[Bob Reynolds]

Hi Bob,

Yes the "rule of thumb" is a minimum of x 10. I personally believe x 50 would be better and while "x 10" removes most of any problems to do with capacitance, it does not take away all of them. x 50 is kindof a "safety factor".

Given my preamp is a hybrid, having a unity-gain tube output stage which still has an output impedance of <200 ohms, I cannot understand why tube designers produce preamps with a Zout of 3,500 ohms?  It just produces the potential to cause matching problems.

Paul, if you go for a low C design - with a low C of around 13pF/ft like the one Bob suggested (although I think you don't need/can't have balanced ICs) - a 3m IC shouldn't give you any troubles. That is, assuming the Prima Luna is able to drive it (I have no idea!  ).

Hi Andy,

I agree with you that > 10X is better. Your 50X should take the drive issue out of the equation. With careful routing, a 10ft. unbalanced (shielding may or may not be necessary) should not cause any audible problems. Ground loops are also more of a possibility with longer unbalanced cables, so the transformer conversion to balanced may be something to keep in mind.

I also am confused with all the tube preamps with what seems to be unnecessarily high Zout. It's possible that they are designing for their own power amps only, which should have very high Zin.

Paul - yes Blue Jeans does sell Belden speaker cable, but they don't have much in the way of terminations. I use their Belden 5T00UP (10 ga.) cable in both my home ($17K) and office ($2.5K) systems. Zebra Cable uses the same Belden speaker wire, but does a nicer job of "dressing" the cable. I'm using Vampire banana plugs at home and bare wire at the office. Note that 10 ga. cable as bare wire may not fit some components.

-- Bob

[andyr]

... If your system is single-ended I'd stick with long speaker cables due to the propensity of unbalanced inputs to pick up hum as they get longer. ...

Mmmm ... there's no hum in my system with my shielded ICs.  

Regards,

Andy

[Robert57]

It seems this discussion is coming down mostly in favor of the long IC's, short speaker cables. An earlier thread on New Years Day seemed to have the opposite consensus, with some rather compelling accounts of trial and error yielding decidedly better results with the longer speaker runs. Maybe this is yet another system-dependent question? Just when I thought this question was firmly settled....

http://www.audiocircle.com/index.php?topic=24329.msg215567#215567

[PaulFolbrecht]

Well, for one thing, it's more of an issue with SETs, right?  They are already more susceptable to impedence problems, which longer speakers cables can make more pronounced.  So, the topologies of the amps in question needs to be considered, right?

[Bob Reynolds]

In general we can make these statements:

1) the connection between preamp and amp is one of low voltage and high load impedance. Thus, very little current is flowing between the two components and therefore, very little potential drop can occur across the cable.

For example, the output from a CD player is 2V, by redbook std. Assuming unity gain level setting on the preamp, we have 2V on the output of the preamp. The input impedance of an amp is say 20Kohms. So the current computes to be 0.1mA I believe.

2) the connection between amp and speaker is one of high voltage and low load impedance. Thus, high current is flowing between the two components and therefore, a greater potential drop can occur across the cable.

For example, assuming the amp has a gain of 25dB, our 2V input will become 35.5V at the output. Assume we have the classic 8ohm speaker. So the current computes to be 4.4A I believe.

It is this several orders of magnitude difference in current flow bewteen the two scenarios that leads us to want short speaker cable with longer ICs.

The issues of noise contaminating a low level signal in the IC can be addressed by shielding and using balanced lines. The classic example addressing this is the microphone feed in the recording studio.

It's this same argument that says we need large gauge speaker wire (14 ga., 12 ga. etc.). Note that running double runs parallels the resistance and thus, reduces it by half. What's the gauge of your typical IC? For the Belden 1505F it's 21 ga.

-- Bob