Behringer DCX2496 digital X-over

Hi,
using transformer as the coupling certainly gets rid of the DC on the output.

But what would be the repercussions, for the transformer, of feeding the AC signals with a 2.5Vdc bias into it? Saturation? distortion? or no problem? Does the balancing remove the bias with respect to signal level?

Judging by the data I have seen published and the large increase in distortion at low frequency I think current into the transformer and core saturation are very big issues. But I have not seen data when AC and DC are mixed at the input to the transformer.
 
Re: Re: Scheme of Audiophonics

nicko500 said:


For the outputs, this is the same schematics that dcx2496.fr

Full asymetrical passive output with RC filter.

Thanks!

I have read the article (studying technical French is quite difficult for me, but also very rewarding).
Do you yourself own this board? If so, are you happy with it?
Would you recommend it?

I am a bit surprised that so may people still use loudspeaker caps as coupling caps. I know these Mundorffs very well from speaker design. They are quite good for the money and certainly better than the standard Intertechnik, but not so good as the Intertechnik Audyn Plus caps. The Mundorff supremes are better still but they are expensive.

However, I hav always found these massive loudspeaker caps quite unsuitable for the low voltage, low current use at line levels.
They do sound better than bad quality electrolytic caps (less harsh and brite), but they are defintely no match for high quality caps like Blackgate and the likes.

Loudspeaker caps are simply too slow, too mellow, not enough transparent wenn used at line levels, so I am always surprised to see them used here...

Regards,

Lucas
 
giulio said:


I have bought one. It implements a 1st order low pass to filter out digital noise and a 1st order high pass with Mundorf MKP cap to keep DC out.

Jan's pcb has 2nd order low pass. I use transformer volume controls after the DCX so 1st order is enough for me. The board is single ended but can be easily modded to run balanced. Given that you cannot do w/out the coupling caps anyway I fail to see (in my setup) the advantage of running balanced ahead of the TVCs.

The board has two analogue (buffered, DC coupled) and 1 digital input. It was the two analogue inputs (I wanted to be able to use a microphone for the autoalign) that prompted me to go for this one.

All in all, they both look very good products to me. Horse for courses I would say.

Hope it helps
Giulio

Hi Giulio,

Thanks for the info.
I will probably use a passive stepped 6 channel volume control after the DCX. Is a 6db filter a disadvantage here?

I have a few questions about the board:

1. Is the digital input a SDIF input, or still an AES?

2. How is the quality of the analogue input?

3. Are the outputs not the high level (pro) of the originals?

4. Could one get the board without the Mundorffs?


Thanks for all you advice.

Regards,

Lucas


P.S. Did you do any of the other upgrades from the same manufacturer (analogue PSU / Clock upgrade...) They are not so expensive, and I wonder how good they are.
 
AX tech editor
Joined 2002
Paid Member
Re: Re: Re: Scheme of Audiophonics

Lucas_G said:
[snip]However, I hav always found these massive loudspeaker caps quite unsuitable for the low voltage, low current use at line levels.
They do sound better than bad quality electrolytic caps (less harsh and brite), but they are defintely no match for high quality caps like Blackgate and the likes.

Loudspeaker caps are simply too slow, too mellow, not enough transparent wenn used at line levels, so I am always surprised to see them used here...

Regards,

Lucas

Lucas, why is that?? What about an amp playing at low level, or, at very low level music passages? Do those caps then suddenly mess up the sound? I think you suffer from a wrong analogy. The speakers are massive wrt the equipment. The caps are just the caps. Both speaker caps and line level caps have the same requirements. A bad line level cap would also be a bad speaker cap and vice versa.

Jan Didden
 
Hi Giulio,

Thanks for the info.
I will probably use a passive stepped 6 channel volume control after the DCX. Is a 6db filter a disadvantage here?

Pleasure. The transformer adds further 2nd order filtering. For the bass channel I am using a buffered stepped attenuator and I plan to introduce as second 1st order stage as Jan did. Depending on your amp, the out of band noise might be excessive (likely if the amp is solid state).

I have a few questions about the board:

1. Is the digital input a SDIF input, or still an AES?
It is AES, but putting a 200 Ohm resistor in parallel will make it 75. I would rather mount the resistor in the cable connector, rather than on the board so that I can choose between AES and SPDIF (I may be buying an SRC2496).

2. How is the quality of the analogue input?
Not bad at all. OPA2134 opamp and there is a socket, so you can play around with it.

3. Are the outputs not the high level (pro) of the originals?
No, they are passive 2.4V single ended. 4.8V if you decide to mod it run it balanced.

4. Could one get the board without the Mundorffs?
Not that I am aware.

I think the other upgrade you mention are from a different company (selectronic). The clock seems very good value for money but I have no way of judging how got it is.

Best
Giulio
 
Re: Re: Re: Re: Scheme of Audiophonics

janneman said:


Lucas, why is that?? What about an amp playing at low level, or, at very low level music passages? Do those caps then suddenly mess up the sound? I think you suffer from a wrong analogy. The speakers are massive wrt the equipment. The caps are just the caps. Both speaker caps and line level caps have the same requirements. A bad line level cap would also be a bad speaker cap and vice versa.

Jan Didden


Hi Jan,

Loudspeaker caps do have the tendency to make the sound more mellow, more relaxed, more romantic, more soft when used at linelevels. Good caps like Blackgate N (and some others) are far more realistic, with more speed and more detail. Much closer to a mere wire.

This opinion is shared by many others on this forum (Peter Daniel and many other people wrote on it when dealing with output caps on their Dacs). If you can read german you will find similar remarks on the forum www.audiomap.de

Of course there is always a personal and subjective element, so you should always verify this in your own setup. However, I would be surprised if you would feel different after comparing these two groups of caps.

Of course, not so long ago, you could not obtain high quality electrolitic caps. Probably the routine of reaching for loudspeaker caps stems from those days, I guess...

The level in loudspeakers is much much higher than in line-level applications. Loudspeaker caps are most certainly optimised for use at these levels of voltage and current.

Kind regards,

Lucas

P.S. Janneman, thanks by the way for that wonderfully written article on the DCX board. Very clear ans well thought out.
 
AX tech editor
Joined 2002
Paid Member
Re: Re: Re: Re: Re: Scheme of Audiophonics

Lucas_G said:
Hi Jan,

Loudspeaker caps do have the tendency to make the sound more mellow, more relaxed, more romantic, more soft when used at linelevels. Good caps like Blackgate N (and some others) are far more realistic, with more speed and more detail. Much closer to a mere wire.[snip]

It is my believe that if I do a double blind test with a 'speaker cap' (whatever that is) and a 'line cap' (whatever that is) that neither you nor anybody else will be able to identify the difference reliably. But I haven't proven it....

Lucas_G said:
[snip]P.S. Janneman, thanks by the way for that wonderfully written article on the DCX board. Very clear and well thought out.


Thanks. I'm working on an active, remote controlled version. Should be fun to compare the two...

Jan Didden
 
DC offset with balanced amp inputs OK?

In my case, my amp also supports balanced inputs. Would these not be tolerant of DC offsets since XLR provides a differenctial signal and only the difference in the signals matters not the absolute value? If this the case I can definitely go with Jan's kit without the caps instead of the French kit since Jan's kit is differential.

Once again, I can only say that there must be a better way of doing this (no tranformers and no caps). On the other hand a lot of intelligent people have looked at this and come up with those as the best solutions. After looking at the specs, voltage reference offsets will not work with the AK4393 becaase it uses a common analog and digital ground plane.

Philip
 
AX tech editor
Joined 2002
Paid Member
Re: DC offset with balanced amp inputs OK?

pgruebele said:
In my case, my amp also supports balanced inputs. Would these not be tolerant of DC offsets since XLR provides a differenctial signal and only the difference in the signals matters not the absolute value? If this the case I can definitely go with Jan's kit without the caps instead of the French kit since Jan's kit is differential.

Once again, I can only say that there must be a better way of doing this (no tranformers and no caps). On the other hand a lot of intelligent people have looked at this and come up with those as the best solutions. After looking at the specs, voltage reference offsets will not work with the AK4393 becaase it uses a common analog and digital ground plane.

Philip


In principle, your diff input amp supresses common mode input signals, including DC. But this generally depends on the very good matching of the input Z of the amp and the source Z of the DAC. If there is an unbalance, the actual part of that 2.5V at each line that gets to the amp input pins may be slightly different and that difference than is amplified as a differential mode signal and will lead to DC at the output of the amp.

With a passive output, using 0.1% resistors you can match the source Z adequately, but what about the input Z of the amp? You have a circuit diagram or a spec sheet?

Anyway, my preference if I had a differential input amp would be to use input caps rather that going to a single ended mode (and throwing away the advantages of differential input which IMHO far outweight a series cap).

Jan Didden
 
Re: DC offset with balanced amp inputs OK?

Thanks Jan,

Here are the only relevant specs I found on my amp (full manual at http://www.bkcomp.com/pdf/Reference_200_5-7S2_UM_13744_0604.pdf):

1. Differential Input Stage - A current source load assures DC stability and wide bandwidth linearity.

2. Input Impedance 33.2 kÙ

3. Input Sensitivity RCA Unbalanced 1.4 Volt

4. Input Sensitivity XLR Balanced 2.8 Volts | Pin 1 = Gnd, Pin 2 = In +, Pin 3 = In -

Does this tell you enough to have an informed opinion on whether coupling capacitors are needed? It seems to me that the only problem with not using them would be an offset in the signal which should not cause any immediate damage in speakers/amps unless this overdrives the speakers/amp. So, for testing pusposes I could just measure the DC output voltage out of the amp to see if it is close to zero, and only add the caps if there is a considerable offset?

The reason I am perhaps too phobic about the capacitor coupling is that these are exactly the types of components I wanted to remove from the signal path (i.e. the passive crossover). I don't even want to place a capacitor in series with my tweeters because this once again can affect their performance. It seems that the best way to avoid worrying about how something impacts your signal path (types of capacitors etc) is to not have anything in the path, if possible. In the case of capacitors there is certainly a small amount in voodoo involed since there appear to be no clear empirical ways of predicting performance in this application. Ironically one also has to consider that the negative impact on the signal can actually cause a positive impact on the perceived sound :-( There are no free meals...

Thanks very much for you replies...

Philip
 
I can tell you that I plug S/PDIF consumer outputs directly into the DCX and it works perfectly. I don't see a need for any input conversions. The thing I would be worried about with digital inputs is jitter resistance.
Okay,I meen if I switch the input transformer in the DCX for anyone of the above,if that would be an upggrade or is the one used in the DCX good enoug?
 
Hi,
A DC coupled amplifier will amplify all signal within it's HF bandwidth equally.
If the gain were 30times, then 2.5Vdc at the input becomes 75Vdc at the output. There's a speaker up in smoke.

Some manufacturers (and commentators) refer to an amp as DC coupled when the input DC blocking cap is omitted (or bypassed) but forget about the DC blocking cap in the NFB loop. This type of partial DC coupled amplifier has a DC gain of 1times.
This time the output DC offset ~=input DC bias i.e. 2.5Vdc.
Many speakers will not perform at all well with this level of DC on the voice coil and some will overheat till destruction occurs.
 
There have been a lots of talk on this board regarding the input and relationship of AES/EBU and S/pdif. My own experience was: I used small Neutrik transformer that did conversion from RCA S/Pdiff to balannced AES/EBU (ballanced S/PDIFF). It worked sort of OK but I would occationally loose a signal and DCX would start clicking. I have no idea why that would happen, but I know that when I introduced SRC 2496 into the chain all problems were gone. The above mentioned problem was happening when I plugged optical cable to DEQ2496 and than had AES/EBU from DEQ to DCX. So there are a few options, but in my experience the best result is having SRC do all your conversions and send AES/EBU. For $ 120.00 I woudn't bother with transformers or any mods - just go for SRC2496 and as added benefit you will have one more digital input or upsampler from 44.1 to full 96. ( I didn't believe in this but when I tried I got better sound with 96 upsampling)
Problem is SRC is not widely available just like DCX but they are trickling them down slowly. I got mine about two months ago from Muzician Friend I believe.
 
Well I got a Roland M-1000 digital mixer I think it does the same thing as SRC2496,I to get much better sound uppsampling to 96khz,it´s just one issue after the M-1000 i have a DEQ2496 and it works fine with 96khz,after that I have the DCX and I can´t have as high level on 96khz as I can with 48khz,it starts distort if go to high..
Well no uppgrading of input trafo then..;)
what about the outputs then will any of the above mention lundahls work?
 
Ryssen said:
Well I got a Roland M-1000 digital mixer I think it does the same thing as SRC2496,I to get much better sound uppsampling to 96khz,it´s just one issue after the M-1000 i have a DEQ2496 and it works fine with 96khz,after that I have the DCX and I can´t have as high level on 96khz as I can with 48khz,it starts distort if go to high..
Well no uppgrading of input trafo then..;)
what about the outputs then will any of the above mention lundahls work?
Check with Kevin at K & K Audio. He is real expert on Lundahls, and he will tell you what to use. He helped me and he is a pleasure to deal with.
 
Digital output of Sherwood-Newcastle P-965 to DCX 2496?

Has anyone connected the digital output (RCA-type coaxial) of the Sherwood-Newcastle P-965 preamp-processor to the digital input of the DCX-2496? I already have the P-965 but have not bought the Behringer yet.

I am designing a 3-way active speaker, and the versatility of the Behringer is appealing, especially in regard to time delay capability. Another option is an analog passive line-level crossover, namely the Marchand XM-46, but 4 of those boards would cost nearly as much as the DCX 2496, and I would then be stuck with fixed crossover points and slopes and no time delay.

After reading this thread, I have already identified 2 big issues with the DCX: 1) Volume control before DCX = bad, and 2) Converting from RCA coaxial output to XLR input will hurt S/N ratio unless voltage gain is inserted.

Thank you for any comments.
 
Re: Digital output of Sherwood-Newcastle P-965 to DCX 2496?

Javachip said:
2) Converting from RCA coaxial output to XLR input will hurt S/N ratio unless voltage gain is inserted.
I don't understand what you refer to here.
Is the P965 a coax output? 75ohm? S/PDIF?
The DCX is aes/ebu balanced input using XLR connector. But the spec says it is also compatible with S/PDIF but using 110ohm and accepts digital signal from 0.3Vpp to 10Vpp.
There should be no S/N issue with digital input, it is a lock onto signal issue.