This essay is a follow up to the first part The Marantz CD94 – Part1 (the drive mechanism) which described the outstanding quality of the drive mechanism from Philips (CDM1 Pro). In this part (II) I will share some thoughts and tweaks around the DA converter section of this classic vintage player.
Modern days
As you might know the CD 94 MK I is fitted with the Philips TDA 1541 A multibit ladder dac. This is a legendary DA converter – it never got any better, just cheaper. The audiophile magazines told us all the past decades, that every new converter generation gained a much better sound quality, than the previous one – which is completely wrong. The truth is, modern bitstream DA chips already imply everything you need to get a decent analogue signal out of your digital source, be it a CD or streamed digital data. There is not much more to do for the manufacturer as to built a proper power supply, often they use two – one for the analogue section and one for the digital stages. They design some nice casework, a good styling and a tremendous accurate clock – in which some of the more serious companies do a real good job. Any manufacturer uses nearly the same stuff. Some differ in doing excessive upsampling, some offer a different clock design, and most of the serious models do a sort of overkill in the power supply department and multiplying the number of converter chips to get a higher accuracy (Accuphase). Only very, very few models are in existence, which do really something innovative, Playbackdesigns with their FPGA DA converter technology comes to mind, Aqua with its modern implementation of the last multibit DA converter generation (already NOS stuff) and maybe those modern interpreatations of the classic TDA 1541 Philips converter, we will discuss here, done by High End manufactures like Zanden or AMR. Some companies think it is a good idea to construct the analogue output stage with designing a tube circuit – in most of the cases, it is just a buffer stage, which should colour the sound of the player. Serious tube output stages are very, very rare, companies like Acousticplan and Lampizator must be named here.
Some facts
Lets talk about some “digital lies” first…..rubbish that is repeated thousands of times which makes it not more true. Lets start with some facts:
Any conversion process will degrade the digital signal!!!!
That is hard stuff – because what is meant is every change in the source data integrity will degrade the sound. That is a matter of fact for up sampling, down sampling, oversampling (which is NO interpolation!!!!), changes in bit depth were it does not matter if we try to make out of 16 bit 24 or vice versa….EVERY CHANGE messes up the integrity of the original data.
There is no way – even for streaming services, to better the sound of the 1980ties Eurhythmics Album “Sweet Dreams” – other than a new mastering process with the original master tapes. If the original master is digital, which happened throughout the 90ties (last century) with DAT machines, and later hard disc recording workstations, the game is over – because we once converted analogue to digital – and now it is digital – and there is no way – other than converting the files to higher sampling rates and / or bit depth, which seriously degrades the original file in favour of some streaming clients, who think – “ohhhh my…..now my favourite Prince album is available in pure DSD…..It is customer´s fake!!!!
Most of the stuff at streaming platforms produced in the end of the 80ties and through the 90ties have seen a very poor AD conversion, and is just up sampled (converted) – because most of these productions had only a digital master – so there is no other way than a new mix – if the session tapes or data on a hard drive (end of the 90ties) are still available – otherwise a remastering could only deal with what was AD converted in those years – mostly not worth the money and the work!
But why was there such a bad AD conversion quality present in the first yers of the CD era???
My own story
Let me tell you a story from my own experience….I worked 25 years in the music business as a producer and engineer. One day – it was 1993, I left my studio with an analogue master under my arm to enter my favourite mastering facility in Hamburg. The mastering engineer took my analogue tape, and told me the record company wants to have, both vinyl and a CD – so we started to measure the Studer A80 mastering machine and routed their outputs to the mastering desc a very, very high end analogue piece of audio jewelry. I took a seat and enjoyed my mixing work – everything sounded familiar to me – a good sign – because nothing is more confusing, as if you cannot detect your own mixing work while sitting in the mastering process. The competent engineer behind the desc started with some tweaks, correcting mistakes in equalisation I did in my mix, we got rid of some emphasis in the lowest midband, gained some nice sparkle on top, so that the cymbals sounds very open and expensive. It sounded fantastic and I got more and more relaxed, because this guy worked miracles on my mixing work – Ekki was very, very happy. After the whole sound adjustment process we transferred the the signal to the Neumann cutting lathe – and had a very nice result. Lets call it a day….!
But than the whole mess started – because we also had to create a digital master for the pressing plant…..The industry in those days brought out the CD format, an international standard, which is in existence till the present day. Everything seems to be fine – but it was not! Because as we would witness two decades later with the development of the SACD the commercial audio industry forgot the professionals – those people who have to create the content, the industry wanted to sell.
Instead of a very elaborated audio workstation, we had the Sony U Matic system – a technology based on a drive, which Sony originally designed to record video data on it, a processor and a pair of lousy AD and DA converters. We connected the Studer master machine to the U Matic AD converter, the signal went trough the processor and error counter (something which is non existent in modern digital days….) – and digitised the formally analogue master. The output of the Sony U Matic DA converter was now routed to the monitors – and I thought I had to kill myself. Gone was the sparkle, gone was the natural midband, gone was the warmth, gone was the recorded space – I witnessed a flat, cold, dead and ugly sound – but that was the modern high end format of those years – stunning!!!!
In my example above, I did my mixing in the studio fully analogue – so, there would be a good chance, with better high resolution converters available today to remaster it again to develop a true high bit, high sample rate digital stream – which would not have been possible if I had used a digital 16bit / 44.1 master. You get the point???
Today it is a matter of seconds to convert any digital signal to something which looks impressive – modern computer technology with the world class audio software on them is able to generate out of a 16bit/44.1Khz file a 196khz, 24bit data stream – but you do not gain anything – instead of that you loose quality. Lets introduce the term “BIT TRANSPARENT”. To have a bit transparent signal in terms of its original source data is the best quality you can get!!!!!!
A lot of people wonder, while the old 1988 Prince CD sounds much better than the high resolution file available at several streaming platforms – now you no why that is the case…..
So why do we choose the CD 94???
The Maranz is very well built, it has a good platine layout, a fantastic drive mechanism and a lot of space for all the tweaks we have to do, to make a giant killer out of this audio grandpa. What that means is – that NO old Marantz CD 94 from the past in its original condition used as an integrated CD player will sound very good or reference class like – far from that! – So how can we unleash the whole potential….????
The TDA 1541
The Marantz is equipped with this DA converter legend – as I said above, it does not get any better than this – but first lets talk about the different TDA 1541 versions.
Philips had a grading system with their DA converters – the worst specs got the name TDA 1541 R – and the R stands for relaxed! – This Chipset was normally sold to companies who made very cheap players – so they wanted to have a nicer price tag on the very expensive TDA 1541 chips (which is way, way more expensive than ANY converter made today).
The next quality level is the TDA 1541 A – which was the standard grade chipset, which Marantz choose for the CD 94 MK I. – If we open the hood of our CD 94 MK I we will find exactly this chip which is soldered directly onto the board.
The high end converter types from Philips were the single and even more so, the double crown TDA 1541….called S1 and S2 and featured a stamped crown on their surface (or two). Be careful – today are a lot of single crown and even more so double crown fakes on disposal at ebay and other sellin platforms…as such a single crown TDA 1541 will set you back of around €250,- it is a good idea to know for sure if it is an original version or a nice artwork of some freaks….
Keep in mind, the double crown TDA 1541 is very, very rare. The one and only CD player ever made with 1541 double crown TDA chips was the Marantz CD7 – which is in my opinion one of the best players ever made – but if we are finished with our Marantz CD 94, we are soundwise very close to this audio legend – in some parameters better, in other the CD7 is unbeatable!
What we need is a single crown TDA 1541 converter – and that means also, that we have to unsolder the old A version, and instead of directly resolder the single crown version onto the board, we should opt for an IC socket, were we can plug and unplug the new converter chip if needed.
Recapping
The change from a TDA 1541 A to S1 status is not the most important tweak, we have to deal with….far from that. Starting from here, you are dealing with serious electric circuitry – a basic knowledge of electric engineering is essential, and all of what you do from now on is on your own risk – please read that twice!!!!! – If you never managed to get a proper solder point forget all of the following modifications and buy a good record player!
The Marantz CD 94 is now 30 yers old – most of its time it was stored in a dark cellar because his owner upgraded it with something new (more correctly, the Marantz shares half of its “celler time” with other old digital players, which are all “upgrade victims”…think about how many CD players you owned….;-)))) All caps – and I mean ALL of them will be far away from their original specs. Most of the players which were retired, had problems with reading a CD – the owner thought it is a worn out laser mechanism – but it is not – in 99 from 100 cases, it is the electrical side of the drive mechanism – not the laser itself or the mechanism. So we have to unsolder EVERY cap and resolder later on the best quality we can get today, which is not such a complicated task, because the quality of modern electrolytics gained immensely in the last decades. To get the work done in a proper fashion. You need the service manual – which you can find on the web (cost free) and you should be able to read a circuit diagram – of course.
I would opt for Panasonic FC caps in the power supply, in all the digital circuits SANYO OSCON is a very good idea – because they behave very stable at different temperature conditions…Be careful – OSCON cpas are widely available in SMD (surface mounted devices) versions – to find the standard mount version is often not an easy task – but they are existent. In the analogue stage I used Rubicon, and Cerafine, as also Elna Silmic caps – if you can find Black Gates – and have the money – feel free to buy them – but they are nearly vanished even on selling platforms like ebay.
PLEASE – do the recap work section by section…and if you are tired – leave the solder iron alone and go to bed – because every electrolytic capacitor is directional, and if you are tired it happens that……;-)))) Do function tests on a regular basis, if you are finished with one section – if there is a mistake, you will have a much better chance to find the problem fast and easy.
If you replace the big power supply caps (5 of them not only the big twin towers) I would opt for slightly larger capacity – and I mean slightly (10 % ) otherwise the original power transformer cannot handle your nice tuning!
The biggest improvement
After the recapping is done you have the rare opportunity to listen to a Marantz CD 94 in quasi new condition – and you will be impressed – there are more CD players in existence today which sound much, much worse than that old machine. But until you tweak the player any further there is no chance to play in the same league were the best players are at home….
The Marantz CD 94 – as most of the other TDA 1541 players have one serious problem in common, and this is the power supply regulators – or better, the amount of them. One regulator in the original design handles nearly all digital stages – and here is our biggest problem. The TDA 1541 needs a daughter chip which is called 7220 – and this thing pollutes a lot of digital rubbish in our rail which supplies also the DA converter, the clock and digital periphery. So we have to take care of that. As you might noticed, every Philips design in this period of time had those cooling fins at the back of the players. In the original design there are 3 discrete power regulators attached to those cooling fins – and we have to double that! Our goal is to give the TDA 1541 a separate rail – without any influence of the “dirty” 7220 – and we will do the same with the 7220 itself und the clock.
So we have to design 3 new power regulators (discrete) and connect them with the corresponding parts on our circuit board bypassing the original power traces, so that all 3 devices (TDA 1541, 7220 and clock) are directly fed with their supply voltages. We are must mount the hot transistors of our discrete regulators to the same cooling device as the already existent three original regulators. To distribute the clean supply voltage, use good quality solid core copper cables, and twist them tightly to gain some shielding.
Now we have reached a lot more sound quality – the truth is – you wont think it is still the same player….it is that impressive!
Further improvements
If you like, you can install now a new high end clock – there are different manufacturers in existence, I would opt for a Tental Labs solution…which fits perfectly into the CD94. They supply their clock with a good manual and it is an easy task, if you already managed our journey to this point.
The analogue output stage
Now things become very interesting again…You can decide if you want to get rid of the whole de emphasis circuit, because (I bet??) you will have not a single CD in you collection, which is coded with emphasis (which was a very old technology in the first years of CD production to gain SR ratio).
What you seriously have to do, is get rid of the 4 poor op amps soldered directly onto the board. To do so, buy some very good quality IC sockets – because (I know you very well) you want to experiment with different OP amp designs. Keep in mind – just pulling out one Op amp and replace it by another one is not the way that game works, because EVERY high bandwith OP amp today needs its proper implementation – so read the papers….and do your homework!
A good choice will be the Burr Brown OPA 627 single (all 4 OP amps are single types). You can also go full throttle and buy some discrete OP amps from companies like Burson Audio – and if that is not enough, you can use Daniel Weiss devices from Switzerland – they are very, very good, but also very expensive – and they need proper implementation, which means a complete redesign of the IV stage and filter design!!!! – Which is a lot of work and needs serious know how.
I opted for Burson Audio, which are way better than any integrated OP in this player (not always the case) and not even in the same galaxy, than what was used in the original design of the CD94!
All Philips TDA 1541 implementations (there are two exceptions) have a DC offset at the analogue output section of the IV stage – which we have to get rid of. The way to do so ist a coupling capacitor directly in front of our output sockets. Here you have to calculate the capacity of the cap – which determine the lowest frequency your player will be able to produce…a figure around 22mF would be perfect. You can search for some really good caps in the high end accessory market. Keep in mind, that you need for the best of them (and the most expensive ones) a lot of space, which you don´t have. And – if you use a big, big silver foil mega hyper cap – you will have very long connection legs, which is not a good thing! I used Mundorf tin foil caps, which are not so big – and fits very nicely into the player – but there are much more options on the market – do not go mad here – it is just one single detail!
Next step would be to get rid of the RCA terminals Marantz used in this player – do not ask me why – but that is the cheapest rubbish you ever have seen. Buy some good tellur copper RCA sockets and solder the output coupling caps directly to the sockets – and you are done! Further improvements can be achived if you change the diodes in the rectifier section of the player. It is also a good idea to deactivate the headphone amplifier of the CD94.
One last word……as all these modifications means a lot of work that has to be done, do yourself a favour and buy only good original parts, leave out ebay and any cheap offerings, their might be a good chance to get copies of the original stuff…..
The Sound
Now you experienced what it means to implement a TDA 1541 near to perfection – there are way more tweaks you can do – but what I explained here are the most important ones – to go any further is another story maybe for part III???.
I had the chance to compare the Marantz CD 94 to the lates Naim triple 5 streaming DAC (also a ladder dac design!!!), as also the smaller Aqua DAC and the second best Plaback Designs CD player – and guess what – our old Marantz machine is in the same league. It does not sound equal – of course not – but it is the same level of quality. What we always have, if we implement a TDA 1541 near perfection, is a once in a lifetime digital treble experience. The TDA sounds creamy, smooth and has a tremendous resolution. This resolution has nothing to do, with a tilted up, analytic top end – this treble is natural, real, and relaxed. I am not a big fan to modify the Marantz CD94 with a tubed output stage, because this modification makes things too soft, too polite – and as I use a full tubed amplifier system behind the player, it would be too much of a good thing. What you also will detect is a tremendous bass performance. The bass gestalt is powerful and deep – really deep, as also tight and very agile – it is one of the most impressive bass performances of any DA converter which is in existence till now. This bass gives your analogue rig a good workout, only the best cartridges (also very expensive) can muster the bass quality of this vintage (modified) CD player. In the midrange we will miss some of the 3D magic our analogue front end is capable of – but that is often a matter of the different mastering techniques used with the media CD – because it has to be very loud (loudness war – another story, I will write about in the future) – so you will face a completely compressed sound on the CD, which is not so much existent on the vinyl mastering of the same production, because such loud and compressed signals are not easy to cut into vinyl. What you also miss is the attack accuracy a Lyra cartridge can give you – but there is a tweak for that at the end of this essay. To make a long story short, the Marantz CD 94 in its modified incarnation sounds close to an EMT TSD 15 cart – BOOOOOM!
The last tweak
Everyone is talking about Non Over Sampling (NOS) today – it is the newest fashion to get rid of oversampling artefacts, which degrade the pure digital signal (time domain). Oversampling was at the very beginning of the CD era a trick from Philips, as they had just a 14 bit DA converter to offer, while Sony could develop a true 16 bit chip. This is the reason why Philips introduced oversampling – to get 16 bit resolution out of a native 14 bit DAC – and they succeeded. Later on, the situation was comparable to the run for the highest megapixel camera. People do not understand digital technology, they have no knowledge about pixel pitch and all the problems a higher number of megapixel will bring to the game – more is always better – and so we will face in the next years 48mp digital cameras – and nobody knows why a 24mp file looks much better.
The same was going on in the late 80ties and beginning of the 90ties – as multibit DAC technology was the latest fashion. We witnessed the oversampling race…
It starts with 4 times and ends with 16 times oversampling processes…what the people did not realise: With every increase of the oversampling factor, the industry had to work seriously on the speed and accuracy of the clock. DSP technology was not developed in those yers, so instead of that the industry used quarz PLL clock systems. The sound of these player generation got worse – because of clocking inaccuracy (jitter) – but the consumer had his higher oversampling rate – GREAT!
We can modify the Marantz CD 94 / TDA 1541 to be converted into a NOS DAC without any digital filtering – it is an easy modification, and many, many websites describe, how that modification is done. What you gain is more timing accuracy – because every steep filter (digital or analogue) will deliver a sort of pre ringing (and ripple). You get 3 cycles (worst) in front of the natural leading edge of a tone. Once you compared the same player with and without oversampling, you will be stunned. Interestingly enough the US company Wadia discovered exactly those phenomena already a long time ago….
But nothing is gained without something we loose – the frequency linearity of our player will suffer after we converted the CD94 to NOS – which means we get a drop in treble energy around 3db at 20Khz with a soft roll off. So you have to choose, accuracy in the time domain versus frequency linearity in the treble section. I opted for the timing accuracy and did everything I could do, to get the last drop of openes out of the player, which is also the reason, why I did not imply a tubed output stage at this point. Finally you get a very, very serious CD machine with all the above discussed tweaks and modifications. If there is no vinyl available, you can easily buy the CD and enjoy it immensely….and that is, why I did all that work….
Happy listening
E. Strauss
Medialux 