DCX2496 Level Problem...how to confirm

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
My DCX recently seems to be providing different levels on the outs when they are set to the same value differing by as much as 6dB. For example output when 5 and 6 (my left and right tweeter outputs) are set to 0dB, one plays louder.

I've done several tests to eliminate drivers, cables, amps and preamps but how can I test the DCX on the bench and verify with DMM?

Has anyone RMA'd their DCX?
 
While trying to isolate the problem I switched from the default 1,2,3 / 4,5,6 to 1,3,5 / 2,4,6. This only helped to suggest that the issue pertains to specific output channels and not the DCX configuration.

I can get pink, white, sine, triangle, square, etc. from TrueRTA.

If I were to guess I would think I would turn off all filters and crossovers for all channels and set all input and output levels to 0.

Then I would input a sine wave at maybe 1k and measure the output by using my DMM on pins 1-2 and then 1-3 which should provide the same value. I'm not really certain how to measure that, VAC on my DMM?

Then I'd repeat this for each channel and for a handful of other frequencies.
 
DIYNewbie ..... i see you have a DEQ along with your DCX .. does it daisy chain and control the crossover with the serial link ? I want to buy one hook it all up and let it auto level my room THEN play with it ... is that a possibility ? Have you done anything similar ?:confused:
 
Madmike2 said:
Why dont you hook up a Mic and let the auto setup function take care of all the annoying stuff :D ?

I don't believe the DCX has that kind of auto setup. It just does that for time alignnment. Maybe you are confusing it with the DEQX?

I've unbound the channels so I can adjust each one individually and get both speakers measuring the same but I went to a lot trouble to build a fixed attenuator for each channel so I could run with all my levels at 0dB.
 
Madmike2 said:
DIYNewbie ..... i see you have a DEQ along with your DCX .. does it daisy chain and control the crossover with the serial link ? I want to buy one hook it all up and let it auto level my room THEN play with it ... is that a possibility ? Have you done anything similar ?:confused:

I have not tried anything like this.. The problem is that the DCX has a rs232 control port, wheras the DEQ's port is MIDI..

The method I used is as follows :

1) set cross over points for each speaker on DCX
2) check polarity with DCX auto-align function selecting only the "polarity" option.
3) time align drivers with DCX auto-align function selecting only the "short delay" option.
4) run auto EQ function on the DEQ to equalize the room response (be sure not to set below the roll off of the speakers or you'll en up with a really nasty boost on the bottom which degrades the sound.

I'm connecting mine Digital into the DEQ, Digital between DEX and DCX which IMHO adds the least amount of noise...

Hope This Helps

--Chris
 
The outputs are balanced; are you using them balanced or single ended?

If you are using them single ended, and one channel has the unused output connected to ground, but the other channel has it left floating, then there will be a difference in level.

Just a thought, but it's an easy fix if that is your problem :)
 
All you need for a source is a signal generator or sinewave test tone from a test CD of some sort. Any frequency will do.
Program your DCX as a direct feedthrough with no processing on all six channels. Use your DMM set to read AC volts and take some readings.

Your external output configuration schematic looks fine. It's configured to attenuate 6db on both the + and - outputs and it's not connecting pin 1 to pin 3 (which is good.)

All six channels should read exactly the same measuring between pin 2 or 3 to ground (pin 1.) If they don't then you've got some kind of internal hardware problem.

Cheers,

Davey.
 
Hi,
my CD player, outputing a 0db maximum of 2Vac, needs +8db on the input gain setting to be just short of clipping on the output leds (+9db clips occasionally on some CDs but I do not hear the clip?). This is way below 9.75Vac.
Q1.
When DCX is set to 0db on all ins & outs.
Is the gain 0db? i.e. times one.
Q2.
When set to +8db on the input.
Is this an output of 6Vac when driven to 0db?
Q3.
The maximum input and output is +22dbu (9.75V). Is this across pins 2 & 3 or pin 1 to 2 and pin 1 to 3?
Q4.
Is the +22dbu = 9.75Vac or is it = 9.75Vpk or is it = 9.75V peak to peak?
 
Andrew,

You've got a lot of questions there. :)

Let's talk just AC volts RMS. If you program the DCX with no processing at all and 0db settings on all sliders the DCX gain will be unity (0 db) if utilizing balanced inputs/outputs. It will also be unity in an unbalanced configuration, but only if you use an XLR/RCA adaptor on the output jack that connects pin 1-3. If you pickup the signal directly from pin 2 (leaving 3 floating) and pin 1 you'll drop 6db going through the DCX. Example: 2 volts RMS input yields 1 volt RMS output.

The "clipping" point of the DCX is a bit less than the rated specification....at least on my unit. The output waveform will start clipping with inputs of approximately 8.4 volts RMS (unbalanced) which yields about 4.2 volts RMS on the outputs. (The red clipping indicators are fairly accurate and they'll be flashing or close to it at this point.) This "clipping" is not originating in the analog output sections, but farther upstream.

So, if you apply a 6db external pad (as ultrachrome did) you should see a 12db total reduction relative to the input.


In your case I suspect your CD player is outputting a bit more than 2 volts RMS for 0dbFS recorded data. (One of my Sony players is like this.) However, assuming a 2 volt signal you should be able to bump up the input gain settings by at least 12db before the onset of clipping. So, your situation doesn't quite make sense to me. Hmmm.

Anyway, initially it seemed most users were operating the unit well below the levels they should have, but now some users are obsessing a bit too much and pushing it to the limit. I don't think you need to set your input gains anywhere above 0db. It really doesn't help the situation much anyway.

+22dbu is 9.75 volts RMS.....13.78 volts peak.....27.5 volts peak-to-peak.

I hope that answers your questions. :)

Cheers,

Davey.
 
I think one other thing you need to consider is that you're slicing the energy of th incoming signal into three seperate bands.

If you have 0 dbU input on channel A, and split it into 3 channels the sum of the 3 channels will add up to 0dbU, not 0dbU on each channel. Do a simple experiment, send in pink noise to your DCX, with all the channel gains set to 0db. Adjust the level till you get 6 green bars on A/B inputs. You're outputs will be around 2 green leds (assuming almsot even split octaves)

I've been trying to work on this a bit on my own.. I think you might be able to get better signal to noise levels but adjusting the individual gains on the DCX so that channel 1-6 is at 0dbU with a 0dbU input signal on A/B (this would give you the best signal to noise ratio/dynamic range at the D/A convertor.. You would just have to really turn down the gains on your amplifiers

--Chris
 
Yeah, when you start applying filters and looking at the indications from the green and amber leds it can get a bit more confusing, but the red, clipping indicators are fairly well defined however....at least the output ones. :) They're simply looking for peaks within the waveform that are outside the defined limits. It doesn't matter how you have the unit setup or whether you're using noise or sinusoidal inputs. If that red light is flashing the waveform is clipping.

Also, on a side note, filters change the crest factor of waveforms (higher peak level, same average level) so it's possible that when you apply a filter or switch to a steeper one you might transition into a clipping condition. It's something to be aware of, but probably not a concern for most users.

Cheers,

Davey.
 
Finally, a night off!

So I measured my outputs. Here are the conditions:

- DCX settings all zeroed out. Input/Output levels at 0dB
- Input A routed to outputs 1-6. (also tested using input B to 1-6)
- TrueRTA sending 1kHz sine wave to preamp
- Preamp feeding input A using RCA-XLR cable with tip on pin 2 and ring on 1 and 3.
- DMM on ouput pins 1-2 and alternately 1-3

Output 1-2 / 1-3 (VAC)
1 .590 / .608
2 .642 / .553
3 .603 / .593
4 .618 / .574
5 .614 / .583
6 .587 / .607

Tip to ring of sound card output was .332 VAC.

I don't have a frame of reference. Is this unit an RMA candidate?

Sanity Checking:

After noticing this problem a month ago, I changed my input/output settings from stock. Since I could not figure out how to reset the device I decided to drastically change the config. So instead of my tweeters being on 3 and 6, they are on 5 and 6. In other words, the left channel feeds the odd outputs, right feeds the evens.

So using TrueRTA I measured pink noise with a mic but used the same preamp channel and the same amp channel and tweeter for each test, setting the output levels for 5 and 6 both to 0dB. Formerly they were 5dB apart. So I physically moved the input and output cables for each test.

Looking at TrueRTA, the output of 6 is lower than 5 by:
6dB @ 5kHz
5dB @ 10kHz
3dB @ 20kHz

So it's not just a level issue but also looks like a frequency response issue.

I feel like I'm going crazy or doing something wrong. Is this normal? Should I expect better performance than this?
 
I don't see any problem with those readings. The worst case channel difference is between 2 (.642) and 6 (.587)....that's 20*LOG(.642/.587) equals 0.78db. That does seem a bit worse than expected, but still way less than what you noted previously. (I'll check mine and see if it's similar.)

It's still not clear to me where your problem is.

Why don't you PM me and we can talk further and try to figure it out? Or if you're in the Bremerton area sometime stop by and I can check it for you.

Cheers,

Davey.
 
So I regrouped this morning and used TrueRTA and the frequency sweep function.

Sound card out to Preamp
Preamp set to 0dB to inputs A or B
DCX zeroed out with on input feeding all outputs
Same output cable used on each output, tested one at at time.

I calibrated with my preamp in the loop as I found that my preamp is -3db @ 20kHz. I'll have to figure out why later.

With the preamp in the loop but calibration only involving the sound card, the sweeps of outputs 5 and 6 match what I saw when I did the same test with pink noise and a mic.

THe attached screenshot is with the preamp response calibrated so this should be a better representation of the DCX's true performance.

Here you can see the vast difference between output levels that I've been fighting.

Again, the same cables were used. I ran the sweep, saved the results and then moved the cable to the next output and repeated.
 

Attachments

  • dcx2496_freqsweep_preampcalib.png
    dcx2496_freqsweep_preampcalib.png
    31.5 KB · Views: 151
Yeah, you've got some kind of a problem with your testing setup because those curves don't agree with the numbers you were getting from your DMM.

The frequency response should be ruler flat and there shouldn't be any of that "rippling" between 100-300Hz.

Cheers,

Davey.
 
I ignored the ripple. I was concentrating on the relative difference between the levels and the high end response variations.

However, I wouldn't doubt my test is flawed. However the tests are all done using the same cables and are repeatable.

I just got an RMA number from Behringer so I'll take it the local service center next week and see what comes of it.

Thanks for thelp.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.