Just starting to change
Where did you read goes to hell in a handbasket? I think I wrote that with -12db digital volume control the sound is "just starting" to turn for the worse.
It's an easy test to do. Doesn't even involve soldering. Maybe just plugging some cables around and turning some knobs. Although, we see that the performance of the DCX is worse at full scale so it would be easier to observe my findings by using a dac with more consistent performance as I did.
I agree that the signal is "clean enough" around 1v output. So if that's all you need, it's fine. Actually, I need a bit more, reaching up to 3v RMS, and it's still fine, though could be better.
But if you take the existing DCX and apply 20dB analog attenuator to the output, now the DCX needs to put out 10v for you to get your 1v out of the attenuator. So by using analog attenuator, you are increasing the percentage of distortion in the output, because the percentage of distortion increases at levels above 1v coming out of the DCX itself, whether you are attenuating that voltage or not.
If you need 15dB or so of attenuation for some reason (though I can't see why, actually, perhaps you should explain why), you can just use the 15dB of digital attenuation provided for in the setup menu for each output. In most cases, unlike using an outboard analog attenuator, using the built-in 15dB of attenuation will be harmless, in my opinion. Whereas most likely the external analog attenuation will not be harmless.
Sorry about the overstatement. I just see no reason why small amounts of digital attenuation should affect things for the worse, under the circumstances like I imagine. Not seeing any reason for it, I'm somewhat uninclined to put the effort into the sort of blind objective test that would be required to get me to believe otherwise. However, I am determined to show that under circumstances that are directly applicable to me, the digital attenuation that I am currently using does not result in reduced dynamic range. I use the digital output of a Sonos system, and control the volume by the digital volume control of that system, possibly using as much as 25dB of attenuation. Now if this works the way I think it should, if I attenuate by, say, 30dB, and then re-inflate 30dB through level controls in the DCX, there should be no change in measured dynamic range. I verified several years ago that Sonos does put out 24bit signal when it uses attenuation. So I think this should work, and I'll report results here though it is only partly related to the DCX itself.
How exactly did you do your test of digital attenuation?
In my case with the DCX I run just a little trim here and there. But I also boost the outputs by 4dB. Why? Because once the input signal gets sliced into 2 or 3 parts, each output section is pretty low. Running +6dB is too much, the rogue peak gets thru here and there.
My volume control is done in the player. It's calculated at 32 bits, then reduced to 24 or 16 as I choose. I will admit that I have not done rigorous testing to find if digital volume in the player is better or worse than analog attenuation after the DAC. But I don't hear the difference right away. I mean it does not strike me. But it's worth further investigation.
My volume control is done in the player. It's calculated at 32 bits, then reduced to 24 or 16 as I choose. I will admit that I have not done rigorous testing to find if digital volume in the player is better or worse than analog attenuation after the DAC. But I don't hear the difference right away. I mean it does not strike me. But it's worth further investigation.
Test set up
I used the above set up by starting with -12 digital and selecting a loud listening level with the attenuators wide open at -4db which is as loud as they go. Increasing the digital to -0 and turning the attenuators down 12db, 6 steps, gives me the same listening level. Back and forth very easily. This test requires analog volume control between the dac and the amps so is easiest with passive cross speakers although the repeatability of 4 stepped attenuators would allow me to make the same test when running active which I haven't been doing often for the last few years as my main focus has been on trying different affordable amps 2 channels at a time. It would also be very easy with a Didden modded DCX providing the repeatable analog attenuation if anyone with one of those happens to be reading this.
.
I don't know if your scenario of comparing -15db digital into +15db digital with -0db digital into +0db digital is really the same thing. I might be able to try it with two DEQs so will let you know what I find.
Just to copy from my original post:
I used the above set up by starting with -12 digital and selecting a loud listening level with the attenuators wide open at -4db which is as loud as they go. Increasing the digital to -0 and turning the attenuators down 12db, 6 steps, gives me the same listening level. Back and forth very easily. This test requires analog volume control between the dac and the amps so is easiest with passive cross speakers although the repeatability of 4 stepped attenuators would allow me to make the same test when running active which I haven't been doing often for the last few years as my main focus has been on trying different affordable amps 2 channels at a time. It would also be very easy with a Didden modded DCX providing the repeatable analog attenuation if anyone with one of those happens to be reading this.
.
I don't know if your scenario of comparing -15db digital into +15db digital with -0db digital into +0db digital is really the same thing. I might be able to try it with two DEQs so will let you know what I find.
Part of the problem is that the DCX has some 14dB gain in the analog output stage to make it compatible with pro levels. This is too high for home audio so indeed you have to either attenuate it in the digital domain (up to -15dB available) which in my experience is audible once you get below -6dB or so. Or you need to have some analog attenuation after the output, but because some of the distortion is caused in that 14dB analog gain stage, the analog attenuation doesn't lower that distortion, as you say.
The optimum solution is to get rid of that 14dB analog gain after the DACs. Lower distortion, no need for digital attenuation, no (or less) attenuation on the analog output.
One solution is the total removal of that output gain board and go with passive output to your power amps. I wrote an article about that for AudioXpress, using a passive filter. Other people have used a xformer here.
My final solution is a replacement output board with filtering and a 6-channel. remote control analog level control, although technically it is somewhat involved and not for everyone I guess.
But again, the optimum solution is to chuck that 14dB output gain on the DCX.
jd
If you want to write it, Ill proof it. Can't help with pics as I'm an old fashioned sketch on paper engineer.
Behringer meets Krell, single ended seems better here
This is now just my limited experience, based on a couple nights of listening, but it appears that my Krell FPB 300 is much happier running on singled ended output of Behringer DCX 2496 than balanced. To get single ended, I'm now using a high quality audiophile cable with balanced adapters on both sides (makes for easy replacement with double XLR cable). All adapters have pins 3 and 1 connected, which is pretty standard, and keeps the gain level constant, and I need lots of volts to drive my amplifier.
Running full balanced, the Krell is clearly being pushed into higher and higher bias plateaus, causing the lights on the front to dim, and more annoying clinks than usual from amplifier thermal expansion during and after peak bias draws.
Running single ended, the Krell is totally unflustered. Same sound and fury, but the amp is unbothered.
I have hypothesis that Behringer's "servo balanced" outputs have slow levels of DC drift. Another possibility is that while hum and noise are "canceled" through the balanced operation of the Krell, equal and opposite noise in both positive and negative portions of the amplifier causes work to be done needlessly.
I have always thought of balanced connection as being always superior, but here may be one case in which it is not.
It also seems that separate behringer outputs may intereact somewhat, especially regarding ground if there are ground loops (as there are when, for example, you put amplifiers on dedicated outlets). After some experimentation prior to deciding to run Krell single ended, I had previously decided to run Krell on it's own dedicated circuit, not sharing that circuit with other amps or electronics (which seemed to be compounding the problems). That necessitated using an isolation transformer for the signal on one other amplifer (an Acurus A250) even though that amplifier is on the same circuit as the Behringer, but not on the Krell, which is on its own dedicated circuit, and I had worried that single ended connection would never work with the Krell for that reason, but it does work fine so long as the OTHER amplifier is isolated.
I had used Behringer on separate circuit with Krell with balanced connection for 8 months last year (before 10 year old Krell started developing distortion problems and got sent back to factory). But there was always a lot more clinking and stress signs from the amplifier than there should have been, even with power sucking Acoustat 1+1 speakers.
This is now just my limited experience, based on a couple nights of listening, but it appears that my Krell FPB 300 is much happier running on singled ended output of Behringer DCX 2496 than balanced. To get single ended, I'm now using a high quality audiophile cable with balanced adapters on both sides (makes for easy replacement with double XLR cable). All adapters have pins 3 and 1 connected, which is pretty standard, and keeps the gain level constant, and I need lots of volts to drive my amplifier.
Running full balanced, the Krell is clearly being pushed into higher and higher bias plateaus, causing the lights on the front to dim, and more annoying clinks than usual from amplifier thermal expansion during and after peak bias draws.
Running single ended, the Krell is totally unflustered. Same sound and fury, but the amp is unbothered.
I have hypothesis that Behringer's "servo balanced" outputs have slow levels of DC drift. Another possibility is that while hum and noise are "canceled" through the balanced operation of the Krell, equal and opposite noise in both positive and negative portions of the amplifier causes work to be done needlessly.
I have always thought of balanced connection as being always superior, but here may be one case in which it is not.
It also seems that separate behringer outputs may intereact somewhat, especially regarding ground if there are ground loops (as there are when, for example, you put amplifiers on dedicated outlets). After some experimentation prior to deciding to run Krell single ended, I had previously decided to run Krell on it's own dedicated circuit, not sharing that circuit with other amps or electronics (which seemed to be compounding the problems). That necessitated using an isolation transformer for the signal on one other amplifer (an Acurus A250) even though that amplifier is on the same circuit as the Behringer, but not on the Krell, which is on its own dedicated circuit, and I had worried that single ended connection would never work with the Krell for that reason, but it does work fine so long as the OTHER amplifier is isolated.
I had used Behringer on separate circuit with Krell with balanced connection for 8 months last year (before 10 year old Krell started developing distortion problems and got sent back to factory). But there was always a lot more clinking and stress signs from the amplifier than there should have been, even with power sucking Acoustat 1+1 speakers.
CS8420 bypass
Anyone tried to bypass CS8420 in DCX. I want to feed my DCX directly with
94/24 AES3 signal from my PC audio player. I don't have schematic for DCX so I'am not sure if bypassing will work. Receiving of AES signal in CS8420 pins RXP 4 and RXN 5 should be wired to output side TXP 26 and TXN 25. This mod will avoid CS8420 replacement. I need any opinion from experienced DCX mod people.
Anyone tried to bypass CS8420 in DCX. I want to feed my DCX directly with
94/24 AES3 signal from my PC audio player. I don't have schematic for DCX so I'am not sure if bypassing will work. Receiving of AES signal in CS8420 pins RXP 4 and RXN 5 should be wired to output side TXP 26 and TXN 25. This mod will avoid CS8420 replacement. I need any opinion from experienced DCX mod people.
Hello,
I have recently bought the DCX2496, it is yet to arrive. I have couple of questions:
1) Have somebody run some tests on unmodded dcx2495? Is the measurement data available somewhere?
2) Have somebody run some tests on (somehow) modded dcx2496 and compared them to unmodded version? Is the measurement data available somewhere?
3) I will use dcx's analog inputs in 3-way system. Signal is routed from XTZ D3 integrated amplifier, which is also used for volume control. Cables are going to be of a good quality. Is this setup going to introduce hiss to my system? Which way I could minimize the hiss? Would it help to reduce Behringer's gain in inputs so I would be forced to use greater volume in XTZ-preamp in order to get the same sound pressure level?
4) Some good tricks to get the most out of unmodded dcx? Can include some minor easy-assemble mods like ferrites (which I'm going to use anyway).
Regards,
Legis
I have recently bought the DCX2496, it is yet to arrive. I have couple of questions:
1) Have somebody run some tests on unmodded dcx2495? Is the measurement data available somewhere?
2) Have somebody run some tests on (somehow) modded dcx2496 and compared them to unmodded version? Is the measurement data available somewhere?
3) I will use dcx's analog inputs in 3-way system. Signal is routed from XTZ D3 integrated amplifier, which is also used for volume control. Cables are going to be of a good quality. Is this setup going to introduce hiss to my system? Which way I could minimize the hiss? Would it help to reduce Behringer's gain in inputs so I would be forced to use greater volume in XTZ-preamp in order to get the same sound pressure level?
4) Some good tricks to get the most out of unmodded dcx? Can include some minor easy-assemble mods like ferrites (which I'm going to use anyway).
Regards,
Legis
I do not recommend a two channel attenuator in front of the DCX.
Try to find a 6channel attenuator after the DCX.
So I have read also elsewhere. What is the reason behind this recommendation, more hiss?
