Signal degration over length, source vs. output ?

Which signal is more likely to degrade over length, source to amp, or amp to speaker? I have a pair of 300B monos on the way and don't know where to place them relative to preamp and speakers. Is it generally better to place the amps close to the pre-amp and use a short interconnect and a longer speaker cable or the other way around? Any insight would be appreciated.
Generally longer speaker wires are better than longer interconnects.

What distances we're talking about? Where is your stand and where is your speakers?
Generally longer interconnects are better than longer speaker wires.

What distances we're talking about? Where is your stand and where is your speakers? What are the speakers and what are the input/output impedances of your preamp and amps?
Generally, you can read a lot of generalizations about this subject : )

Balanced lines are less susceptible to running into problems with long line lengths. They are still subject to degrading / losing signal via dielectric absorption and the impedances that they bring with them into the system.

Single ended interconnects ( RCA's ) are more susceptible to signal degradation due to their design and are also more susceptible to picking up RFI / EMI interference. They too are subject to degrading / losing signal via dielectric absorption and altering system response / loading characteristics due to their impedances and component matching.

Having said that, all interconnects deal with a signal that is sometimes measured in millivolts and of extremely low current levels. Once a signal is lost / distorted prior to high level amplification, that signal is then passed on in lesser, more distorted form for further amplification. The effects of this signal loss / distortion are compounded as they go through each link in the signal amplification chain.

Speaker cables typically deal with higher level signals i.e. increased voltage and current levels. All of the above criteria regarding interconnects applies here too ( dielectric absorption, impedances altering loading characteristics, etc... ), but since we are already dealing with a higher level signal that can no longer be amplified, losing / distorting the signal is not as impactful on system performance.

To put things in English using money as compared to signal, if you had a dime's worth of signal going through an interconnect and lost a penny, that 10% distortion would probably be quite noticeable. This is because the signal distortion ( loss is a form of distortion ) was noticeable AND the effects of that signal distortion are further compounded as the signal is amplified as it passes through each gain stage.

On the other hand, if you had a dollars worth of signal ( fully amplified with no further amplification taking place ), losing that same penny would result in a 1% distortion figure. While we are still at the same amount of loss ( one penny's worth ), those losses are no longer compounded through further amplification. Since we have more signal to "play with" and there is no further chance for errors to be further amplified, it makes the most sense to "take your chances" with longer speaker cables than with longer interconnects.

Obviously, if one has a mild load in terms of speaker impedances and levels of reactance, you will always be better off regardless of what path you take. If one has a highly reactive load and / or a load that has erratic impedances presented to the amp, speaker cable length and impedance can play very important roles in what you hear.

While most will say that shorter cables allow the amp to get a better "grip" on a highly reactive speaker, the reverse is also true. Since there is less "buffer zone" provided by the impedance that a longer length of speaker cable would provide, high levels of reactance from the speaker can more easily "modulate" or influence the amplifier's output stage, power supply and other devices ( like the preamp ) further up in the signal change. It is in these types of situations that using what one might consider "less than optimum" cables can result in actual improvements to the system. In some cases, changing cables can make the difference between normal operation via buffering the load that the amp sees and sending the amp into protection mode due to being completely revealing of a radical speaker load.

Given all of the above, common sense would dictate that the weakest signals are the most delicate and any distortion introduced into the system at that point will only further compound those distortions as they are amplified. As such, one should "take the most care" of the signals prior to final amplification. This is not to say that speaker cables / lengths are "less important", as a system is only as strong as its' weakest link. Obviously, there is enough variance from system to system, components used and personal preference that you may like / prefer something different than what i like, so use what works best in your system. As usual, this means trial and error, which can be as cheap or expensive as you care to make it. Sean
With tube amps, I would favor keeping the speaker cables on the short side. Two reasons, mainly: first is that tube amps give their push with voltage instead of current, so limiting the voltage drop from amp out to speaker will conserve power - which is meager to begin with.
Second is that tubes have a very low damping factor. Long run speaker cables decrease the damping factor which makes it more likely that voltage will be reflected back along the speaker cables. Long cables are capacitive and will store and reflect the voltage back to the speaker, resulting in a loss of coherence as the voltage starts to oscillate.

Other factors do come into play and there are many permutations with length, common mode noise, ground loops, impedance swings etc., etc., so there's no easy answer IMO.

You gave us no indication as to the source, but if the pre amp is a low output impedance/high gain solid state, then I think the IC's will have less of an impact than the speaker cables with respect to length only. Again, this is VERY general - with no particular equipment as a reference.
Here's the system I'm talking about:
CD - Mission PCM4000 (1987) to be updated when cash reserves return
Pre-amp - Cary SLP-74
Power - Cary CAD 300SE monos
Room is 20'X12' but to preserve the design, (or so she insists), the system will have to be placed against a side wall so 1 speaker will be about 8 to 10 feet away from the pre-amp and source.
The lengths you describe are not a real concern either way but I still prefer shorter speaker cables.

The reason for my preference is that the connection between the preamp and amp is impedance-defined for the transfer of voltage and, assuming decent cable and appropriate impedances of the preamp out and power amp in, the only loss will be voltage. That is easily compensated for by a little touch on the VC.

The amp to speaker interface needs to transfer power (voltage and current) and the input of the speaker (almost any of them) varies widely in impedance. Increasing cable length increases resistance and reduces power transfer but also does so in a manner that varies with frequency due to the varying load.

All that assumes near ideal equipment and really long runs. Your setup is not stressful.
You might just try it both ways. If you notice no differance, then just do what is convenient. having said that, I agree with the above post. 8 to 10 feet is no big deal either way.
Kr4: Since your concern is with loss of power transfer, which do you think "loses" more power / signal ?

A) A long run of 20 gauge or thinner wire ( as commonly found in most interconnects )


B) A long run of 14 gauge or heavier wire ( as commonly found in most speaker cables )

As far as the "voltage losses" that can be compensated for with the volume control, that is not just "voltage" that you've lost, it is a dynamic part of the signal. Since the losses are most likely to take place when the least amount of signal / voltage is present, the likely effect is that one will lose low level information. This results in the masking of subtle details. While some may mistake this loss of signal or noise transfer as a reduction in the systems' noise floor i.e. a "blacker background" due to NO noise or signal being present at very low levels, it is in all actuality, a reduction in resolution and dynamic range.

Obviously, there are pro's and con's to each method. If your system is carefully thought out and uses conductors that are suitable for passing the quantity of signal that will be in operation without incurring measurable amounts of series resistance, chances are, either method will work "okay". Sean
We have seen gain in our long cable runs; Why? Because signal degeneration is mostly a product of field contamination and poor resonance tuning NOT CABLE LENGTH. Remenber: alternating current is a complimentary technology that's in diametric opposition. Think outside the box!
Actually, my concern is the harmonic change due to very long speaker cables rather than power or voltage losses. In addition, living with a 10meter run for one or the other, I have experimented with both and found my preference. Of course, I generally use a preamp with a <50ohm output impedance and excellent voltage output.

In the OP's case, (and indeed in any other), best advice is to try both ways if one can.
If you have a solid state preamp with output impedance in the 50-100 ohm range, interconnects are not all that critical. As Kr4 says, impedances for this interface (preamp to power amp) are well defined, whereas the power amp-to-speaker impedance relationship is variable. Therefore go with long interconnects. I note that at recording sessions it is not uncommon to have a hundred feet of interconnect wire between microphones the mic preamp and the recorder.

If you have a tube preamp, with the typical 600ohm output impedance, keep the interconnects short.

These general rules apply for "decent" (affordable)wire. The disadvantages of long wires of either type can be overcome (so they say) by various kinds of exotic wire configurations, for which you will pay dearly.
I have a tube preamp with <50ohm output impedance as well.
speaking of 8...10' you loose almost nothing with longer speaker cables.
I agree with what El and Kr4 are saying in terms of low output impedances being highly beneficial to this situation, but El is tossing out statements about long mic cables while forgetting to mention that these are all primarily runs of balanced / xlr cables. BIG differences in performance between xlr's and rca's on a long run.

Corona: If music consisted of a series of non-distorted sinusoidal waveforms, your comments would make more sense. Since it is quite possible to have notes that are not of equal amplitude / duration in positive and negative attributes, "diametric opposition" or having an "AC flywheel" doesn't come into play quite as much or as linearly as one might think. Then again, i might be misinterpreting what you are saying as you tend to "fly" with your "cloaking devices" pretty well maxed out.

Having said that, I do agree with your comments pertaining to field contamination. I have tried to get this point across to Audioengr in terms of shielding of cables only being bad when it is improperly applied. You may / may not agree with this, but judging from what i think you're getting at, i think that you would agree. Sean
Kr4...I didn't know that you could get 50 ohms output impedance with a tube output stage (usually a 12AU7 cathode follower: 600ohms). What kind of preamp do you have, and did they slip in a transistor line driver?

sean...Yes: balanced lines are better for long runs, but that's another discussion. I only meant to suggest that a long run of low level signal is not the end of the world.
Sonic Frontiers Line3. Output impedance 45ohms SE, 90ohms bal.
Kr4...Your SFI preamp looks like a nice unit. They confim that the output stage is "parallel tube with feedback" (not a transistorized line driver as I suspected) providing output impedance which they agree is low for a tube preamp. Looks like you get any tube advantages without one of the drawbacks.