This is true. I am trying to get to the source of this hiss and remove it. It would be nice if my mains' power amps would have gain control, but such hifi power blocks don't usually have them. Besides, potentiometer is just another sound detoriating component in the signal path
. If there is any way to remove/reduce the hiss that comes from the DCX, I would gladly try it.Do you see any benefit in trying a linear PSU or change a better voltage regulators to regulate DAC's power.
Quote from audiosmile:
Is this hard to do by one's self?
More transformer measurments
Ok, I spent a bunch of time yesterday figuring out how to configure my test system for measuring transformers in a manner that will line up with mige0's measurements. By re-routing some cables and one mod to my sound card measurement input buffer I can test the transformers at 1V, .1V and .01 volt just by flipping switches. Nice and repeatable. Here is the basic test setup.
First, the loop back measurements to show the limitations of the test system.
@1Vrms
@.1Vrms
@.01Vrms
The first transformer measured was the Lundahl LL1676 amorphous core line level transformer configured for 1:1
@1Vrms
@.1Vrms
@.01Vrms
Next up was a TRW UTC PC0-25 2K/500 to 600/150 ohm nickle core transformer found in a surplus shop. I've used these in several projects so far. This transformer is configured for 2:1 ratio, the way I've been using them in my alternate input for the Behringer.
@1Vrms
@.1Vrms
@.01Vrms
Next is the Sowter 3575 10K bridging transformer
@1Vrms
@.1Vrms
@.01Vrms
Lastly the Sowter 9100 20K-20K interstage transformer. So for this is the transformer to beat in the distortion race...
@1Vrms
@.1Vrms
@.01Vrms
Gary
Ok, I spent a bunch of time yesterday figuring out how to configure my test system for measuring transformers in a manner that will line up with mige0's measurements. By re-routing some cables and one mod to my sound card measurement input buffer I can test the transformers at 1V, .1V and .01 volt just by flipping switches. Nice and repeatable. Here is the basic test setup.
An externally hosted image should be here but it was not working when we last tested it.
First, the loop back measurements to show the limitations of the test system.
@1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.01Vrms
An externally hosted image should be here but it was not working when we last tested it.
The first transformer measured was the Lundahl LL1676 amorphous core line level transformer configured for 1:1
@1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.01Vrms
An externally hosted image should be here but it was not working when we last tested it.
Next up was a TRW UTC PC0-25 2K/500 to 600/150 ohm nickle core transformer found in a surplus shop. I've used these in several projects so far. This transformer is configured for 2:1 ratio, the way I've been using them in my alternate input for the Behringer.
@1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.01Vrms
An externally hosted image should be here but it was not working when we last tested it.
Next is the Sowter 3575 10K bridging transformer
@1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.01Vrms
An externally hosted image should be here but it was not working when we last tested it.
Lastly the Sowter 9100 20K-20K interstage transformer. So for this is the transformer to beat in the distortion race...
@1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.1Vrms
An externally hosted image should be here but it was not working when we last tested it.
@.01Vrms
An externally hosted image should be here but it was not working when we last tested it.
Gary
Thanks Gary, nice work!
Again I see that rising H3 on the low end that I've seen on every signal transfo I've tested so far. The smaller the core, the more the rise. Size does seem to matter.
Interesting to see how well behaved they all are at 0.01V. Very clean down there.
I also see the rising H3 on the Sowter 3575. Maybe that's what I heard and didn't like on my Sowthers.
Any further thoughts on your measurements?
Again I see that rising H3 on the low end that I've seen on every signal transfo I've tested so far. The smaller the core, the more the rise. Size does seem to matter.
Interesting to see how well behaved they all are at 0.01V. Very clean down there.
I also see the rising H3 on the Sowter 3575. Maybe that's what I heard and didn't like on my Sowthers.
Any further thoughts on your measurements?
Hi Pano,
I don't know what to think so far. Once again, what one measures and what one hears are 2 different things. Of the transformers in this set I have many listening hours on the LL1676, the 3575, and the UTC.
My personal opinion- The Lundahl was harsh and grainy, the Sowter was much smoother but not very clear in the way that low level details were missing. The best of the 3 is the little UTC. Distortion measures higher but they have a very transparent sound to them. I use them as outputs in the phono preamp and the modified input of the Behringer.
Gary
I don't know what to think so far. Once again, what one measures and what one hears are 2 different things. Of the transformers in this set I have many listening hours on the LL1676, the 3575, and the UTC.
My personal opinion- The Lundahl was harsh and grainy, the Sowter was much smoother but not very clear in the way that low level details were missing. The best of the 3 is the little UTC. Distortion measures higher but they have a very transparent sound to them. I use them as outputs in the phono preamp and the modified input of the Behringer.
Gary
Yeah Gary, I remember seeing those at your house. Small, but the sound was nice.
I've been trying to come up with some sort of technique to demonstrate loss of fine or low level detail. Bud and I were talking about it. Basically I've just been recording a musical signal and comparing the original file to the recording, bit by bit. It's proving very hard to do!
I've been trying to come up with some sort of technique to demonstrate loss of fine or low level detail. Bud and I were talking about it. Basically I've just been recording a musical signal and comparing the original file to the recording, bit by bit. It's proving very hard to do!
Great measurements Gary – thanks a million !
I guess you are right that size really matters at least for the "measurable" distortion - in our case it translates to bigger core (reducing flux > down in th more more linear region)
If you have a close look – 3rd is dominant with all transformers at low frequency distortion rise
Again - thanks for linking to your findings.
Not having your auditioning experience with such a veriety of transformers – very much appreciated !
Michael
Last edited:
New test signal needed
That's the problem we have with measuring. Steady state and swept sines don't challenge the equipment nearly as much as real acoustic sound does. Computers are fast enough now to do the before and after comparison. We need a new test signal. Maybe something like a multi tone square wave would give results that correlate better with listening tests although the ideal way would be to compare before and after of a real audio sample that you hear differences on. But even then, you may find the limits of the DACs/ADACs, buffers, cables, of the measurement hardware before you reach the limits of your ear/brain.
24 bit, symmetric IN and symmetric OUT - external is my favorite - but seems there are more and more internals that are stunning.
There are a lot available nowadays - actually it might be even hard to find 16bit anyway
Check out good driver support and compatibility with ARTA or whatever you use!
Would have recommended the Terratec Phase 24FW as this one is now for around EUR 100.- but have heard that there can be compatibility problems on the fire wire hardware level.
I use Mackie Onyx 400f but this one is outdated (and was at the expensive side anyway)
Michael
I prefer external soundcards too for a lot of reasons. I was looking at the EMU 404M USB, but I think it doesn't have an adjustable mic input
Do you think it could be good for this kind of measurements?Thanks!
Ciao
Paolo
@Legis: OK, I see your problem. I'm so used to pro power amps that I tend to forget that hi-fi power amps don't have volume controls. Heck, I even have some here that I helped design! That's because they are made for typical hi-fi levels like 1V. Most preamps used to drive them will be quiet enough not to be heard.
So your best solution is a voltage divider, as SQL mentions. Has the same attenuation functions as a pot, but it's just a few resistors. You could even go crazy and use premium resistors if you like. You're not going to hurt the signal. BTW, what levels do you see on the DCX meters when you are playing at your highest comfortable volume?
@Sendler: Quite agree! I can use a swept sine and see that two circuits have very similar frequency response, maybe even similar THD. But to me they don't sound the same. Why? Well I can see some of why that might be in the harmonic content. Harmonics are said to be somewhat important in music. But what about the other stuff I "think" I hear? Like low level detail, spaciousness, attack and decay. How do you measure that?
That's why I've begun trying to compare the original wave file to a recorded version of itself. If the re-recorded version is identical to the original, why worry? It's not easy to do, there are a number of hurtles. The 1st hurtle is amplitude - you want that to be exact. The second hurtle is time. That's a little easier, as I can put a marker in the file for time alignment.
Third hurtle is the limitations of the ADC used to record. And I just don't know what those limits are.
I've used 3 simple signals so far. A sawtooth waveform, a single note from a solo cello and a single note from a bass drum. If I do it right, the differences between original and recorded are very tiny indeed. It may be time to move on the more complex waveforms, or to improve my technique. Or both.
So your best solution is a voltage divider, as SQL mentions. Has the same attenuation functions as a pot, but it's just a few resistors. You could even go crazy and use premium resistors if you like. You're not going to hurt the signal. BTW, what levels do you see on the DCX meters when you are playing at your highest comfortable volume?
@Sendler: Quite agree! I can use a swept sine and see that two circuits have very similar frequency response, maybe even similar THD. But to me they don't sound the same. Why? Well I can see some of why that might be in the harmonic content. Harmonics are said to be somewhat important in music.
But what about the other stuff I "think" I hear? Like low level detail, spaciousness, attack and decay. How do you measure that? That's why I've begun trying to compare the original wave file to a recorded version of itself. If the re-recorded version is identical to the original, why worry? It's not easy to do, there are a number of hurtles. The 1st hurtle is amplitude - you want that to be exact. The second hurtle is time. That's a little easier, as I can put a marker in the file for time alignment.
Third hurtle is the limitations of the ADC used to record. And I just don't know what those limits are.
I've used 3 simple signals so far. A sawtooth waveform, a single note from a solo cello and a single note from a bass drum. If I do it right, the differences between original and recorded are very tiny indeed. It may be time to move on the more complex waveforms, or to improve my technique. Or both.
I think this kind of device would be good for use with sound cards when measuring.
Sound Card Interface / AC RMS voltmeter
Soundcard Interface
Jim
Sound Card Interface / AC RMS voltmeter
Soundcard Interface
Jim
Yesterday when I tried the DCX with full -15db attenuation at the outputs (inputs at 0dB) I got -30 and maybe -20db led to light for a short period. My preamp would have something like +10dB more, but I'm very sure I won't never listen to higher levels than that at least with music. Maybe with movies, and that is very uncertain maybe.
I read somewhere that driving DCX at the -3dB - 0dB will actually cause more distortion than with lower gain. I'm using analog inputs so I won't lose bits either if the gain is low. I quess there is no harm done with this low input level setup?
Btw, I tried it again today with full attenuation at outputs and silenced my whole room. The hiss has actually reduced quite a bit, almost inaudible to listening point. If listen straight from tweeter the difference seems smaller than from the listening point. I think I could be happy with full attenuated outputs if there ain't no "final solution" to hiss that would cure the hiss even with normal gain.
I got LM4562s today from the UPS but my silver solder and other equipment I ordered have not come yet. I read somewhere that Behringer have actually glued stock op-amps to their place, is this true? I dunno if this comment was aimed to DCX and it was from year 2008.
edit: Input signal levels corrected, the latter values were taken from FBQ2496's scale by mistake
Last edited:
Well that's interesting. You get less hiss when you use the digital attenuation on the output channels of the DCX? If so, I'd say the hiss is coming from the inputs, not the outputs.
What happens if you disconnect the inputs to the DCX? Or mute the inputs, or mute the outputs? What differences do you hear? That would help trace the hiss.
FWIW, I've never had hiss problems with the DCX, but I run attenuation at the amps. (Stepped attenuator or pot).
What happens if you disconnect the inputs to the DCX? Or mute the inputs, or mute the outputs? What differences do you hear? That would help trace the hiss.
FWIW, I've never had hiss problems with the DCX, but I run attenuation at the amps. (Stepped attenuator or pot).
Most noisy --- > least noisy:
tied 1) and 2): DCX ON, input and output gain 0dB OR inputs disconnected
3): DCX ON, outputs or inputs attenuated
4): DCX ON, mute ON
tied 5) and 6): DCX turned OFF or outputs disconnected from DCX
So, unpluggin inputs from DCX wont affect hiss in anyway. On the other hand unplugging outputs have the most impact, which is logical since my amps are very, very silent by themselves.
BTW, muting DCX and setting outputs to -15dB have same effect. If I mute DCX while outputs are at -15dB, I won't hear any difference (ear near the tweeter).
Noise if definitely coming from DCX.
Again, i must say that the hiss is not loud, but it is still there. It is not even irritating, but I if I could choose I would get as silent system as my bare amps are. You get the point...
This was good test actually, now I know that it won't do any good to drop the gain below -15dB. I tested that the point that I can not hear difference between "attenuated hiss" and "total mute hiss" is somewhere between -10dB to -12dB actually.
And since removing inputs won't do anything to the hiss, the current divider resistor thing in the input cable won't help.
Here is some interesting story about how changing stock op-amps (4580) to LM4562s reduced the output noise/hiss dramatically. It was not DCX that was modded though. Maybe there is hope after all
http://www.intensemetal.com/~aortizjr/projects/ADA8000/behringer-upgrade
http://www.intensemetal.com/~aortizjr/projects/ADA8000/behringer-upgrade
I prefer external soundcards too for a lot of reasons. I was looking at the EMU 404M USB, but I think it doesn't have an adjustable mic input
Thanks!
Ciao
Paolo
Looks good -
seems to have even a knob for gain adjustment at the upper side ...
Michael
What do you mean by Mute On?
There are two mutes: mute at power up is done by the mute transistors that shunt the signal after the LP filter; mute with the Mute button is done in the DACs.
If you're talking about muting with the Mute button, then the noise is coming from the DACs, at the latest, or from something further upstream in or before the DCX. However, since you indicated that you still have full hiss with the DCX on but inputs disconnected, the noise, or at least some of the noise, must be coming from within the DCX. This basically narrows it down to the ADCs, the DACs, or the input buffers.
I'd try next disconnecting the A input, and switching the unit to Digital In. If the noise goes away, or mostly goes away, then you'll know it's in the input buffers or ADCs. If the noise stays the same, it's in the DACs.
Cheers,
Paul
There are two mutes: mute at power up is done by the mute transistors that shunt the signal after the LP filter; mute with the Mute button is done in the DACs.
If you're talking about muting with the Mute button, then the noise is coming from the DACs, at the latest, or from something further upstream in or before the DCX. However, since you indicated that you still have full hiss with the DCX on but inputs disconnected, the noise, or at least some of the noise, must be coming from within the DCX. This basically narrows it down to the ADCs, the DACs, or the input buffers.
I'd try next disconnecting the A input, and switching the unit to Digital In. If the noise goes away, or mostly goes away, then you'll know it's in the input buffers or ADCs. If the noise stays the same, it's in the DACs.
Cheers,
Paul
Hi Pano , Legis
All that hiss problems gets mute once you do signal level optimization throughout the chain.
Meaning - measure at what signal level the amp is gonna clip with the DCX at full signal level and set a voltage divider in front of the amp accordingly - no need to mod DCX at all.
If you do that step by step for any device working backwards the signal chain - well thats close to the best one can have.
Michael
All that hiss problems gets mute once you do signal level optimization throughout the chain.
Meaning - measure at what signal level the amp is gonna clip with the DCX at full signal level and set a voltage divider in front of the amp accordingly - no need to mod DCX at all.
If you do that step by step for any device working backwards the signal chain - well thats close to the best one can have.
Michael
Yes, the hiss reduced by half by setting the input mode from analog to ASE/EBU. This narrows the hiss to analog in op-amps, ADC and analog input buffer?
The hiss might be completely normal in DCXs. It might be a feature of a unmodded DCX? I find out when my other DCX arrives later this week.