Gain structure problem with DSP active crossover and P3a

[brunogiovs]

Hello

I had to put a buffer between my tweeter amp that is a p3a and my behringer dcx2496 to reduce the noise floor.
I'm not sure if noise floor reduction is the right term , but my problem is that I have a reasonable sensitive tweeter and when I have the dcx2496 on there is
a considerable noise going out from my tweeters.
So to solve it I put a preamp/Buffer in the middle and it has a 12db attenuation on the signal , so I had to increase the gain in 12db to my tweeter on the dcx2496.
and happiness !! no more high noise floor.

my questions are :

do I need a preamp/buffer to my p3a or I could just put a resistor voltage divider of - 12db on the signal path to my p3a and I should be good no?
the p3a input impedance is ~22k so it should be good no? or I will loose high frequency?

Thanks

 

[dotneck335]

It definitely seems that the gain structure is way off somewhere, or your DCX2496 is broken. Behringer quote a -90bdu output noise figure, which SHOULD be completely inaudible.

 

[brunogiovs]

will all inputs and outputs off , just by turning my DCX2496 causes a considerable noise floor .
my tweeter is very sensitive .. 93db .

I read a lot of comments about DCX2496 mods and replacement of the analog output.
So decided to play a bit with the gain structure to my tweeter and it worked very well ..

do you have a DCX2496 with a sensitive tweeter? and have no audible noise floor?

Thanks

 

[dotneck335]

I DO NOT have a DCX2496; but if the specs are accurate it shouldn't be noisy. Do you mean that if you attenuate the signal IN the DCX2496 signal path that it improves the signal-to-noise ratio? It seems odd that just attenuating it and amplifying it again AFTER the the DCX2496 would improve things at all.

 

[sreten]

Hi,

You'll be fine with a resistive divider tacked into to the input sockets of the P3a.

rgds, sreten.

 

[brunogiovs]

I DO NOT have a DCX2496; but if the specs are accurate it shouldn't be noisy. Do you mean that if you attenuate the signal IN the DCX2496 signal path that it improves the signal-to-noise ratio? It seems odd that just attenuating it and amplifying it again AFTER the the DCX2496 would improve things at all.

I don't attenuate the signal on dcx2496 , quite the opposite. I attenuate by 12db right before the amp so the signal needs to be 12db higher to the same power output , so so the signal-to-noise gets divided by 3 .

Thanks

 

[brunogiovs]

Hi,

You'll be fine with a resistive divider tacked into to the input sockets of the P3a.

rgds, sreten.

Thanks, I did that to all my amps , one more I learned . how important is the gain structure.

 

[AndrewT]

I don't attenuate the signal on dcx2496 , quite the opposite. I attenuate by 12db right before the amp so the signal needs to be 12db higher to the same power output , so so the signal-to-noise gets divided by 3 .

Thanks

This is the only way (from an unmodified CX) to get the low noise output required for domestic style power amplifiers that need signals of ~1Vac to reach clipping.
You must use passive attenuation at the output of the DCX.

Even better is to remove all the excess gain built into the DCX output stages that was required to meet the professional >+20dBu voltages.
You can manage with maximum outputs of around +3dBu to +6dBu for most domestic style listening. For a very sensitive tweeter a max output of 0dBu to-6dBu would probably be more useful.