diyAudio (
-   Digital Line Level (
-   -   Settig up active Speakers with Behringer DCX2496 (

ThorstenL 1st September 2005 02:25 PM

Settig up active Speakers with Behringer DCX2496
2 Attachment(s)

In a different thread a gentleman was struggeling to get the levels involved, attenuation etc. right. I feel that his plight might be a good illustration example as to how to combine different sensitivity drivers, amplifiers, a DCX and and a preamp in a system that makes best use of the available dynamic range and all.

I have taken the liberty to "normalise" the input sensitivites of amplifiers and other levels to dbu and the power levels to dbw (all rounded to nearest db for ease of use) in order to allow simple calculations of attenuations etc.

If you need to understand db, dbu and dbw better, there are many websites covering this in detail. The 2.83V/1m sensitivity rating for drivers BTW can also be said to be taken at 0dbw.

Any specifications are derived from manufactirers data to the best of my knowledge and research, but claim no specific accuracy, so doublecheck before using these numbers.

Given are :

LF Amplifier - Adcom GFA-545II

(150W/4R or +19dbw @ 1.125V or +3dbu)

MF Amplifier - Panasonic XR-45

(100W/6R or +19dbw @ 0.2V or -12dbu)

HF Amplifier - Panasonic XR-45

(100W/6R or +19dbw @ 0.2V or -12dbu)

Crossover DCX-2496

(Inputs and outputs at full-scale with 9.75V or +22dbu)

LF Drivers - Lambda Dipole 12 2pcs in parallel

(the TS parameters from Lambda suggest a pair in parallel will produce 94db/2.83V/1m or 94db/1m with 0dbw)

MF Driver - Audax PR170M0 (on dipole?)

(while the gentleman lists them as "100db/W/m" the TS data, manufacturers datasheets etc suggest however a midband sensitivity of 95db/1m @ 0dbw with a rise in SPL for the upper midrange and dipole use may be able to add a broad 3db boost in the lower midrange, so for our experiment we will take 98db/1m @ 0dbw as "the gospel")

HF Driver - Fountek NeoPro5i

(the manufacturers published data suggests 102db/1m @ 0dbw and the graphs support this)

Form the above we can see that we have many mismatches in sensitivity etc.

First, the Panasonic Receiver will develop full power with a signal 15db lower than the Adcom Amp.

Secondly, the DCX will attain full output at a level 19db HIGHER than the signal that will clip the Adcom Amplifier and 31db HIGHER than the signal that will clip the Panasonic multichannel receiver.

If we apply no analogue domain attenuation to the DCX outputs we would loose these 19 ... 31db dynamic range or in other words, noise would be increased by that degree AT LEAST.

So, we need to attenuate the outputs of the DCX.

First, in order to allow a little headroom I would leave between 3 - 6db at the top of the dynamic range of the DCX unused. So we will take +19dbu as our "clipping on the amplifier" point.

The lowest sensitivity speaker section and the lowest gain are present in the LF channel, which will limit the SPL our system can achieve overall. There we have 94db/1m @ 0dbw and 113db/1m @ +19dbw and with an input signal to the Amplifier of +3dbu.

At the same time we have a +19dbu signal from the Crossover, so we need to attenuate 19db-3db = 16db to ensure best dynamic range in the LF channel.

Before we calculate this (and the other) attenuators let us continue calculating the respective attenuation for the respective channels.

The MF channel will show 98db/1m @ 0dbw and is thus 4db "louder" than the LF section. So, respective to the LF section we would need 4db more attenuation. HOWEVER, at the same time the Amplifier in the MF channel has 15db MORE GAIN than that in the LF channel, so we need to attenuate 15db + 4db MORE than the LF channel, where we have 16db attenuation. Running the numbers suggests 35db!!!! attenuation to be needed to match the system.

Finally, the HF channel will show 102db/1m @ 0dbw and is thus 8db "louder" than the LF section. Again, the Amplifier is also 15db more sensitive, so to bring the whole shooting match in line a total of 39db!!!! attenuation is needed in the HF channel.

In a different thread I suggested a very different way of attenuating the output of the DCX and converting the outputs to SE to the commonly used one (I have good reasons too for the differences).

The schematic is in this post:

As we want to make a fixed level attenuator, therefore we omit the switch and R2 and only wire in R1 fixed.

The circuit as drawn with a high input impedance amplifier (>>10k) will in effect attenuate by 12db if R1 is left out.

The traditional approach to get our attenuators right would be to convert the attenuation levels back to from db to linear attenuation and then to work out the resistive dividers.

Being perennially lazy, I made a little Excel Spreadsheet to calculate the attenuation by just typing in resistor values in Ohm. I also added Voltage/dbu and power/dbw conversions for convenience. I attach the zip file to the post.

The bottom line is that we need Values (NPV selected) of R! for the desired attanuation as follows:

3K9 for 16db
160R for 35db
100R for 39db


ultrachrome 1st September 2005 08:00 PM

Excellent spreadsheet. I'll gather the necessary resistors and let you know how close to 0 I get.

catapult 2nd September 2005 03:43 AM

Why make the DCX feed into such a low impedance? Are the output stages of the DCX more happy with that?

ThorstenL 2nd September 2005 08:39 AM



Originally posted by catapult
Why make the DCX feed into such a low impedance? Are the output stages of the DCX more happy with that?
The DCX is desiged to feed several 10K "bridging" loads in parallel and even down to 600 Ohm.

If you add an attenuator somewhere the overall impedance becomes an issue, get it too high or too low and the sound suffers.

But you can use my spreadsheet to plug in other values of R, so use whatever makes you happy.


All times are GMT. The time now is 08:16 AM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2