CD960 CD94 TDA151A obsessive meddling

[UV101]

OK! Not wanting to completely take over this thread, I decided that I'd try to move what was ending up off topic to a new thread so here it is.

I think I may be more than a little obsessed by this rather magic multi-bit DAC from the late 80's. I was surprised to read about the AMR CD-77 that was still utilising this awesome chip some 20 years after it 1st hit the audio scene.

My journey began when my Marantz CD63 modifications were proving to be a little forward, bright and a little sterile for my liking. After some internet research, I decided that I wanted to see what this TDA1541A was all about. Fortunately, I knew where I could get my hands on a philips CD960. The 960 is almost identical to the Marantz CD94MKI. The only difference so far as I am aware is that the 960 uses the 14bit TDA1541 and the Marantz the 16bit "A" version and that the 960 is by far the ugly duckling!!! What Philips were thinking, I do not know!!!!!!

Out of the box, the transformer VA appears to be massive. The transformer is rather large! There are a total of 6 separate voltage rails. The I/V and analogue out stage, headphones and CCS for DC offset are treated to their own transformer windings, large Elna for audio and cerefine caps and fully discrete regulation. The rest of the player runs from 4 additional regulators. The most surprising thing for me was the all the +5v rails including the decoder, filter HF amp and digital uProcessor share a single 7805 regulator! I guess that's just the way they did things back then!

The Servo control section is fully discrete! I'm not aware of another player that can boast this! All this to control the infamous fully die cast, glass laser'd CDM-1 transport!

Sounds good already! However, the caps were all very tired and based on what I'd learned messing with the CD63 I knew there was much that could be improved. I set about separating all of the supplies in order to 1 at a time treat to some nice low noise regulation.

Over the past 2 1/2 years, I have managed to separate many supplies. I have tried various I/V and output stage configurations including an MF tube buffer clone. The DAC itself has been replaced 1st with an A version and finally at some considerable expense with what I believe is an original S2 double crown. The clocking for the player has been reworked with direct feeds for all devices requiring a feed and a direct feed with a derivative of the master clock to run the TDA DEM clock which is normally dealt with internally to the TDA by way of a cap connected across pins 16 & 17. (I think this could be the "hidden ‘cheat’ in the TDA1541 circuit which could be utilized to further improve the chip’s performance" utilised in the AMR CD-77??).

The latest round of modifications saw the 960 changed into a high quality I2S transport bringing the TDA and output stage outside the player housing. The TDA PSU's are fully discrete Salas shunt regulators and the DAC layout incorporates DEM reclocking and I2S attenuation all developed on the DIYAudio forum. The I/V stage is a simple Burson discrete opamp stage with feedback filtering and a CCS to null the output DC offset negating the requirement for dc blocking caps on the output. Inside the original player the transport uses the original windings for the analogue stage to run the CDM1 transport. This now comprises of Hexfred diodes, Audio upgrades (Mundorf SI) 22,000uF low esr capacitors and SPower regulation. The original servo windings still power the servo section again via Hexfred, Audio upgrade SI caps and Audiocom Q-Power regs for the +/-12v rails. The +/-5v rails again use the original windings into hexfred, Audio upgrades and Panasonic TSUP low esr caps and SPower regulation. The Decoder chip SAA7210, RAM HF & HF amp have their own winding, Hexfred, Mundorf SI caps running an SPower for the decoder and Audiocom super regs on the ram and HF. Another winding into Hexfred, SI cap and SPower runs the filter chip SAA7220P/B. Local decoupling is a mix of Black gate, Rubycon ZA and OsCon SEPC solid polymer.
The clocking comprises of an Audio Upgrades C2 o its own PSU again with Hexfred and Mundorf caps is responsible for direct clocking of the Decoder, Filter, DAC and DAC DEM. The servo is also clocked with my own clock powered from a basic early version of my own noise cancelling feedback reg on the same PSU as the main clock.

I dread to think how much time I've invested s far but its far from over. All I know is its actually quite good if I do say so myself!!

Transport PSU caps


Decoder, Filter, RAM and HF with own regs and Clocks
Not all the clock feeds are currently running direct as there is more work in progress!!!


CDM1 including SPower regulation


Transport


offboard I2S DAC


Early stages with S2 and Valve buffer

[Tweedle-Dee]

Very, very impressive work Ian.

TD

[Spectral Morn]

Indeed very impressive


Regards D S D L

[UV101]

Thanks guys, It’s definitely been a labour of love!!!

There are still a few things that could do with a tweak! I should really drop a separate transformer for the clocks into the transport and I need to get the distributed clock sorted out to the external DAC. I can then finally drop the S2 back into place. I never work on the machine with the S2 in place as I very much doubt I could replace it.

I know where I can lay my hands on a CD94 case. I may very well perform the transplant in favour of the better looking front panel. I also need to house the DAC section properly but right now I'm not convinced that the longer I2S bus lines are the way to go. It could be that I go back to internal DAC and external PSU's.

Anyway there is a lot of blood sweat a tears in that there player but I have to say its all worth it!

Apologies for some of the rather long sentences in the original post too!! It makes it a little difficult to read in places!!!!