Modding a Marantz CD40

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi ppl.

Just got myself a CD40 in great condition (with a nice TDA1541A in it), except for the tray loading gear wheel (like usually...). But cheap replacement part is already on its way (thanks, epos!). Now I'm in the process of modding the CDP. (yes, I do have the schematics) :D
I want to replace the output stage (so to have a better opamp, not too much passive filtering, and no unnecessary components) and build a better regulated supply board for this.
I also started to "recap" the whole board (at least the 'lytics, as they are 16 years old by now...) and want to add a few decoupling caps or change them.
Of course, it'll get new gold plated RCA outputs aswell as some kind of metal shield around the boards of the output stage and it's PSU so to reduce HF interference from the digital circuits.
And last but not least, I will change it to NOS mode.
For now those are all the mods I plan to do. But I am open to every good suggestion, as long as it seems possible (and worthwhile) for me to realize them. I mean, I don't want to do overkill, you see... and I am no EE... lol

I've already done a little search on this forum and also on others, and found some good ideas, but still decided to open a new thread, because I couldn't find answers for everything. So, for a start, here comes the schematic of my planned output stage and it's PSU. I kept the same design as in the original schematic, simply excluding passive filters and unnecessary stuff (like the deemp. circuit). If you see any mistakes in it (hopefully not :rotfl: ) or anything that could be improved easily, please tell me :)

About the NOS, I found a little schematic explaining it, what they do is:

they connect WSAB, CLAB and DAAB directly to TDA1541. Now I am wondering if it's better to cut off the leads of the SAA7220, or to cut the traces on the PCB...
They also describe a Mute mod, so that without the muting transistors, one won't run into trouble with clicks and pops:
connect pin 23 (MUSB) of SAA7220 to pin 11 (MUTE) of SAA7210. The thing is, I have SAA7310 in my player... schematic says pin 18 is called MUTE (with trace over the word, as to say it's inverting). I suppose it'll work just like the SAA7210, no? Any experience with that?
Any other change recommended for NOS?
Read somewhere about changing the way the clock signal goes... but for now I don't want to build a new clock for it (would be the best, I suppose, but difficult to do), so I think the best is to leave things like they are, right?

In regards of replacing the 14 decoupling caps of TDA1541, is it worth to increase the size, to, let's say, 220nF?

Also, I was wondering if I should implement a simple RC lowpass filter in the last opamp stage... to filter out any rest of HF noise... keeping in mind that I want to do NOS mod... I know I can't get rid of harmonics (as that's the goal with the Digital Filter...), but that's ok... just thinking I'd like to avoid injecting HF noise to my amp :rolleyes:

And finally, to simplify things for everyone, what I'm not going to do (too much hassle for too less a difference, and too much cost in a CDP that's not that good afterall...):

I am not going to replace the power supplies of the whole CD player (although I would do a few tweaks to them, if you have any suggestions. They are nothing fancy, but aren't bad either).
I am not going to make any physical alterations to the player (change the case, make it a top loader, change the transport's mounting, clamping or whatso on), because I think it's quite ok as it is...
And I am not going to do any "crazy" stuff like glueing hundreds of cork or whatever strange materials all over the player, or fitting any real "boutique" componentes, etc, I think you see what I mean ;) .

Thanks in advance.

P.S.: if anyone needs the schematics, f.ex. of the PSU of the player, so to suggest mods, I'll post them here. I also can send the whole PDF per e-mail (it's of CD50, which is almost the same as CD40).


  • marantz2.png
    31 KB · Views: 1,195
Hi JoyDivision,

Sounds like a fun project.
When you bypass the SAA7220, you might want to remove it and place a socket in. Then, you can run the I2S from the input pins to the output pins of the socket with wire jumpers. That way, you'll be able to use the SAA7220 again some time.
Looking at the output schematic, there's a great big 100uF electrolytic in the signal path. I think that could come out.

That is a good suggestion. Also it would mean, I could install a switch or jumpers, making the modes (NOS and OS) changeable, so I could do direct sound comparisons...

Regarding the cap, I want to have a cap there, because I can't be sure today, where I'll connect my cdplayer to, tomorrow... so no dc-coupling here at home ;)
actually I have been using Panasonic FC caps for that purpose lately, in my amp and also in another project, and they sound really good, nothing to complain about.
I guess the value could be different, though... for now I chose the same like it is in the original circuit, probably meant to be ok for quite every possible strange input impedances of the connected device... but sure 10uF would be enough for standard 10k impedances... probably I'll end up with that value :D

Also, after talking over this project with some other people and reading a bit more, I decided to put in a simple analog filter in the 2nd opamp stage, and also later, I'll afterall will do a clock mod to it... it just won't be the 1st thing to do for me... I'll go step by step and see how the sound evolves :D

So, updated schematics to come soon.

P.S: any idea if it's important for the opamps to have +-15V, like they do in the original circuit? I know normally it's not, but here, I mean, regarding the I/V conversion? Or is it irrelevant? Because, as you can see in schematic, it's at +-12.5V so that I can use resistors I already have at home...
Use soft recovery diodes, especially on the anallogue PSU.
For now just dampen the crystal can with a rubber band or grommet.
AND, big improvement, if you keep the filter (7220) then lift the safety resistor on the supply line to it and feed it with its own regulator taken from the raw digital supply. This makes a big difference. Heatsink the reg, the filter draws quite a bit of current.
OK, so I've been doing some more progress. Updated the schematic of the output section, and also draw a schematic of what I'm planning to do with the digital section. Any comments/suggestions are highly appreciated.

Additionally to having a new clock (from Elso Kwak), it modifies the circuit so that basically it will be switchable between NOS and OS modes, allowing direct A/B comparison (that's my main aim. I've read a lot from both "sides", saw the measurements, etc, but now I want to hear for myself the difference...).
Also, when in OS mode, the DAC get's a direct clock signal (master clock divided by 2). I alwo implemented the MUTE mod. Main question I still have, is: will it be better to use good quality relays (like in my schematic), or something like 74LS157 (data line switcher)?

thanks for the interesting suggestions. I think, that safety resistor you say, is a ferrite bead, or am I wrong? So, what you say is, disconnect that from the main +5V, and connect it to a new +5V regulated supply, with for example a low noise regulator, well filtered, etc... right? I'll try that.


  • cd.gif
    49.7 KB · Views: 1,013
Indicate me your email adress and I'll send you the manual of CD50, which I've been using and which is almost 95% the same like CD40.

Btw, regarding the mods, in the meantime I've almost finished them, and the first results have been very good. I've a problem with NOS mode, though. As soon as I find the fault and get the whole thing finished, I'll post a complete report on here, for those who might be interested.
Cd40 Mods

Hello, I just bought CD40 & want to do some mods to player! I'm very interested in the non & os mods . I want to compare for myself the difference in sound. I would like to have a copy of schematics for the CD50 please. Can you post some pictures when you have finished the mods & don't forget to report sound improvements. Are you going to come of player thru analog jacks or are you going to install a digtial out jack as well? Ronnie
Ok, so I just finished the project.
Player sounds great so far. NOS mode is working now aswell (the problem is an error I did in the schematic: CLK signal for NOS mode has to be taken from 7310 (CLAB) and not from after 7220 like in my schematic... will correct post corrected schematic one day when I have time lol...

So, I now can switch between OS and NOS and back on the fly while playing, with no interruptions, nothing. Funny thing though is, that the sound seems pretty much the same :bigeyes: :rofl:
Important (and for me not easy to explain) is that it does have a different amplitude: 1.86V in OS, and 2.09V in NOS. So at first listening it does appear slightly different, but then it probably comes down to difference in amplitudes... I have to find a way of equaling the amplitude after the player, so I can do direct A-B comparison...

Anyway, for now I'm happy with the results, works fine, sounds great and looks ok

I've only done the analog output, as I don't have any external DAC at home. But if you want, you can implement it, CD50 schematics shows how.
As for the mods, feel free to start with what I've done and described here in the topic. Or go all along your own way :)
If well done, in the end you'll be rewarded with a much better sounding player. At least here it happened ;)

Just don't believe what some people say about NOS, because after hearing and comparing for myself, the difference seems to be very little. And I highly suspect: for worse... only explanation that some people judge their CD player better sounding after doing the very simple NOS mod (without changing anything else) could be proved by a theory I've heard: in OS mode, due to higher frequency, it's more jitter sensitive, so without a new good master clock, simply putting it to NOS mode might improve the sound, as there would be less jitter. I don't know if that's for real, but could be the explanation.
Now, as soon as we upgrade all the rest (new clock, output stage, etc), OS mode will most probably sound better, just like in the measurements: it has less staircase artifacts in the waveform...
But don't believe the other extreme side, either: NOS doesn't sound bad... actually as I said, it's very hard to hear the difference... it only looks worse on the scope, then probably all the "staircase things" get "anti-aliased" by the preamp and amp and finally speakers and our ear, the same happening to HF images of the sampling frequency, etc... but sure, you need at least a low pass filter in the output stage, so not to put at risk your amp (there are amps who react badly to HF).
So, good luck with the mods :)
Give me your email adress, and I'll send you the CD50 pdf.

Here come some pics of the finished player:


OK, just sent out the manual to you guys.

Ronnie: yes, perfectly possible. Although I've yet to hear a case where you can really distinguish the difference between "not-broken-in" and "broken-in" :D
But until any evidence against it, yes I do believe every _analog_ new electronic components have to "break in" to get to their full potential. What I do think though, based on experience, is that the difference has to be quite small. Otherwise I'd have noticed it more obviously.

Anyway, today just compared extensively my modded Marantz in O.S. mode to a classic Technics CDP with MASH DAC I have (and have known as quite good sounding. It's modified aswell, but only in the output stage. No new clock.). The Marantz wins, although with a small margin only consisting - and that's very interesting, probably being caused by the nice, jitter reduced, Kwak Clock - in better sound image and higher resolution. For all the rest they sound almost identical. But that little difference is enough to make it more pleasant and musical to listen to the Marantz than to the Technics.

Regarding comparison with my special Playstation CDP, I still have to do a few more tests. But the last conclusion I took was that the Playstation still sounds a little better than the Marantz. Have to check it again. The main problem is that output level in O.S. mode is little inferior to that of the playstation cdp... same thing is valid for the comparison between OS and NOS modes in the player... but I gonna find a way and then report about my conclusions.

So, good luck to you guys with your own experiments! :) Be sure to report back here, it'd be interesting for me aswell.
hi joydivision,
Glad to see you've been busy!
It think the hole OS vs NOS argument is fascinating - especially how people can have such strong opinions. I'm glad you tried both.
I first of all started off building a NOS DAC, mainly because it would be simpler (mainly because its dealing with lower frequencies), and less sensitive to jitter even though I was feeding it from something like a Kwak Clock.
I think that, and the fact that it can be used unfiltered, appeals to a lot of people. It does have a poor treble response (its really obvious if you have a test CD and oscilloscope - shocking in fact), but I feel that characterises the sound that appeals to some. A very 'laid back' 'analogue' sound perhaps. I found (especially when moving to OS and noticing the difference) NOS, due the lack of treble, seems to have a relative pronouced bass response that I really liked. But I don;t think it was a particularly truthful reproduction.
I was very pleased with the results of NOS, but wanted to try OS.
OS gets a bad rap (by the NOS fans) because, I think, it is trickier to implement, and requires more thought in circuit layout, and because the oversampler chip (especially the SAA7220) can pollute the power rails (hence the suggestion of using a seperate regulator - did you do that mod?) and introduce jitter into the I2S signal.
However, having tried both (alas I can't swap over on the fly yet), I was surprised by 4x OS. The relative reduction in bass was noticeable, but this was sacrificed in return for a far better frequency response and dynamics.
I think they're both different, and I can see how some prefer each other.
But I do prefer OS.

Also, I think the IV stage and filter can affect this. I doubt my anti imaging filter is optimum for instance - in relation to the SAA7220 and I want to go back and rework it some day. And I'm still working on a discrete IV stage, which should help.

Nice work!
joydivision said:
Ok, so I just finished the project.
Player sounds great so far. NOS mode is working now aswell (the problem is an error I did in the schematic: CLK signal for NOS mode has to be taken from 7310 (CLAB) and not from after 7220 like in my schematic... will correct post corrected schematic one day when I have time lol...

So, I now can switch between OS and NOS and back on the fly while playing, with no interruptions, nothing. Funny thing though is, that the sound seems pretty much the same :bigeyes: :rofl:
Important (and for me not easy to explain) is that it does have a different amplitude: 1.86V in OS, and 2.09V in NOS. So at first listening it does appear slightly different, but then it probably comes down to difference in amplitudes... I have to find a way of equaling the amplitude after the player, so I can do direct A-B comparison...



Nice work !

I have moded several players now. Marantz CD40 and 50 are a good choice in this regard.
There is still lot of room for improvement especially at servo part. That TCA0372xx Motorola chips can have a better supply etc.
Have a look also on laser supply. ;) A lot of "digital" garbage is coming from there.

I have a question.
In OS mode, do you know how an interpolated sample is calculated. How many master clock circles are necessary to recover a C1 error and how many for each interpolated sample ?

Just sent the manual to you.
I wouldn't say "much" to what I did :D I mean, I didn't change the whole digital power supply, for example, or even the trafo for a toroid, like some people do, etc... you can do much more mods... the question remaining then, is if it's worth the "trouble"... ;)

What you said about OS and NOS, is a very concise resumée :), I'd only add to it, that NOS also has its troubles, and they are not that small: the "stair-case" artifacts generated by the sampling frequency... yes, a normal analog active filter behind the I/V stage solves most severe problems, but leaves always a rest of "modulation", unless it's really complex multi-pole, but then you lose sound quality and phase linearity.
Anyway, not only high frequency loss is the consequence... not that you'd hear much of it, but well, it clearly is there... I can't say how tricky it is to get OS right, as I didn't do the mods sequencially in my player, but what I suppose is that you need a clock upgrade like I did, in order to have good OS sound, because it's more jitter sensitive. Probably I'd find NOS better sounding if I hadn't do the clock upgrade to my player... but like it is, suprisingly they really sound very similar... even the high frequency loss of NOS isn't very perceptible, probably due to the very nature of most music: it seldomly reaches out to over 15 khz... finally I do prefer OS also, if not only for the fact that while in A/B comparison sounding at least as good, I get a "theoretically" cleaner signal ;)

What I find strange though, is the difference of level: 1.86V for OS, and 2.09V for NOS?? :confused: I can't see how that happens... would be much better for comparing A/B if they had the same amplitude... maybe the 7220 reduces the amplitude while filtering in digital domain? Anyone knows that?

Sure that the I/V stage and filter do a lot to the sound... I wouldn't be suprised if a really well designed discrete I/V stage would outperform a standard opamp design... and even in opamp version, there is the choice between VFB (which I used) and CFB (some recommended it to me, others say it's worse... gotta try it one day!).

Regarding mods of the digital supplies, I haven't done anything there, except for better decoupling and exchanging all the caps for new ones. That would be something to try in the future: better regulator ICs (I have got to like one very special: the LT1963, it has ultra-low noise output...), seperate supplies, etc... one day when I have time ;)

About improving the servo section, don't think it's worth, at least with the CD40. Unless you mean doing it to avoid induction of digital noise to the rest of the circuit. Because the CD40, thanks to the CDM4, reads very well and quickly, almost every CD I through at it... much different than the old Technics (has a Sanyo laser unit) I have, it refuses to read some CDs, and I even had to do a complete readjustment to get it working acceptably again with normal CD-Rs (actually it's not that hard: find out where the Pot's are, mark there original position, put a good commercial CD in it, and start adjusting them, always looking to improvement of reading and tracking, etc... then repeat the process with a CD-R. Many a player that seems "laser dead" can be brought back to normal operation like this.)

can't answer your question. I am a little confused, though: what does OS mode have to do with error recovery? I thought that was done in a completely different IC (the 7310, if I'm not in error), so it should be exactly the same process, regardless of having the rest of the circuit in OS or NOS, no?:confused:
One of the biggest changes I heard in my TDA1541A DAC was when I replaced the 78/9xx regulators with LM3x7 regulation followed by a further TL431 regulation stage. Regulation, I think, makes a huge difference!
I also feel regulation, and power rail seperation for other parts of the player contribute greatly as well. So I can understand why seperating the power supply for the servo is important.
The CDM 4 is a great transport isn't it? I use a CDM9 these days, but there are discs that I reckon my old CDM4 would've read better.

I prefer OS, as I do think the treble elements in the signal, especially things like percussion, and also soundstage have are more detail.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.