fabrice63 , do you understand ad hominem, and I do know what the "we" are talking about.
I recognize public appeals (as you have done in this forum) to at minimum cause consideration of censuring others to potential threaten their freedom of speech and/or expression. Your claim to do so seems reliant upon your irrefutable conclusion that the dialog is outside the scope of the thread, a condition that no-one has any responsibility to prove to you, being true or false, except perhaps the moderators. So why didn't you report this to them in lieu of creating the countless off topic texts that you are now responsible for... as notwithstanding suggesting that my submissions were disingenuous?
Remember , it is not about me , it is about the forum , the more you give to it the more it lives , and words are not always enought ....
would you like to read comments without knowing if it is based on real experiences or not , free speech is not the deal , truth is , I am maybe wrong but I thought that this was what we were all here for ......
again , the forum is wide open for all your thought , respect this thread and don't do it for me , do it for all the other menbers ...
.
would you like to read comments without knowing if it is based on real experiences or not , free speech is not the deal , truth is , I am maybe wrong but I thought that this was what we were all here for ......
again , the forum is wide open for all your thought , respect this thread and don't do it for me , do it for all the other menbers ...
.
Last edited:
Let's be back to the subject 🙂
Here is my next pcb , for the simultaneous version , including 50 Hz DEM and some others features
it is battery powered , the logic schematic is from John
as before there is an usb to spdif CM6631A board on top of the dir9001 , copper pour on both side are not shown
and as usual any constructive thought welcome 🙂
.
Here is my next pcb , for the simultaneous version , including 50 Hz DEM and some others features
it is battery powered , the logic schematic is from John
as before there is an usb to spdif CM6631A board on top of the dir9001 , copper pour on both side are not shown
and as usual any constructive thought welcome 🙂
.
Last edited:
Apologies for the veer, it is however very tenuously TDA1541 related.
The reason to use thermometer code and not binary weighting is to use DEM error scrambling. Thjat is exactly what Philips did with TDA1541.
By doing the 4 MSB this way instead of R2R we have much better precision for the downstream It's an extension of this principle design:
This means we use a random selection of individual unitary weighted elements (all 16) and by updating the selection at a high multiple of the sample rate (say 32 Fs using BCK) and thus "average out" resistor tolerances.
In effect using 15 Resistors averaged twice we get a 8 times improvement of tolerance even for a single weight. So our 0.01% resistors become in effect 0.00125% Resistors.
Of course to create a bitstream at such a rate we need in effect 16 Bit (switches) from something like a FPGA, serial update is not workable. I guess we can use Bit 16 bit switch for the LSB of the R2R DAC which will receive a 32FS DS modulated signal with the lower bits. The lower bit's could be loaded serially, otherwise we would need 32 Pin's on an FPGA (or 64 to do it balanced) which is quite manageable.
So I would suggest using an FPGA with 64 Bit output that handles all the logic and if followed by "bitswitch" drivers that differential, that is each IC is used half for positive phase and second half for negative, which should largely also reduce the Vcc/Vss current from charging/discharging the gates inside the IC).
I do not see a sensible way of doing this well without FPGA or a rather high power, large pin number CPU.
Thor
Attachments
"Apologies for the veer, it is however very tenuously TDA1541 related."
Who knows... you could use this technique to add a few more bits to the TDA1541 and turn it into an "Ultimate NOS DAC using the TDA1541A".
Who knows... you could use this technique to add a few more bits to the TDA1541 and turn it into an "Ultimate NOS DAC using the TDA1541A".
Yes, I remarked before, we could use 4 PC's TDA1541 per Channel, differential.
With an FPGA we could randomly select one of these for 16 LSB and use the remaining DAC's (just turn each DAC to full or zero) to add two MSB for an 18 bit DAC.
We could probably use some more interesting scheme to distribute our 16 bit across our four DAC Chips so each DAC Chip contributes about equally to the output.
We could then use a simple thermometer decoded resistor set to get two more bits for a 20 Bit input, ~116dB SNR (or ignore that and stick to 18 bit).
Many options exist once we get to include FPGA's.
Thor
With an FPGA we could randomly select one of these for 16 LSB and use the remaining DAC's (just turn each DAC to full or zero) to add two MSB for an 18 bit DAC.
We could probably use some more interesting scheme to distribute our 16 bit across our four DAC Chips so each DAC Chip contributes about equally to the output.
We could then use a simple thermometer decoded resistor set to get two more bits for a 20 Bit input, ~116dB SNR (or ignore that and stick to 18 bit).
Many options exist once we get to include FPGA's.
Thor
Hi Thor,
Thanks for sharing your schematics- just wondering why you didn't use a pass PNP for the -5, +5, and -15 TL431 shunt supplies?
Also curious if you ever tried running the DEM at 50Hz? and if so what did you think? For me it was a similar performance gain as attenuating the digital inputs.
Nice to read your posts.
Thanks for sharing your schematics- just wondering why you didn't use a pass PNP for the -5, +5, and -15 TL431 shunt supplies?
Also curious if you ever tried running the DEM at 50Hz? and if so what did you think? For me it was a similar performance gain as attenuating the digital inputs.
Nice to read your posts.
I haven't tried it yet, but i do belive decoupling below the chip next to pin, with something like rubycon MU (preferably this one), would bring more benefit than 50hz DEM caps. The sooner it hits the gnd, the better. Soldering rather than socket as well.
Hi,
I have used PNP "gain boosted" shunts at a later design, that was not put into production, as the 77 series was discontinued. This was more to get more dissipation and use SOT-23 431 as the bigger ones became hard to get. It did not produce a different sound IIRC.
And no, I did not try DEM at 50Hz, as this seems to make no more sense than disabling DEM. It can only add 50Hz IMD, which it does (based on measurements posted) and I do not want such an effect. If you like it, more power to you.
Thor
I have used PNP "gain boosted" shunts at a later design, that was not put into production, as the 77 series was discontinued. This was more to get more dissipation and use SOT-23 431 as the bigger ones became hard to get. It did not produce a different sound IIRC.
And no, I did not try DEM at 50Hz, as this seems to make no more sense than disabling DEM. It can only add 50Hz IMD, which it does (based on measurements posted) and I do not want such an effect. If you like it, more power to you.
Thor
never worked for me (the 50 hz) after the listening test. I came back to 0.1 uF everywhere on the DEMs with 0805 size Panasonic or CDE plastic caps. Short inductance.
Well , if you look at the previous measurements there is no 50 Hz spike ( -105 dbv ) , the one you see is an ambient pick up , very common at that level , especially in a home setup 😉
.
Last edited:
DEM seems to be operating just fine in this 4 x 1541 deferential setup i tested, this is a 1k sine at -72.5dBFS. Surely without the DEM operating correctly this wouldn't be possible?
According to LTspice using a pass PNP with a bypass cap lowers the output impedance down to 3 milliohms in the audible frequency band, I sure heard the improvement, especially when using Rubycons PML caps.
According to LTspice using a pass PNP with a bypass cap lowers the output impedance down to 3 milliohms in the audible frequency band, I sure heard the improvement, especially when using Rubycons PML caps.
Attachments
With all due respect Thorsten, if low jitter is the goal here (its been my goal at least), then removing RFI/EMI noise generated on the same die as sensitive DAC circuits- surely this would be beneficial? I can understand from your point of view that a 50 Hz modulation will occur and probably other issues, but its so low it doesn't seem likely to be of any concern. Please forgive my ignorance, im just a diyer trying to learn and have a bit of fun. To my ears at least the trade off seems worth it (if there is any?)
It would be a very interesting test to measure the 'jitter' at the 1541 analog outputs with 50 Hz DEM, 705khz DEM, and some type of DEM sync with BCK (Grundig DEM sync for example, or CD-77 solution). I dont have the necessary hardware or expertise,(is a test like this even possible?) but im sure many of the experts here would be capable. Just an idea from a backyard solder slinger. Definitely beyond my limits.
Anyway, im a bit out of my league questioning the CD-77 designer, forgive me Thorsten, just curious.
Happy holidays everyone.
It would be a very interesting test to measure the 'jitter' at the 1541 analog outputs with 50 Hz DEM, 705khz DEM, and some type of DEM sync with BCK (Grundig DEM sync for example, or CD-77 solution). I dont have the necessary hardware or expertise,(is a test like this even possible?) but im sure many of the experts here would be capable. Just an idea from a backyard solder slinger. Definitely beyond my limits.
Anyway, im a bit out of my league questioning the CD-77 designer, forgive me Thorsten, just curious.
Happy holidays everyone.
Just did a quick search and found an interesting article:
TDA1541(A) Dynamic-Element-Matching Slowed Down
TDA1541(A) Dynamic-Element-Matching Slowed Down
It is an interesting article , among other in he's pages , wrote by someone who has great practical experience with TDA1541A
that is what we need here , people who really knows what they are talking about 😉
.
that is what we need here , people who really knows what they are talking about 😉
.
That's funny, no one was knowing about MVAudiolines till thorsten had talken about it few posts ago. ANd now anyyonne is finding it after "small" searchs crying at the moon as an obviouness, getting lesson to the same person who knows and gave the name !
Just laughing but makes me vomit ! As for the ones able to know whom is knowing having themselves no clue to judge it really but the sense of the wind of the day ! What a stupid world of big mouths, really !
Literally A FTCH one whom are solfing 50 bucks the designs of others !
Of course I will be the faulty one to tell again such an evidence ! What a world full of not sincere personns just able to stole other one's ideas and being in the false visible moral of things, sorta wokism!
Well if the fools were flying, some will be squadron leaders here ! And I already know they plot it against sincere people whom will be banned to tell the truth !
As JokoHomo said : I hate you guys lessons givers (I assume it was against the unsincere people whom breath in the rigth wind direction)
Just laughing but makes me vomit ! As for the ones able to know whom is knowing having themselves no clue to judge it really but the sense of the wind of the day ! What a stupid world of big mouths, really !
Literally A FTCH one whom are solfing 50 bucks the designs of others !
Of course I will be the faulty one to tell again such an evidence ! What a world full of not sincere personns just able to stole other one's ideas and being in the false visible moral of things, sorta wokism!
Well if the fools were flying, some will be squadron leaders here ! And I already know they plot it against sincere people whom will be banned to tell the truth !
As JokoHomo said : I hate you guys lessons givers (I assume it was against the unsincere people whom breath in the rigth wind direction)
Last edited: