Updated: 11-Jan-03 17:59
Doede's Audio Tweaks...

Tweak of the week: The Sony 777 SACD Player: Transformer Output, By Passing the OPAMP Filter / Buffer Stage

Updated Jan 11: click here to go straight down to the final solution !!


Introduction (March 16th)
Who read my column, now I was already very satisfied with the sound of my new toy, the Sony 777 SACD player. Actually this is an understatement. de "777" actually is VERY GOOD !! I am still breaking my head over the fact how they could make an OPAMP Stage sound so good.... But deep in my heart, I still know what the normal effect of opamp low pass filters is; due to the negative feedback to create a 2nd order filter (there are 2 stages in de signal path !!), you are always too late.... Not having negative feedback normally gives you a cleaner sound and more spatial soundstage. If the OPAMP is very good, the effect will be less. The used Burr Brown OPAMPS (OPA2132) belongs  definitively to this group. So why bother ???? Very simply, I just NEED to try this!! The DDDAC 2000 proves how good a simple Transformer Output can be...... For this purpose I already had 2 Sowter Transformers, the 9360, which is a 1:2 step up with a bandwidth over 100kHz.


This little project is not as simple as it looks like. The DAC used in the 777 is not a PCM63 like with just 1 single ended current output. The PCM1738 used here actually is a stereo DAC, but here it is used in double symmetric output. So you have 4 current outputs per Channel...... To be honest, I still need to think on to to do this best. Do not forget, that the 777, for stereo, is paralleling 3 DAC's (the 3 DAC's you otherwise would use for 5.1 multi channel mode). So per channel you will end up with 12 current outputs to be made 1 single current output.......This will be the next tweak. Back to the DAC stage: In the schematic below you see the first part of the DAC conversion and filtering. You immediately recognize the DAC, the I/V resistors, the I/V converter with the first OPAMP followed by 2nd order Butterworth filter and balanced-to-single-ended-converter. After this stage there are another 2 OPAMPS with a summing amplifier (remember the 3 DAC's) and another 2nd order filter. Than the signal goes to the cinch connectors through an Elco in parallel with a film capacitor. WOWWWW, that is a lot of OPAMP and negative feedback !!

So, for the first try, I will make my life simple and use the I/V converter with the first OPAMP. After this converter, the double balanced signal is single balanced and this can be made single ended with the Sowter Transformer ! Followed by a simple 6dB filter at 100kHz, this must be sufficient, I thought. Just have a look below where I tap of the signal for the transformer. PLEASE NOTE, that the output signal is completely FLOATING from the CD-Player, so take care for ground loops through the Power cables. They can induce HUMMM. I do not use the power grounding, so my signal is as quiet as , uhhh, as what? Well, quiet! I hear nothing, so that must do......s


Click to Download PDF File


Details on how to do this:
Pictures can tell much more than words, so have a look for your self. Click the little pictures for a large one with more detail.


sony777tweaks1.jpg (9769 bytes) sony777tweaks2.jpg (14757 bytes) sony777tweaks3.jpg (15150 bytes)
At the left, an overview. The RC-filter is wrapped in the Shrink Tube. At the right: Space enough, no problem !!


How does it sound????????
Almost forgotten: How does it sound, bypassing 3 Opamp stages and a lot of components ?? You would expect BIG differences, eh? Well to be honest I did as well...... and? Guess what, the difference is SMALL !! Why? because this analogue stage is so well designed, it hardly messes the signal up. So it is hard to gain giant leaps here. BUT it does sound BETTER and I will not reverse this tweak.  I have been listening to both CD and SACD and the "only" difference it makes, is that the last bit of sharp "SSS", which is so typical for OPAMP feedback is strongly reduced. This means 2 things:

  1. The whole sound stage is gaining rest, the music sounds more relaxed (not brown, laid back sound, no no no! it stays dynamic and transparent)
  2. As the little details gain on explicit sound reproduction (no more sssss) the whole sound stage opens up a little bit. Focus is getting better.

Big big thing ??? No, another class? no, Worth it? Ohhhh YES !!!!  This little tweaks goes into your subconscious. After listening a while you can blindly recognize this. If you switch back (just plugging the cinch cable in the other connector) you immediately hear the little harshness coming back and you know that you want to go back to the new situation and stay there.


Ready ????
Are we finished now with the 777 Tweaking ??? No, I still want to by pass the "last" OPAMP as well (the I/V one) and make it totally passive. But I still need to think on how to do this best..........


Update March 17th
That's pretty quick, eh? Well, after I finished and uploaded the article, I have had some brainwaves. So this afternoon I spent some interesting time with my scope/FFT analyzer and a test Disc in the 777....... Before I start spitting my idea (not built yet...) on the final OPAMP less TX output, let me share some interesting measurements on the solution above (which is any way simpler, than what I have in mind for the other one, by the way, so it is still a valid tweak...).


sony777_tx_vs_org1.jpg (11908 bytes) sony777_tx_vs_org2.jpg (12295 bytes)

Left: FFT picture from a 1kHz Sine wave taken from the standard Sony output, Right the TX output. Need to say more ??? hahaha!

sony777_tx_vs_org3.jpg (11861 bytes) sony777_tx_vs_org6.jpg (11749 bytes)

The same, but no the FFT window goes to max 4Mhz, In the Sony Output at the left , there is still a H.F. component !! In the TX Output at the right, quiet as..........

 sony777_tx_vs_org4.jpg (12407 bytes) sony777_tx_vs_org7.jpg (12577 bytes)

Ok, here the Sony output wins..... clearly the TX output does not go that low in frequency. Actually 25Hz -3dB. By the way, the terrible (pre) ringing is typical for the CD filter. I only had a test CD, no SACD. on SACD you will not find ringing like this. Mhhh need a SACD test disk........

Conclusion ??
This very simple tweak is absolutely worth it. The measurements proof, that every thing is ok. A little drop in Bass ? well I could not tell before I saw this pictures.... or may be I did not play the right CD yet..... but straight till 30Hz is not bad at all.


And now the serious stuff: the Final Solution without OPAMPS (March 17th)
OK, to be able to set up the design on how to get the all the 24 ( !!! ) current signals out of the all the 6 PCM1738's in the 777, we still need to do some measurements first, to know what we have in hand here:


pcm1738_Iout_600ohm1.jpg (10210 bytes) pcm1738_Iout_600ohm2.jpg (13175 bytes)

What I did, is isolating one of the 4 current outputs from the DAC, by very carefully desoldering one of the 47 SMD resistors...... pfffffff....... Now I loaded the output with a 2 kOhm trimmer and turned around a bit whilst playing the 1kHz Sinewave tone.

Top: This is where there is still no distortion. R = 600 Ohm Output is 450mV RMS
Bottom: OK, here we see total mess from the curentsource, R=2kOhm Output is 610mV RMS

pcm1738_Iout1.jpg (10664 bytes) pcm1738_Iout2.jpg (13681 bytes)


What do we learn? In theory we can load the positive and negative Iout from the PCM1738 with both a 600 Ohm resistor and get 900mV RMS out of it. This could be done, BUT I have a few reasons NOT to do so and use this measurement only as an indication of where the safety limits are:

At this point I had my brainwave ! Just have a look at the circuit below and see what I have in mind....

  1. We actually use the 47 Ohm resistors which are on the audioboards already and are very easy to access as "patch-point"

  2. At the DAC side of the 47 Ohm resistor the Output voltage from the DAC is available IF we ground the other side. You could argue that the minus input from the OPAMP is already ground level..... Yes it is, but it is a virtual ground, so the OPAMP is STILL in the signal path, No, no, no, by the relays we just connect this point to REAL ground. The OPAMP has no problem with that, and by doing so, we more or less ISOLATE the whole OPAMP stage from the DAC, without desoldering or cutting PCB routes.........

  3. Through a whole series (24) of 100 Ohm resistors we more or less "add" the 6 positive outputs and the 6 negative outputs per channel. Why 100 Ohm ?? just a round value, could have been 47 or 150 as well. Just a hard wire could also work, but I want to make sure, that the little offset differences in DC at the Current outputs are NOT leading to too high offset currents. 100 Ohm seems a good value to me, so let it be.

  4. The positive and negative are now loaded by the TX which makes it single ended and is removing all common mode error signals and filtering the high frequencies as well.

  5. The source Output will be 150mV eff. This is all very nicely within a wide range of the supply rails and with the 1:18.7 this will bring the output to 2,8Volt RMS. More than enough for my Line amp.

Click to download PDF document

This is what I will built soon..... a lot of relays, so need to check if there is enough current available in the 777. Otherwise I have to build in an extra power supply.

transformprim.gif (3034 bytes) transformsec.gif (3088 bytes)

What the heck is this here? Transformations??? Well it is to show the output circuit from 2 points of reference:

  1. From the DAC perspective, to see if the load is not too high. As you see it is not and the filtering is done smoothly and outside the audio band

  2. Perspective from the input circuit from the Line amp. Do we have enough Voltage ?? Do we have a low impedance source to drive the line amp ??

The answers on the questions are, that the load for the DAC is absolutely ok and that the source impedance for MY LINE AMP is fine. But be aware !!! If you want to drive a 20kOhm Potentiometer in an integrated amplifier or so, FORGET it. You will need to put an extra (tube) stage in between !!!! So I tweaked the whole thing for my personal audio. This is a not a 100% overall solution, working with every thinkable amp in the world !!!

If you want to know if it is appropriate for you(r amplifier)? As a rule of Thumb, do not load the TX output below 100kOhm and do not use interlinks with a capacitance of more than 200pF. Well, that isn't quite impossible, eh? I have ordered the Relays, so I can start building soon. Than I will be able to report on the final audible impact......


Update March 19th
Just a quick one..... below is the circuit I am going to build, not the easy patch as the other solution, but I have great confidence in the outcome of this Tweak (more project than Tweak, isn't it??) You can click the picture to navigate around at double resolution (or copy and paste it......).

An interesting solution to enable Frontpanel switching between the Sony output (handy for A-B comparison AND to play 5.1 mode...) and the TX mode is the little circuit at the top right. I use the existing Volume Potentiometer for the Headphones , which I do not use anyway, as a variable voltage divider. The voltage from the tap is lead to a comparator with hysteresis. If the pot is turned to one end, the comparator will switch a relays, which on its turn switch all the other relays and actually putting the whole TX-Output in place. Actually the original output at the back is switched as well between Sony and TX...... So no need to switch cables over and over again !!! The indication which one actually is playing (I do not want to have blind testing all the time...hahaha) is a LED which I will stick into the 6.3 mm Jack connector for the headphone. So no drilling, no nothing and the 777 stays intact. Weird idea? may be yes, but it complies exactly with what I wanted..........

To be cont'd !!!!


Finalization (January 10th)
I finally find some time to actually start building this tweak, which is quite some work to be honest. The principle is extremely simple: Through a bunch of relays the Current Output from all 6 PCM1738 DAC's are bundled into 4 signals, Left / Right and for both POS and MIN. This combined signal (therefore all the resistors) is being up lifted by the Sowter Transformer, Case ready. The most precise work is the wiring in the Sony 777, not to mention the disassembly of the chassis. I will limit my self now to the image details of the construction and (of course) the sound impression.......

sony777circuits.GIF (19851 bytes)

Actually Built Version 2.3 shows here:        Click the circuit for double Resolution !!!!

Or click here to get the full version 2.3 PDF document


The 777 Sound  Motherboard, location very close to the stereo output cinch


The 777 Sound  Motherboard, but now with the relay , switching the output terminal between Factory output (OPAMPS) and the transformer output. The yellow wires are the wires coming from the Comparator circuit which is driven by the headphone volume control....

This is how the PCB is hard wired, All Silver Wire !!!

Preparations for the Silver Wiring of the DAC boards to the PCB...

At the front you can clearly see that the DAC boards # 2 and #3 are taken of the motherboard. This gives room to solder the wires...

Detail of where the signals are being "tapped" and lead to the relays PCB... Very careful soldering here !!! Before you know the small SMD resistors fall of !! Quite some work to get them back.... Speaking from experience, ahem....

OK, at the end, when all 3 DAC boards are wired, this is the result !

The same, but from a different perspective... The Sowter Transformers can still be recognized. They are glued at the chassis, right under the signal wires...

This is how it works with the headphone: When the Phone Level Control is turned clockwise, the LM311 comparator will change state and drive the relays to take the Transformer "position" and the signal LED will shine through the Phones jacket  (just glued the LED behind the 6.3mm jack) CLICK here for details on how to make the connections.....  When turned the other way, the relays  will fall off and the original factory circuit is complete back in place without any interference (well, one relays contact than....)


The burning Question: How does it SOUND ????
When I bought the Sony 777, I was  wondering how OPAMPS could sound so good. Actually they really do in the 777.... So I was very curious if the transformer tweak could do any good here !! The day the tweak   was ready I plugged  in and was a bit disappointed..... there was hardly any difference !!! All this work for nothing and are OPAMPS really ok than ??? I stopped   listening and let the 777 play day and night in the transformer position. After a few weeks this really made a difference now !!! It seems that the MU metal in the Sowter transformers needs quite some time to get burned  in. Good to know for the next time...

As know by now (yes ??) the transformer output can be selected by the former headphone volume control, so I can try it over and over again, real A-B comparison. Compared to the transformers, the OPAMPS sound harsher. Remind: RELATIVE !! OPAMPS can sound much much sharper !!! It is a very relative world we live in.... The 777 with the opamp output is indeed very good, actually I have not heard of any thing better in this price class (and above...) so far. Than you turn the knob and hear how good it becomes with a transformer output. The gain is actually in the pinpoint definition, depth/width of the soundstage and the clarity of the high frequencies, very very smooth, very analogue..... This was definitively worth it !!! For a short while I thought OPAMPS could do the job, but now I know better, they are a limiting factor  at the  end after all.... In the  meantime I do not need the signal LED, Sometimes, when I sit down to listen some music and up goes the eyebrow, what is wrong here??? MMHHHH, missing something, check the LED and yes..... need some turn clockwise.....


Happy Listening and Building !!!!

Doede Douma


IMPORTANT: The information provided on this page is intended as guide for DIY activities and therefore free to copy and or publish. If any one wishes to use any of the information from my WEB site, please make sure to refer and footnote to my URL Link as source! Doede Douma