I got the following kits from Selectronic:
I/O board (with Blackgate Capacitors)
Linear Power Supply
Low Jitter Clock 24MHz
The assembly of the clock and power supply boards was straight forward.
The assembly of the I/O board was somewhat more demanding, as there is no silk screen on the board to aid component placement. The information supplied is sufficient, but know what you are doing before diving in. if you use the BG caps, you don't need the 2-pin jumpers next to each OPA134 op amp.
Once the boards are assembled, the entire DCX chassis needs to be pulled to pieces, and the holes for the switchmode supply extended to allow fitting of the new linear psu's dedicated transformer.
The mains power wiring removed from the switch and power socket, and re-wired as per instructions provided. Later, the mains socket ends up being upside down (earth pin at the top of the case), as Behringer's original switches the neutral, not the active wire. This is changed with the modification.
once you have the CPU board out, remove the old 7805 regulators and attach wires to go to the linear supply board (not fitted yet)
Not included in the clock instructions:
remove the resistor R33 or R34 (depending on the revision of the board) and remove and short the capacitor - you can remove the crystal if you want to. This will allow the new clock to work without the original one interfering.
At this point, I was forced to digress from other instructions (www.dcx2496.fr) as the solder pads are not firmly attached to the board. I connected the clock to pins 7 (GND) and 21 (CLK24) of IC1. this worked fine in the end....
The power supply transformer has no outer markings of which is primary and secondary, so measure and mark it on the tranny or you will blow the fuse (and maybe more) on first power-up. (hahaha)
ok, so we screw the I/O board in, and then mount the new tranny with double-sided tape, and then re-fit the switchmode supply, socket and switch.
check all component values and orientation
test the 8V output from the linear supply.
check everything again (especially the I/O board) as it's easy to get things in the wrong place... I originally had 2 op-amps in each others' place.. would have worked, but they were wrong.
free up the wires from the switchmode and cut/splice and connect as instructed to the I/O board and the new linear supply board to the processor board connector
carefully reassemble everything.
All my original settings were still in there.
when the clock isn't connected properly, you get 'INPUT A' and 'OUTPUT 1' reading full-scale and the display is lit up, but blank segments (lights on, nobody home).
The supplied fitting instructions for the PSU and I/O board have lots of pictures - each of which tells many more words than what is written.
The clock instructions were not specific to the DCX.
Hope this helps.
As for audio quality - very nice, the mechanical sound in the highs is gone, the black is definitely blacker, and the clarity and transparency improved manifold. dynamics are improved considerably, and the sound is so much smoother and liquid. The stage is wider and the instruments have moved back into the room where they were recorded.
I like it.
I/O board (with Blackgate Capacitors)
Linear Power Supply
Low Jitter Clock 24MHz
The assembly of the clock and power supply boards was straight forward.
The assembly of the I/O board was somewhat more demanding, as there is no silk screen on the board to aid component placement. The information supplied is sufficient, but know what you are doing before diving in. if you use the BG caps, you don't need the 2-pin jumpers next to each OPA134 op amp.
Once the boards are assembled, the entire DCX chassis needs to be pulled to pieces, and the holes for the switchmode supply extended to allow fitting of the new linear psu's dedicated transformer.
The mains power wiring removed from the switch and power socket, and re-wired as per instructions provided. Later, the mains socket ends up being upside down (earth pin at the top of the case), as Behringer's original switches the neutral, not the active wire. This is changed with the modification.
once you have the CPU board out, remove the old 7805 regulators and attach wires to go to the linear supply board (not fitted yet)
Not included in the clock instructions:
remove the resistor R33 or R34 (depending on the revision of the board) and remove and short the capacitor - you can remove the crystal if you want to. This will allow the new clock to work without the original one interfering.
At this point, I was forced to digress from other instructions (www.dcx2496.fr) as the solder pads are not firmly attached to the board. I connected the clock to pins 7 (GND) and 21 (CLK24) of IC1. this worked fine in the end....
The power supply transformer has no outer markings of which is primary and secondary, so measure and mark it on the tranny or you will blow the fuse (and maybe more) on first power-up. (hahaha)
ok, so we screw the I/O board in, and then mount the new tranny with double-sided tape, and then re-fit the switchmode supply, socket and switch.
check all component values and orientation
test the 8V output from the linear supply.
check everything again (especially the I/O board) as it's easy to get things in the wrong place... I originally had 2 op-amps in each others' place.. would have worked, but they were wrong.
free up the wires from the switchmode and cut/splice and connect as instructed to the I/O board and the new linear supply board to the processor board connector
carefully reassemble everything.
All my original settings were still in there.
when the clock isn't connected properly, you get 'INPUT A' and 'OUTPUT 1' reading full-scale and the display is lit up, but blank segments (lights on, nobody home).
The supplied fitting instructions for the PSU and I/O board have lots of pictures - each of which tells many more words than what is written.
The clock instructions were not specific to the DCX.
Hope this helps.
As for audio quality - very nice, the mechanical sound in the highs is gone, the black is definitely blacker, and the clarity and transparency improved manifold. dynamics are improved considerably, and the sound is so much smoother and liquid. The stage is wider and the instruments have moved back into the room where they were recorded.
I like it.
I used two cables to begin with (from the local music shop - "proper" AES/EBU cables, and removed the female from one end, and fitted that cable into the female of the other cable for the DEQ end and one cable end to each DCX.
edit - I did get some "Processor" type noise from the two unmodified DCX's analog outs when they were both running, and switching either of them off by the mains switch saw the problem go away (but now only running one DCX. I have not tried it since I modified one of the DCX's. Also, this was in-car running from a sinewave inverter (stereo 6-way was fun to 'tune up'). I never tried it on house power.
once I have finished testing this unit, and modified the other, I may put them back in the car.
edit - I did get some "Processor" type noise from the two unmodified DCX's analog outs when they were both running, and switching either of them off by the mains switch saw the problem go away (but now only running one DCX. I have not tried it since I modified one of the DCX's. Also, this was in-car running from a sinewave inverter (stereo 6-way was fun to 'tune up'). I never tried it on house power.
once I have finished testing this unit, and modified the other, I may put them back in the car.
correct me if I'm wrong, but as the DCX is transofrmer coupled in the AES input, then switching one of them off would not affect the impedance of the coupling.
My suspicion is that the interference was caused by the proximity of the units (mounted together in a rack) and possibly the feedback of slightly different oscillator frequencies in the power supply and/or digital circuits 'coupling' together.
If I happen upon another unmodified DCX, I'll do some more testing.
maybe I should have tried unplugging the AES from one unit, whilst still leaving them both connected on the outputs, and powered on....
My suspicion is that the interference was caused by the proximity of the units (mounted together in a rack) and possibly the feedback of slightly different oscillator frequencies in the power supply and/or digital circuits 'coupling' together.
If I happen upon another unmodified DCX, I'll do some more testing.
maybe I should have tried unplugging the AES from one unit, whilst still leaving them both connected on the outputs, and powered on....
Anyone use a balanced to unbalanced output with this setup without introducing noise! 2 of three of my amps are RCA only.Hi,
there is so much voltage available at the output that noise is not a problem.
The balanced will give 6db more output than unbalanced, but that 6db loss is relatively unimportant. We are not talking spare headroom here, the DCX will put out 9.75Vac max from CDs and high enegy recordings will average about 6db to 10db below this.
CDs with more dynamic range will still average 20db below 9.75Vac or about 975mVac. Take off 6db for unbalanced and you still have ~500mVac average output to the next stage.
I see that the unbalanced output should increase by 6db when the other polarity is grounded, but I did not find this with my version.
What are you using to control volume after the DCX?
there is so much voltage available at the output that noise is not a problem.
The balanced will give 6db more output than unbalanced, but that 6db loss is relatively unimportant. We are not talking spare headroom here, the DCX will put out 9.75Vac max from CDs and high enegy recordings will average about 6db to 10db below this.
CDs with more dynamic range will still average 20db below 9.75Vac or about 975mVac. Take off 6db for unbalanced and you still have ~500mVac average output to the next stage.
I see that the unbalanced output should increase by 6db when the other polarity is grounded, but I did not find this with my version.
What are you using to control volume after the DCX?