hehe, yeah these guys make it sound like we can not listen and we do not own good enough equipment. They may be right and they are just better then us but 90% of them can fit their whole house in my living room
full bucket of partybrakers ........
this thread is certainly for people willing to mod their DCX-es , and to share ways how to do it ...... subjectively or objectively , who cares .
Gentlemans .....with proper measuring equipment , probably also with proper pair of ears .... but without proper attitude to other fellow members ........ feel free to hop out in any other subjective-objective thread ........ and let these ppl to do something .
and - yes - I don't have DCX ...... but I repair gadgets like this for living .
this thread is certainly for people willing to mod their DCX-es , and to share ways how to do it ...... subjectively or objectively , who cares .
Gentlemans .....with proper measuring equipment , probably also with proper pair of ears .... but without proper attitude to other fellow members ........ feel free to hop out in any other subjective-objective thread ........ and let these ppl to do something .
and - yes - I don't have DCX ...... but I repair gadgets like this for living .
These measurements of an original DCX 2496 show frequency responses (page 2) and distortions (page 3 and 4), which certainly are well enough for me.
Does this looooong thread contain any measurements which show better values after some 'tuning'? Or measurements comparing other relevant data? I confess that I'm too lazy to look through all those pages myself.
Typical for many modding descriptions is Jan Diddens 'Greening of the Behringer ...'. He tells us what is wrong with the original output section and how to improve it. The only technical 'proof' of his mods is a simulation of the frequency and phase response before and after. His own conclusion: 'At any rate, I think the differences are too small to be of audible consequence'. Apart from that we have to rely on his subjective hearing impressions. Is that really proof enough?
Yes, I DO care whether the proof is subjectively or objectively.
Rudolf
Does this looooong thread contain any measurements which show better values after some 'tuning'? Or measurements comparing other relevant data? I confess that I'm too lazy to look through all those pages myself.
Typical for many modding descriptions is Jan Diddens 'Greening of the Behringer ...'. He tells us what is wrong with the original output section and how to improve it. The only technical 'proof' of his mods is a simulation of the frequency and phase response before and after. His own conclusion: 'At any rate, I think the differences are too small to be of audible consequence'. Apart from that we have to rely on his subjective hearing impressions. Is that really proof enough?
Yes, I DO care whether the proof is subjectively or objectively.
Rudolf
WARNING: Read below at your own risk.
You'll have to put me in the "cable huggers" camp on this one.
I do NOT like the sound of stock DCX running an unbalanced load. (Ditto the DEQ). Balanced is much better to my ears. Balanced to unbalanced done from the stock outputs via transformer sounds pretty good to me.
But transformer right out of the DAC sounds better. It's what I like and I can stand to listen to the thing without fatigue.
Is the difference "Night & Day"? No. But big enough in the right direction that I'd never want to go back. The Stock unit is "OK" when using the balanced outs.
These are COMPLETLY subjective results. And no controlled tests, either. So shoot me.
Am I deluding myself? Possible, but I doubt it.
But anyone is welcome to come to my house and have a listen. We can easily do a modded vs. stock blind test. If you can't hear a difference, or like the stock version better, that's OK! We will post the results here.
As far as clock upgrades - I have no comment - have not tried them. Don't know if I could hear them. Might be interesting to try.
You'll have to put me in the "cable huggers" camp on this one.
I do NOT like the sound of stock DCX running an unbalanced load. (Ditto the DEQ). Balanced is much better to my ears. Balanced to unbalanced done from the stock outputs via transformer sounds pretty good to me.
But transformer right out of the DAC sounds better. It's what I like and I can stand to listen to the thing without fatigue.
Is the difference "Night & Day"? No. But big enough in the right direction that I'd never want to go back. The Stock unit is "OK" when using the balanced outs.
These are COMPLETLY subjective results. And no controlled tests, either. So shoot me.
Am I deluding myself? Possible, but I doubt it.But anyone is welcome to come to my house and have a listen. We can easily do a modded vs. stock blind test. If you can't hear a difference, or like the stock version better, that's OK! We will post the results here.
As far as clock upgrades - I have no comment - have not tried them. Don't know if I could hear them. Might be interesting to try.
hehe, yeah these guys make it sound like we can not listen and we do not own good enough equipment. They may be right and they are just better then us but 90% of them can fit their whole house in my living room
Mornin' guys!
Sorry that I didn't respond earlier, I was passed out from smoking crack in my little bitty house.
For the sake of good conversation let me set this straight. On one side there is a group of people that:
1. didn't do any modifications
2. didn't perform any measurements
3. didn't share with the group any of their A / B findings or scientific measurements
4. didn't even read this thread partially or completely
5. didn't add anything of substance that will elevate overall quality of the discussion and knowledge
6. Demand that someone performs "scientific" measurement, even though that was already done and posted within the very same thread (Sorry that I am not going to search for you and post link)
On another side there are people that:
1. Did modifications and shared with the group their findings and experience
2. Did A / B measurements and listening tests
3. Posted their results here in the thread
4. Provided anyone with knowhow, schematics, kits, answers to group questions...
5. As added bonus this group of people are declared by the above group to be: delusional, crack smoking and because of all that prone to live in small house.
I do not think we have a good ground to carry out any debate unless group 1 comes back and does exactly what they demand to be done and what Group 2 already did. Scientific measurement and A /B tests. Only than I would consider that we have a good conversation. Without that it is just you telling us that off shelf DCX sounds very good to you, and you do not understand why we would change something that is as good, while you never had a chance to listen to it after mods are done.
Bonde Testing
Well AR2, you do have a point there in your last sentence!
I realized that I have not shared the testing I've done. Subjective, mostly, but controlled. No air tight procedures, I'm afraid. The Nobel committee will not be phoning anytime soon.
TEST 1 Transformers vs active. Maui Hawaii, circa 2007.
Thought I'd put one of my transformer output DACs to the test. Could I really tell the difference? Was it better or worse?
Up for test was my stock Sony CD/DVD/SACD player vs Citypulse DAC vs my transformer DAC (EDCOR iron, at the time).
I measured the FR and levels 20-20K on the o'scope. 50K termination. Sweeps burned to CD. All flat as a pancake except the transfos that rolled off a little at the bottom.
Built an A/B/C switch box. Retested levels at 100Hz, 1Khz, 10KHz. IIRC.
Had my wife (the blonde tester) connect the three sources so that I could not see. Wires under cloth. Listened to many cuts of music and some spoken voice.
Source C was a standout. Dull, lifeless. At least compared to A&B. Very easy to hear. But between A&B? No difference that I could hear. I could not even hear the glitch with the switch! As far as I could tell, identical. If I were not doing the switching, I would never have heard it.
But a funny thing started to happen. I noticed that at times my heart beat would speed up a little, my foot would start tapping, I would sit back. After a few tries I began to notice this effect was always with source B. I could not hear it, no way. But I could feel it. Would have easily put money on it.
The results? Yeah, you guessed. A=Citypulse, B=Transformers, C=Sony.
TEST 2 Portland, Oregon early 2009.
Having found that better transformers sound better, I went listening for the "best" I could find.
Up for test were - Onetics, Jensen, Edcor, Sowther and Auricaps direct.
All were put in plain white boxes with 9 pins connectors for easy connections and swapping. I even put ballast in the boxes so they weighed the same. Then lettered them so they could be identified, but only by letter.
Listened again. One was clearly bad, one was very direct but "lean" one was nice, one was great but with a tiny bit of edge and the final one had all the good and none of the bad. It's the one I kept coming back to.
Ranking?
Bad = Sowther
Direct but lean = Auricaps
OK = Edcor
Great = Jensen
All good, no bad = Onetics (very close between those last 2)
So there you go. My Rube Goldburg testing. Not bullet proof by any means, but at lest I did it! I did other tests, but for the sake of brevity, they are omitted here.
Well AR2, you do have a point there in your last sentence!
I realized that I have not shared the testing I've done. Subjective, mostly, but controlled. No air tight procedures, I'm afraid. The Nobel committee will not be phoning anytime soon.
TEST 1 Transformers vs active. Maui Hawaii, circa 2007.
Thought I'd put one of my transformer output DACs to the test. Could I really tell the difference? Was it better or worse?
Up for test was my stock Sony CD/DVD/SACD player vs Citypulse DAC vs my transformer DAC (EDCOR iron, at the time).
I measured the FR and levels 20-20K on the o'scope. 50K termination. Sweeps burned to CD. All flat as a pancake except the transfos that rolled off a little at the bottom.
Built an A/B/C switch box. Retested levels at 100Hz, 1Khz, 10KHz. IIRC.
Had my wife (the blonde tester) connect the three sources so that I could not see. Wires under cloth. Listened to many cuts of music and some spoken voice.
Source C was a standout. Dull, lifeless. At least compared to A&B. Very easy to hear. But between A&B? No difference that I could hear. I could not even hear the glitch with the switch! As far as I could tell, identical. If I were not doing the switching, I would never have heard it.
But a funny thing started to happen. I noticed that at times my heart beat would speed up a little, my foot would start tapping, I would sit back. After a few tries I began to notice this effect was always with source B. I could not hear it, no way. But I could feel it. Would have easily put money on it.
The results? Yeah, you guessed. A=Citypulse, B=Transformers, C=Sony.
TEST 2 Portland, Oregon early 2009.
Having found that better transformers sound better, I went listening for the "best" I could find.
Up for test were - Onetics, Jensen, Edcor, Sowther and Auricaps direct.
All were put in plain white boxes with 9 pins connectors for easy connections and swapping. I even put ballast in the boxes so they weighed the same. Then lettered them so they could be identified, but only by letter.
Listened again. One was clearly bad, one was very direct but "lean" one was nice, one was great but with a tiny bit of edge and the final one had all the good and none of the bad. It's the one I kept coming back to.
Ranking?
Bad = Sowther
Direct but lean = Auricaps
OK = Edcor
Great = Jensen
All good, no bad = Onetics (very close between those last 2)
So there you go. My Rube Goldburg testing. Not bullet proof by any means, but at lest I did it! I did other tests, but for the sake of brevity, they are omitted here.
I did do some modifications: I took the rack mounting brackets off.
I have considered upgrading the PS as Jan Didden did and I may in the future do this, if I'm sufficiently bored.
I did build this fancy balanced line receiver to connect the DCX to my 6 channel (unbalanced) amp:
But you guys wouldn't be interested in that - it's putting more of those filthy opamps in the signal path. Oh NO! What's that? Look close and you'll see the dreaded ceramic caps! Shocking! Hey, it works and sound comes out the mother, what do I care?
Have fun fellas.
Hey all. My 2 cents here.
Biggest OBSERVABLE change to my system (for the better) was a better DAC. More than any amp change or speaker (to an extent). So with this finding, I set out (based on opinions from this site no less) to purchase a better DAC, so I bought two items. A DCX 2496 and Benchmark DAC.
I was unimpressed with both (more so with the Benchmark). Since I spent over $750 for the Benchmark, I immediately sold it to some fine chap in England before I somehow got stuck with it. The DCX I still have and keep around to experiment with. It's kinda cool - all the things I can do with it - and believe the analog section has some redeeming value as a crossover (and 'cause I got it real cheap). I wouldn't mind messing with it, except it not high on my list at the moment (finishing my 6SN7 Aikido, CS4398 based DAC, JLH, Hiraga, Aleph X) now are my higher priority
Now I have two friends that absolutely LOVE their Benchmark DACs, and I respect that. I don't think anything less of their opinion on audio subject matter because they like it, and they respect my opinion for disliking it.
Makes me smile that my $50 CS4397 based DAC bested that Benchmark on my system tho.
I might add that the real reason I disliked the sound of the Benchmark can be attributed to those lame 5532 opamps, and absolutely believe a better opamp like the LM4562 or LM6172 would change my opinion of it. Too bad the 5532's were soldered, and didn't want to chance being out $750.
I don't hate opamps, btw. I do have a pair of Sescom Transformers I wanna play with...
Biggest OBSERVABLE change to my system (for the better) was a better DAC. More than any amp change or speaker (to an extent). So with this finding, I set out (based on opinions from this site no less) to purchase a better DAC, so I bought two items. A DCX 2496 and Benchmark DAC.
I was unimpressed with both (more so with the Benchmark). Since I spent over $750 for the Benchmark, I immediately sold it to some fine chap in England before I somehow got stuck with it. The DCX I still have and keep around to experiment with. It's kinda cool - all the things I can do with it - and believe the analog section has some redeeming value as a crossover (and 'cause I got it real cheap). I wouldn't mind messing with it, except it not high on my list at the moment (finishing my 6SN7 Aikido, CS4398 based DAC, JLH, Hiraga, Aleph X) now are my higher priority
Now I have two friends that absolutely LOVE their Benchmark DACs, and I respect that. I don't think anything less of their opinion on audio subject matter because they like it, and they respect my opinion for disliking it.
Makes me smile that my $50 CS4397 based DAC bested that Benchmark on my system tho.
I might add that the real reason I disliked the sound of the Benchmark can be attributed to those lame 5532 opamps, and absolutely believe a better opamp like the LM4562 or LM6172 would change my opinion of it. Too bad the 5532's were soldered, and didn't want to chance being out $750.
I don't hate opamps, btw. I do have a pair of Sescom Transformers I wanna play with...
Last edited:
Hey all, this is a rather large thread, and I must admit I have not read all 2000+ posts, but by the looks of it the best mods to do are to the output stage.
I was wondering what my options are with my current setup. I'm using a DCX for a simple BI-amped system, using 2 x Sure Tripath TC2000 TK2050 based amps, one for each speaker, and the amp will be mounted in the speaker. more info on these here:
http://www.diyaudio.com/forums/clas...onics-new-tripath-board-tc2000-tp2050-16.html
Now, the TC2000 datasheet says that the input impedance should be 2000ohms, and the output impedance of the AKM4393 is 600ohms. Now I'm hoping to wire the output pins of the AKM4393 directly to the XLR pins, bypassing the DCX output stage.
I'm guessing this should also fix the infamous 'egg-frying' problem? I've already insulated below the output PCB, so I'm 99% sure it's the ribbon cable causing my DCX to fry eggs!
Now, what are options with going from differential to single ended output?
I'm guessing there are 3 options here:
Op-amp buffer
Tube
Transformer
Now, I need some sort of volume control also, I was hoping to use either a simple pot or maybe a stepped ladder attenuator... I'm guessing stepped would be better, and I don't really need that many steps as I can use the DCX's menu to attenuate up to 10Db or so (my source has 16bit output).
Now, I'm not really sure which option is best, and often there is no best, but I was wondering has anyone already done what I'm planning to do - a combined output stage and volume control? I don't want to re-invert the wheel after all!
I'm kind of leaning towards a trafo based output right now as I think it will be the simplest and give good results, but I'm not really sure where to start. Will I need a 600:2000 trafo? Or can I use a 1:1 wound trafo and add resistors in order to please the amp and DAC with it's impedance? Also, is it possible to do away with the input capacitor on my amp is a trafo is used?
Suggestions welcome, the egg-frying isn't too bad right now, it's quite low level, and I'm using a pot between my DCX and amps right now, which makes it much less noticeable.
Finally, are there any issues have the output stage near the amp, and not near the DAC physically?
I was wondering what my options are with my current setup. I'm using a DCX for a simple BI-amped system, using 2 x Sure Tripath TC2000 TK2050 based amps, one for each speaker, and the amp will be mounted in the speaker. more info on these here:
http://www.diyaudio.com/forums/clas...onics-new-tripath-board-tc2000-tp2050-16.html
Now, the TC2000 datasheet says that the input impedance should be 2000ohms, and the output impedance of the AKM4393 is 600ohms. Now I'm hoping to wire the output pins of the AKM4393 directly to the XLR pins, bypassing the DCX output stage.
I'm guessing this should also fix the infamous 'egg-frying' problem? I've already insulated below the output PCB, so I'm 99% sure it's the ribbon cable causing my DCX to fry eggs!
Now, what are options with going from differential to single ended output?
I'm guessing there are 3 options here:
Op-amp buffer
Tube
Transformer
Now, I need some sort of volume control also, I was hoping to use either a simple pot or maybe a stepped ladder attenuator... I'm guessing stepped would be better, and I don't really need that many steps as I can use the DCX's menu to attenuate up to 10Db or so (my source has 16bit output).
Now, I'm not really sure which option is best, and often there is no best, but I was wondering has anyone already done what I'm planning to do - a combined output stage and volume control? I don't want to re-invert the wheel after all!
I'm kind of leaning towards a trafo based output right now as I think it will be the simplest and give good results, but I'm not really sure where to start. Will I need a 600:2000 trafo? Or can I use a 1:1 wound trafo and add resistors in order to please the amp and DAC with it's impedance? Also, is it possible to do away with the input capacitor on my amp is a trafo is used?
Suggestions welcome, the egg-frying isn't too bad right now, it's quite low level, and I'm using a pot between my DCX and amps right now, which makes it much less noticeable.
Finally, are there any issues have the output stage near the amp, and not near the DAC physically?
Last edited:
I did do some modifications: I took the rack mounting brackets off.
I have considered upgrading the PS as Jan Didden did and I may in the future do this, if I'm sufficiently bored.
I did build this fancy balanced line receiver to connect the DCX to my 6 channel (unbalanced) amp:
View attachment 149519
But you guys wouldn't be interested in that - it's putting more of those filthy opamps in the signal path. Oh NO! What's that? Look close and you'll see the dreaded ceramic caps! Shocking! Hey, it works and sound comes out the mother, what do I care?
Have fun fellas.
That board looks really neatly done.
Many pages ago Ergo did really extensive reports and tests and posted all the results. Frank did exceptional design on the digital side where he exchanged CS8420 digital receiver that even manufacturer admitted in its paper has fault. That meant that every unit off the shelf had a problem. Franks receiver board exchanged not just faulty CS840 receiver for CS8416, but added complete new very low jitter clock as well as SRC chip. That board turned DCX to be asynchronus DAC that completely eliminated any incoming jitter.
Many pages ago we shared the results and we valued what mod did the most and what did the least. In my experience the biggest difference was elimination of all circuitry after DACs and replacing it with Lundahl line transformers. On par with that I would say was digital input board mod by Frank that made really big difference in sound. I exchanged DAC chips AKM 4393 for 4396 and I would rate that as really minimal advancement, but still noticeable. I exchanged bypass caps on chips for Oscons, and I have no idea if I made any improvement. Most likely I did but I cannot say that I could hear that.
The last and least was in my case power supply mod, that really didn't change much, but keep in mind that I do not have analog stage, so my power supply mod affected only digital section.
So if you decide to do any mod, here are my experiences confirmed by several other members. That will tell you where to except the biggest impact and what to expect.
source (Rs), load (Rload), input (Rin) output (Rout) impedances confused?
I posted this somewhere else, but it seems I might get some better answers here.
I assembled a class D amp to control my subwoofers (with two of these modules). I built them with no connection between the signal and chassis grounds, which the manufacturer told me would lessen the chance of noise issues (not ground loop possible).
I have a DCX2496 on order to use as follows: receiver preamp (unbalanced) > DCX (balanced) > amps (unbalanced).
Based on what I have seen in this thread, it seems that the best interconnect method would be the following (figures taken from a Rane note document): #17 for the preamp>dcx and #6 for the dcx>amps with a 15-20k ohm resistor between pin 3 and 1 instead of a direct connection. Is this correct?
The amps are rated at up to 3V for the input level. Am I going to have any issues using the DCX here with the amps wide open (they do not have any volume control)? I am interested in Jan's output stage mod but will see how things work with the stock unit first.
-Doug
I assembled a class D amp to control my subwoofers (with two of these modules). I built them with no connection between the signal and chassis grounds, which the manufacturer told me would lessen the chance of noise issues (not ground loop possible).
I have a DCX2496 on order to use as follows: receiver preamp (unbalanced) > DCX (balanced) > amps (unbalanced).
Based on what I have seen in this thread, it seems that the best interconnect method would be the following (figures taken from a Rane note document): #17 for the preamp>dcx and #6 for the dcx>amps with a 15-20k ohm resistor between pin 3 and 1 instead of a direct connection. Is this correct?
An externally hosted image should be here but it was not working when we last tested it.
The amps are rated at up to 3V for the input level. Am I going to have any issues using the DCX here with the amps wide open (they do not have any volume control)? I am interested in Jan's output stage mod but will see how things work with the stock unit first.
-Doug
to enable the DCX to perform well (to specification) the analogue inputs must get the clipping leds almost on, but never quite lighting.
This will push the digital circuits to near 0dBfs.
At this level the output will drive your power amps way too high, probably into clipping and your speakers will be unlisten-able.
You must fit some form of 6channel attenuation between the DCX and the 6amplifiers.
What you must NOT do is reduce the analogue input to bring the digital processing to a maximum throughput of -10dBfs nor -20dBfs nor -30dBfs nor -40dBfs.
Yes, that is the typical range of attenuation you may need. somewhere between -10dB and -40dB.
This will push the digital circuits to near 0dBfs.
At this level the output will drive your power amps way too high, probably into clipping and your speakers will be unlisten-able.
You must fit some form of 6channel attenuation between the DCX and the 6amplifiers.
What you must NOT do is reduce the analogue input to bring the digital processing to a maximum throughput of -10dBfs nor -20dBfs nor -30dBfs nor -40dBfs.
Yes, that is the typical range of attenuation you may need. somewhere between -10dB and -40dB.
Last edited:
Confused is not a bad way of putting it, hitting a brick wall would be another!
Line-level impedances just seem to make no sense to me.I know ohm's law inside out, and am more than comfortable when working with speaker impedances, bridging amps, etc. but when it comes to line level impedances and designs I'm totally lost.
I think I'm going to have to do some more reading up, but would I be right in saying the Rload would be equivalent to a speakers impedance (eg 8ohms) and Rsource is equivalent to an amplifiers load rating (upto 4ohms)?
The most effective and unobtrusive mode is using high-quality 80-100V trafos of 5-10Watt.
For unballanced amps could be used also such kind of device:
APart-Audio - Quality Audio Solutions - 19-VOL650
My choices in order of increasing "betterness" (and cost)
The way I do it is with stepped attenuators on the inputs of the amps to set the gains where they need to be. I use digital volume control (32bit) in my player, so overall volume is taken care of.
Running attenuation at the amp inputs takes care of the noise floor because I'm forcing the DCX into the top of its range.
- multi gang pot. 4 gang for a stereo 2 way, 6 gang for stereo 3 way of L/C/R 2 way.
- Multi gang stepped attenuator
- Mulitap transformers. Great, but very $$$
The way I do it is with stepped attenuators on the inputs of the amps to set the gains where they need to be. I use digital volume control (32bit) in my player, so overall volume is taken care of.
Running attenuation at the amp inputs takes care of the noise floor because I'm forcing the DCX into the top of its range.