XLR / Balanced cable chipamp

[gmarsh]

You can't go wrong with an instrumentation amplifier approach for receiving balanced input... see attachment. Ignore the pin numbers, I just used a generic op-amp part.

Based on the output level of the DCX, you'll have to make a few resistor value changes. You can increase gain by adding R7, and you can decrease gain by adding R9.

Remember that for U1, R1=R2 and R3=R4 and the gain is R3/R1. Adding R9 effectively changes the value of R1 to the parallel resistance of R1 and R9/2... so when you add R9, increase the value of R1, R2 so that U1's gain is constant. I know LM-series chips don't like operating at low gains.

Connect pin 1 of the XLR to chassis ground. Don't use it as any kind of signal ground.

[akunec]

I would like to drive a chip amp with XLR cables. Meaning i would like the pin 2 on the xlr to go to the + on the chipamp, pin 3 to go the the - on the chipamp and pin 3 or ground just to go to the case of the amp. I would also like to have a dc blocking cap in the inputs and i dont know how to implement that either.

[poobah]

Use the circuit gmarsh posted above. Substitute the LM3886 for the last op-amp in the drawing.

Are you sure you need blocking caps??? they will be very large in 600 Ohm circuit.

[carlosfm]

Originally posted by XELB
But Carlos, you can also minimize noise and lower the distortion if the ouput is differential...

I think it's better that akunec explain the context.

XELB, this is basic op-amp stuff...
Only when you convert to unbalanced you are rejecting noise, and the LM3886 has a typical CMRR of 110db (at low frequencies).

Loot at figure 2 here:
http://sound.westhost.com/project51.htm

[XELB]

Carlos, I am not saying that you are wrong.

I just said that, using Differential output allows, in theory, to reduce the distortion and noise....
Single ended it's more sensible to bad ground techniques.

Like akunec wants, differential output "minimize noise"(his words) and distortion.
Or am I not right ?

[heater]

gmarsh - Whats with R7 and R9 in that instrumentation amp schematic ? R7 I think I understand as half its value set the gain as normal A = 1 + Rf/Ri
R9 I don't understand as in my simulation I get the same gain with it in or out!

In my sims I tried to keep the resistors small, inline with the ideas carlosfm, and keep a gain of the chipamp to 10 or 20.

[carlosfm]

Originally posted by XELB
Carlos, I am not saying that you are wrong.

I just said that, using Differential output allows, in theory, to reduce the distortion and noise....
Single ended it's more sensible to bad ground techniques.

Like akunec wants, differential output "minimize noise"(his words) and distortion.
Or am I not right ?

You are making a lot of confusion about such a simple thing.
Akunec wanted differential between the DCX2496 and the amp, he doesn't necessarily want a bridge amp, with differential output to the speakers.
The 'output' word is yours, not his.

Originally posted by akunec
does anyone have a design to use a chipamp differentially with XLR cables... I have a DCX 2496 and would like to stay balanced to minimize noise.

What you suggest, XELB, doesn't take advantage of differential signal transmission, which at some stage has to be converted to single ended to take advantage of the high CMRR of the chips.
I can't explain myself better than what I already did, sorry.

[Panelhead]

My preamp has balanced outputs, converted to single ended to work with all of my amps.
How would you configure the input of the LM3886 to drive it balanced without adding any buffers or drivers.? Single neded output is fine. I have never seen a circuit here that did it without either two LM3886 or more chips.
My preamp has an output transformer, so it is very flexible, it is balanced out thru a 3 pin locking din. The interconnects are wired to convert to siingle ended RCA.

George

[carlosfm]

Gainclone with pseudo-balanced input?

[gmarsh]

Originally posted by heater
gmarsh - Whats with R7 and R9 in that instrumentation amp schematic ? R7 I think I understand as half its value set the gain as normal A = 1 + Rf/Ri
R9 I don't understand as in my simulation I get the same gain with it in or out!

*takes another look at schematics*
*laughs at own stupidity... voltage across R9 is always zero*

I guess R1, R2 should each be split into two series resistors with R9 connecting the junction of each pair. I'd just leave R9 out and replace the input 600 ohm resistor with a balanced attenuator network.