Best opamp for I/V conversion? (DAC)

[Elso Kwak]

Originally posted by janneman
Elso,

The opamp can only be too fast for the buffer if the SIGNAL is too fast. The opamp itself has no opinion on this.., no?

Jan Didden

Hi Jan,
No, I don't think so. See page 87 and further in the Application Note. I am not going to explain this again. This is about loop oscillation.

[Elso Kwak]

Hi IpsilonSound,
My approach to audio is very empirical.
The westhostsite by Elliot is one of those sites I cannot stand....
I have tried dozens of opamps as IV in my NON-OS TDA1543 KWAK-DAC like in random order: LM6171, OP249, AD844, AD811, OPA604, AD711, AD744, AD8620, OPA603, LT1026, AD815Y, AD 826, AD817, AD827, AD847, AD823, OPA 627, OP27, OPA134, OP275.
To my ears the OP275 was the best. I did not perform any distortion measurement.
What else can I say..........
I will try the THS4082 which has the same simplified schematic as the AD817 by the way.

[Gordon McGregor]

IpsilonSound, it is interesting that you came to the same conclusions with Alex Nikitin, former chief engineer of Creek Audio (known in this forum as x-pro), concerning THS op amps, though he used THS4062 instead of THS4082, but they are quite similar.
IMO the good sounding of those op amps are in their output stage in A class (or almost A class), as with some loadings they run quite hot, plus in their superiority of even harmonics structure vs odd ones, which makes the sounding softer or warmer.
Concerning OPA2604, I don't really like it. OP275 has very low level of distortions, but majority of them are odd ones. I never tried those op amps in TDA1543 DAC, only with PCM1704 though, but that PCM1704 works in x8 oversampling mode with DF1706 ...
For me OPA627 is superior in comparison with OPA2604, and even OPA2134 is superior to it. OP275 is OK, but I liked OPA627 and AD843 better.
The difference between non-os TDA1543 stage and OS DF1706/PCM1704 is that in the first case we have to deal with some glitches ... and in the second case any HF above the DF frequency ranges are filtered. I did the distortions measurements for 49TXi, used as AV processor (signal path is SPDIF output Lynx L22->SPDIF input 49TXi->DF1706->PCM1704->OPA627->OPA2134->OPA2134->OPA2134->input Lynx L22), in Direct mode with RMAA software and through Lynx L22 soundcard:
http://www.angelfire.com/clone/gordon0/49TXi.htm
"Frequency response (from 40 Hz to 15 kHz), dB: +0.00, -0.01 Excellent
Noise level, dB (A): -110.8 Excellent
Dynamic range, dB (A): 109.5 Excellent
THD, %: 0.0042 Very good (it is equal to -87dB for the whole AV processor!)
IMD, %: 0.0073 Excellent (equal to -83dB for the whole AV processor)
Stereo crosstalk, dB: -105.3 Excellent"
Those results are superior even in comparison with the good professional soundcard, for example ESI Waveterminal 192X:
http://audio.rightmark.org/test/esi-...192x-2496.html
Or AudioTrack ProDigy 192 (semi-professional):
http://audio.rightmark.org/test/audi...y192-2496.html
I am not even talking about such devices as Audigy 2 ...
http://www.extremetech.com/article2/...1151975,00.asp
Noise level, dB (A): -96.5 Excellent
Dynamic range, dB (A): 94.1 Very Good
THD, %: 0.0037 Very Good
IMD, %: 0.0096 Very Good
Stereo crosstalk, dB: -88.4 Excellent

Test conditions:
Output level of volume control in 49TXi was +12dB.
Connection: SPDIF output of Lynx L22-> SPDIF in 49TXi (used DVD/LD digital input 1), stable clock of Lynx L22 is almost equal to i-link (though anyway worse, because of PLL clock restoring in the receiver, it means the results through i-link should be even better). Analog outputs of 49TXi (pre-out, FR and FL) -> analog inputs of Lynx L22; this is the required loop. Level of the input signal -3.1dB below FS.

http://diyparadise.com/TDA1543.pdf -70dB THD+N in TDA1543 plus their limitation to support SACD and DVD-A were always problems for me ...

[robbie b]

In the end it is the sound what is important.
i like the simple method; a resistor as I/V and a Sowther transformator 9545.
Sounds beautiful smooth and relaxed.

But then again,.... I like that

[Gordon McGregor]

I have an offer to you - let's record through my audio signal path and trhough yours the same VERY high quality musical track (from NAIM or TELARC samplers, for example) and using ABX or ABX plugin to foobar, compare them ... which one will sound better ... Unfortunately you can't do that with Telarc SACD multichannel sampler 2004 or using very revealing DVD-A Advanced resolution Metallica 2001, which I usually use for system's quality testings, but OK, I agree to make a comparison at the CD quality level ...
Seriously, I don't claim that at 44100Hz*8=352.8 kHz DF1706+PCM1704+OPA627+OPA2134 will win vs TDA1543+Sowther 9545 at the CD level (44100Hz non-oversampled), but still such comparison will be interesting for a lot of people and at least with DBT will be more or less objective ... Plus, don't forget that I will use i-link with PQLS on and you will use SPDIF ... (i.e difference in the level of jitter)

[robbie b]

Originally posted by Gordon McGregor
I have an offer to you - let's record through my audio signal path and trhough yours the same VERY high quality musical track (from NAIM or TELARC samplers, for example) and using ABX or ABX plugin to foobar, compare them ... which one will sound better ... Unfortunately you can't do that with Telarc SACD multichannel sampler 2004 or using very revealing DVD-A Advanced resolution Metallica 2001, which I usually use for system's quality testings, but OK, I agree to make a comparison at the CD quality level ...
Seriously, I don't claim that at 44100Hz*8=352.8 kHz DF1706+PCM1704+OPA627+OPA2134 will win vs TDA1543+Sowther 9545 at the CD level (44100Hz non-oversampled), but still such comparison will be interesting for a lot of people and at least with DBT will be more or less objective ... Plus, don't forget that I will use i-link with PQLS on and you will use SPDIF ... (i.e difference in the level of jitter)

Gordon I guess your question was adressed to me?
Just to make something clear. I'm not using a TDA 1543 but using the following:
CS 8420 24 bit 96 kHz; DF1704E 8 times OS; PCM 1704; resistor and Sowter.
I started with a standard Sony XA 50 ES player wich was not satisfying to me.
I first implemented the ZAPP filter from LC audio to the existing Sony DAC. This was already a big improvement. The ZAPP filter is a Class A discrete I/V, filter and amplifier, look at LCaudio.com.
Then I build the DAC with the upsampler and it outperformed the Sony DAC by far. Also the upsampler, 24 bit 96 kHz, brought some extra air.
Just to be curious I connected the new DAC with the standard I/V, filter and amplifier of the SONY player. This was a big dissapointment, the OPA 2604 for I/V and filtering and the OPA 2132 for amplifying where not as near as good as the ZAPP filter. Unmusical, flat, sharp edges en very fatigueing. Changing back to the standard SONY DAC made it even worse, I never thought that this player sounded that unpleasant in the standard situation. How a person can adjust to a certain soundquality, you must have heared an improvement to believe what is possible.

After reading a story about NOS and the effect of filtering I thought that one could use an transformer to use this natural high roll-off as a filter, done a bit of searching I got to Sowter and got in to discussion with Brain Sowter. He suggested to try the 9545 transformers with a 100 ohm resistor. I must say it worked very well. The music flows in to the room, it sounds all natural. I must say that I listen to blues jazz vocals and classical most of the time. With hard rock you miss a little bite, I agree. But when I want to listen to that I can switch back to the ZAPP filter which will give me this extra kick.
When using a transformer just try to increase the input impedance from the preamp. I increased it from 47K to 150 K to correct the otherwise to early rolloff. All these above adjustments where made with no measuring devices other than an DMM and my ears.
To come back to your proposal, I have no means to record music except on an old cassettedeck. So I'm sorry, but I find your idea very interesting so perhaps I can find someone in my neighbourhood with either a cd recorder or a DAT recorder, but for the moment I can not help you.

Regards

Rob

[fmak]

[QUOTE]Originally posted by Gordon McGregor
[B]I have an offer to you - let's record through my audio signal path and trhough yours the same VERY high quality musical track (from NAIM or TELARC samplers, for example)
--------------------------------------------------------------------------------
Telarc sampler may be ok, but the Naim sampler is nothing to shout about--the music for a start. Many of the give away samplers with magazines are not musical, I think simply because they are cheaply produced.

[ThorstenL]

Konnichiwa,

Originally posted by fmak
Telarc sampler may be ok, but the Naim sampler is nothing to shout about--the music for a start. Many of the give away samplers with magazines are not musical, I think simply because they are cheaply produced.

Funny, I feel the opposite to be true. Much of the Telarc stuff is musically bland and boring (if techncially competent) and recording techniques seem to emphasis effects and gizmos at the expense of realsitic sound.

On Naim's Recordings I will again severely criticise the recording quality, however many pieces are musically very good, intreagung and make me want to listen more, which is not what Telarc recordings do to me.

Also, a lot of the free samplers on MUSIC MAGAZINES are quite good, but one must take care what one buy's. I have some excellent CD's from these covermounts, among them a Linn Records (BBC Music Mag) with the Orchestra of the Age of Enlightenment playing Vivaldi, Scarletti, Haendel, Corelli and others. Exceptional musicmanship, high production values and faiely realistic sound.

Also, for Rock/Pop several of the Q Magazine covermounts are excellent, as are many of the Classic FM Magazine Samplers for classical (especially for tasters). Last but not least, the "attached" magazines often make excellent reading.

BTW, does anyone posess the Linn Series "Songs of Robert Burns"? Drop me a line please....

Sayonara

[carlosfm]

Anyone knows the two (I have two, I think there's more) CDs from Focal/JMLabs - "Le grand spectacle su son"?
Beautiful.
Joe Jackson as never before on CD.
And much, much more...

[Gordon McGregor]

[QUOTE]Originally posted by robbie b
Gordon I guess your question was adressed to me?[quote]
Yes, Sir.

Just to make something clear. I'm not using a TDA 1543 but using the following:
CS 8420 24 bit 96 kHz; DF1704E 8 times OS; PCM 1704; resistor and Sowter.

Wow that is even better, almost the same config as mine, but I use OPA627 instead of Sowther and Icy-Lynx Micro (i-link with A&M http://www.xilinx.com/esp/consumer/h...e/complete.pdf) instead of CS8420 SPDIF, but it is OK. That comparison may be very interesting.

I started with a standard Sony XA 50 ES player wich was not satisfying to me.

It is understandable

I first implemented the ZAPP filter from LC audio to the existing Sony DAC. This was already a big improvement. The ZAPP filter is a Class A discrete I/V, filter and amplifier, look at LCaudio.com.
Then I build the DAC with the upsampler and it outperformed the Sony DAC by far. Also the upsampler, 24 bit 96 kHz, brought some extra air.
Just to be curious I connected the new DAC with the standard I/V, filter and amplifier of the SONY player. This was a big dissapointment, the OPA 2604 for I/V and filtering and the OPA 2132 for amplifying where not as near as good as the ZAPP filter. Unmusical, flat, sharp edges en very fatigueing. Changing back to the standard SONY DAC made it even worse, I never thought that this player sounded that unpleasant in the standard situation. How a person can adjust to a certain soundquality, you must have heared an improvement to believe what is possible.

Understandable again

After reading a story about NOS and the effect of filtering I thought that one could use an transformer to use this natural high roll-off as a filter, done a bit of searching I got to Sowter and got in to discussion with Brain Sowter. He suggested to try the 9545 transformers with a 100 ohm resistor. I must say it worked very well. The music flows in to the room, it sounds all natural. I must say that I listen to blues jazz vocals and classical most of the time.

The same as I do, and I got the same feelings with OPA627 ... but I use M&K 150THX speakers+M&K 350 MkII sub. What speakers do you use?

With hard rock you miss a little bite, I agree. But when I want to listen to that I can switch back to the ZAPP filter which will give me this extra kick.
When using a transformer just try to increase the input impedance from the preamp. I increased it from 47K to 150 K to correct the otherwise to early rolloff. All these above adjustments where made with no measuring devices other than an DMM and my ears.

I used measures to be sure that everything is OK. I can't trust to my ears after Double Blind testings ABX and op amps replacements ... some of them were so close that I could not hear the difference ... and I could see it in measurements. It is not concerning OPA627 though.

To come back to your proposal, I have no means to record music except on an old cassettedeck. So I'm sorry, but I find your idea very interesting so perhaps I can find someone in my neighbourhood with either a cd recorder or a DAT recorder, but for the moment I can not help you.
Regards
Rob

Thanks for your wish, Rob. The best stuff to use for recording, as anyway we need to analyze the signals in PC, is to use good soundcard such as Lynx L22, Two B or E-mu 1212M/1820M. I hope that sometimes we will be able to record and compare oversampled and non-oversampled DACs to finish the discussion ... or at least to have more or less objective listening evaluations.

Here is the link to Linn's "Songs of Robert Burns" http://www.linnrecords.com/cd_inform...ecordingID=232

I don't want to argue about the quality of Telarc's samplers or Naim's or Chesky Record's ... we can choose the one which will satisfy everyone. As we will record just a small most revealing part of the song for testing, we will have no problems with copyrights here. We can freely distribute such small samples to let everyone compare them in ABX software or through ABX plugin to Foobar.

Joe Jackson as never before on CD.
And much, much more...

Carlos, I prefer SACD multi-channel or DVD-A Advanced resolution to any CD, even to any HDCD, sorry about that ...