Interconnects and internal signal wiring - How long is too long?

[Deke609]

I'm finally wiring up the Kaiju rebuild and I'm not happy with the signal wiring layout that I've committed to. The internal signal input wiring will need to be almost 20" -- from rca jacks at the back to a stepped attenuator at the front and then to the drivers near the center of the amp.  On top of the 20" of internal input signal wiring is another 20" of interconnect between the Beepre and the Kaiju. So approx. 40" total per channel.

I did a bit of research on cable capacitance, inductance and impedance but found so much contradictory information that I gave up: one says cable capacitance has no impact on audio frequencies over lengths less than a couple meters, and that it's all about impedance; another says impedance doesn't matter and that it's all about capacitance. So I am confused.

All signal wiring will be copper shielded -- internally with solid copper tubing and the interconnects with thick copper braid. Unless I re-jig the layout by moving the attenuator close to the driver tubes, all I really have to play with is wire diameter and cable structure.  My default approach is use 20 ga silver plated copper wire, twisted 3 times per inch, and then shielded with shield grounded at one end.  But I also have some 30 ga wire on hand that I cold use if that might make a positive difference.

Question for the BH experts: any educated or experienced-based guesses as to my best approach (other than moving the attenuator)?

MTIA, Derek

[Paul Birkeland]

I did a bit of research on cable capacitance, inductance and impedance but found so much contradictory information that I gave up: one says cable capacitance has no impact on audio frequencies over lengths less than a couple meters, and that it's all about impedance; another says impedance doesn't matter and that it's all about capacitance. So I am confused.

They are both wrong.  If you have 1 ohm of output impedance from your preamp, cable capacitance will tend not to matter.  If you have 1pF of cable capacitance, driving impedance will tend not to matter.  There is miller capacitance at the input of the Kaiju to consider as well.

Question for the BH experts: any educated or experienced-based guessed as to my best approach (other than moving the attenuator)?

Moving the attenuator was the first thing that came to mind.  You don't want much cabling between the attenuator and the 5670.  The cable between the input and the attenuator is less important.

[Deke609]

Many thanks PB. I was afraid you were going to say that. I'll do so head scratching and see what's possible. Unlike the Beepre rebuild, there's not a lot of spare room in the chassis.

cheers and thanks, Derek

[Paul Birkeland]

So, do you have the stepped attenuators?  can you use them elsewhere?

Maybe consider getting a 3 position attenuator and set it up for -0, -2, and -4dB at 50K.  That will let you do a little trimming without running into issues driving a lot of cabling.

[Deke609]

Yeah, I already have it and it's a single unit that I wired up myself. A tiny 4 pole, 2 deck Grayhill that I've wired as nominal 100K with -0dB, and approx -6, -9, -12 and -15 db steps.  So I should redo it and make it 50K?

[Edit: this is in place of the PEC trimmers, which I won;t be using]

[Edit #2: The only reason for the attenuator is a concern that when switched for 64R output impedance, I might have to turn the Beepre down quite low -- the attenuator is meant to give me some wiggle room to find the sweet spot between attenuation at the Beepre and attenuation at the Kaiju].

Many thanks, Derek

[Paul Birkeland]

You'll probably have to try these experiments on your own.

[Deke609]

Thanks PB. Yup. Figured as much. Just thought I'd check to see if there was a straightforward answer -- ideally, "Don't worry about it b/c it doesn't matter".  Unfortunately, I don;t think moving the attenuator is workable. As it stands, it is about 6.5" from the drivers.  I could move it to within 2" of one driver, but it would still be 6" from the other.  I think my only option is to test the amp with the attenuator, and without it using 100K resistors near the drivers to load the Beepre.

cheers and thanks, Derek

[Paul Birkeland]

Just use low capacitance wire from the output of the attenuator to the 5670.  I would use an unshielded twisted pair.

[Deke609]

Excellent. many thanks PB.

[Deke609]

Found a solution. I can mount the attenuator on a L-bracket close to the drivers.  It'll be tight, but I think I can make it work. Thanks for the nudge in the right direction.

cheers, Derek