Stock vs. modded

Created 06.01.2008

As I have made numerous mods to my DCX2496 unit over the years. It is of course desirable to be able to objectively evaluate if the performance has increased.

DIY audiophiles are known to believe to hear subtle differences in any small change - some do agree later that some mods did not work as well as first impression and original hope suggested, some seem to do only mods that they praise as positive. May be they do, but it is somewhat unlikely. From my own experience - I have done and undone many mods and also modified the modifications if needed. Ok enough of blabbering.

So lets get started - this is a picture of a stock unit and my modified unit side by side
(stock unit is about 2.5 years newer production - notice the power supply transformer etc..)
orig-vs-modded_1.png

Well I will try to summarize what are the mods that I have in my current DCX2496 versus the stock unit:

* 75 ohm BNC input socket for coax cable continued with thin 75 ohm coax cable upto DSP PCB where it is terminated with 75ohm resistor and Schott transformer that feeds the digital receiver IC. The original 110 ohm termination resistor is removed from PCB.

* Electrolytic caps in switching power supply changed for Panasonic FC series caps

* Electrolytic bypassing the DAC chips changed with Oscon dry electrolytic caps.

* DAC 7805 regulator IC changed with Frank Oettle’s regulator placed at the DAC bypass cap pins.

* CS8420 changed with Frank Oettle’s digital input / AD resampling module.

* Input C ADC chip removed as considered unnecessary.

* All digital IC’s are shielded with Copper shielding tape.

* Outputs use Unbalanced Neutrik RCA sockets.

* Original analog in/out module removed and replaced with DC removal caps and RC filters forming a passive output stage.

* In real life this DCX setup is followed with my own version of 6 channel APLS volume pot + single stage of opamp per channel. Opamp stage has additional two times passive RC filter built around it.

6ch-pre.png

…time to get to measurements….

First a picture of measurement setup for reference.

meas-setup.png

!!!!!
On all graphs the
yellow line = unmodified DCX2496
green line = Ergo’s modified version of DCX2496
the 0dBFS output level is always about 3dB higher than modded unit as the passive out mod can not play as loud.
!!!!!

1. Noisefloor with 44.1kHz digital silence track played to DCX inputs
noisefloor-yelloworig-greenergos.png

2. THD with 997Hz sine wave - 44.1kHz -3dBFs peak test tone played to DCX inputs
1khz-thd-yelloworig-greenergos.png

3. IMD with 9kHz and 10kHz sine waves 1:1 - 44.1kHz -3dBFs peak test tone played to DCX inputs
9k-10k-imd-yelloworig-greenergos.png

4. IMD with 19kHz and 20kHz sine waves 1:1 - 44.1kHz -3dBFs peak test tone played to DCX inputs
19k-20k-imd-yelloworig-greenergos.png

5. Multitone sine waves - 48kHz test tone played to DCX inputs
multitone-yelloworig-greenergos.png

6. ARTA jitter test signal - 44.1kHz test tone played to DCX inputs
jitter-arta-yelloworig-greenergos.png

7. ARTA jitter test signal - 48kHz test tone played to DCX inputs
jitter-arta-48k-samplerate-yelloworig-greenergos.png

Conclusion:

Last time I measured performance pre and after installing Frank’s mods to my DCX - there were subtle differences but overall the measurements did not indicate much change.

BUT - one has to remember that in that case I remodified my already quite extensively modified version. So most of the performance I already had and just managed to take it a small step further.

NOW in this case where the modified unit is compared against the original I think there is no question at all that there are very clear indications of performance increase visible in all the tests done above.

I would perhaps point out the following as the most important ones:

* Noise floor drop of as much as 5dB at high frequencies. (Low end is limited by something other than PS or analog stage.)

* IMD graphs show much lower “hash” around the two test sines
imd-detail.png

* Jitter measurement shows much improved performance with both 44.1k and 48k sampling rates.
jitter-detail.png

I think this should make it very clear that the best mods out there for DCX2496 do actually help to get a better performance out of this relatively low cost unit. I do not know by hart how much all this has cost me - especially as many expensive items I have bought for my DCX have in the end moved back out and lie on shelf, but still the sound I have today I very good and the flexibility it has for DIY crossover development is such a big plus in itself.

So I hope you enjoyed reading this “good news” as much as I did creating them ;)