BA-3 crosstalk level?

I was doing some room measurements, and I discovered that when feeding a signal just to the right channel of my BA-3 amp, the left channel speaker would produce output at around 17 dB below the right channel. This seemed like not much separation to me. Is that typical for this amp?
 
Is the BA-3 amp both BA-3 pre + BA-3 power amp?
I think it depends how BA-3 is implemented with signal wiring + how the power supply is implemented.
I would guess that it has something to do with implementation rather than the amp design itself.
 
Is the BA-3 amp both BA-3 pre + BA-3 power amp?
I think it depends how BA-3 is implemented with signal wiring + how the power supply is implemented.
I would guess that it has something to do with implementation rather than the amp design itself.

Yes, it's has both the front end board and the power boards. I was wondering if it was implementation related.

The main thing I wanted to know was, what is the expected crosstalk for this amp? What's my "reference" to compare to?

@Zen: Pics of what? The graph? It was two nearly identical curves, one approximately 17 dB higher than the other.
@Ben Mah: I connected preamp input to the right channel of the amp, with the right channel speaker wire connected. Nothing was connected to the left channel. I ran a 20-20k sweep and measured the results with an in-room microphone at the listening position. I then disconnected the right channel speaker wire and connected the left channel speaker wire. Still the only preamp connection remained on the right channel. I ran the same sweep and recorded the results. I observed a curve produced by the left speaker approximately 17 dB lower than the first curve, and roughly identical to it in shape.

Before we begin addressing potential problems, I want to know what the expected crosstalk should be for this amp, so I have a point of reference for everything else going forward.
 
Member
Joined 2012
Paid Member
I don't have a BA-3 so I cannot do any measurements.

I looked up some Stereophile reviews and they have reviewed the J2 and SIT-3. Measured channel separation was greater than 100dB below 1kHz, 80dB at 20kHz for the J3, and 74dB at 20kHz for the SIT-3.

Zen Mod asked for pictures of your BA-3, as layout of components and wiring can strongly impact channel separation.
 
These two pics show my signal routing and power supply setup. The opposite side is just the output stage.
 

Attachments

  • IMG_4084.jpeg
    IMG_4084.jpeg
    608.9 KB · Views: 56
  • IMG_4085.jpeg
    IMG_4085.jpeg
    619.9 KB · Views: 55
Moderator
Joined 2011
Multiple tightly-twisted pairs should be spaced a bit apart, at least an inch or two if possible.
If you are only getting 17dB separation at lower frequencies, there might be a wiring error.

Are you shorting the input of the undriven channel when measuring the crosstalk?
If not try that. The source normally supplies a low impedance at the input node,
but if nothing is plugged into the RCA, the impedance level at the input goes way up,
increasing the possible amount of crosstalk.
 
Last edited:
Multiple tightly-twisted pairs should be spaced a bit apart, at least an inch or two if possible.

Gotcha. Question: How does Audio Research get away with bundling all their input wires together like this? https://www.arcdb.ws/Database/SP11/ARC_SP11MKII_top.jpg (SP-11 Preamp)


Are you shorting the input of the undriven channel when measuring the crosstalk?

No. It was an ad-hoc measurement.

Keep each channel twisted, run the left channel down the left side, the right channel down the right side.

I understood that I should minimize the area between wires, so I kept everything on the same side for that reason.
 
Member
Joined 2012
Paid Member
Minimizing area between wires applies only to wires in a circuit. The left signal + and ground form a circuit. The right signal + and ground form a circuit. The left channel and right channel are separate audio circuits.

There may be benefit to keeping the left and right channel input wires together if there is a noise problem. In my amplifiers I separate the twisted left and twisted right input wires and I hear no noise from my 103dB speakers.
 
Moderator
Joined 2011
How does Audio Research get away with bundling all their input wires together like this? https://www.arcdb.ws/Database/SP11/ARC_SP11MKII_top.jpg (SP-11 Preamp)
No. It was an ad-hoc measurement.
I understood that I should minimize the area between wires, so I kept everything on the same side for that reason.

Those are shielded cables in the SP11, not twisted pairs.
Repeat the measurement with the undriven channel input shorted.

You are right about the inter-channel loop area. If there still is a real crosstalk problem,
it might be simplest in this case to keep the routing, but to replace the twisted pairs with shielded cables.
 
Member
Joined 2012
Paid Member
I just realized that my stereo amplifiers are either dual mono or have a shared transformer with separate transformer secondaries, rectifiers, and power supply filters to each channel. So the two channels within the chassis are separate circuits, and separated left and right channel input wires are not an issue. With a shared power supply for the two channels, the loop area within the left and right channel input wires may be an issue.

But as Zen Mod and rayma mentioned, shielded cables is the solution.

By the way, if you do not have any shielded cable handy, you can still separate the left and right input wires and check to see if the channel separation improves.
 
Last edited:
  • Like
Reactions: 1 users