It is not I that says it is an improvement, it is ecdesigns, I only visualized the problem he was talking about and that the the signed magnitude actually eliminates that.
If it is an improvement , soundwise, I cannot comment on as I haven't heard it yet.
I do not have an AP or other sofisticated measuring tools, so I cannot perform a measurement as you show, everything drowns in noise in my measure equipment at those levels.
@ dreamths
I would not ignore such a spike even though the level of energy stored in there seems small. I think there is a reason why a lot of the highly rated multibit DACs do use signed magnitude. They have other issues as they for the most part are R2R DACs. And I think there is a reason why John says this is the best way to use a TDA1541A. I personally do not care if the 2. order harmonic is only down 60 dB, but it could be an indicator of something else wrong. I would really like to hear what level of 2. harmonic John experienced , when he made that circuit.
Thank God i was born with the cassette player generation...If you listened to at least 100 cassettes in your lifetime you will never have a problem with any reasonably well built cd player except that you might miss that tape noise and its associated distortions which were 100 times higher than of any cd player ever built...
The only advices i listened to were to use those bit filtering capacitors around that dac according to the current needed to be delivered and abraxalito's opinion on lm6172 as i/v stage...It is indeed the best op-amp as i/v for tda1541 and i completely dumped the idea of using ad844 as i/v although i bought 4 of them...
The only advices i listened to were to use those bit filtering capacitors around that dac according to the current needed to be delivered and abraxalito's opinion on lm6172 as i/v stage...It is indeed the best op-amp as i/v for tda1541 and i completely dumped the idea of using ad844 as i/v although i bought 4 of them...
Last edited:
That's why I said that it is a nice technical challenge, but seen from an audio point of view it makes less sense.
But this is a DIY forum where all possible combinations are being tested, giving great satisfaction when a step forward is achieved.
Hans
Well ..it seems that those steps forward were greatly achieved by the more modern dacs...As i'm looking to improve 9.5 cm /sec cassette players...i think that i can understand "the niche", but my experience tells me that the audiophile market is actually looking for higher THD or noise not lower...I repeatedly find evidence of it in the diy world where claims of improved sound resides in much higher THD and noise than the worse dac ever delivered , but with specific patterns that can fool our ears.
Last edited:
I was born before the cassette recorder, and have never used it as a source for serious music listening. Reel to reel or vinyl was the only ways to listen then IMHO. And they are still challenging the digtal audio...
And IMHO opamp I/V is not the best sounding solution either..
How do you come to that conclusion?
1 LSB , to my calculations, with 1 V out FS corresponds to 15 uV so 0,25 LSB would be less than 4 uV.
Clearly the spike is much bigger than that. It is 15 mV , but I don't have the level of the 300 Hz sinus I was measuring, but I can guaranty it was not more that 5 V.
You skipped the " Thank God for that"...
Just try lm6172 or 2x lm6171 on a simple tda1541 for both channels and tell me about that after!
I don't care about the vibe surrounding the discrete non-NFB as long as i'm happy with the result of this simple IC.
What's really important is the ENERGY of that spike which is close to nothing compared to a full step energy while the step itself isn't much...
Last edited:
This is written in the article: "A Monolithic Dual 16-Bit D/A Converter", IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. SC-21, NO. 3, JUNE 1986
Thanks. Well if you are happy with that, fine by me. I am not sure I have tried that particular opamp, but a lot of others that people told were the way to go (LM4562, LME49720 and many more not present in my memory), and still I prefer discrete solutions and in TDA1541A especially a passive solution. At some point I will try the LM6172 though.
Have you ever tried a passive solution with a Sowter transformer?
Yes I saw that... But still my question is, what is the spike I measure then?
And it is not present in signed magnitude..
Probably got an unsuccessful in terms of timing instance of the chip. Or something is wrong in the scheme.
I haven't tried any transformer or buffered resistor loading solution, most probably, the transformer must be the most interesting one from the weirdest.
I did try lm4562 vs lm6172 in one last mod i did to a Sony player which has the same chipset as the Nakamichi cdp-2e :
https://www.diyaudio.com/forums/dig...cd-players-enhancing-noise-7.html#post5718722
I simply replaced the lm4562 with lm6172 and the difference was instantaneous.Although the lm4562 gave me the impression of a deeper base while the lm6172 sounded more on the bright sight first, i quickly realized that actually the lm6172 had the exact amount of base that you need to have while the highs were just clearer .It's not a mellow combination, lm6172 is just very fast and powerful. Its noise is the only downside, but i think it's the right amount of noise that can hide any other digital noise, while it's not generating any distortion because it's so fast that it can amplify anything that the whole digital system can throw at .As this player has no low pass filter i just added a first order rc filter that i usually put everywhere in my toys (2k4-1...3.9nF).After that i have a tpa6120 headphones amplifier , another extremely fast circuit, and that is all i have...i have no place for speakers in my room. All i can say is that, even though it's on the bright side, i don't need any deeper base.My headphones are known to have a little more base than usual so it might be that the 6172 is even brighter .I have another player based on pcm1701 i am working on right now and i hope to be able to hear some differences after i finish the ada4898-1 i/v +lm4562 buffer and supply mods . Sorry for not being able to give you a proper feedback as i have no speaker setup at hand.
You don't need an AP to measure this.
You have a soundcard, so you can record a 1kHz test tone.
Bring this with whatever sound program to a level of -90.31dBFS and burn a new CD with this signal.
Now you have most likely a Riaa preamp.
Disable temporarily the Riaa network and turn it into a straight amp, a matter of one or two components.
Now you are ready to go, that's all, no AP.
As an example, here is a -90.31dbFS signal that I made sometime ago.
Dropbox - Sinus2, −30-90.3 dB, 1 kHz.aiff
Hans
Thanks I will try that tomorrow..
No riaa preamp though. but a lot of opamps,so I can easily make an amp..
Alternatively, use my linearity test tracks:
DAC linearity test CD
You need to use a x100 low noise amplifier (battery powered) and an oscilloscope.
DAC linearity test CD
You need to use a x100 low noise amplifier (battery powered) and an oscilloscope.