About PSU refinement... someone mentions here about the L6, for making a better filtered supply for it, since is the main 5V clock translator? I think, since it drains little current, if a simple series R and more cap will not make the day here... bigger cap for filtering LF noise, and perhaps C18 already helps, with additional R in series of the L6?
If it suffices... normally, for my own DAC's, I make a separated shunt low noise reg for the clock section.
If it suffices... normally, for my own DAC's, I make a separated shunt low noise reg for the clock section.
Mine are PWM8 and the steep filter, but I must admit I actually don't hear any difference between the filter settings.
About hearing differences... recently I have various opportunities to listen my most recent DAC addition, the one with TDA1541A in NOS (for using digital filtering in the PC). Then I put it against my trusted DAC with DF1704+PCM1704, using a trafo for I/V, followed by an Aikido stage. The TDA1541 I ended using subminiature tubes for gain stage.
Some photos of it, below... is very messy, since I used some old parts and boards that I have, and put all toguether.
I know that the correct method to evaluate a PC digital filtering vs a on-board digital filtering is to bypass it and compare, but I don't want to mess with my PCM1704 DAC, since it have a terrible but very short air connections, with a solid ground plane and short local decouplings.
Even so, I'm haapy to hear that they have much more in common than differs; the DF1704 filtering more than suffice to me. Basically I'm listening to sum of parts differences, but get some idea. I bet that the digital filtering differences are lees than the DAC parts and topologies (the analog part differs a lot). Also is admiable how good a ancient DAC part can sounds with some attentions.
Far from scientific, but enough to me. Soon, will the time for the Valve DAC, but this I can test in a another system, the one mentioned using Stax 'phones.
Below: some progress with the Valve DAC board:
I'm using some Tropical Fish caps, and some ancient silicon Brazilian Philco transistors for the peripheral functions.
All of this, fighting with speaker issues: I needed to change the suspension of my Jantzen Audio Troel's JA-8008 driver:
And numerous family (the father of my wife dead recently...) and work (things changing...) issues/fights in between... very busy times!
Some photos of it, below... is very messy, since I used some old parts and boards that I have, and put all toguether.
I know that the correct method to evaluate a PC digital filtering vs a on-board digital filtering is to bypass it and compare, but I don't want to mess with my PCM1704 DAC, since it have a terrible but very short air connections, with a solid ground plane and short local decouplings.
Even so, I'm haapy to hear that they have much more in common than differs; the DF1704 filtering more than suffice to me. Basically I'm listening to sum of parts differences, but get some idea. I bet that the digital filtering differences are lees than the DAC parts and topologies (the analog part differs a lot). Also is admiable how good a ancient DAC part can sounds with some attentions.
Far from scientific, but enough to me. Soon, will the time for the Valve DAC, but this I can test in a another system, the one mentioned using Stax 'phones.
Below: some progress with the Valve DAC board:
I'm using some Tropical Fish caps, and some ancient silicon Brazilian Philco transistors for the peripheral functions.
All of this, fighting with speaker issues: I needed to change the suspension of my Jantzen Audio Troel's JA-8008 driver:
And numerous family (the father of my wife dead recently...) and work (things changing...) issues/fights in between... very busy times!
My condolences to you and your wife.
You are the first to replace the high-voltage PNPs that normally drive the neon lamps with coloured LEDs (not the first to use LEDs instead of neon lamps). It should work well as long as the forward voltage is low enough (well below the 3.3 V high level of the FPGA IOs).
You are the first to replace the high-voltage PNPs that normally drive the neon lamps with coloured LEDs (not the first to use LEDs instead of neon lamps). It should work well as long as the forward voltage is low enough (well below the 3.3 V high level of the FPGA IOs).
Thanks about the condolence.
Yes, the blue LED's, I confess that I put it here without fully checking it... only remembered that I had some that worked well with near 3V (not sure if are these ones now...)
Yes, the blue LED's, I confess that I put it here without fully checking it... only remembered that I had some that worked well with near 3V (not sure if are these ones now...)
A little OT, for clarifying this post: About the JA-8008 suspension, for ones not familiar with this model, it have something gooey in the original suspension, for ultimate correct damping. I don't have this just now, but it seems to responded very well, and response in the TQWT have a crossover near 1.5kHz, so I can live without it for now.About hearing differences... recently I have various opportunities to listen my most recent DAC addition, the one with TDA1541A in NOS (for using digital filtering in the PC). Then I put it against my trusted DAC with DF1704+PCM1704, using a trafo for I/V, followed by an Aikido stage. The TDA1541 I ended using subminiature tubes for gain stage.
Some photos of it, below... is very messy, since I used some old parts and boards that I have, and put all toguether.
I know that the correct method to evaluate a PC digital filtering vs a on-board digital filtering is to bypass it and compare, but I don't want to mess with my PCM1704 DAC, since it have a terrible but very short air connections, with a solid ground plane and short local decouplings.
Even so, I'm haapy to hear that they have much more in common than differs; the DF1704 filtering more than suffice to me. Basically I'm listening to sum of parts differences, but get some idea. I bet that the digital filtering differences are lees than the DAC parts and topologies (the analog part differs a lot). Also is admiable how good a ancient DAC part can sounds with some attentions.
Far from scientific, but enough to me. Soon, will the time for the Valve DAC, but this I can test in a another system, the one mentioned using Stax 'phones.
View attachment 1344831View attachment 1344832
Below: some progress with the Valve DAC board:
View attachment 1344834
I'm using some Tropical Fish caps, and some ancient silicon Brazilian Philco transistors for the peripheral functions.
All of this, fighting with speaker issues: I needed to change the suspension of my Jantzen Audio Troel's JA-8008 driver:
View attachment 1344836
View attachment 1344835
And numerous family (the father of my wife dead recently...) and work (things changing...) issues/fights in between... very busy times!
Then, I need to change for the other side soon.
BTW: the clipping one will be a Ne lamp; justo these, since is more functional for the DAC usage, and I didn't have 7 Ne lamps just now...
Is working!!!!!
Surprisingly (!), almost all things worked flawlessy from start but the crystal oscillator: this works only at certain trimmer's range, and with adjustment rather low (~330R in the cathodes). But I'm able to achieve the -500mV at the test point, with a little effort.
About the clipping indicator: it lighted only once for <1sec, when I'm watching it working with a very horrible recording clipping almost all time. Otherwise, it remained off.
The status LED's lights with my choice. Is normal? Or I don't understood it clearly (I think it lights only using the surprise mode)?
And, a good news, the blue LED's I'm put works with 3V3.
BTW: the heater's PSU are indicating 6.79V; I adjusted for compensating the small gauge cable test I used there.
Surprisingly (!), almost all things worked flawlessy from start but the crystal oscillator: this works only at certain trimmer's range, and with adjustment rather low (~330R in the cathodes). But I'm able to achieve the -500mV at the test point, with a little effort.
About the clipping indicator: it lighted only once for <1sec, when I'm watching it working with a very horrible recording clipping almost all time. Otherwise, it remained off.
The status LED's lights with my choice. Is normal? Or I don't understood it clearly (I think it lights only using the surprise mode)?
And, a good news, the blue LED's I'm put works with 3V3.
BTW: the heater's PSU are indicating 6.79V; I adjusted for compensating the small gauge cable test I used there.
Congratulations! What modulator did you use for the horrible recording? The chaotic one?
The functionality of the status lights has changed in version 3, which is what you are using, so that is normal.
The functionality of the status lights has changed in version 3, which is what you are using, so that is normal.
Thanks!
I'm using the PWM8. I'm presume that the recording produced some high crest and then a high intersample over, or something like that. 1h after that, with all other testings, including waveform generators (like ARTA) remained off.
Great that are working ok the indicators, then.
I'm using the PWM8. I'm presume that the recording produced some high crest and then a high intersample over, or something like that. 1h after that, with all other testings, including waveform generators (like ARTA) remained off.
Great that are working ok the indicators, then.
So far, very happy with the results, and eager to liten something! But, due to very small cores I choose for the filter, is very difficult to work with it... and, even with same Al, I think that the huge size difference are accounting for some difference in the measured value... The assembled 60.5 turns one gave me the result for the original 57.5 turns (~825µH)... At least, are a small difference, easily compensated with some more turns.
OOPS: nevermind: my old meter had some issues...
OOPS: nevermind: my old meter had some issues...
Last edited:
Some more comments:
I noticed that the waveform in the scope is good, even without the correct load. Of course, the response is not correct, due to reactance or first filter cap vs the resulting impedance.
But, about the noise: even comparing with the barely filtered out, without the filter bank, is LESS noisy than I first think. Since I'm used to measure el-cheapo China pen-drive solutions, and DVD players (years ago). Some DVD players try to filter the large noise from cheapo DACs with the sloppy 4558. Sometimes, the noise at the 4558 out is HIGHER than it's input. Hence, these things are noisy, some have a lot of HF noise.
Even DAC's like the raw ES9018 are not ultra silent, either (are better than el-cheapo solutions, but naturally, direct out will have the expected HF noise).
I noticed that the waveform in the scope is good, even without the correct load. Of course, the response is not correct, due to reactance or first filter cap vs the resulting impedance.
But, about the noise: even comparing with the barely filtered out, without the filter bank, is LESS noisy than I first think. Since I'm used to measure el-cheapo China pen-drive solutions, and DVD players (years ago). Some DVD players try to filter the large noise from cheapo DACs with the sloppy 4558. Sometimes, the noise at the 4558 out is HIGHER than it's input. Hence, these things are noisy, some have a lot of HF noise.
Even DAC's like the raw ES9018 are not ultra silent, either (are better than el-cheapo solutions, but naturally, direct out will have the expected HF noise).
I want to repeat this test, to see if is a issue of some interference at that moment, or really is the track. I'm testing this randomly at that time, so I need to find the mentioned track.
NOTE: no complaining about that. Certainly was a very uncommon phonomenon.
Interesting. The PWM9 mode (it was PWM8 in versions 1 and 2.1, PWM9 in version 3) has the best headroom for intersample overshoots compared to the other two sigma-delta modes. I never saw the clipping lamp light up in PWM9 unless I played a special, way too loud DSD track that I generated specifically for this purpose.
Anyway, when the clipping lamp lights up, you can decide to put the DAC in medium rather than loud mode. It gives you 6 dB of extra headroom at the expense of 6 dB less signal level and 6 dB less signal-to-noise ratio. If it's a high sample rate recording with lots of treble, you could also try loud mode with the apodizing filter.
Anyway, when the clipping lamp lights up, you can decide to put the DAC in medium rather than loud mode. It gives you 6 dB of extra headroom at the expense of 6 dB less signal level and 6 dB less signal-to-noise ratio. If it's a high sample rate recording with lots of treble, you could also try loud mode with the apodizing filter.
Last edited:
Too bad, is something difficult to find this track... when the DAC was working for the first time, I put some random tracks from a playlist containing ~ 3300 tracks, just for see if ot works... and, the lamp lit not at start of the music...
Until now, it not occurred anymore.
But, for curiosity, if it occurs again with some music, I will take note of the track and the elapsed time.
For now, based on your explanation, I'm under impression that it will not occur again for a long time...
Until now, it not occurred anymore.
But, for curiosity, if it occurs again with some music, I will take note of the track and the elapsed time.
For now, based on your explanation, I'm under impression that it will not occur again for a long time...
About the clipping behaviour, I found the offending condition, even without finding the track. When I put the square wave oscillation with 0dB output (maximum signal) using ARTA, the clipping lighted continuously. But is the only condition that's trigger it (basically, the bad track are full of clipping AFAIK).
Last edited:
I'm finished the coils, and soon I can able to install the filter board and finally listen the DAC, and test the results!
When I finish the DAC, I will include the funcions switches. For testing, I put only wires, for not messing with loose parts in the bench.
When I finish the DAC, I will include the funcions switches. For testing, I put only wires, for not messing with loose parts in the bench.
Ah, of course. Re-reading the DAC document, pages 51-52, is indeed one of the worst signals to feed. Will be interesting and fun to play with when I install the swiches.
Finally, I measured and listened to it!
Distortion is in the vicinity of the described in article. Noise, is far lower than a reel-to-reel, quality FM strong signal radio, vinyl or suchlike. Excellent!
My listening... for now was using a small "portable" amplifier that I made with valves, just for testing, because my trafo that converts to unbalanced have a reduced output (is 1+1:1), and this amp have low gain, so I listened with relatively low volumes. Even so, I liked that I hear. To be fair, I don't expected about that I will be hearing from this very peculiar DAC, and I'm a little skeptical about reviews (but can be motivational). It seems really to be something special. Excellent! More to come, when I install it with my DIY Stax system (this can take some days, due of these busy times). Also, I need to solve some minor issues like eg. the different reference valve and the CCS modification to feed it, and the final assembling (where I can put the DAC...).