Friday 8 February 2013


Quick update to explain what those regs are doing! Probably repeating myself a bit here but hopefully this will help in following the principle.
In the standard player, we have a distributed 5v rail fed from a single regulator. The reg converts an input voltage of approx 10v to a regulated 5v. There are 3 connections for the reg. IN/GND/OUT. THe std reg is a 7805 which is a fairly noisy device and hasn't changed in many years. Newer regs are much quieter which can only be a good thing. What we are doing by adding multiple regs, is to distribute the unregulated 10v rail from the IN pin of the single 7805, feeding into a separate regulator and using this individual supply to feed a descrete pin on a device. In this example we are talking about pin 12 of the SAA7350 DAC chip. The pin was originally fed from the shared 5v supply via a low ohm resistor. After the resistor, we had a 47uF cap to ground and in parallel a 4.7nF ceramic SMT cap to ground.

The small value cap is designed to shunt any trace noise to gnd. Ideally it should be as close as possible to the supply pins of the device and ideally it should span the relevant supply and gnd. Looking at the datasheet for the SAA7350, we can see the digial supply pin is pin12 and the digital gnd pin is 14 which is quite handy as we can attach an SMT cap (100nF PPS but C0G would also be good here) directly accross the pins with excellent effect. Next we can remove the original 47nF ceramic cap and replace with a larger value (the effect the original cap has on the rail has been superceeded by the 0.1uF directly on the DAC chip). I opted for a ceramic 2.2uF here. Finally the original 47uF electrolytic has been replaced for a much larger solid polymer cap (solid polymer have astoundingly low ESR).

Texus Instruments REG104

The regulator I have chosen to use here is not the std 3 pin type but don't let that worry you, its not difficult to impliment. The TI REG104 is a 1A reg with an impressive noise figure. It actually has 5 connections. In order pins 1 to 5 IN/OUT/GND/NOISE REDUCTION/ENABLE. The tab is also GND whihc i'm connecting to the top screen of the main PCB as this is player gnd in this example.

Hopefully the 1st 3 pins are obvious. The other 2 are pretty straight forward too.
NOISE REDUCTION - connect via 0.1uF (100nf) to GND - This reduces the noise on the output pin
ENABLE - Connet via 175k to IN - the resistor is required where the input voltage exceeds 10v

Before and after diagram




The process has been repeated for DAC Digital, DAC Analogue Left and DAC Analogue Right so far with the only difference being the use of normal electrolytics for the analogue rails as they tend to sound better then solid polymer.

Saturday 2 February 2013

DAC PSU - new caps and separate voltage regulators


Packing to move and a few other issues have meant progress is slow. I've now got separate regulators on the DAC digital, analogue left and analogue right. I'm using TI reg 103 and 104's which quite approx 33uV noise which is very low. 

The digital reg is on the top side. I've replaced the 47uF elco with a 330uf solid polymer nichicon cap with exceptionally low esr. The 47nf on the PCB side is replaced with a 2.2uF ceramic and I've got 100nf pps right on the supply pins on the dac itself. The analogue L&R have similar treatment with the regs fitted to the bottom of the board. The 47uF elco being replaced with 470uF rubycon ZLG, 2.2uf in place of the original 47nf and 100nf PPS on the pins of that dac itself.

It's definitely sounding like a completely different machine now 









Thursday 24 January 2013

Reverse engineering the 5v rail!


Did some studying of the service manual today. Carried on with the investigation of the 5v rail. Rather than try to explain with the service manual I thought I'd draw a connectivity diagram. Its not complete but it does identify key components that would benefit from independent regulation. If you want the manual, you can download it HERE.

By identifying the components in the digital signal path, you can find out how they are powered and therefore which psu components can be changed and consider what you might change them to. By using a block diagram approach to the digital signal path through the player we list the devices in the chain prior to the analogue output.

HF -> Decoder and RAM -> oversampling filter and SPDIF encoder -> DAC 
At the DAC the analogue signal is created before going to the audio output stage which runs from the +&-15v rails

With respect to the DAC chip SAA7350, the datasheet for the chip confirms 5 separate supply pins. Digital, Analogue Timing Left, Analogue Timing Right, Analogue Left and Analogue Right. Looking at the service manual, each of these pins is connected to the common 5v rail via a low value feed resistor. Each supply also has a local 47uF electrolytic decoupling cap along with a surface mount 47nF ceramic under the board.

By doing the same, for each of the interesting chips, we can build a list of supply pins. Most other chips only have 1 supply pin.

Anyway, this is the DAC so you can see what I'm talking about in the manual.



By following the same process for each device we end up with this (which is not complete but 95% there)


The devices including the interesting capacitors are listed here 

SAA7350 - DAC
Pin
12 VDDD- Via 3611 feed resistor, Cap 2615 - 47uF + 2614 smt 47nF
18 VDDAR- Via 3614 feed resistor, Cap 2619 - 47uF + 2621 smt 47nF
27 VDDATR- Via 3626 feed resistor, Cap 2639 - 47uF + 2641 smt 47nF
29 VDDATL- Via 3625 feed resistor, Cap 2638 - 47uF + 2640 smt 47nF
39 VDDAL- Via 3613 feed resistor, Cap 2618 - 47uF + 2620 smt 47nF

SM5840 - Filter
14 VDD- Via 3646 feed resistor, Cap 2611 - 47uF + 2610 smt 47nF

SAA7310 - Decoder
28- Via 3602 feed resistor, Cap 2604 - 33uF + 2603 smt 22nF

MN4262-15 - RAM (for decoder)
9- Via 3601 feed resistor, Cap 2602 - 33uF + 2601 smt 22nF

We also need to do the same for the main CPU and the SPDIF driver chip (if we are keeping it). The remainder can potentially run from the original shared 5v reg.

Now we have identified each supply pin, we can look to regulate separately




****** do not just copy this last image as the reg in the picture will fail when used with a supply voltage greater than 10 volts!!!!

I'll explain more tomorrow

Wednesday 23 January 2013

Why would you??? CD52 mods!!! SOme Theory!


One of the great things is that these players can be got for next to nothing. I'm trying to make this thread about time rather than boutique products just to see where it ends up. hopefully it will inspire others to discover how good some of the older players actually are. Hopefully my explaining how the power supplies work and are distributed to the various components around the player, it will give others the tools to try the same.

You can draw analogies for what I am trying to do with the 5v rail.

Imagine if you had 8 fridges in your house all on the same circuit as your hifi. You'd constantly hear pops as the pumps start and stop. Now imagine if put all the fridges and each individual hifi component on their own circuit from the fuseboard or better still on their own service supply from the electriciy board via their own dedicated consumer units. Scale that down and to whats inside the CD player and we have a single 5v regulator feeding many critical components, each introducing its own noise to the common supply rail. Obviously the same applies to all the internal rail but for now, I'm concentrating just on the +5v......the rest will come later

Replacement loading gear has arrived!!!


I'm in shock!!!! Exactly 1 week after ordering the loading gear, it arrived from China!

Only took a few seconds to fit and now we have a fully functioning tray 
I cleaned and re greased the teeth and gear with some silicon grease.





I've also been working on the distribution of the +5v rail within the whole player. I stated previously that it would be good to try and seperate the "noise domain" by individually regulating each power supply pin rather than having everything share the same single regulator. In this player, each pin is separated by a feed resistor in order to address this issue (well to a very small degree).

I've now got the details together to seperate the 5 supply pins for the DAC, the digital O/S filter (if it stays) the Decoder, the RAM for the decoder, the SPDIF driver (if it stays) and the CPU. I'll use low noise regs for everything except the CPU which just needs putting in it own noise domain. I'll sort some pictures explaining what I mean a bit later.

Monday 21 January 2013

In Response to some questions relating to the effort vs reliability of players of this age


I've not had a CDM4 fail on me yet. This is the 7th or 8th player with that mech that I've had including the CD850MKII I had from new which I finally sold about 4 years ago. They do all suffer with the loading gear issue and also a failed cap on some of the earlier machines (a blue 33uF Philips axial jobbie) that causes read and tracking issues. The cap is a different make on this model but I will get to it later when I recap the rest of the player. 

With respect to the opamps, I've changed opamps on many players now and by the time the PSU including local decoupling caps have been changed and the filter treated to some better components, I've yet to hear an adverse reaction to these HA's that I've got fitted. The LME47920 is one of the best opamps for audio in my opinion but as I have stated, I've fitted gold rolled pin sockets to allow for opamp rolling and I've got a few options that I'll try when I've finished the rest. I do appreciate that if the filter is poorly designed then there is potential for issue. The HA's are expensive and may not ultimately fit the budget and depending on other work in the whole output stage my be a little OTT 

I'm trying to keep the thread as informative as possible just in case someone else fancies tackling something similar. The principles are the same for pretty much everything. Its all about the PSU, reducing noise and separating PSU rails to keep the noise from one part of the player effecting everything that shares that rail. As an example, there are several things on the +5v rail including the CPU, DAC analogue, DAC digital, digital filter and SPDIF driver to name but a few. Digital rails are especially noisy. Even fitting separate basic regulators (7805 @ 22p each) to each individual device supply pin in place of the original shared for everything will stop noise from one device being transferred around the original 5v rail. I describe this as separating the noise domain. In many players each device is fed via a supply resistor which can be removed making it easy to fit independent regs. There are some chip type regulators available now with very good noise figures which are very cheap. These are what I will use in this player. I have my own discrete low noise regs but I'm looking to complete as cheaply as possible and I'm very confident I can achieve a very high level of performance using good quality LDO low noise chip type regs.

I'll post the details of each section I change and show how I work out that cap does what job (analogue/Digital) and what I change it to along with my opinion on why. I'm now waiting for supplies before I move on but I'll show some of my working's out over the next few evenings :-)

PSU Diagram

Here's the PSU updates on the circuit diagram



Sunday 20 January 2013

Power Supply Mods

Made a few changes to the power supply.


First off I identified the audio analogue supply. There is a separate winding on the transformer dedicated to this which is pretty normal. The winding supplies the rectifiers which in turn supply +/-24v dc to the +/-15v regulators via smoothing capacitors. I've seen anything from 470uF right up to 9400uF per rail as standard for this job in CD players and DACs. This is a quick win normally governed by space. Here I've gone from standard 470uF up to 2200uF Panasonic FC (great pre reg where space is an issue) The rectifier is replaced with 11DQ09 schottky type. Post the regulators were some rather weedy 47uF caps. These have been replaced with 470uF Rubycon ZLH low impedance, low ESR caps which I often use on analogue rails.

Next to address is the +/-10v. The same principle can be applied here but due to the heavy load put onto the positive rail (its used to supply the 5v reg which supplies most semionductors in the player) I've gone from 4700uF to 6800uF. Again we could go bigger but space is a issue. On the negative rail from 1000uF to 3300uF. Schottky diodes replace the original bridge too. Post regs I've gone from 47uF on the negative rail and 220uF (Black Gate!!) on the positive to 470uF ZLG again. These mods bring the PSU's closer to that of a much more expensive player. The restrictive part of the PSU is now the transformer which we may address later depending on budget. We'll save the Black gate for later!


I've also identified all of the local PSU caps around the DAC, filter, receiver, ram etc for swap out later. AS usual, I'll go for solid polymer on any digital rails and standard construction low ESR (probably Rubycon ZLG) on analogue rails.

Next easy win is the opamps. As standard, the PSU caps are Elna Simlic so not bad. I've swapped out the standard opamps for gold rolled pin sockets so I can play with opamps. I've popped some LME49720HA for now just to see where we end up. I've also swapped the green Nichicon BP elco's blocking DC on the output for some MKP type. It may be that we don't actually need any at all but I'll measure the DC offset and check later.



I also have a few 11.2896Mhz oscillators so I paired one up with one of my regs and couple of nice caps a 470uF Nichicon solid polymer bypassed with a 1uF PPS on the reg and 0.1uF PPS on the oscillator. Should be a pretty good clock. 




Its starting to sound quite acceptable! The clock makes a big difference on this chipset and right now, its only clocking the DAC. I will direct clock the filter and decoder chips later which should improve things again.

I've also been looking at low noise LDO chip type regulators. I would like to separately regulate various components including isolating the DAC analogue and digital rails as I think this will bring vast improvements, but thats another job for another day. I'll leave this lot to settle for a few days

Friday 18 January 2013

Pre Modification investigation

OK Now to get the main board out ready for some pre mod investigation.


From the top, we need to disconnect the loading motor.





From the back, there are torx screw to undo by the output phono's in the middle and by the mains socket. From the bottom there are torx screw in the PCB and 2 larger black cross heads either side of the transformer to remove. Once these are out there are 2 tabs to press to allow to board to move out. Press the tabs to start the board moving but only move it enough to ensure its mechanically free. There is still a cable to the spin motor and ribbon the laser to disconnect.

This picture shows 1 of the tabs.




When the board is free, we need to locate the ribbon. You can access the ribbon from the side. The white collar on the plug needs to lift away from the the PCB to release the cable. Its a bit of a fiddle the 1st time you try to do it. Be very careful not to pull on ribbon until the collar releases or you could damage it. Its very fragile! Once release you need to access the spin motor connector from the front. Its got a tab on the PCB side that needs pressing to release the plug.




With both of these released, the board will come out of the player and leave you with this


I would now put all the mechanical bits back together for safe keeping.
You will notice a white mech under the transport. This will just fall out. Make sure you dont loose it because its part of the drawer proximity mechanism!

I've now got the main PCB out. The first thing I noticed is that the transformer runs warm. This suggests to me that its working quite hard. This is a sure sign that things can be improved with bigger transformers. I've yet to decide if I'll take this route due to the likely value of the player after some mods.




Now we can start to identify components and possible upgrades!


The Teardown - Replacing the loading gear (all CDM4/19 )


To get to the cog for replacement and the main board for mods,we need to completely dismantle the player. Same principle applies to most of the CDM4/19 based players by Marantz and Philips.

There are 5 screws in the top cover which we need to remove 1st. 4 cross heads on the side and 1 torx on the back in the middle. Undo all 5 and remove the lid lifting the back of the cover 1st up and backwards.

Next turn the player upside down. We need to remove the bottom cover.
There are 8 torx to remove to release the bottom plate.



Next we need to remove the on/off switch bar. This is done by disconnecting the bar from the actual switch then lifting up and back to remove it from the front panel.











Next we need to remove the front cover from the CD tray. From the top of the player, you can just push the tray out a bit. You need access to the underside of the tray to remove the front of it. There is a tab that needs to be pushed out very gently to allow the front to slide across and off. Should be reasonably obvious from the pictures. The tab is at the top of the 1st 2 pictures.








From the top again, you can now unplug the 2 white plugs from the display and headphones

Now you need to remove the 3 torx screws from the top of the main front panel assembly. Ensure the CD tray is fully retracted inside the player. Release the tabs on each side of the player and release the front forward from the top. once free, it should just lift out from the bottom.





Now you need to remove the CD tray. Lift the clamp mech to the vertical position and just withdraw the tray from the player. You may need to move the clamp up and down in order for the tray to clear the transport and come tight out. 





Now you can see the broken cog which needs to be removed. By applying pressure to the mounting on the transport side, you can move it enough to free the cog......easy!!!







Putting it back is just the reverse! I don't suppose my replacement will be here for several weeks so i'll just keep the screws safe and proceed to remove the main board for mods........TBC

CD52 MKII Mods part 1


This is one that I've not had a play with at all yet. I've had several goes at the CD50 which uses the TDA1541 dac, CD850 and I owned a CD850MKII from new for about 18 years all of which use the later swing arm type laser assembly from philips, the CDM4.

I thought i'd start this thread to show the mods progress and explain some of the rationale behind some of the changes that will be made. I'll lay my cards on the table now, I've no intention of keeping the player at the end and I will offer it up for sale (in the appropriate forum section) for a sensible price that reflects the cost of the parts plus a few beers. I'm not looking to spend hundreds but I do expect to make significant improvements within a reasonable budget.

Hopefully the thread will be informative and therefore my approach acceptable to the community. As previously stated in other threads, I do this because I can and I really enjoy it! If I can fund my obsession by selling at the end of the project then happy days 

There is plenty of information relating to these players on the internet. The player is Pre the KI Signature badge. That doesn't mean "he" wasn't involved! Indeed all of the SE range are his work. From experience I can tell you that in general, the SE versions are mainly about a few capacitor changes and bracing the actual cases. I have read that in some SE models the transformer is constructed using higher quality materials but I cannot confirm this. Certainly in KI models, the mods are SE plus a toroid transformer most likely of a higher VA plus copper chassis plating and a shiny badge!

Since I'm likely to change and improve most of what the SE would offer, I've started with a basic CD52MKII as the only reason I can see for using an SE would be for perception.

I've got hold of a reasonably priced player that works other than the drawer loading cog which is a common problem on all CDM4/19 transports. A replacement made from a more modern plastic is readily available for £6 on ebay. This is one time I'm glad our friends from the other side of the world are happy to 'clone' everything! The original part went out of production many years ago! For a test listen, the drawer can be manually operated. 

I'm not going to washy word audio reviews at every stage of this project but I will give a very brief overview at the start and finish of the project. To start I need to verify that everything work as it should and that there are no actual faults with the player (other than the loading cog). I have a couple of test CD's I know very will with a mix of tracks on them so having listened to them to verify functionality I can also comment generally on the sound. What I will say at this point is that the sound is quote flat compared to what I'm used to. Its quite a warm sound but lacks sparkle and accuracy. Something I'm sure I can improve to a reasonable degree. Don't get me wrong, I'm sure the standard sound would suit many, its just not good enough for me out of the box and I know it can do better.

Here's the player in the flesh during test this morning. I've already removed the offending loading cog and ordered a replacement which will come with a new belt too.




Here's the offending cog. As you can see, its stripped most of its teeth. The theory offered on the internet is that the grease philips used has reacted with the plastic and made it go soft. I can tell you it feels more like hard wax now than plastic although I'm not convinced by the theory about the grease! Having owned an 850MKII from new, I can tell you that you push the drawer to close. This is bound sure to stress the gear as its work ratio means that the motor spins hundreds of times to move the drawer from end to end. just pushing the drawer 1 cm is likely to require the motor to spin 3 or 4 times. This will put far too much pressure on the fragile teeth of the old cog. Armed with this information, even with the new cog, I'd recommend using the play button to close the drawer.


Next up will be some investigative work in the core power supply to see what can be done to improve things immediately before and after the voltage regulators............