advertisement


Honest no-foo speaker cable question

Mullardman said:
Neither do I.. but I'd query 'old' editions.

'Old' editions are from the 1960s.. except that the BBC in it's infinite lack of wisdom wiped most of them.

Click to expand...

Dave Clark bought a lot of the broadcasting rights to the rest
 
Arkless Electronics said:
I'm afraid much of that is quite wrong...

It's all down to NFB (Negative Feedback, the most important "thing" in amplification and largely ignored by non technical people).

Emitter followers are not used in all SS amplifiers and the instability you mention is a negative resistance oscillation that can happen in certain conditions only and is not really any issue in audio power amplifiers. It is also a local to the actual follower phenomena and not global, plus in the case of output transistors if it were to happen it would be very much lower in frequency than UHF! As I say it's not an issue though and even if it were, in the case of Naim amps they are quasi-complementary and so the "push" and "pull" halves of the output stage are not the same... one is an emitter follower and one is common emitter.

The network you mention is not actually a Zobel network... although it is often misnamed as such! A Zobel network consists of a resistor and capacitor in series and in the case of an audio power amp goes from the output to ground. Naim amps DO have these. The network in line at the output (not used in Naim and quite a few other amps) is usually an inductor of a few uH with a damping resistor in parallel, it is sometimes known as a Thiele Network (yes as in Thiele-Small parameters) but this appellation is not widely used and it is usually not referred to by any name... or wrongly called a Zobel Network!

The purpose of the Thiele network is to prevent the load, including cable capacitance, from affecting the NFB loop parameters in such a way as to cause oscillation (instability) and ringing due to "gain peaking" as the point of instability is approached. This network can actually ring separately from any NFB issues though (L-C resonance) hence the damping resistor as a part of it...

The Zobel Network has more to do with keeping the amp stable without a load and/or with a load of rising impedance with frequency but is much less relevant to this discussion.

Now primarily the effect of this Thiele Network is that it's inductance causes rising impedance with frequency so as to partially isolate the feedback point at the output of the amplifier from capacitive loading. It should be understood that we are generally talking in terms of frequencies well above human hearing when considering all these effects.
The beauty of the Thiele network is that at low frequencies it has very low resistance and so has negligible effect on speaker control and damping factor but as frequency rises its AC impedance starts to come into play and it provides the required isolation at high frequencies.

Now Naim and several other companies (for certain models anyway) don't use a Thiele Network but put a resistor at the output to give the required isolation... and in Naim's case still get it wrong as there is insufficient isolation for the characteristics of the NFB used!
The only advantage of using a resistor here is that it can give some reverse isolation from the back EMF of speakers coming into the amplifiers NFB network via its output. Disadvantages are that it adds to the resistance in line with the speaker and reduces damping factor. In Naim's case the resistor is 0.22Ohm which is about the biggest resistance found in such an application and on its own limits damping factor to 18 with a 4 Ohm speaker...

In the case of Naim amps it is simply very bad engineering that the 0.22 resistor on its own is not enough isolation to guarantee stability with all cables and loads. There are no advantages in things being as they are and it does not in any way shape or form "remove the speaker cable" between amp and speaker. The only amps that make any attempt to do this that I know of were the Trio/Kenwood "Sigma Drive" models and some later models from DPA that did the same thing. They both worked by having extra cores to the speaker cables to allow the NFB to be derived from the speaker terminals directly and hence the NFB theoretically compensates for the speaker cable.

In the Avondale amps, which are closely related to the Naim circuitry and in effect a reworking of them to remove the more glaring faults, a Thiele Network IS used and the resistor removed... jobs a gud 'n.
The only reason I can think of as to why Naim have never corrected this glaring engineering error is that it has become a part of "Naim lore" and to many non technical people possibly looks like the unit is "so special", "such a thoroughbred" that it needs "special cable" :rolleyes:

A corollary to all this is that in the case of low or no NFB amps often no Thiele Network or resistor is required as the feedback is low/non existent and so it can't be "messed with" by the load/cable to cause any instability.
Click to expand...

I have no idea what you are going on about..

Tell me.. in simple terms.. Is this why I think NAIM amps mostly sound crap? :D
 
Last edited:
Used with a normal cable of only a couple metres long, Naims are likely to oscillate a bit when the output stage switches. The result is harsh sound
 
herb said:
Dave Clark bought a lot of the broadcasting rights to the rest
Click to expand...

I believe so. I can't recall the detail. but ISTM that he is hanging on to some stuff which will lose value if he doesn't release it while those who remember it are still alive.

In a wider sense.. it was just coincidence that I came across the DC5 to make a point about what really constitutes 'old' TOTP performances. There are many 'gems' still about. This vid has a real mixed bag. a lot of what I'd call dross... but also some great stuff, some 'foxy chicks' and a real sense of the 60's 'vibe'. And of course the late Stuart Henry.

 
Arkless Electronics said:
A corollary to all this is that in the case of low or no NFB amps often no Thiele Network or resistor is required as the feedback is low/non existent and so it can't be "messed with" by the load/cable to cause any instability.
Click to expand...
Emitter and source followers oscillate into a range of capacitance even if there is no overall negative feedback. The only sure way to prevent this is either an output network or a series base/gate resistor. You can draw the stage, with its strays, as a Colpitts oscillator.
 
davidsrsb said:
Emitter and source followers oscillate into a range of capacitance even if there is no overall negative feedback. The only sure way to prevent this is either an output network or a series base/gate resistor. You can draw the stage, with its strays, as a Colpitts oscillator.
Click to expand...

Yes I'm well aware of this and it's exactly what I was talking about... More usually it is a negative resistance oscillator. It is not relevant to this and not a problem in power amplifiers where the issue is very much related to NFB.
 
Got lazy and ordered 2 x 6m terminated lengths of VD Blue 2.5mm from Mark Grant, so I can easily compare it to the Chord stuff, then post a way OTT flamewar-starting thread about night and day differences :p
 
You put your speakers out of phase and facing each other to cancel the sound. Then you play really HOT music continually for weeks. Day and night. No listening. You need to completely forget what your system sounded like, so that you can truly claim it sounds different after burn in. Cool Jazz is no good. It doesn't burn the wires in. The Red Hot Chilli Peppers are better... especially using this method, because you don't have to hear them.
Classical listeners report good results from Handel's Fireworks and De Falla's Ritual Fire Dance.

Under no circumstances have the wires coiled, or even crossing, as the heat of the burn in process could build up in hot spots and cause a fire.
 
deebster said:
Got lazy and ordered 2 x 6m terminated lengths of VD Blue 2.5mm from Mark Grant, so I can easily compare it to the Chord stuff, then post a way OTT flamewar-starting thread about night and day differences :p
Click to expand...
Chord Sarum is only 0.8mm2, so at the edge of high enough resistance that you just might be able to hear it with an extreme low Z load, like some B&W in the deep bass or a Martin Logan at the top end.
 
Darth Vader said:
However some amplifiers e.g early Naim are affected by these small values. Transistor amps employ emitter followers in their output stages and these are unstable and tend to oscillate in the UHF. To counter this a Zobel network is placed at the output that connects to the speaker. Naim thought they were being 'clever' and removed a critical component from the Zobel network a series inductor of around 5μH. So with a 'normal' speaker cable the output stage would oscillate in the UHF and heat up and might actually burn out. This loss of power to UHF may cause distorted output.


DV
Click to expand...


IIRC the oscillation bursts for the Naim amp I tested mumble decades ago was at more like a couple of MHz, not UHF. To me that's an overally feedback loop + load 'Barkhausen' problem. So a matter of overall design, not just the output topology.

The simple solution is to shove a few microH inductor in series with the output close to the output. Standard part of the signal/load conditioning network. Can make a mess of squarewaves in the ultrasonic, but ensures stability for an amp that isn't inherently uncoditionally stable.
 
Jim Audiomisc said:
IIRC the oscillation bursts for the Naim amp I tested mumble decades ago was at more like a couple of MHz, not UHF. To me that's an overally feedback loop + load 'Barkhausen' problem. So a matter of overall design, not just the output topology.

The simple solution is to shove a few microH inductor in series with the output close to the output. Standard part of the signal/load conditioning network. Can make a mess of squarewaves in the ultrasonic, but ensures stability for an amp that isn't inherently uncoditionally stable.
Click to expand...

Yep. I covered it in some detail up thread...

I would disagree though that adding a Thiele network can make any amplifier unconditionally stable. Semantics really but it's almost always a part of the overall strategy to obtain stability. Very few amplifiers have neither a Thiele network or a series resistor.... Those that don't have very low or even zero NFB or in rare cases just a heap of phase margin.

It can be worthwhile doing empirical testing of the required size of damping resistor when it comes to square wave ringing as you can sometimes get away with 1R or less and this pretty much kills the ultrasonic ringing :)
 
You must log in or register to reply here.


advertisement


Back
Top