Denon DCD-2700 CD Player Mod Thread

[Bratwurst7s]

There are a total of 16 carbon film resistors directly in the signal path. 18 if you count the headphone section. I replaced all of them except for the headphone pair with Vishay-Dale 1/4w 1% metal film resistors.

R-117, R-234, R-369~372: 1k ohm,
R-373~376: 3.6k ohm,
R-381, 382: 6.8k ohm,
R-385, 386: 1.3k ohm,
R335, 336: 150 ohm.



This isn't all of them but shows a sample.

Cheers,
James

 

[Bratwurst7s]

Each channel has 2x PCM-1702 DAC's, each of which feeds a PC4570 op-amp. After each of these 4570's there is a 3rd 4570 that sums the output of the first 2 4570's. The signal then goes past a single BA15218 op-amp and then feeds a single PC4570 that goes to the output. The BA15218 op-amps,if if I'm looking at the schematic correctly (always in question), sit in parallel to the signal path but isn't directly in it. Because I didn't know the purpose of the 15218 I didn't touch it at first.

So the first round of op-amp replacement I just addressed the 8x 4570's. They were removed and sockets installed in their positions. Here is a photo that is actually from the 2nd player that I bought and shows all 10 of the op-amps removed and socketed. But at this point in the first player you would only see 8 sockets.



Cheers,
James

Edit: corrected info about the BA15218's+.
edit #2: Corrected info incorrect. See post #38.

 

[Bratwurst7s]

To replace all of the PC4570 op-amps I choose LME49720HA/NOPB in the TO-99 metal can. I also ordered some little heat sinks for them from Mouser. After that I closed up the player and used it for a couple of weeks while I decided what to do with those BA15218.

Compared to the "before" config there was a serious improvement in sound, mainly in the highs and mids. Much more clarity than before, but still not what I'd like. Voices (like in singing) seemed a bit shadowed. There just wasn't that "in your face" clarity that I love. In the background I found out that apparently the BA15218 op-amps are part of a "GIC Filter" circuit. I see it variously being described as a "low-pass" and as a "band-pass" filter. In the end I just pulled them and installed "plain Jane" LME49720NA/NOPB op-amps.

Replacing those filter op-amps realized another very nice improvement in sound. The veil was lifted so to speak and sound clarity really jumped out. The player was now sounding very good, really a night and day improvement over the untouched "new" configuration.



edit: I forgot to mention. Being over here in Europe it's a bit of a PITA to order those nice Brown Dog adaptors to mount the TO-99 package op-amps. So as a field expedient measure I just carefully bent the legs of the op-amps, shortened them, and mounted/soldered them in a Dip-8 socket. That works just fine and also saves a bit of money.

Cheers,
James

 

[Bratwurst7s]

After having done all of that I got into a discussion with HypnoToad about adding a 0.1µF film cap across the supply lines of the op-amps to stabilize them and damp oscillations. I tried that first on a DCD-1290 and liked the results. There was a measurable drop in temp in the op-amps and another small improvement in overall clarity, including a little tighter bass. So I did the same to the 2700 units.

Actually I figured that if adding the film caps to the op-amps in tha analog section, would it be of use to add them to the op-amps in the digital section that are acting as motor controllers? It didn't seem like that it could hurt and so I added the 0.1µF/63v Wima MKS2 caps to all of the op-amps in the entire player.



I like this. Another incremental sound improvement. And when I was working on the 2nd player and paying a bit more attention I could see that the pcb was discolored from heat under all of the op-amps in the analog section. So I think that adding sockets for the op-amps might act as an additional bit of heat sinking to protect the pcb from heat damage, as well as the stabilization film caps.

Cheers,
James

 

[Bratwurst7s]

So after having done all of that, and using the player for a couple of weeks I finally decided on a filter cap to use in the power supply. Elna LAO Tonerex 4700µF/50v, available from Digi-Key at about €4.50 each. And here is why I'm glad as to why I did this last. Because changing those caps as a single step after having done everything else let me hear specificly what the change made. And the change surprised me a lot.

There was an very nice improvement in bass/mid-bass, both in amount and clarity. A much nicer, tighter bass, and imo an overall improvement in sound clarity. I really hadn't expected this just from replacing those filter caps with larger value ones. But the improvement is very nice! And there is no question that that one change is responsible. Worth every penny of the cost of those caps.

So this is where I'm at at now. The player is sounding very very good now. The music has a lot of depth, imaging is great, overall clarity is great, with the highs and mids being fantasticly deatialed and bass is very tight.

So where do I go from here? I don't have unlimited funds and have a number of other projects running in the background, so have to go forward in small steps. Somewhere during the journey through the DCD-1290's I stumbled across "Ray's Audio Page". He has a number of interesting projects there, and they are nicely oriented towards the DIYer and are very affordable.

http://www.raylectronics.nl/index_en.html

So I'm thinking that the next mod will be to try out the LM317/337 voltage regulator project. It looks like something that's inside my skill level and won't cost too much.
http://www.raylectronics.nl/frame_it_en.html?http://www.raylectronics.nl/lm317regs_en.html

After trying that I'll probably do one of the Pink Fish Media "Flea" low jitter clock kits. Again, entirely doable within my budget and skill level, and I can buy the parts incrementally. I've already bought a few of the little power supplies from Conrad that are to be used with the clock kits, although I still need to buy the components to mod them.
http://www.raylectronics.nl/frame_it_en.html?http://www.raylectronics.nl/flea_en.html

I've also read about people replacing some of the ceramic caps in the PSU & digital sections with styroflex caps. I'll have to look into this.

I'm also saving funds to buy an oscilloscope and function generator with the intention to eventually learn how to calibrate focus and tracking in the players.

I'll post more as/when it happens.



Also, when I get a little time I'll post a list of Mouser & Digi-Key P/N's & links for all of the parts that I used, as a service to anyone getting ready to work on one of these fine cd players.

Cheers,
James

edit: I forgot to mention that to install the Tonerex filter caps I had to open up the mounting holes slightly. A 1.6mm (1/16") drill bit works just right.

 

[monkas]

Thank you very much for your sharing, Today I bought DCD 2700 and found this topic, I will follow your path. I'm waiting for more

Greetings from Poland.
Slawek

 

[Bratwurst7s]

Hi Slawek. I'm happy to be of service. And Welcome to Audiokarma.

Cheers,
James

edit: It will probably be a month or 2 before I try anything new with the player(s), but I'll post the results when I do. But I do plan to make a parts list in the next week or 2.

edit 2: If you do any of the same things that I did please post your impressions. It always helps to have more opinions as to the value of any changes.

 

[unfairlane]

First, upgrade the clock-unit. Then; there are four op-amps after the four 1702-dacs, you can take the signal (balanced if from all four) directly from (leg 7?) the op-amps and skip the following "filter". Then; you can upgrade the trafo. Think big.

You can do a lot of changes to this player but this here is what really makes a differense. A big differense, even if the DCD-200 is not as capable as the DCD-3560

 

[monkas]

Thank you James, now I noticed/realized interesting connection in the name of Audiokarma with myself. I love music, I love audio equipment, I dream about improving/renovating it( I need to learn a lot). This implies I need this kind of "karma" audio-electronic karma so It's good I'm here, I feel it will be helpful..

Ok, I'll do it. I have also DCD 920, maybe I'll modify it also and compare, time will show.

Cheers,
Slawek

 

[monkas]

First, upgrade the clock-unit. Then; there are four op-amps after the four 1702-dacs, you can take the signal (balanced if from all four) directly from (leg 7?) the op-amps and skip the following "filter". Then; you can upgrade the trafo. Think big.

Hi, can you describe how to do this, how it improves sound quality, what parts did you use ?

Thanks
Slawek

 

[Bratwurst7s]

You can do a lot of changes to this player but this here is what really makes a differense. A big differense, even if the DCD-200 is not as capable as the DCD-3560

OK. I'm listening. Give us some details please.

I'd love to have a DCD-3560 to try out but afaik it has a laser that is unobtainable.

Cheers,
James

 

[unfairlane]

While all multibit cd-players can be easily modifyed, it is the bigger four-dac players that are worth modifying, specially if you can run them into a balanced preamp.
Not so many parts needed but a better clock-unit is Worth the Money, a must. I used this one from a Danish digita-doctor:

http://lcaudio.com/index.php?page=4 (he gives all neccesary guidance on his site, just read)

About clocking: http://www.tnt-audio.com/clinica/diginterf2_e.html

it is important to chose a clock with the correct clock-frequense:

http://lcaudio.com/index.php?page=39

The easiest way to improve any multibit player is to simply bypass the analoge-filters. That means taking out the signal directly from the op-amps.
Still remember how this blew me off my feets when I first tryed it on my old DCD-3560. One modified and one unmodified output at the rear recommended, to compare.


Another way to make the players grow is to upgrade the trafo. You need a bigger trafo, minimum 100va, can be mounted at the rear panel if there is no room inside.
Check out internal secondary voltage, my Denon runs at +/-15V. A 200va toroid is recommended. Remember to Upgrade the fuse too.

I could continue but these are the mods that does most, or should I mention changing the powercord to a solid-core too. Another jaw-drop.

 

[monkas]

http://lcaudio.com/index.php?page=4 (he gives all neccesary guidance on his site, just read)

Very interesting but not cheap, more than half of my CD player price

The easiest way to improve any multibit player is to simply bypass the analoge-filters. That means taking out the signal directly from the op-amps.
Still remember how this blew me off my feets when I first tryed it on my old DCD-3560. One modified and one unmodified output at the rear recommended, to compare

You mean taking out the signal from the op-amps directly to output connectors ? btw, if so why they use it (analog filters) ?

please continue, I'm hungry for knowledge

thank you
Slawek

 

[unfairlane]

Yep Monkas, diectly from the opamps as long as it is a multibit player like the Denons. And do not underestimate the effect of putting in a real trafo. These two + mybe a solid powercord would lift even your cheaper player a few steps, in clearity as well as dynamics

 

[Bratwurst7s]

OK. This is interesting.

I've read in other places that those filter circuits are not doing the sound a lot of good. I'm still learning to understand schematics but it looks to me that the circuit goes like this: Each Dac feeds an op-amp, 2x dacs & 2x op-amps per channel. After each of those op-amps there is a BA15218 op-amp/filter but it doesn't appear to be directly in the signal path and I don't understand yet how it is filtering the signal. After the filters each pair of signals in each channel are fed into a single op-amp per channel. The output of those last op-amps then goes through the coupling caps directly to the output jacks.

Since I socketed those filter op-amps I'm wondering if I could just pull them and jumper the path the signal would be taking. I could simply solder a jumper wire into a dip socket and plug it in and do a listening test. The question would then be, Which traces do I jumper. I'll have to sit down with the schematic and try to figure this out.

As far as replacing the clock with a low jitter clock that is definitely in my plans but I can't afford to do it for a couple of months at least. I have to work in small steps. A set of upgraded voltage regulators is actually the next thing that I plan to do. But I think that working in small steps is best for me as it lets me listen to the player after each mod and evaluate the results.

Upgrading the trafo sounds interesting. But one thing is that the original trafo has separate windings to supply the digital and analog sections. What I've been thinking about is finding out what I would have to do to add a second torroidal transformer to supply just the analog section. I would think that this would also lessen the load on the existing trafo and let it supply the digital section better. And since the supply to the analog section goes through separate wires/plugs it would be relatively easy to unplug those wires and run the new supply wires there.

BTW, In post #22 I made a mistake in describing the number and position of the filter op-amps. Now corrected.

Cheers,
James

edit: No. The above is wrong. Looking at the schematic again, each dac feeds a 4570. the pairs of dac/4570's then feed a single 4570 that sums their signal for that channel. The path goes past (but not directly through) a single BA15218 and then feeds a single final 4570 per channel, and then goes to the coupling caps & output jacks.

So perhaps I will try pulling the (replacement for) the BA15218's and replace the final 4570 with a socket that jumpers pins 1&3 and leaves the rest empty. This would effectively bypass both of those last 2 op-amps and will be very easy to do. Perhaps I'll try this on the weekend.

edit #2: Darn it, I got the layout wrong again. See post #38.

 

[monkas]

Thank you Guys, and sorry for my english errors.

 

[Bratwurst7s]

Thank you Guys, and sorry for my english errors.

Your English is good enough. I have no problem understanding you.

Cheers,
James

 

[Bratwurst7s]

OK. So I've spent the last couple of days like a dog chasing his tail over these op-amps and have now realized why. I have been talking about what the schematic shows and had forgotten that what is installed in the players was different than the schematic. There were actually 2 of the BA15218's installed in each channel. So, one more time hopefully correctly this time.

Using the left channel only as an example.
IC301 & 303 are the 1702 dac's. Each of them feeds op-amps IC315 & 317 respectively. The schematic shows PC4570's installed there and that is what was installed in the players. Those 2 4570's feed into IC311, where their signals are summed. The schematic shows PC4570's installed there but the player actually had BA15218's.

The signal path then goes past IC309, which is shown in the schematic as a BA15218 and that is what was installed in the player. IC309 sits in parallel to the signal path, which then goes into IC313, shown as a PC4570 in the schematic and that is what was installed in the players. From there the signal goes directly to the coupling caps and then to the output jack.

I looked over the schematic today with a couple of people who actually know what they are doing and the consensus was that there is no electrical reason that IC309 can't simply be pulled from the board. Whatever it is doing it taps from and feeds to the signal path in parallel to it. Unless the schematic is wrong (again). Looking at the flowchart/block diagram it doesn't show both the summing op-amp and the filter op-amp as separate objects but rather as "Active L.P.F". Which I'm guessing means "Low Pass Filter".

Then comes the question, just why is there another op-amp after this filter stage? Is the filter stage lowering the signal strength so that it needs to be boosted again? I'm going to find out. I just prepared a dip socket with a jumper between pins 1 & 3, which the schematic shows as the signal path. Tomorrow or Friday I'm going to first pull IC309 and listen to a CD. Then I'm going to replace IC313 with the blank/jumpered socket and listen again. Should be interesting.

Cheers,
James

 

[monkas]

Then comes the question, just why is there another op-amp after this filter stage? Is the filter stage lowering the signal strength so that it needs to be boosted again? I'm going to find out. I just prepared a dip socket with a jumper between pins 1 & 3, which the schematic shows as the signal path. Tomorrow or Friday I'm going to first pull IC309 and listen to a CD. Then I'm going to replace IC313 with the blank/jumpered socket and listen again. Should be interesting.

Yes, very interesting, maybe you'll discover the reason behind the use of this filter.

Best
Slawek

 

[Bratwurst7s]

Yes, very interesting, maybe you'll discover the reason behind the use of this filter.

Best
Slawek

From what I've been able to find it's called a "GIC" (Generalized Impedance Filter) and it was something that Burr Brown was pushing at the time that the PCM1702 dac's came out. It was only used for a short time and dropped by the time the next product line upgrade came out, so for perhaps 2 years. I'm guessing that it wasn't successful or it was decided that it wasn't needed. In any case I'll be finding out what removing it does in a couple of hours.

Here are a few links to look over...
http://calhoun.nps.edu/bitstream/handle/10945/2024/05Sep_Cheong.pdf?sequence=1
http://www.ti.com/lit/an/sbaa001/sbaa001.pdf
http://www.diyaudio.com/forums/digital-source/2787-gic-low-pass-filter.html

Cheers,
James