Behringer DCX2496 digital X-over

[Pano]

Strange.....
My Altec 291 drivers (on 803 horns) are roughly as sensitive as the 909s. I hear NO hiss at all from the stock DCX outputs - unless I put my ear right up to the horns.

I'm running attenuation in front of the amp (volume control) and after the amp (L-Pad) to keep the levels reasonable. The DCX outputs are set to +3dB.

Are you running straight from your DCX into the amp? Any attenuation or volume control before our after the amp?

[Cobra2]

Setup: source->Pre (CS3310 & jFet-buffer)->DCX(full analog output volume)->PowerAmps(LM3886)->Altec 515-8G-HP(x4) + Altex 909-8B(x2)
No L-pads

Arne K

[mige0]

Wow - someone with stock unit should do a THD&N measurement over frequency then...
Thanks, Cobra2 .


Michael

[Gary P]

Originally Posted by mige0

Wow - someone with stock unit should do a THD&N measurement over frequency then...
Thanks, Cobra2 .


Michael

This measurement is on my list of things to do. Just need more hours in a day to get caught up on all the actively running projects - audio, home improvement, race car, day job...

Gary

[Legis]

Originally Posted by mige0

Wow - someone with stock unit should do a THD&N measurement over frequency then...
Thanks, Cobra2 .


Michael

THD% has been measured over frequency already in that many times mentioned Germany test: Google Käännä

The THD% keeps constantly below (big deal) 0,01 and does not vary between 20hz - 20 000hz as you can see from the first picture from light blue/turqoise line from previous link. There has spread some rumour (based on flawed measuring) that DCX would add THD% in the high frequencies, but obviously this is not the case. (underlined so that this misconception would go away)


I have measured DCX with RMAA using E-MU 0404 PCI sound card (DAC = AK4395). Result were very good when comparing to E-MU. I tested it with two 1m (3,3 feet) long good RCA cables.

For some reason I always get 10dB worse Noise level with everything (incl: E-MU 0404, E-MU 1212M, E-MU 1616). I have tested all these cards and compared my result to the various test that are on the net, and always the same. With my setup noine of the cards don't achieve the same values as they have achieved in other valid tests. Maybe the PSU and Graphics card inside the PC that are very near the sound cards tend drop the DR and raise noise. There is an unexplained peak with both DCX and EMU in 3kHz, although DCX is not inside my computer .

Keeping in mind that these measures do not make justice for either of the measured candidates, the measured noise level is:
E-MU: -106,4dB
DCX: -96dB.

If I add 10dB to both they are they very near the specs that the manufacturer promises (DCX: 109dB and E-MU: 116dB). With E-MU I naturally need only one metre of cable from outputs to inputs, that might explain some of the "extra 3dB" with DCX comparing to E-MU.

Other values:

THD%:
E-MU is 0,0008%
DCX 0,0083%.

IMD% + noise:
E-MU: 0,0025
DCX: 0,010%

With the results from that Germany test and these measurements you can deduct the THD% + noise, at least somehow. It does not vary greatly over frequency, and is actually quite good. The promised 109dB dynamic range and -109dB noise are achievable, and THD% + noise would then be something like 100 - 105dB maybe? Maybe you can estimate it better.

If you paste some of your results please paste alse the result of the sound card that was used in measuring.

[mige0]

Hi Legis RCA asymmetric connects is not the best to use for precise measuring - change to symmetric and you will most certainly get more reliable results. Also - depending on balanced versus mere complimentary out, you waste some dB in S/N with asymmetric operation due to different summing of correlated and uncorrelated signals.

Anyway - the linked AP measurements are flawless and do not indicate *any* noise that's worth a mod (reflecting my own feeling back then too).

Wonder what Arne possibly has overlooked or if there are such differences in DCX samples, which is hard to believe ...

Lets wait what Gary will come up with.

Michael

[Pano]

Originally Posted by Cobra2

DCX(full analog output volume)->PowerAmps(LM3886)->Altec 515-8G-HP(x4) + Altex 909-8B(x2)

Hmmmm.... OK. So no attenuation between the DCX and the amps? None after the amps.

That setup does not work for me here. Too noisy in my rig.

My signal flow is like this:

• AES digital in / 0dB gain.
• Low Out - stock DCX pins 1 & 2. ->IcePower amp with stepped attenuator. ->Altec 416A
• Mid out - Transformer on AKM chip +4dB digital gain. -> Stepped attenuator on Tripath TA2024 amp -> L Pad ->Altec 291 driver
• High out - Transformer on AKM chip +3dB digital gain. -> Volume pot on Tripath TA2024 amp. -> Heppner horn tweeter.

The reason I use an L-Pad on the mid horns is for S/N ratio. The L-Pad allows me to drive the midrange amp hot enough to keep up "out of the dirt". The damping offered by the L-Pad doesn't hurt, either.

I think I need an L-Pad on the tweeters, there is some noise there. It comes from the amp, not the DCX.

The reason for the volume control on the amps is to adjust level. I need so little gain that the DCX signal must be attenuated before hitting the amps. It also keeps the DCX outputs up "out of the dirt" for better S/N. For general listening the output VU meters run about 1/2 scale, or more.

A typical log pot has about 20dB attenuation at mid scale - 12:00. If I can use that, I've just gained 20dB S/N.

[Pano]

Originally Posted by Legis

There has spread some rumour (based on flawed measuring) that DCX would add THD% in the high frequencies, but obviously this is not the case.

Well my measurements certainly showed rising distortion on the stock DCX outputs. THD is less useful than looking at the harmonics themselves. In dual-tone and multitone tests, it's obvious that the stock output circuit contains more distortion at the high frequencies.

It will be interesting to see what Gary finds.

[SY]

Originally Posted by panomaniac

My signal flow is like this:

• AES digital in / 0dB gain.
• Low Out - stock DCX pins 1 & 2. ->IcePower amp with stepped attenuator. ->Altec 416A
• Mid out - Transformer on AKM chip +4dB digital gain. -> Stepped attenuator on Tripath TA2024 amp -> L Pad ->Altec 291 driver
• High out - Transformer on AKM chip +3dB digital gain. -> Volume pot on Tripath TA2024 amp. -> Heppner horn tweeter.

How do you do overall volume? I can't imagine you adjust each of the amps individually every time you want to change the volume...

[Pano]

Gee Sy, that's what the menehunes are for.

It's a good question, actually. I use the volume control in my media player. The volume is done in the software at 32 bits, then output at 16 or 24 bits as I choose. It seems to work well - and I use ASIO drivers. So far, I have not been able to measure or hear any bad effects doing it this way. Gain structure is set so that most tracks play at 60-75% of the player volume. Classical tracks get 100% as the average level is lower.

One could use a 6 gang volume control after the DCX. I have a 6 gang motorized ALPs and some multi-gang stepped attenuators. So far, I have not felt the need for them.