Ok, so 78xx would be better choice for my 4 parallel TDA1541A dac, also easyer to implement. But seeing is believing, so sent your pictures at the time then.
Could you please give a theoratical explanation for this? For as far as i know this DAC is quit linear. It only transforms digital samples from the time discrete domain to the time domain. As long as the time between every sample is constant there can't be a problem with that 🙂Bernhard said:
It realy doesn't matter...Bernhard said:
MidiMaze said:
Beside that I am not sure if I understand how they get 16 bit from 14 bit DAC resolution, it is only a 14 bit DAC chip and linearity is max. 14 bit.
If 14 bit is interpolated to 16 bit it can never be 16 bit resolution because the error of the interpolated bits is too much.
You can interpolate bits in the digital domain to a higher number because there are errorless values, but after conversion it will only smoothen the curve and increase noise ratio but not linearity.
Imagine you have two codes, both with 1 LSB error in the same direction, if you interpolate them, you still have a 14bit-1LSB error in your expected higher resolution/linearity.
If the errors of the interpolated codes go into opposite directions they could cancel, maybe that is why it works not too bad.
But a TDA1540 is not the same as a TDA1541.
See my -60dB measurement s here:
http://www.diyaudio.com/forums/showthread.php?postid=570869#post570869
Same thing is interpolating 16 bit data to 20 bit data and feed to a 20 bit DAC that will not make a 16bit CD become a 20 bit CD.
I would also be interested in measurements TL431 and 78XX.
They don't get 16 bit resolution. The DAC is stil 14 bits and is lineair. The only trick Philips performs is noise-shaping. You can read it in the datasheet of the SAA7030. It says that upsampling and interpolation of the bits gives you the same in-band quantizing signal-to-noise-ratio as if it was a 16-bit DAC. That's a totally different thing. You can't get 16 bit resolution with a 14 bit DAC.Bernhard said:
The DACs are not the sames, that's true. The trick is done in the digital filters before the DAC 🙂Bernhard said:
Thrue, but: By oversampling you increase the sampling frequency and thus gives you the opportunity to interpolate samples. By doing this in a smart way you can increase the resolution, to say 20 bits (you have to round off the interpolated samples). Now the quantization errors are smaller than just oversample and interpolating while not increasing the bit length.Bernhard said:
In the CD-300 the TDA1540 maybe selected for maximal linearity or minimal distorsion because of suffix "D" but i am not sure about that.
The measure-story is almost the same as the story with
tube vs transistors and non overall-feedback vs overall-feedback
Measuring is very good as a guideline, the rest is to the ears.
The measure-story is almost the same as the story with
tube vs transistors and non overall-feedback vs overall-feedback
Measuring is very good as a guideline, the rest is to the ears.
If there is distortion, there is distortion and if there is no distortion, there is no distortion.
Very simple.
Very simple.
TDA1540 wrote:
I kept thinking about this for my CDP with four TDA1541's parallel project.
Digged a hole in the MDF underneath CDM-2. The Servo PCB is now mounted under the base of CDM-2 and i got rid of 30 cm (possible jitter-inducing?) extension cables. Thanks TDA1540 for reminding 🙂
Haven't made a hole yet for the heatsink: wat a luck, because the power/signal PCB and FTS PCB must be re-arranged in cabinet. Now have to check if its working....
I kept thinking about this for my CDP with four TDA1541's parallel project.
Digged a hole in the MDF underneath CDM-2. The Servo PCB is now mounted under the base of CDM-2 and i got rid of 30 cm (possible jitter-inducing?) extension cables. Thanks TDA1540 for reminding 🙂
Haven't made a hole yet for the heatsink: wat a luck, because the power/signal PCB and FTS PCB must be re-arranged in cabinet. Now have to check if its working....
Oh not to forget: Non-Os vs Os
My experiences by CDM-0 + dual 1540
Hello,
have now the best Player, what have never heard-Philips CD101 with CDM-0 transport and dual TDA 1540 D/A converters.
I have made a lot of experiments with the upgrades and must say that the best opamps are original NE5532 opamps. BB makes localization worse. if you will have a good player, then please don't make any big changes in schematics. Only change all electrolytic caps to better ( ELNA Cerafine ). After opas are good: metal film caps by filter part, better resistors by analog part, BLACKGATE NX by analog part, better RCA connectors and better high-end copper wire by analog part, better copper wire by HF signal between cdm 0 and digital input to converter part, 2 x 5 0,022mf film caps for TDA 1540.
My bad experiences are: to change original power diodes to schottky, film coupling caps for electrolytic power supply caps, BB opas, to change ceramic coupling caps by all digital ships power supply to film caps was bad idea... the best are ceramic caps.
If you do all the right upgrades, you get the best CD Player for this money. Please send me ideas, how i can get SPDIF from CDM-0? Or this is impossible!
Hello,
have now the best Player, what have never heard-Philips CD101 with CDM-0 transport and dual TDA 1540 D/A converters.
I have made a lot of experiments with the upgrades and must say that the best opamps are original NE5532 opamps. BB makes localization worse. if you will have a good player, then please don't make any big changes in schematics. Only change all electrolytic caps to better ( ELNA Cerafine ). After opas are good: metal film caps by filter part, better resistors by analog part, BLACKGATE NX by analog part, better RCA connectors and better high-end copper wire by analog part, better copper wire by HF signal between cdm 0 and digital input to converter part, 2 x 5 0,022mf film caps for TDA 1540.
My bad experiences are: to change original power diodes to schottky, film coupling caps for electrolytic power supply caps, BB opas, to change ceramic coupling caps by all digital ships power supply to film caps was bad idea... the best are ceramic caps.
If you do all the right upgrades, you get the best CD Player for this money. Please send me ideas, how i can get SPDIF from CDM-0? Or this is impossible!
Agree, same bad experience so far here also, better keep slower diodes, and even reduce the slewing by a small resistor (10 ohm for line use, 0.1 ohm with poweramps) before the diodes. The filtercaps after the diodes can do their work then easier (less HF from diodes self to filter away)
Have tried fet-like OP275 opamps in a 1541 player also, but for bass performance and more neutral sound i think 5532's still are better, or the use of a simple 4 transistor schematic or something like that.
Agree, look out for extra oscillation with foil caps. There exists a better quality ceramic cap then the standard ones, but where to get them...
Now the right values, but tweaking around the right value might still neccecary, depending on current draw from PS:
*10 ohms before the rectifier diodes in PS of tube pre/line amplifier
*1 ohm before diodes in CD player or any other line-level PS
*0.1 ohm in solid-state power amps.
No huge changes in sound but worth it anyhow.
Simple mod, try it and let us know!
I kept thinking of the TL431 vs LM78xx PS questions, see the comments much earlier about it in this thread.
When you measure the spectrum of a 78xx in a testcicuit there could be measured 48kHz or so and that can easy filtered away by a cap. But what happens when 78xx will be used in real conditions: RF pollution from PS before regulator, variable loads(music transients)?
The TL431 is a parallel (shunt) regulator, the 78xx is a series regulator. I can think situations in which the shunt TL431 is much "quicker" in reaction/regulation compared to a series 78xx, for example with those variable loads and RF pollution. The standard A suffix TL431 has a low resistance of 0.22 ohm, pretty low. The TL431 would be still better choice then i assume, but as i am not a electronic designer i cannot say which regulator is best.
So maybe this would be a good subject for a new thread, to get some answers, and because i want to finish my four 1541 parallel project in a good way.
more comments?😉
When you measure the spectrum of a 78xx in a testcicuit there could be measured 48kHz or so and that can easy filtered away by a cap. But what happens when 78xx will be used in real conditions: RF pollution from PS before regulator, variable loads(music transients)?
The TL431 is a parallel (shunt) regulator, the 78xx is a series regulator. I can think situations in which the shunt TL431 is much "quicker" in reaction/regulation compared to a series 78xx, for example with those variable loads and RF pollution. The standard A suffix TL431 has a low resistance of 0.22 ohm, pretty low. The TL431 would be still better choice then i assume, but as i am not a electronic designer i cannot say which regulator is best.
So maybe this would be a good subject for a new thread, to get some answers, and because i want to finish my four 1541 parallel project in a good way.
more comments?😉
Lourens said:[ Not easy to put a tda1541 in this pre I2S machine. Maybe someone with his brains into glue logic can.
Sure: http://www.diyaudio.com/forums/showthread.php?s=&threadid=58700
but no white smoke yet
Found a CD204 on a fleamarket for five euro's. Strange player, metal inside everywhere (chassis,cdm,drawer) but a cheapy plastic front. Shame, it's build to last for ever.
Was told it worked
After 15 minutes at home it did 😀 Just resoldered the power regs and cleaned the lens. Bit slow in reacting, but it works perfect now.. cdm0/1 machine with 1540's.
Also had a 1541A dac module lying around (read: piece of pcb of a marantz cd40 with all the smd caps on still on it).
So connected some wires and hooked it up to the SAA7030 (input side 16 bit no oversampling yet):
STR1 (21) to LE (1)
CLCF (18) to BCK (2)
DLCF (20) to DATAL (3)
DRCF (17) to DATAR (4)
Guess this info is also in the other thread.
Works fine! (at least with a test cd on the scope, just used R's for i/v conversion and watched the tube. Left/right is also ok, checked with some test tracks).
Dont forget to put the 1541 in parallel mode (pen 27 to 26, -5V)!
End of the experiment and the evening, going to remove it now.
Got more than enough 1541 machines and i dont need another one 🙂
So no comments on sound quality and i dont know if you could hook up the 1541 after the dig filter (and feed it with 14 bits only).
Did not try.
Greetings,
- Status
- Not open for further replies.