I don't believe cables make a difference, any input?

[haldor]

Originally posted by MBK
Example 2. A confusing one for me. In relay specs I read not just a maximum rated V and I, but also a minimum rated V and I. That makes me uncomfortable: does that mean now that under AC, that contact will actually have a nonlinear behaviour at the zero crossing? It seems so. And this is just an ideal zero resistance contact we're talking about.

Originally posted by MBK
The ratings are for the signal portion, not the coil portion. For those relays (signal level of course) where these data were given the maximum rating was typically around 125V, 1 A, the minimum rating about 1-10 mV, 10 uA.

Oh boy, Relays. I actually know something about that this.

Relays are catagorised as either being power relays or small signal relays (or both, but you have to pick one mode and never use it for the other).

Power relays have silver alloy contacts that tarnish (oxidise) which increases contact resistance. At some current level, electron flow has a cleaning effect on the contact which removes the tarnish. Below that current flow, the contact is not cleaned and the contact becomes unreliable.

Small signal relays have a noble metal coating over the silver alloy contacts (usually gold) which prevents tarnishing and oxidation. The problem is that the gold coating is very thin and if you ever pass too much current through the contact the gold plating is destroyed and the relay is no longer reliable for small signal use.

This is why relays have both a minimum and a maximum current rating. The maximum rating is how much it can safely handle and the minimum is how much current is required to keep the contacts clean.

So yes, outside it's ratings a relay is a non-linear device. However this is only over a fairly long time scale. If you are passing an AC signal through a relay there is insufficient time for an oxide coating to form during the zero crossing.

Phil

[SY]

Phil, great info. Where do mercury-wetted relays fit in this universe?

[MBK]

Thanks Phil,

nice to have a clear and definitive solution to that one! I was really wondering what to make of these relay specs...

[dkemppai]

Originally posted by IanHarvey
You cannot create nonlinearities by adding parasitic L's, C's and R's about a component - it requires something more exotic (such as capacitance which changes with applied voltage).

Cheers
IH

...or inductance which changes with current. Such as any inductor based on a core material that does not exihibit a linear BH curve (Like most used for inductors and transformers).

Also the nonlinearities depend on the domain you're in. If you're looking at a time domain of a single frequency, what you say may be true. However, if looking at frequency domain, the parasitics can wreak havoc... ...the L's and C's become frequency dependent, thus nonlinear depending on frequency. Think of why you don't use speaker cable to transmit Radio frequency energy!


-Dan

[IanHarvey]

Originally posted by dkemppai


...the L's and C's become frequency dependent, thus nonlinear depending on frequency. Think of why you don't use speaker cable to transmit Radio frequency energy!

Not sure I understood this bit. You can vary any parameter you like with frequency; just so long as it obeys the 'superposition' principle, its effect will remain linear.

(The main problems with using speaker wire for RF outputs is the difficulty of getting stuff of the right characteristic impedance which terminates neatly into commonly used connectors. Consider the use of a bent wire coathanger as a TV ariel - it's not even sold for its electrical properties, but conveys a useful amount of signal even at >400Mhz).

Cheers
IH

[SY]

Actually, there's a lot of zip cord that has a characteristic impedance of about 75 ohms- as a teenager, I found that useful in emergencies for hookup to my 40 meter dipole. And a few of us oldsters will remember 300 ohm twin lead, which is constructed like zip cord (PVC extruded over stranded copper), but with the conductors further apart. Worked well at a few hundred megahertz.

[mrfeedback]

Naim amplifiers specify Naim speaker cable (NAC-A5) which looks like oversize 300 ohm TV aerial ribbon cable.

By definition this is relatively highly series inductive and very low shunt capacitive, and indeed Naim amplifiers require this inductive loading to ensure stability - overly capacitive cables will cause Naim amplifiers to oscillate.

Naim NAC-05 Page
Owners who say that cables DO make a difference
Naim saying that cables Do make a difference

Eric.

[mrfeedback]

And....

[MBK]

Speaking about nonlinearities - does noise have any well defined relationship to signal level? For all I know, voltage noise and current noise in resistors is related to R and to f. Incidentally such constant, amplitude-independent noise would then be a nonlinear effect in such a simple thing as a resistor... or not?

Anyway - listening to actual recordings I sometimes have the impression that with rising amplitude the signal - say, a loud piano note - is like "enveloped in a noise cloud". That of course may be a) an illusion b) derive from active components (rather than passive ones).

Anyway if the noise would truly rise with amplitude it could at least possibly do so linearly ... yet if this is so, then that would complicate the assessment of real world S/N ratios enormously, wouldn't it?

[jh6you]

Originally posted by mrfeedback

Naim amplifiers specify Naim speaker cable...

In a way somewhat disapointing as Naim is shy of strangers.

JH