal trafo, AK dac chip and hardly tweaked separated analogic PS and carrefully choosed caps for digitals : very very natural, even the good and more modern Subbu is less natural despite a slighty better precision. We can have subjective analogic feeling with S-S ihmo, but sure : heard so many times than a fully tubes system with nos tubes (not the modern ones) with less microphonics and better specs than todays are the best way (if the speakers is not a never finished Crompression drivers system !
)I surmise the P; Rodgic flagship with S-S but output trafo maid just for it to be a killer also (Cheung is happy with it !)
like very much the pictures of your blog : because it's true diy, not just light tweaking as I do... but have to say I think I have a good experience with caps mixing... spending hours like a fool with its ! Like food : can be very subtle and make miracles... as people said to me ! I maid better than a famous man of my country who tweaks many cd players ! I'm persidtent !
cheers !
Not sure that DIY are the most rational people and sometimes its just because you want to do it a certain way.
If you want really low output impedance (<600 ohm at >50Hz) with a tube stage you're left with options like variations on Mu follower, cathode follower, SS buffer or step down transformer. When you step down with the transformer you lose voltage gain, and you need a decent amount of primary inductance to get full BW with even 2k Rp types.. and then phase shift at lower frequencies may or may not be acceptable.
Of course with the Audial gear and its current injection and low output impedance, not having the need for an air gapped transformer, or the need for step down ratio… things become simpler, less expensive and more compact.
If you want really low output impedance (<600 ohm at >50Hz) with a tube stage you're left with options like variations on Mu follower, cathode follower, SS buffer or step down transformer. When you step down with the transformer you lose voltage gain, and you need a decent amount of primary inductance to get full BW with even 2k Rp types.. and then phase shift at lower frequencies may or may not be acceptable.
Of course with the Audial gear and its current injection and low output impedance, not having the need for an air gapped transformer, or the need for step down ratio… things become simpler, less expensive and more compact.
Thanks Shane for the inputs, I'm really ashamed about my lacking of knowledge with all those things. I just believed that the low current maid the things more easier for an output transformer here in relation to an amp... and than the problems was more to find the right ratio or maid it tailored by hand ("sur mesure") !
I really have to do the effort to open books !
My amp has a 100 k Ohms input stage ! But don't know if it's a real avantage as the impedance matching in voltage seems less important than with curent ! Maybe for single ended wires I use (75 cm max... sorry for metric).
Hummmm short night
....
see you, cheers
I really have to do the effort to open books
!My amp has a 100 k Ohms input stage !
But don't know if it's a real avantage as the impedance matching in voltage seems less important than with curent ! Maybe for single ended wires I use (75 cm max... sorry for metric).Hummmm short night
.... see you, cheers
With 100k input impedance, you really don't need to worry so much.. the rule of thumb is 1:10, so even 10k might be acceptable and 2k from 5687 is certainly no problem. If your amp has high enough input sensitivity then mu 20 from 5687 and an appropriate I/V resistor is probably enough, too.
Hi Shane, Andrea,
What about a tube as the Bendix 6900 in NOS ? (not existing anymore ?)
or a modern 6SN7 (or military 5692) ? Sovtek 6H30PI ? What NOS tubes T. uses in him CD-77 ? Notice I know absolutly nothing about tubes ! (Shane you have a bad influence on me about tubes
Andrea, I have a look at your work about shunt PS... impressive. Did you have look at Peufeu measurement about low noise PS in the S-S Thread (I linked two month ago elsewhere in this thread in the beginning when we talked about shunt with Alexiis fellow ?)
Why more a shuntt PS in relation to more modern chip like : TPS7A4700 e.g. : 4.17uV Ultralow noise DAC power supply regulator +-12/15V 1A - DIYINHK for a - 15 V.
More subjective transcient and punch with a shunt ?
(I ask here not for the outputstages but for the 3 Voltage rails of the TDA1541 and eventually for the XO board and inputs stages...). I saw also Abraxalito has a simple Shunt PS shematic with low noise on his blog but seems more involved in the same time by working to reduce dramaticly the noise before the regs and go with ultra low ESR ( around 0.5 Mohms !!! iirc) with caps towers (this is art !).
A question comes in mind about ESL and caps for PS : is ist better to have a strong ESL factor to filter stronger the noise harmonics from the wall for our dacs ?
kind regards
Eldam
As Shane said, I think the D3A triode strapped is a good enough choice. Not yet tried, but suggested to me from Thorsten.
The only problem seems the D3A has not sufficient gain to keep very low the I to V resistor value, I believe at least 18-20 ohm, so the voltage swing at the output of the DAC exceeds the maximum value of +/- 25 mV.
My regulator is not properly a shunt type, it's a series regulator using a push pull as the output device. Never tried the TPS7A4700, so I cannot tell nothing about.
Personally, I prefer series regulator rather than shunt type, at least to supply the analog section of the DAC.
But this is a subjective opinion.
For 'standard' 2vRMS output, assuming CCS load D3A will get you 98% of the way there with 18R I/V.
4mAAC pp x 18 ohm = 0.072vpp
0.072vpp x 77 = 5.4vpp
(5.4vpp/2) x 0.707 = 1.909vRMS
If you were wanting to strictly stick to 25mVDC offset value, I/V = 12R will have you at -24mVDC with -2mADC from the 1541A, D3A would have you with 1.3vRMS output.. depending on your overall gain structure and requirements, this could be more than enough. (or not).
Out of interest, Zanden 5000 using half 6922 and Vout 1.6vRMS
(1.6vRMS/0.707)x2 = 4.53vpp
4.53vpp / mu 33 = 0.137vpp
137mvpp /4mApp has I/V =~34R.
Shane
4mAAC pp x 18 ohm = 0.072vpp
0.072vpp x 77 = 5.4vpp
(5.4vpp/2) x 0.707 = 1.909vRMS
If you were wanting to strictly stick to 25mVDC offset value, I/V = 12R will have you at -24mVDC with -2mADC from the 1541A, D3A would have you with 1.3vRMS output.. depending on your overall gain structure and requirements, this could be more than enough. (or not).
Out of interest, Zanden 5000 using half 6922 and Vout 1.6vRMS
(1.6vRMS/0.707)x2 = 4.53vpp
4.53vpp / mu 33 = 0.137vpp
137mvpp /4mApp has I/V =~34R.
Shane
Last edited:
Hi guys,
Yes will start because the cost with the single ended Cen.
Will learn the tubes... seems more complicate because dispersions of values, drift, stronger devices for measurement and so on... and because todays my pre and amp are full silicon (yes... strongly artificial)
But tubes is a futur temptation... I have time... it's a hobby (wankenoby!)...time and wise are my allies !
Ryan, I will say as I receive the board, hope this week !
cheers
Yes will start because the cost with the single ended Cen.
Will learn the tubes... seems more complicate because dispersions of values, drift, stronger devices for measurement and so on... and because todays my pre and amp are full silicon (yes... strongly artificial
)But tubes is a futur temptation... I have time... it's a hobby (wankenoby!)...time and wise are my allies !
Ryan, I will say as I receive the board, hope this week !
cheers
Attachments
If yes, the Queen is fat ! Remenber me Lampizatruc ! Did he try also with 845 ?... I'm going to put a wire into a bottle of Champagne if the size matter !
Thank you Andrea.
Wow : as If my understanding is correct : i/v is maid by a 6.8 ohms ? Not saw much often a so low value.... nice.... with a LL transformer, seems the resistor sould be at least a Rohpoint.
... I'm going to put a wire into a bottle of Champagne if the size matter !Thank you Andrea.
Wow : as If my understanding is correct : i/v is maid by a 6.8 ohms ? Not saw much often a so low value.... nice.... with a LL transformer, seems the resistor sould be at least a Rohpoint.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.