[stvnharr]

Hello all,
Two new subjects have come up, and I can only add my own experiences to the body of experiences/observations.

I use a shunt stepped attenuator with fancy resistors.  Can't say I really have done any comparing to anything else.  The stock cermet attenuator is a good one, but if you allow some space behind the front plate you can always use attenuators now or later.

I run a sub out after the volume attenuator.  I have a short run to a sub.  I don't need a buffer.  There is no sound degradation to the buffered tube output.  Hugh and Marty have heard my GK-1 in Marty's system.

Steve

[andyr]

Andy,

when I connect the sub out, the tube out signal is degraded. The tube sound goes flat and looses all the top end. I have a vague recollection that Hugh had a reason for this, I think?  I am not sure if the buffer fixes this issue as well?

Theo

Aah, that's interesting to know!    I'm certain the Sub Out buffer fixes this problem too ... after all, did you notice "flat sound with no top end" at my place the other night? 

However, the maestro will need to confirm this.

Regards,

Andy

[jules]

Interesting stuff ...

I'm wondering how a Burson Buffer might fit in with this sub matter. The usual placement for these is before the pre-amp [or in the pre-amp if you can manage it  ]. The BB gives a 6 db gain which could be very useful here. It's also possible to order them with a higher gain though I'm slightly doubtful about this path. I use a BB myself and it's one of the best additions I've made to my system but I don't have a sub in there yet so maybe someone else can comment.

Another option for the sub would be to have the sub amp separate from the sub itself and right next to the rest of your amps/preamps etc. I'm surprised this isn't done more often as it has the logic of reducing the length of a signal interconnects by increasing the length of the "speaker" wires. It also gets the sub amp away from the vibrations of the sub but this might not be significant. This change wouldn't help if the problem is that the sub somehow steals part of the signal to the rest of the system.

On attenuators ... I believe the TDKs have become very expensive lately. There are several DIY options for those who want to practice their soldering. I've used a full ladder Elma [from Percy Audio] and these have the advantage of using only one resistor per channel at any level you have them set.

Space is always a good idea whatever you happen to be doing, so yes, leave some space for an attenuator and leave some more space to help avoid problems with fields surrounding various bits of wiring. 

jules

[Grumpy_Git]

There has been talk in the past about changing cap values to affect the system f3. can anyone comment on where this would lie with the suggested C1, C19, C21 changes?

cheers

Nick

[andyr]

I'm surprised this isn't done more often as it has the logic of reducing the length of a signal interconnects by increasing the length of the "speaker" wires.

IMO there is no "logic" in reducing the length of signal ICs, so as to increase the length of spkr cables.

With ICs, there's bugger-all current flowing, so L and R are irrelevant; C is the only parameter which can affect the sound ... apart from the dielectric around the copper/silver wire.  The amount of C which is acceptable depends on the ratio of the Zout (output impedance of your source/preamp) to Zin of your power amp.  If you have a ratio of 100+ (my own is about x350) you can have quite high C (read "long ICs") without any sonic detriment.  ) 
 
Spkr cables OTOH carry large currents - so low:
* R is important, otherwise the resistance of the cable relative to the driver impedance becomes too high for the amp to be able to exert the best control over the driver, and
* L is important, as it acts as a "brake" against changes in current flow.

Dielectric behaviour is also important (whereas C is only important to the power amp driving the spkr cables).

Achieving low R and L with spkr cables is a much more dificult exercise to get right than just low C on an interconnect.

Regards,

Andy

[kyrill]

Andy
.. Achieving low R and L with spkr cables is a much more dificult exercise to get right than just low C on an interconnect. ..

wouldn't it make sense to minimise therefore the length of the SC at the "expense" of longer ICs?
so have 2x mono poweramps

[andyr]

Andy
.. Achieving low R and L with spkr cables is a much more dificult exercise to get right than just low C on an interconnect. ..

wouldn't it make sense to minimise therefore the length of the SC at the "expense" of longer ICs?
so have 2x mono poweramps

Absolutely! 

My own spkr cables are 200mm long! 

Regards,

Andy

PS: In my case it's 6x mono poweramps! 

[kyrill]

Andyr

with such short lenghts, can you discern differences between different brands/self made cables?

[gerado]

[/quote]

Aah, that's interesting to know!    I'm certain the Sub Out buffer fixes this problem too ... after all, did you notice "flat sound with no top end" at my place the other night? 

However, the maestro will need to confirm this.

Regards,

Andy
[/quote]




Andy,

Absolutely not.
However you are not running a powered sub off the sub out of the GK. You are running the sub out of the GK into a lifeforce into to base panels. I do not know if that actually makes a difference with impedance matching?

theo

[andyr]

Andy,

Absolutely not.
However you are not running a powered sub off the sub out of the GK. You are running the sub out of the GK into a lifeforce into to base panels. I do not know if that actually makes a difference with impedance matching?

theo

No, I'm not running a sub ... and the Zin of my bass XO is 100K, which is probably a lot higher than many sub plate amps.  So mebbe you're correct ... "Sub Out" is affected with some (very low) Sub plate amp Zins?

If you heard a degraded tube out signal when you connected your sub then what Zin is its plate amp?

Regards,

Andy

[AKSA]

The sub-out feature of the GK1 was a consequence of an unusual attenuation topology.  Let me explain according to my understanding of this important component.

Level controls, in potentiometer (three terminal) form, ideally require two parameters for best sonics.  One is a low source impedance to drive the pot, otherwise attenuation is highly variable with the source chosen;  and the second is an extremely high target impedance for the circuit block being driven from the wiper.

The problem is the wiper.  At the point where it slides on the carbon/plastic track, there is spring pressure, and this means that the wiper contacts the track, distorting it physically at a number of microscopic points, setting up a small number of alternate paths into the circular resistive substrate.  These multiple paths will be pretty much at the same voltage if no current is flowing, but if a current is flowing into the wiper, then they will all differ in their voltage drops and the output will tend to jarble the sonics.  The only way to avoid this problem is to draw no current through the wiper, and this mandates a very high target impedance.

Thinking on this, I felt that since we could not control the source impedance driving the preamp (digital and tube sources are different, as you'd expect), and since the input impedance into the GK1 was around 45K, fixed by the input stage topology, it might be better to place the volume control between the SS and tube sections.  The Zout of the SS sections is around 30R, nice and low, and the Zin of the tube stage, with its bootstrapped, self biased cathode follower, is 4.3M (calculated), extremely high.   This would mean that the SS section operates at full line level, but the tube operates only at the listening level set by the volume control.

This 'enclosed' environment thus defines the ideal point for a level control.  Source and target impedances are optimally controlled, and almost no current flows from the wiper into the tube grid.

However, if we connect a sub amp to the wiper, then it must deliver a small current into the amp input via the wiper contact, blurring the waveforms as described above.  This is the problem, and it flattens the sonics for the tube.

The only way around this is to insert a buffer, an opamp of good quality is fine, between the wiper output (sub-out) and the sub woofer amp.  Andy has done this to good effect.  This prevents the sonic effect on the tube output, as Gerado has noted, by restoring the wiper current to negligible levels.

It's worth noting, too, that the gain of the tube is around 0.92, that is, 1Vrms in will yield 0.92Vrms out, and if we take the sub-out at the 1Vrms point you can see that the bass frequencies will enjoy a small lift.  It's very low, however, around 0.72dB, and I doubt most would hear it, (1.5dB is acknowledged to be around the finest resolution of the human ear), but it's nice to have it......

Hope this is helpful,

Cheers,

Hugh

[jules]

Ah ha ... most helpful, thank you Hugh

Andy, a small point of order:

IMO there is no  "logic" in reducing the length of signal ICs, so as to increase the length of spkr cables.

I agree completely. My post was badly worded. It would be very silly to have long SCs as a goal. 

jules

[aurelius]

Hi all,

I don't want to start an off-topic debate, but there are several people whose opinions direclty contradict Andy's (I would hate for someone to spend $$$ on modding their system without being exposed to a cross-section of views).  This is directly from Mapleshade audio's website:

NEVER use speaker cables shorter than 8'. Amazingly, 4' sounds much worse than 8'. Contrary to common belief, shorter interconnects (2 m or less) and longer speaker cables sound WAY BETTER than the opposite—based on extensive head-to-head tests.

http://www.mapleshaderecords.com/audioproducts/freeaudiotips.php

I don't have an experienced-based view, so I will say no more.

[andyr]

This is directly from Mapleshade audio's website:

NEVER use speaker cables shorter than 8'. Amazingly, 4' sounds much worse than 8'. Contrary to common belief, shorter interconnects (2 m or less) and longer speaker cables sound WAY BETTER than the opposite—based on extensive head-to-head tests.

But marcus, where does that leave all us ATC/Orion/etc. active speaker owners?  Fer shure there ain't 8' of cable per driver coiled up inside each active ATC speaker box!   

Regards,

Andy

[whubbard]

Amazingly, 4' sounds much worse than 8'.

I just don't see how this is any way possible. It just doesn't make any sense. There must have been another factor here.

Contrary to common belief, shorter interconnects (2 m or less) and longer speaker cables sound WAY BETTER than the opposite—based on extensive head-to-head tests.

This however, seems like it COULD make sense...

I don't have anything new to add, but I just don't see how 4' can sound any better than 8'

-West

[RonR]

Re Speaker cable lengths:

Most amplifiers need to 'see' some inductance on their outputs. Some manufacturers (e.g. Naim) provide this by using inductive cable (NAC A5), which means that the inductance is proportional to the length of the cable! I think the accepted best sounding length for a Naim 250 / NAC A5 combination is around 3-4 Metres.
So longer cables can sound better than short cables.

Hugh's designs include a wound inductor on the output already, so they are much less dependant on the cable inductance (but it still has some effect).

Cheers,

Ron

[aurelius]

I have read somewhere in the distant past, someone suggesting that the assumed-ideal of zero resistance & reactance for a speaker cable is in fact an error of judgement because it in turn assumes a perfect amp (performance independent on load reactances) and a perfect speaker (presumably totally resistive => non resonant and frequency independent).  In a real world, the compex load characteristics of drivers interact in a nasty way with global negative feedback amps.  The hypothesis is that having some degree of "electrical separation" can actually benefit the amplifier/speaker interface.  Bollocks?  Who knows...?

Back on topic, has anyone tried to repeat Philip's experience with Solens or other large film caps?  Seems there's been a lot of talk about tantalums, small WIMA films, etc, but I don't recall anyone giving Philip's specific mod a go? 

[andyr]

I have read somewhere in the distant past, someone suggesting that the assumed-ideal of zero resistance & reactance for a speaker cable is in fact an error of judgement because it in turn assumes a perfect amp (performance independent on load reactances) and a perfect speaker (presumably totally resistive => non resonant and frequency independent).  In a real world, the compex load characteristics of drivers interact in a nasty way with global negative feedback amps.  The hypothesis is that having some degree of "electrical separation" can actually benefit the amplifier/speaker interface.  Bollocks?  Who knows...?

I suggest this proposition is relevant to "conventional" passive-XO speakers ... not (active) situations where the same amp is connected to just one driver?  I say this because IMO, the inductive load which a passive XO presents to an amp is more complex than that offered by a driver (whose reactance can be corrected anyway by a Zobel, if need be) ?

Regards,

Andy

[whubbard]

Re Speaker cable lengths:

Most amplifiers need to 'see' some inductance on their outputs. Some manufacturers (e.g. Naim) provide this by using inductive cable (NAC A5), which means that the inductance is proportional to the length of the cable! I think the accepted best sounding length for a Naim 250 / NAC A5 combination is around 3-4 Metres.
So longer cables can sound better than short cables.

Hugh's designs include a wound inductor on the output already, so they are much less dependant on the cable inductance (but it still has some effect).

Cheers,

Ron

I guess I stand somewhat corrected. However, I did think that was a possible explanation, but the difference of inductance in a 4' to 8' would not be that big of a change from the amps point of view, no?

Also, I think we should try and get back to the Gk-1 as we have slowly drifted away.

[aurelius]

I suspect you're right Andy, but in a genuine attempt to kill off-topic conversation, I won't comment further.

If any Aussie's want to try Solens at C4 & C7 and would like to ammortise international postage across a small group buy, PM me.  I am curious enough to give it a go (I am much more of a "capacitor believer" than a "wire believer" anyways).