It is not magical thinking.
Passive as a rule has orders of magnitude less distortion than passive and passive circuits.
It becomes magical thinking only if we do not reference this to actual quantifiable qualities.
Thor
Thanks I will try.
What type for the PNP i am not that familiar with chrs of BJT sorry... 🙁
I am reassured, it was not Far More Jitter, finally ! I remember this cd player and its complex digital filter embeded !
Last edited:
I heard one and was not so amazed, and anyway for what it worths in the mouth of the guy as it was said it means it was like there is no active after rigth to the ears ! This what I call magical thinking ! Same to me with passive pre, the soundstage is always less good to me for the few I have heard and actives when good has as much resolution ! Of course we know the quantifiable qualities, but they are meant not obe enough "alone".
Try BC557C.
BC327/337 are overhyped. For noise 2N4401/03 or SS8550/SS8050 are better, for beta and bandwidth MPSA18 rules for NPN, but no real complement exists.
Thor
remember me the old good day, I still have a LL 1450 (iirc the ref) iwas using as output of some good Philips DVD/SACD voltage output with akm chips and bigh FGPA, neede some smal tht ceramic to filter "on the can"... sounds was nice... but was output line one...
Last edited:
Yes i did this it working up to 192KHz S
BUT does not matter - I2S USB interfaces ususalu come with 22.xx and 24.xx MHz MCK, and BCK is equal to with 382Kz SR.
Any way it can be find some solutions for that...
I audtioned the Zanden once. It only sounded good with light music output voltage as I recalled was low so needed a preamp with higer gains.
Btw that was like 20 yrs ago. Lol. I questioned the used of so many passive filters at its output, perhaps Thors sinX could be a better solution back than.
I think that is actually zanden is using Sin(x)/x passive filter? As the parallel row of 4-5 notch RLC filters for
n x FS. You can find more in the Zanden patent papers?
.
I must try it one day. Already make the calcs... You can also do it active way emulating L and C with OP amps...
Last edited:
Or better to undetstand if NOS and they like squares w/o timing altération or if It is a myth and figthing against back interpolation is better for S/Q?
But I digress, sorry.
But I digress, sorry.
Simultanious should do 384k. But the BCK runs at the absolute maximum allowed.
If the system is programmed to actually use 16Bit data and then silence, BCK run's actually at a faster clock than official maximum and that is a bad idea.
This here shows 16 Bit clock cycles and 16 Cycles silence.
This means more than 192kHz cannot be supported. Nothing to do with the clock frequencies upstream.
Thor
I run 384k absolutely fine with TDA1541 (without A) in simultaneous mode + 1.536Mhz DEM (reclocked). I use jlsounds USB board. The only thing I do is making LE wider and reclocking it with 256Fs MCK.
Here is the question. Since the TDA1541 starts conversion on the positive edge of LE and needs 0.5uS to settle, I would move the positive edge of LE as close as possible to the data train on the left to increase the "silence" time. Another option is to make LE wide up until the first next bit clock comes in. It will make the LE spectrum narrower and lower HF noise. What do you think?
Here is the question. Since the TDA1541 starts conversion on the positive edge of LE and needs 0.5uS to settle, I would move the positive edge of LE as close as possible to the data train on the left to increase the "silence" time. Another option is to make LE wide up until the first next bit clock comes in. It will make the LE spectrum narrower and lower HF noise. What do you think?
Tubee's circuit long got lost for easy access, let me drop it here:
2 X 1nF in series = 500pF, the 330R is low enough that oscillator will still oscillate or be at the edge of oscillation.
The differentail clock is fairly high impedance and is tuned for reliable lock and a stable clock.
Thor
Ok, let's treat this fairly:
Transformer I/V TDA1541A s1
First, 1kHz THD &N
Seen worse, seen better, but seeing that H3 & H2 are at the same level, this is poor. And more audible and unlike dominant H2, not "nice" distortion.
At similar levels a S&B transformer will have two more zeros in the THD.
How about HD vs Frequency?
Above 2kHz and below 300Hz HD takes off seriously. At 100Hz and 4kHz we are above -60dB THD and at 30Hz we see around -30dB. Really?
These are pretty poorly performance transformers when applied like this. Clearly levels are already near maximum, there is no headroom.
To get sensible LF HD at least 20dB more level handling is needed and 30dB wouldn't go amiss. Now that is simply correct application, not the fault of the transformer if the application is incorrect. The transformer is likely not bad.
The HD rise in the upper midrange is puzzling. A competent audio transformer design should not have it. Maybe some unpredictable interaction with the TDA 1541?
Will this provide a neutral, uncoloured sound, or will it be "colourful"? The levels of HD certainly become audible at high digital levels and this is absolutely unnecessary for using transformers.
I'm not against transformers, but this example is not showing the kind of performance I'm used to from the kind of transformers I have worked with, specified and co-developed.
As said, part at least is a problematic application, at by far higher levels than appropriate.
If set up for 0.1V output (instead of 1V) distortion, with the current/voltage conversion resistance split 50/50 between primary and secondary, it would look much better and we would also have very low HD from the TDA1541.
Of course, now an analog stage with a gain of 10-20 (20-26dB) is needed to get normal signal levels.
Thor
Last edited:
Do you have some scope traces of the different signals at 384kHz?
This is useful information, I need to re-read the datasheet, I must have missed this
The edge rate is what causes trouble, not the width of the pulse. You already have BCK and data to make trouble.
I'm not buying into all that "silent time" stuff. More magic thinking.
The logic inside TDA 1541 is differential current mode, so ground bounce etc. do not apply internally and there is no ground bounce on DGND, except from the actual input signals.
Key is to limit the voltage swing on the inputs and avoid the input stages going into saturation or slewing and to limit signal coupling into the substrate.
Thor
This man is really unfair , he seek on purpose the worst case out there to show it up , I have 0.0038 % at the same level and freq with the 1495 that are not the best at Sowters , he should build a NOS and purely passive setup and show us what is able to get from it ???
finally he is showing he's true face , the man who think he knows it all and wants to be the one and only , fact is there is already one in the place 😎
he doesn't even know the datasheet of that chip , he has been remind few times 😒
.
Last edited:
I don't use them at 1V so I'm afraid your measurements are not relevant to me.
Please repeat with a 100r source resistance and max output of 400mv p-p which are more typical of their use for I/V conversion.