or 'The Perfect-one Theory'

La justice est aveugle : my 20$ remote blindfold tester .The one and only truth .

Intro- Conditions of the test and some thinking

All the contenders were tested first blind A/B on my stereo system as transports / drives with my jacques-dac ( gigawork+mykonos) and then plugged into my oscilloscope from spdif output , both waiting and playing a record. Some of them gave the same signal both ways but others developped a mirror image of the 'square' wave or some other weird designs. I think these 'ghosts' come from interferences between electronics and laser. I could be wrong of course.
So , what about the 'sound' of the laser cell and the mechanism ? So much attentions are paid in high-end models for the laser assembly , it's too bad we can only refer to our ears to analyse it. Anyway, here are my scopes of the square wave, which I think only show us something about how the spdif signal is made by the chips---
sorry for the poor pics but I can't use flash here and I drink too much coffee to have a steady hand.

1- My fave Siemens RA306 . Signal taken directly from pin 27 of CXD1135 chip

I just love that signal just as I love the tone of the transport itself. Real strong and perfectly plane on the -1- . This is where I can throw my 'Perfect-one' theory. As the Siemens has the best sound to my ears, I came to think that the most important part of the square wave is its -1-, not its -0-...Of course this is pure fantasy, no science here. The cheapness of my oscilloscope anyway clearly shows here the best -1- of all the bunch and that's the main difference between my fave and the others. So why not ? The second important thing for me is its force. Compare this to a stock classic Philips spdif :

This trace comes from the stock spdif of my Marantz CD75 mkII but you get the same trace with any classic Philips stock spdif using SA7220. Both weak and sloppy -1-. Really average on the ears too. So let's steal the spdif from SA7220 :

2- Marantz CD75 mkII direct from pin14 of SAA7220 with resistor network

Better I will say. But it still keep its slope PLUS some sort of standard chaos on both -0- and -1-. I thought this wonderful player would make a perfect drive but I was wrong both trace-wise and ear-wise.

3-My second fave : SONY X222ES direct from pin60 of CXD2500

Nearly as good as the Siemens , same that I noticed listening to it A/B with the Siemens. Anyway, the -1- is not as linear as the Siemens, so, if my fantasy theory is right, that is why I always can tell the Siemens blindfold with its incredible airy reality. Worth noting that both Siemens and sony X222 did not developped 'ghosts' when playing.

4-Bronze medal : Sony CDP-770 direct from pin27 of cxd1135

A wonderful trace this is. Still the -1- is not as linear as the -0- and in fact this player does not sound as good as the 2 first. Another point to my theory.

5- Best PHILIPS : CD340 direct from pin32 of saa7341

The CD340 is so ugly I had to make this 'artistic' picture to swallow the pill.

The SAA7341 chip is a jack-of-all-trades : motor servo, ram , filter, DAC and EBU. Only the last one is of any interest. The player itself is very similar to the following CD115 , with the great drawer-laser , while it's made in Austria. Go figure .

Green bus rca : I don't know what they used them for , so I isolate the upper one for direct spdif.

Stock spdif was exactly the same than the marantz. No matter how the original spdif signal is, they always managed to turn it into this infamous trace. Those Philips guys must love standard . I immediately wired pin 32 direct to the green output for a good test. Incredible tone , VERY difficult to tell apart from the siemens. Maybe the siemens have more decay on piano, but it's a close, close shave . This signal has to be good.

That's exactly what I thought. A real perfect signal to MY standards. A very good -1- and great amplitude. This player is a must have for transport, while I really don't know if it's easy to find. Maybe all players with SAA7341. One thing for sure, the laser assembly is not responsible as we will see just after.

6-Flying dutchman on a zero over Portugal : Philips CD115 direct from pin4 of an obscure Mitsubishi DSP

I had great expectations from this player. First it has the very cool all-in-one drawer-laser assembly combo same as CD340. I love that.
Then it was ultra easy to fit it with a spdif output while there was none stock. See the pic :

Sound was ok but no magic as I expected. Up on the scale, champion :

My theory again : a very ugly -1- here, with no horizontal at all. Fly bbaack empty-handed to the wanabe pile. Obrigado.

7- TOSLINK steal from a Sony CDP-L3

The CDP-L3 was made for heavy use , like DJ stuff or something. Weights a ton, but only because its box is doubled with steel sheets. Good heavy KSS250. I could not trace the DOTX easily so I laid back on a prong from the TOSLINK thingie :

Promised I won't do it again : I fear shark teeth.

So I came back on it to locate the digital out : it was marked D/O on the board , originating from the green resistor ---see the 2 red circles --. Wired on a new cinch on the back to get the true spdif --notice the VERY simple cheap board and I could not locate the DSP...where is it???---

A very good trace indeed, similar to the best Sony here , but same thing happened as on the JVC : through my long spdif cable I got the shark teeth again . The good signal you see here is taken directly to the oscillospe through a short test cable. Maybe an impedance mismatch or something similar.
Not a good choice for transport if it has to be used with a very short cable.

8-Modern stuff : stock DVD players, SACD players or newer NAD CD players, etc...

Well all they have more or less the same trace, which is :

Seems to me they were expecting the oscilloscope test. Look like they taylored that trace with the chips companies.
Sound-wise they all are ok middle-of-the-roads. I can't help but thinking this is not strong enough and that there is something artificial in this trace. Still a good choice to experiment with DACs as you can find that sort of apparatus anywhere very cheap. Once again, a no-name brand new 35euros DVD player WITH USB input for your key full of mp3, with nice remote WITH volume control on the spdif, will do this job perfectly.

9- Crude computer CD-Rom

I like making transports with garbage CD-Rom from ancient computers. Like a revenge on consumerism. Here is my ugliest stuff :

Only the controler costed me some money. And it was not obligatory. The rest was found among friends surplus from old times. So, what would you expect from this tramp :

Well, already better than a modified classic CD player from Marantz, don't you find ? I just got this from the digital out of this ancient ASUS. Really a serious option if you don't mind the looks and the smell. And how can you beat quality/price ratio with something that's free ?

10- Tweakers fave CD722 stole at pin 2 of SAA7345 ( EBU )

It had to be a Philips player among good drives and a good DSP chip among their great production. So, instead of insisting on SA7220 , I tried the SAA7345 famous for being present on a lot a players of the '90s, like the early CD723 which is supposed to be a tweaker fave.
I had this nice CD722 with tda1545 not used --- for stock CD players I prefer the tda1549-- so here we go with some thin wire and my soldering iron :

At least that was an easy dismantling : two screws and everything goes out. The pins on the SAA7345 are bigger than on the CXDs so no problem soldering it. I took ground nearby so I can do some sort of quick shielding. Pin2 on the 7345 is EBU, which is not exactly spdif, but can work if you have a permissive DAC, not the Ultramatch. So here I go direct from pin2 to my gigawork. And it works. Great tone, still no cigars. Seems powerful, strong as my fave Sony...let's look at the scope now :

Heck , powerful it is. Certainly the highest voltage on this page. Does not look that good anyway, no horizontals but great amplitude, typically crude Philips DSP ( compare it to the direct SA7220 ). We've got to change that. It will certainly follows the CD930 and CD618 rule. Let's see this here. This player can be a very good choice, moreover if you already have a CD72X which seems ubiquitous and easy stealing. Prepare yourself to loud rotating noise as its motor runs full speed at any track. Not a good point.

11- From a predicted failure to a 99% success : spdif from a Playstation 1

I am getting good at soldering CXD pins thanks to useful adviece from a guitarist friend who makes a living of hand surgery. If this guy can sew back a human hand , I have no right complaining about how tiny are those pins to solder. So here we go with the legendary psx , web champion of low budget high-end cd player. I immediately emailed Mick Feuerbacher, maitre es ps1, who was kind enough to discourage me to try and get digital out from his beloved toy. Nobody succeeded before . But I was really eager to get a trace from it, considering its top class sony circuitry and funny laser arrangement, resting freely on rubber pods.

First attempt : the optimistic way.
My psx is a SCPH-5552 which differs from the acclaimed 1st model SCPH-1002 ( 2 is for Europe ) by the point it's almost the same. Anyway, the audio DSP here is a CXD2545 instead of the 1st model CXD2510 . So here I go wiring pin71 ( digital out ) directly to a cinch plug. No signal. Nada. Niente. Here in Marseille we would say degun , which is bad latin for nobody. My guess is that the digital switch pin is off.

Second attempt : the SONY way.
I just go and fetch the MD2 pin ( digital on/off switch ) . It was open. Open can be High or Low depending of the chip structure, I guess. So I wired it H just like on Sony schematics, with a 1K resistor going to Vcc. Still nothing. I am really upset and I really want to get this square wave out !
As a consolation, I put a CD in it. Great sound indeed thanks to its AKM4309 dac and simple buffer if you use the A/V MULTIOUT, while not in the ballpark of upsampling. And my ps1 makes glitch background noises with CDr.

This is only a drawing because I wasn't so sure and did not take a pic.

Third attempt : Broadway the hard way.
So I turned the problem in my head, searching far fetched solutions ( programming the $8X commands on the ram !) and all the time I was looking at this 1K resistor. And I said to myself : why a resistor ? what for ?? No logical answers---some heat for the winter ?---. So I simply cut it, and hardwired Vcc to the MD2 switch. Eureka !!!! All of a sudden a signal appeared on my oscillo and shut the unlock led on the minidac : SPDIF !!! That will teach me a lesson : when you want to put a pin on H, nothing beats full voltage . Here's view :

A nice spdif from a psx for the 1st time in the history of mankind----I get carried away here. Anyway, silly enthousiasm apart, this trace has nothing special to my eyes. It looks like the small brother of the Philips CD115 with the mitsubishi DSP. So I A/B tested it with my siemens, and nothing special happened. The Siemens eats PSX for breakfast , with much, much superior image of the singer in the mix. And still the psx have problems with CDr : it could read my test CDr for 10mn and then made weird noises in the background. I had to return to commercial CD.
Why did I say 99% in the title ? Because the psx as transport makes a LOUD high noise when the laser starts. I could not get rid of it with any resistors/caps network. Once started it's ok. I come to think this is why Sony always used TOSLINK instead of rca spdif on all players that had the CXD2545Q for DSP.
If you want to do it on your SCPH-100x playstation, locate the CXD2510Q chip,
just put the MD2 ( pin 59 ) on 3,6v and connect DIGITAL out ( pin 60 ) to cinch. Ground the cinch anywhere and you're done.

12- Messy outsider : Philips CD618 from pin13 of PCF3523

A high-end transport for 5 euros, a piece of plastic and 3 dremel blades . One if you're skilled with a saw .

I got this player for 5 euros because the tray did not work. After looking inside I realized the yellow wheel was completely roten and broke like thin glass . Strange time-chemical reaction. Buying a new wheel was excluded considering it will cost me more than the whole player . So I took off the complete drawer and decided to make it a top loading player, high-end style. I simply put a piece of wire pipe in the drawer switch ( red circle ) and sawed the top .

This player, easy to find cheap , has a SAA7310 for DSP which has no spdif function, so they use here this PCF3523 chip for digital out, pin13.

Stock spdif had exactly the same shape as the marantz or any old philips I hate . So I wired a twisted pair from pin13 to DIG OUT cinch isolated from the board...a very easy job as the CD618 can be opened on top OR bottom..I wish all players had this option....( while at it I also hooked SAA7321 direct to output. This bitstream DAC has no reputation for good sound, but opamps sure don't do it any good. ) So let's see what we have :

At first I thought I soldered the wrong pin. The messiest trace I ever had. But it locked on my test DAC with no problem. Worth a listening test. Blind A/B tests revealed a very good stereo spreading. Not as defined as the Siemens but really wide. In fact, it was very easy to recognize CD618, a whole new category of its own. I could believe I was using an old tda1541 with wider stereo image, while I was just listening to the CD618 as transport through my reference 24bit upsampling DAC. I wish I could have made this test when I was sure all transports sounded the same.
A real bargain, easy to mod---the laser don't care if you steal spdif direct--- , reading anything perfectly and any remote will control it.
If you don't believe me, see what you get with a simple 75 ohms load at the spdif output :

 That's the spdif signal from a 5euros cd618 from Philips , playing a cd, with a 5mn mod. Beat that.

13- ASUS DVDROM E818A3 from its digital audio output

What can I say : perfect trace and perfect sound. You need the DVDrom itself ( 20euros new ) , the chinese controler with remote ( this time obligatory : no start button - 20euros ) and a power supply 12v + 5v ( from china 5euros, you can see it in the background...I also tap 5v from it for the controler ) , a 45euros total for a GREAT transport almost impossible to tell from my best Siemens...maybe a very little harshness...very little...
and here is the trace you get for 45euros :

Perfect and quite silent too , once started , thanks to its 'Quietrack' system. If it would not need a casing , I could not see a better solution.

14- KENWOOD DP730 don't want no spdif

What seems a very easy mod at start turned out in a no-way : I easily spotted the spdif signal, as everything was ready for the upper model . But everytime I connected a wire on the stealing point ( red circle ) , the player stopped playing. No resistor or caps could fix that. Furthermore, this cd player won't work correctly without its hood. Too much mysteries : I gave up, but not before connecting its two PCM56 directly to ouput for a nice overall tone.

15- Very cool JVC XL-E53 from the Japan Victor Company

What a speech, while I don't understand a word...seems high-tech and certainly sincere...Let's see what is this PE.M. D.D. technology :

Well at least, it's not the usual OEM : JVC own chips & DAC , JVC laser mechanism and Rubycon caps almost everywhere---see the K on their top .
I personally think Rubycon are the best electrolytics ever -- I use them in mykonos -- and besides that reminds me of Tangerine Dream and Julius Caesar Alea jacta est . Great name.There was an easy stealing point close to the back---red circle--, very useful for short wiring but something weird happened : when touched with the oscilloscope probe, I got a great spdif signal . But on the cinch I wired up , I got the toslink signature : shark teeth. Why ? I can't understand .. just a wire and a cinch made a difference . No logic . So I decided to take the spdif from its source and not waste time on thinking.

Please tilt your screen 45° clockwise. Thank you.

It worked and I got a GREAT signal this time . I suspect the DSP to be Sony, but I did not look under . This cheap player -- 10 euros -- makes a perfect transport if I managed to find a remote for it. I also tried the JVC DAC direct but it was weak. So I decide to use the JVC opamps for once . They look good . And the sound was very musical . Great player in all ways .

16- Unmodable Philips cd723

The cd723 has a great reputation and it's true, this ugly player sounds very good stock thanks to its tda1545A dac and minimalist output buffer that seems to use only discrete transistors after one opamp for I/V conversion. Seems I have engineered this project. What about making it a good transport ?

This version uses SAA7378 for DSP which has DOBM on pin 31. So I just steal this signal with ground shielding and rewire the digital output . It worked on cd722 , so why not here ? Then I just put that new spdif source on the oscilloscope to see what we get :

Very similar to the cd722 but one big problem occurs : the cd won't play ! It needs a resistor in series, like on the original circuit . I believe, but I could be wrong, there is not enough DC to output a big spdif signal like this one, so when wired directly, the DSP misbehave and is not able anymore to do the servo correctly. Maybe that's what happened in other cd players like the sony X222ES .
I put the 723 back in stock form and away from the transport contest.

17- Philips 900 series : CD931 from PCF2705 pin 12

Here we have the last 'high-end' players from Philips. Nice looking, matching my DCC900 but nothing special soundwise as they use the infamous bitstream tech. Just like in CD618 , they use a special PCF chip to create spdif, here the PCF2705 .The pinout I found on internet was wrong : the spdif comes from pin 12 . I did not want to spoil it, so I took digital signal on a resistor at the end of the circuit . They are not directly in contact, but signal was exactly the same on both end, and really similar to the other PCF user, CD618 :

This signal works great but I wanted to give it a better trace so I try to experiment with all sort of components in the middle or in parallel .
Here is what I got with a 150 ohm resistor in series :

And here is the trace with a 75 ohms resistor :

Very good trace : PCF2705 chip and 75 ohms :

So my theory about shark-teeth trace and impedance I start to come up with the Sony CDP-L3 was quite right : if you go past 75ohm , you get the shark-teeth/TOSLINK trace. Very interesting . Anyway, while being a very good choice for transport, my CD931 is quite lazy with some CDR and there are no inner pots to adjust the laser. Usually I found out that philips with special PCF spdif chip don't need laser adjustment after direct stealing.
But maybe this one does.


I'll go straight to Sony based CD players and steal the DOTX signal where I can. This is a cheap and very efficient solution to own a GREAT transport.
And any 1euros universal remote will control them nicely---unlike Akai !---.
After all , this is similar to the famous Shigaraki transport with no heavy clamps I dislike and no hard work or big money spent.
My improvements : you can go inside and shield the wires coming out of the laser assembly the best you can, shield the wire that steals the DOTX all its way and keep it as far as possible from everything , use heavy copper shield wire from transport to DAC, clean the lens of course, silicon grease the mechanism rods and put ONLY good music in it.

WARNING : some players ( most ? ) will not accept the direct spdif steal as it seems to draw more current ( or something else )
and they will not play correctly anymore .
An adjustment of the servo pots maybe needed and it's quite a dangerous job if you are not used to it.

Philips players that use special PCF chip for spdif are a very good choice too as they usually don't need laser adjustments .
Like my incredible CD618 I strongly suggest.

Or if all this seems quite a pain to do, just buy a ASUS dvd rom with controler.
That's what I did in my homemade CD player.


In the meanwhile, I have tried a lot of CD players for use as transport.... Most of them did not like the mod, and/or did not make good traces on the oscilloscope. I love the cmd4, but most often, the spdif generator is SAA7220, which does not give good results. I tried several other players using the cdm4 in combination of a mitsubishi chip as DSP but taking its spdif directly stopped the player.

So my final conclusion is that the magic chip in this story is PCF3523 , which is specialised for this job. It can be found in philips cd618 and also in Marantz CD72. Both players are not that good for audio, therefore make perfect candidate for drive use.
Taking off the drawer of my 618 gave me the feeling of a better tone, but that could be placebo.
All the other possible mods---power supply, cap changing, clock, rubber damping, pointy feet, etc....--- have not been experienced here, but I am not convinced .
The combination of a great laser system ( cdm 4/19 ) and a philips spdif specialised chip ( pcf3523) seems more logical and away from mojo mods.
I know there are other chips that do this job , such as pcf2705, but I found the result a little less under ... and my cd931 did not work perfectly after the mod. I supposed the laser system and/or the servo circuit is of less quality than good ol'cdm4.
The race is on for these pcf3523 players ! ( p.s. tips on other models using this chip are welcome ! )


back to jacques hifi main page