Denon DVD-2910 mods

[Marmotus]

Dear all,
I made some mods to 3 2910s, only Stereo R/L channels (the most ellaborated in the analog channels) and ended with 3 serious competitos that having different "flavors" beated top models (SACD players, CD players, so on) from NAD, ARCAM, LEXICON ($3.500!!!!!!!), not said by me but from the actual owners.
Very simple, OPs changed to LM4562 I/V (way more warmth and detail at the same time!), output buffers changed to LM4562, LM49723 and OPA1642 in each player, muting transistors taken away, Output Condensers taken away, and shunted one of two output resistors, leaving 150 ohm in series.
All players in general are seriously impaired in the analog output stage (low cost parts), I know they can't leave a product without an output condenser, for instance, but as single mod it is to shunt it, im not aware of a bigger and cheapest mod.
One of the kings I sold to a customer wich has Lexicons, Arcams, Proceeds, beeing his 1st player, forcing Him to play more CDs than other formats is a 1989 SL-P999 Technics... with the same mods described, LM49723 output OPs.... that´s it.
For one of my Denons it started the common issue of non reading CDs, I'm going to ask for a couple of HOP1200R/S pickups, but they don`t will be here till next 2 weeks, so I will ask I someone could tell me wich of the 2 presets at the base of the pickup is for the CD laser??? to see if I can do something in the meantime... (THANKS in advance for this little help!)
As for the Flavors of the output buffers...
LM4562 - Very lineal, everything is at the same level, strong dynamics (a little on the more than enough side), but in this case sound stage is short of right. Voicing is extremely precise with very high definition.
OPA1642 - More Hifish presentation, more bass and highs than mids, very high definition sensation could be highs in the 10k over to be some dB up.
Laid back vocals. And high contrast dynamics.
LM49723 - Somewhere in the middle, but above, other dimension...
you are there Sound Stage, natural crispness of transients, natural voices weight, bass is deeper and makes all blend toguether without loose definition and being overweighted... I can talk of a similitude to my analog old Pioneer RT-2022 open reel here...
Cheers
Marcelo

[SoNic_real_one]

I did the same mods on my 2930 and 2910... I liked the best the sound of LM4562. Especially on the 2930 that has AL24 Processing.
But today I just got in the mail some LT6201 and I can't wait to put them in.

[SoNic_real_one]

OK. I fryed a couple of LT6201...
Didn't realized that Denon put under the OpAmps ground with a dot exposed (to connect to an eventual heat sink?). Anyway, the LT6201 in DD package have the underside heatsink connected internally to V-... Oh, well, those two where all that I had.
So I put there (both of the L/R outputs) some AD8599 that I wanted to test for a while.
Not bad, not bad at all... I don't know which I like better now: LM4562 or AD8599. Hard to tell in such a short time.

[Marcelinho]

I would like to share some other mods:
-Disconnect the scart connector and the audio to it (two resistors)(euro-version).
-place a high quality 5v regulator on the audio board (for the dacs) the pcb is already prepared for it, but is left open.
-I removed the output cap's as my amplifier already has good input cap's.
-I build myself a 27Mhz clock from a tentlab X0 and a 3,3V tentlab regulator.
-clock the mpeg2 chip directly (bypass the pll) A bit speculative but i haven't seen or heard any problems. As this chip takes care of the spdif and audio over hdmi it seems logical.
-clock the faroudja chip with clock/2 (if video is your thing). I didnt dare yet.
-use output two on chip cy2302 instead of one. It has less jitter (90ps over 200ps) and is part of the clockline to dac and dsp. Be carefull it's all smd stuff!

[Marmotus]

Thanks for sharing! SoNic_real_one & Marcelinho (like me )
I had my eyes on those AD8599, too. Let them brake-in and after tell us what happened!!
I'm impressed with your mods Marcelinho, sad for me is I have very little time and no place to do my "experiments", I have to start and end in a few hours, leaving all clear and clean!!!! WAF / SAF for our hobbies is very low...
I share this with speaker building, flying from end to end of the audio chain, I have just finished an EP2500 Behringer Amp, changed OPs (4580) for the LME49723, changed Fan for a Papst one and reversed it, put bypassing polypropylene caps in Power plant (also carry audio signal) and in imput caps.
I've been the last two hours smashing my head with: Phutureprimitive, Emerson Lake & Palmer and Engelbert Humperding to be eclectic!!!
Some of my stuff, below
Cheers
Marcelo

[SoNic_real_one]

I have looked into adding that "missing" regulator on the board, but I didn't think it would do any good - I just added another filtering ELNA capacitor (different color in the picture). Adding the extra reg requires to remove four 0 ohm resitors in order to work properly, adds extra load on the +9V rail for the OpAmps... I think that the ferrite beads are enough "separation" between analog and digital +5V (it is like that also on the TI dataseet).
Next to the OpAmps are some places for extra capacitors - I am thinking to put there something

As for the clock - I don't think that clock is clocking the DAC's. The DSP has it's own clock (18.75MHz) and that will generate from it the signal for the DAC's inputs. The buffer memory inside of the DSP will take care of the incomming jitter at the DSP inputs.

[Marcelinho]

The japanese version has the missing regulator, The point i was making that you can provide the dac's with cleaner power if you place a good quality reg there. I didn't notice the resistors though. Why should they be removed?

With 88,2 an 96Khz sources (sacd and dvd) the jitter from the pll is 450ps. If you can reduce that before it enters the dsp..... If you can hear the difference i don't know. You can do the mod for free, its a smd hassle though.

Maybe Marmotus can compare a modified with an unmodified:-)

[SoNic_real_one]

In the third pic that I did post, you can see the four zero ohm resitors R641-644, between the +5V_D1 (incoming from the regulator on the power supply board) and the +5V_A (going to feed the DAC). If you just install the IC317 in the OPEN position, you just "paralleled" the reg in the PS with the new one. Basically, the one that has a higher output voltage will work, the other will be blocked.
In the forth pic I am showing the place where the +5V_A enters the DAC - as you can see there is a separation with a ferite coil between the analog and digital supply.
The original +5V_A is as clean as the one you install - it is an analog reg also. I decided just to add an ELNA 50uF/50V capacitor in C480 placeholder.
The one that you installed draws the power from the regulated +9V line that feeds the OpAmps - I really don't like that apporach and maybe Denon engineers changed their minds too about it after they manufactured the Japanese board (I have the US version and is missing it too as you can see in the fist pic).

As for the clock, I would rather invest the money in the 18.75MHz clock of the DSP. Yes, the main PLL clock affects the SACD signals, but the DSD is less sensitive to jitter compared with PCM signal - it misses the precise sample and hold requirement to be decoded in the delta sigma DAC, data is already in delta-sigma format.

PS: The AD8599 sounds goood... in all the details, highs, bass, voice. That is a very sensible improvement compared with original OpAmps (NJM2068). I cannot compare A-B with the LM4562 since those are in a DVD-2930 that has supplemental DSP processing (AL24 Processing Plus). I would say there are comparable, very close. And also in my AVR-3805 receiver with the same AL24 processing Plus.

[Marmotus]

I will put that cap and make some comparison, SoNic_real_one.
For regulator, to have in mind, wich one you suggest, Marcelinho?
I made comparisons in every mod step to the stock unit... I have to say that the stock was very good to what you find over units without price considering. Probably someone invest hours at Denon listening to the balance of the presentation, very good soundstage, good voicing and dynamics all considering parts used. The flaws are that highs and lows suffered typical cap distorsion, the former a looping the loop ripple, little one comparing to others but inducing fatigue, the latter "molded" to fit but overhanging.
Please tell me if you know wich preset is for the CD laser at the pickup, I have a suspiction as I found D and C letters but need to confirm, before I turn it like crazy!

[SoNic_real_one]

As you can see on the board pic above and schematics, the caps for the stereo mix output are better than the ones in the multichannel output. Better (and bigger) electrolytic (RSV instead of ASF), paralel with a Polypropylene Film Capacitor (NH2).