Because this thread is all about the TDA1541A , and its is what I am doing , following the path of John from EC-DESIGNS
.
And wht about the spdif is ON, CD running?here are the frame , from the cd payer , NON playing , DOL + and DOR + are low , dont know if it's ok or not , any input will be appreciated
View attachment 1194057
.
.
try to connect the CKSEL pin of DIR9001 to the ERROR pin.
The PLL did not operate when there was no SPDIF signal,
so both the DATA signal and the LRCK/BCK pulse did not come out,
from this link, maybe it would be from help...
https://www.diyaudio.com/community/threads/tda1540-and-le-crosstalk.394835/page-2
.
If the numeric logic is good, there are perhaps some path of improvment seen from my screen, surely details however that can be heard when the proto is finally working :
I am not a fan of the 100 uF DEM decoupling John introduced and I know several people thinking the same after a try. And the caps in your layout are way to far from the TDAs.
Try to look for Pedja Rogic DIR9001 shems on the the net, you may find some usefull shematics.
Year and origin of the production plant of the TDAs matter as well as the quality of the clock (same inputs as above about the layout and parts choice).
Just icing on the cake inputs, of course. A complicate trae off at the end but that matters !
- spidf : I would choose not spidf from a RCA but a true pro connector from the good impedance and directly sold on the board for the futur proto as the source of course.
- from my experience, whatever the quality of those connector there is still some difference and I found the hot wire made from a solid standalone core - the same as TVs' antennas- to work well VS some more exotics
- the quality of the passive parts matters : the spidf traffo : try a direct connection to split it to understand its behavior.
- Caps on all signals, digital as analog... well all is analog for electricity and it can be heard according the quality of your loudspeakers. Can have a role on the final tones, according the way you play with them : inductance - pcb layout- or dielectric of the smd as the lythics.
- resistors on the digital path if they can reduce some bouncing problems can introduce also less transparency according the whole layout, and straigth digital path vs zero ohms vs 25R -for illustratioon- Sussumu resistor should be chose with ears... (I mean at the ebd as a icing of tthe cake).
I am not a fan of the 100 uF DEM decoupling John introduced and I know several people thinking the same after a try. And the caps in your layout are way to far from the TDAs.
Try to look for Pedja Rogic DIR9001 shems on the the net, you may find some usefull shematics.
Year and origin of the production plant of the TDAs matter as well as the quality of the clock (same inputs as above about the layout and parts choice).
Just icing on the cake inputs, of course. A complicate trae off at the end but that matters !
@fabrice63 ,congratulations on the project, it looks good, although there are small birth problems that will be corrected in the future.It would be better if the tda1541 had slightly better and larger capacitors (OS-CON)for decoupling, as well as the shunt regulator being better suited to the tad1541 than the serial one. Decoupling and regulators are very important for DACs because they produce HF interference that really affects the sound.
Sorry but that might be not so good. Because the external connector should be the part of chases. Not a part of PCB. When the connector is a part of PCB could introduce all vibrations via cable mass. In case that on the other side is the same thing, then the 2 PCBs will be connected...
In his case connecter is part of chases only, PCB could be mounted on some anti-vibrational surface and then to the chases...
Bare in mind that on the pcb crystal oscillator is a part of DIR9001 circuit.
And if You take a look at the older devices they in majority way had the isolated connectors to the chases connected with soft vires to the dedicated PCBs
Thank you all for your advice
there wont be any spdif connector at the end , an board usb to spdif cm6631A will be on the top of the DIR9001 connected with 2 very short wire
I will make FFT measurements of the all thing once completed , if needed I will decoupled and /or modified as per measurements
ceramic caps are all COG , and résistor are all 0.1 % 25 ppm , DEM cap are nichicon UKL , all other are panasonic FR , timming cap of the TDA are film cap PPS
forgot to mention that the DAC board will be lipo battery powered too
.
there wont be any spdif connector at the end , an board usb to spdif cm6631A will be on the top of the DIR9001 connected with 2 very short wire
I will make FFT measurements of the all thing once completed , if needed I will decoupled and /or modified as per measurements
ceramic caps are all COG , and résistor are all 0.1 % 25 ppm , DEM cap are nichicon UKL , all other are panasonic FR , timming cap of the TDA are film cap PPS
forgot to mention that the DAC board will be lipo battery powered too
.
Last edited:
Leave SPDIF connector like it is. (As well as any external connector too.)
External connectors are part of the chases not a parts of PCBs...
.
Put the boards from some heavy steel base, 10mm simple steel would be fine. To add some mass against vibrations.
You can put the "mass boards" to chases ground too in one point.
Isolate the mass plate with wood from the chases...
.
Are You tested the construction to check the sound?
At least I tried with the Wolfson equiquavalent chip on the Subbu DAC, same slaved by crystal near the W8805 and a simple streamer I also converted to make it 75 ohms but direct spidf cable on the pcb of the logitechh streamer. It was way better than a rca connector with its wire from the pcb, whatever the vibration.
Improvment in sound quality was so good than you could hear the different spidf cables tried.
An experience that is good to try about clarity and details improvement whatever it is digital flux is to avoid the anti boucing resistors that are often between 20R to 50R and the traffo for a direct connection at the input of the chip : it is very a learning moment on how the digital stream embeded on the analog electrical stream is layout dependant indeed... vibration is comming after in the order of things to cure about the layout imho and at least from the experience I had here. Same applies on the I2S signal after : you may or not need of the anti bouncing resistors, without it is often more resolving on non sota pcbs !
@ fabrice63 : the UKL caps whatever John is saying are not so good caps imho. First, the caps are way too far from the TDA pins. What matter here is to reduce the inductance, so the leg of the cap has to be the nearest of the TDA pin. I would try to decouple the lythics by acrylic 0.1 uF smd caps and not NP0. For the lyhtic, the Pan FC or FR with 2 mm leads gap would be a better choice than those Nichicons imho. Been there already.
my 2 cents though
This project and thread is about what John ( and others ) have made and said , plus my own experience , if you have experience in "TDA1541A Dual Mono 17 bit Signed Magnitude DAC " , it will be welcome
sorry but , I am not interested in generic opinion or point of view , only proven facts in this a particular setup
.
sorry but , I am not interested in generic opinion or point of view , only proven facts in this a particular setup
.
Pffff, your judgment is a bit hard, it is not generic opinion boy, but 25 years of work about TDA1541A, I am btw in the others in this long thread ! If you had that experience you will know it is all but generic opinion, your reaction is positivist imo as focused only on the glue logic . The inductance was said by John btw, that proves you read not entirely this thread !
At looking your pcb it is not hard to be seen how it can be improved just by seinng the first layer. But now I will keep it for me !
Good luck countryfellow !
At looking your pcb it is not hard to be seen how it can be improved just by seinng the first layer. But now I will keep it for me !
Good luck countryfellow !
https://hr.mouser.com/ProductDetail...JN65C1H104JE01D?qs=ufv2HEpFn%2BQKyaRitfoO0A==
0704 footprint 0.1uf c0g 😁 can't get any better than this for non 50hz dem clock modification.
0704 footprint 0.1uf c0g 😁 can't get any better than this for non 50hz dem clock modification.
Nothing personal mate but , til now your words were mostly useless , read yourself before writing , as said if you have good experience in this project , share it , if not keep it
.
I keep it if I want, and I am not your mate and this is not your thread ! Useless, how you can judge it ? You just do not know at seing just the layout ! If you want to avoid some inputs, then open your own thread. I shared it, you 're not happy, so I stop as I said, not usefull to tell me to stop, your attitude was arogant enough for me to stop ! Of course nothing personal.
@ Brijac, you can make shorter with class 2, but as already mentionned in this thread, ceramic is not the best dielectric for the 14 DEM decoupling, acrylic was prefered by many (Loesh, Rogic, or some inputs from me about this dielectric for the crystal decoupling), 0806 is a good trade off then. As for the DEM decoupling capacitance and frequency choice with the 100 uF caps. Many that tried were not so happy, best is to try it before and have a second board for 0.1 uF testing or the way Philips made on the TDA1540 datas
@ Brijac, you can make shorter with class 2, but as already mentionned in this thread, ceramic is not the best dielectric for the 14 DEM decoupling, acrylic was prefered by many (Loesh, Rogic, or some inputs from me about this dielectric for the crystal decoupling), 0806 is a good trade off then. As for the DEM decoupling capacitance and frequency choice with the 100 uF caps. Many that tried were not so happy, best is to try it before and have a second board for 0.1 uF testing or the way Philips made on the TDA1540 datas
Last edited: