AndrewT said:
Well, at the ends of the pot are two signals of equal amplitude but opposite phase, rhight? So, with the wiper at midpoint, that wiper has zero signal - referred to ground. So the top output amp input is effectively at ground level. The bottom output amp gets the full signal. Therefore it seems to me there is no longer a balanced output signal. To do that, two pots in parallel would be needed with the wipers rotating in opposite directions.
Another option would be to use the pot as a variable resistance acrross the outputs so that both outputs are attenuated equally and then run the ends of the pot straight to the output amps.
Jan Didden
Hi Jan,
I hope we talk about the same circuit (post 826 and 827, it's the same). For my understanding due to the two AC coupling caps and the two 100k input resistors of the UcD you have a balanced signal referred to ground.
Has anybody some theory of the optimal filter setting for the tweeter channel (attenuation at 20 kHz, 40 kHz, 80 kHz)? Might depend on whether you have CD or SACD data.
Thanks and Regards,
Frank
I hope we talk about the same circuit (post 826 and 827, it's the same). For my understanding due to the two AC coupling caps and the two 100k input resistors of the UcD you have a balanced signal referred to ground.
Has anybody some theory of the optimal filter setting for the tweeter channel (attenuation at 20 kHz, 40 kHz, 80 kHz)? Might depend on whether you have CD or SACD data.
Thanks and Regards,
Frank
oettle said:
#826 I talk about. The signal to the lower part of the output amp is always max, whatever the pot setting. It is directly coupled from the 750 ohms to the 100 ohms and into the input via the coupling cap The signal to the top output amp goes from +max to zero to -max depending on the pot.
Jan Didden
Hi,
out of phase does not mean the voltages cancel at the midpoint.
The balanced signal is opposite phase from each line and that just gives double the voltage across the pot.
The output end taps off some of this voltage using one line as the reference. There is no connection to ground on the left and the connection to ground later on is symetrical.
So, as I see it, the output end sees an input that varies with the pot setting just like a normal volume pot. This behaviour is just like a conventional unbalanced attenuator working like a variable filter each time the volume is changed.
Now, the filter sees the pot as a source resistance in addition to the resistance of the line (100r) resistors. It also sees some of the preceeding circuit and these should be taken account of otherwise the model will be inaccurate.
The 100pF & 100R give 0.01uS (-3db@64MHz), not a lot of filtering in there.
When the pot is set to zero the filter is doing very little.
But add in half the pot setting to each half of the output stage filter and the 100r +Rs=2000+100 & 100pF >=0.21uS (3MHz). That of itself is probably inaudible. but is it combined with any other filters creating a very slow roll-off that droops down at the top end of the audio spectrum? If you understand where the filters are being generated then you should be able to get the sym to tell you the predicted numbers in the model.
If you don't understand HOW the filters work then you have no hope of recognising the numbers that come out or whether they make any sense.
out of phase does not mean the voltages cancel at the midpoint.
The balanced signal is opposite phase from each line and that just gives double the voltage across the pot.
The output end taps off some of this voltage using one line as the reference. There is no connection to ground on the left and the connection to ground later on is symetrical.
So, as I see it, the output end sees an input that varies with the pot setting just like a normal volume pot. This behaviour is just like a conventional unbalanced attenuator working like a variable filter each time the volume is changed.
Now, the filter sees the pot as a source resistance in addition to the resistance of the line (100r) resistors. It also sees some of the preceeding circuit and these should be taken account of otherwise the model will be inaccurate.
The 100pF & 100R give 0.01uS (-3db@64MHz), not a lot of filtering in there.
When the pot is set to zero the filter is doing very little.
But add in half the pot setting to each half of the output stage filter and the 100r +Rs=2000+100 & 100pF >=0.21uS (3MHz). That of itself is probably inaudible. but is it combined with any other filters creating a very slow roll-off that droops down at the top end of the audio spectrum? If you understand where the filters are being generated then you should be able to get the sym to tell you the predicted numbers in the model.
If you don't understand HOW the filters work then you have no hope of recognising the numbers that come out or whether they make any sense.
Hi it's me again,
I would say the passive filter has a slow rolloff problem. So I modified the circuit:
firts order filter: R500, 1000pF
second order filter omitted at all (R250, 1000pF)
instead two 680pF caps in parallel to the 1k gain setting resistors of the instrumental amp on the UcD.
The values are based on a 0.1dB attenuation at 20kHz.
This solution is now rather similar to the active solution, but roll off is still slower and attenuation at 20kHz is higher. On the other hand phase shift is better.
Perhaps someone has an idea to mod the instrumental amp on the Ucd to improve rolloff further?
Regards,
Frank
I would say the passive filter has a slow rolloff problem. So I modified the circuit:
firts order filter: R500, 1000pF
second order filter omitted at all (R250, 1000pF)
instead two 680pF caps in parallel to the 1k gain setting resistors of the instrumental amp on the UcD.
The values are based on a 0.1dB attenuation at 20kHz.
This solution is now rather similar to the active solution, but roll off is still slower and attenuation at 20kHz is higher. On the other hand phase shift is better.
Perhaps someone has an idea to mod the instrumental amp on the Ucd to improve rolloff further?
Regards,
Frank
Been reading a little about the clock of DCX,and it says tha it does NOT get the clock signal from the S/PDIF -AES/EBU input but uses it´s own clock.
Anyway the frevkensy for the DAC is 24.576 and for the ADC is 12.288.If I am NOT going to use the analog side of DCX Is it enoug to uppgrade the DAC clock(I have a kwack clock with 24.574Mhz)
Anyway the frevkensy for the DAC is 24.576 and for the ADC is 12.288.If I am NOT going to use the analog side of DCX Is it enoug to uppgrade the DAC clock(I have a kwack clock with 24.574Mhz)question on cap sizes
I have 2 questions/comments here:
1)
I just got the french DCX kit and am somewhat disappointed that it did not appear to improve things. If anything it may have made the sound worse! Of course it's difficult to do A/B testing so I can't say for sure yet. Having consumer voltage levels is, of course nice as I don't have to worry about blowing things up. This is with digital source and B&K 200.7 amp and Monitor Autio GR60 speakers. Perhaps the amp is my bottleneck?
The reason I got the french kit is because it has a cleaned up analog in circuit in addition to digital input. Does anyone else have positive experiences with this kit?
2)
Anyway, I have a general question about all the kits out there. They all use the same size caps for all 6 output channels. It seems to me that the mid and high channels could use increasingly smaller cap values since they will be fed much higher frequencies than the low channel (in my case the crossovers are 250 and 3000 Hz). It also seems to me that these smaller caps would be more like wire in their tranparency.
Is this a correct assumption? If yes, what cap values/brands would be good for my crossover frequencies (to insure that there will be no rolloff caused by the caps given 48db/octave slopes).
Thanks for any comments
Philip
I have 2 questions/comments here:
1)
I just got the french DCX kit and am somewhat disappointed that it did not appear to improve things. If anything it may have made the sound worse! Of course it's difficult to do A/B testing so I can't say for sure yet. Having consumer voltage levels is, of course nice as I don't have to worry about blowing things up. This is with digital source and B&K 200.7 amp and Monitor Autio GR60 speakers. Perhaps the amp is my bottleneck?
The reason I got the french kit is because it has a cleaned up analog in circuit in addition to digital input. Does anyone else have positive experiences with this kit?
2)
Anyway, I have a general question about all the kits out there. They all use the same size caps for all 6 output channels. It seems to me that the mid and high channels could use increasingly smaller cap values since they will be fed much higher frequencies than the low channel (in my case the crossovers are 250 and 3000 Hz). It also seems to me that these smaller caps would be more like wire in their tranparency.
Is this a correct assumption? If yes, what cap values/brands would be good for my crossover frequencies (to insure that there will be no rolloff caused by the caps given 48db/octave slopes).
Thanks for any comments
Philip
Hi,
the problem regarding sound quality aren't only the 6 caps, but also the opamps. If you need unbalanced (single ended) outputs, you can ommit the second dual opamp of each channel at all. Only the R64 protection resistor is neccessary. Take better opamps (15V types).
With a 47uF cap you have an 0.1dB attenuation and 10° phase shift at 20Hz. With a 4.7uF cap you get similar values at 250 Hz and with a 0.22uF cap at 3000 Hz.
Take caps with low dissipation factor. Polyprop caps are best but large. Silmic from Elna might be a solution for the 47uF value. See post 829 for the whole circuit.
Good luck,
Frank
the problem regarding sound quality aren't only the 6 caps, but also the opamps. If you need unbalanced (single ended) outputs, you can ommit the second dual opamp of each channel at all. Only the R64 protection resistor is neccessary. Take better opamps (15V types).
With a 47uF cap you have an 0.1dB attenuation and 10° phase shift at 20Hz. With a 4.7uF cap you get similar values at 250 Hz and with a 0.22uF cap at 3000 Hz.
Take caps with low dissipation factor. Polyprop caps are best but large. Silmic from Elna might be a solution for the 47uF value. See post 829 for the whole circuit.
Good luck,
Frank
Thanks Frank.
This kit (and none of the others, I believe) does not use opamps in the output circuit, so I don't know what you mean...
Thanks for the advice on the caps. Do you think this would improve things? I mostly want to improve the clarity of the midrange.
Philip
This kit (and none of the others, I believe) does not use opamps in the output circuit, so I don't know what you mean...
Thanks for the advice on the caps. Do you think this would improve things? I mostly want to improve the clarity of the midrange.
Philip
Hi Philip,
obviously we are talking about different things. I thought you are using http://www.dcx2496.fr/alimentation.html which is based on the original Behringer analog PCB.
Are you using the dutch mod http://www.dcx2496.fr/alimentation.html or something else?
Frank
obviously we are talking about different things. I thought you are using http://www.dcx2496.fr/alimentation.html which is based on the original Behringer analog PCB.
Are you using the dutch mod http://www.dcx2496.fr/alimentation.html or something else?
Frank
Sorry something went wrong with the dutch link http://www.audioxpress.com/magsdirx/ax/addenda/media/didden2657.pdf
Hi,
some misunderstanding here?
posts 849 & 850 are referring to the french (non understanding the lingo) text for mods to the PSU and it's voltages.
Whereas everyone else, at the moment, is referring to the output stage. e.g.Oettle & Didden.
some misunderstanding here?
posts 849 & 850 are referring to the french (non understanding the lingo) text for mods to the PSU and it's voltages.
Whereas everyone else, at the moment, is referring to the output stage. e.g.Oettle & Didden.
Hi,
for fast check you can bypass the output caps with 1 Ohm smd resistors on the bottom side. You can alter this back very easily, but the sound is now very different, sounds great.
If you do not have dc coupled power amplifiers, there is no risk.
It works now since one year with my DCX.
Harald
for fast check you can bypass the output caps with 1 Ohm smd resistors on the bottom side. You can alter this back very easily, but the sound is now very different, sounds great.
If you do not have dc coupled power amplifiers, there is no risk.
It works now since one year with my DCX.
Harald
Frank,
I was referring to the modified output state from http://www.audiophonics.fr/carte-entreesortie-audiophile-pour-behringer-dcx2496-p-1081.html
The sound quality does not appear to have improved. My first impression is that it may have gotten worse...
Philip
I was referring to the modified output state from http://www.audiophonics.fr/carte-entreesortie-audiophile-pour-behringer-dcx2496-p-1081.html
The sound quality does not appear to have improved. My first impression is that it may have gotten worse...
Philip
Hi Pg,
I have no understanding of the French page.
Is your mod using a passive output stage or are the opa2604 still doing some active buffering or amplifying?
I have no understanding of the French page.
Is your mod using a passive output stage or are the opa2604 still doing some active buffering or amplifying?
I have the same card and have never listened to the stock card (thinking it was a waste of my time). I have to say that the sound is an improvement over a modified Pioneer dv575, with direct output (bypassed active output and filtering stage) into a transformer volume control and no dc coupling cap. I am very satisfied.
I have a hard time believing that the OK polyprop coupling caps in the new card are worse than the ghastly electrolitic ones in the original one!
You do not mention a preamp. What is your crossover driving? Mine is driving one Tx102 transformer volume control per mono channel (I am running two way at the moment). The transformers effectively add a second order LP filter to the card's first order one. The out of band noise after the first order filter may be causing problems to your SS amps or preamp.
Best
Giulio
Hi Philip,
Do you have a circuit of the audiophonics mod which you could send? What's the input impedance of your power amp?
Frank
Do you have a circuit of the audiophonics mod which you could send? What's the input impedance of your power amp?
Frank
Frank
but asymmetric and with 6.8u coupling cap.
Giuilio
An externally hosted image should be here but it was not working when we last tested it.
but asymmetric and with 6.8u coupling cap.
Giuilio
http://www.audiophonics.fr/carte-en...496-p-1081.html uses a passive output state. There is no published schematic for this, although I may reverse engineer it...
I use the digital input of this card and have not implemented real volume control yet (just the DCX input volume control of 30db). For testing purposes this should be sufficient since chopping off a few bits for volume should not affect the dynamic range since we are talking 24 bit processing on 16 bit data.
I also have a hard time believing that these caps are worse than the stock ones. That seems to be the only sensible reason why things would not sound better.
I cannot test without the caps because I use a B&K 200.7 seven channel amp which is DC coupled. The 6 outputs of this amp are wired to Audio Monitor GR60 speakers with passive crossover removed.
My first reaction was that all these ouput mods should use smaller caps for the mid and high outputs. I would think that the use of larger caps would be unnecessary and would have bigger potential of muddling the sound.
Generally speaking, my system sounds incredible with things like guitar and vocals. However, when the music gets a little "wilder" and has more instruments and sounds, the sound appears kind of muddled. I am trying to improve this aspect of the system and this mod appears to only have made this marginally worse...
Philip
I use the digital input of this card and have not implemented real volume control yet (just the DCX input volume control of 30db). For testing purposes this should be sufficient since chopping off a few bits for volume should not affect the dynamic range since we are talking 24 bit processing on 16 bit data.
I also have a hard time believing that these caps are worse than the stock ones. That seems to be the only sensible reason why things would not sound better.
I cannot test without the caps because I use a B&K 200.7 seven channel amp which is DC coupled. The 6 outputs of this amp are wired to Audio Monitor GR60 speakers with passive crossover removed.
My first reaction was that all these ouput mods should use smaller caps for the mid and high outputs. I would think that the use of larger caps would be unnecessary and would have bigger potential of muddling the sound.
Generally speaking, my system sounds incredible with things like guitar and vocals. However, when the music gets a little "wilder" and has more instruments and sounds, the sound appears kind of muddled. I am trying to improve this aspect of the system and this mod appears to only have made this marginally worse...
Philip