[95bcwh]

All,
  I have read in many posts that in order for passive preamp to work in our system, one of the criteria is to have short interconnect with low capacitance.  But I haven't been able to find any quantitative description of this low capacitance.. how low is low?

  I now have a 1 meter Zu Oxyfuel IC and intend to use it to connect my source to a resistor-based attenuator before going into a power amp. According to the specs, the capacitance of this IC is 56pF. Is this low enough?

 Thanks for sharing
 barry

[dado5]

56pF is middle of the road if it refers to the full meter. If it is per foot, it's  high. The standard is Belden coax which is around 20pF or so a foot (some types are as low as 15pF/ft).

The extremes would be Nordost/DNM/Mapleshade parallel spaced conductor cables with pF capcitance in the single digits per foot.  The high end would be Goertz  sandwiched flat conductor cables.

The issue is that the capacitance and the resistance of the cable + the amp provide an AC impedence to the source that is not uniform with frequency. It is a low pass filter network. The higher the capacitance and/or resistance the lower the cutoff  frequency of the filter, to the point where it may be in the audio range. So the 'safest'  passive in this regard is a low resistance attenuator + low capacitance cable.

The big question is do you like what you hear? If so f@#k the specs and love the music.  If not  try lower capacitance cables and see how they sound.

[SET Man]

...The big question is do you like what you hear? If so f@#k the specs and love the music.  ...

Hi,

     Exactly! My motto is "Listen first, measure later!" You know what.... I think my system measure like sh*t Well, than again if is all about measurement than I won't be using tubed for sure.

   Anyway, back to the IC... yes low capacitance is very important in cable design well at least to me. And as dado5 said resistence also, this extent to the wire qaulity/ga and shilding.

    I made my own IC based on parallel spaced conductors and they work great with my system... so no more buying $$$$ brand name cable for me!

   But than the bottom line is synergy... so if you can borrow some cables and try them yourself and see.

Take care,
Buddy

[Steve]

Just thought I would mention that I measured some different expensive RCA plugs for capacitance, and they seem to range around 10pf per plug. So two would seem to mean around 20pf minimum with any wire/cable.

Of course, the jacks will add some more, plus any stray capacitance within the components themselves.

Cheaper, less quality plugs might have less or more capacitance, but the tradeoff would be brass with whatever coatings.

Hope this helps.

[95bcwh]

Maybe measurement is indeed not an issue at all. I just checked Zu Gede IC, which is 2 times the price of Oxyfuel, the Gede has capacitance of 242 pF per meter!!  

[dado5]

Just thought I would mention that I measured some different expensive RCA plugs for capacitance,

Good point Steve.  I recall back in late 80's or so one cable company (I want to say Kimber) made a point of the fact that they included the connectors  in their cable specs while most others did not.

I would think the lowest cap connectors would be the 47labs/Eichman type and I would wager the big Cardas and WBT plugs are the highest.

All that aside I must say that IMO the RCA is a poor connector design. DIN is much better and direct solder is best, even wire nuts are preferable. Resale consideration is the only thing that makes me use RCA plugs.

[Roger A. Modjeski]

Barry,

It's the cable FROM the passive preamp TO the power amp that matters. If the source is  CD player, it has low output impedance and can drive most anything.

You need to now just two things.
1. the output impedance of the passive preamp (worst case, i.e. highest value).
2. the total capacitance of the interconnect and the input impedance of the power amp (which is usually small and around 100 pf).

Now the math....    The 3db down point will be F=1/6.3xRxC, where R is the worst case output impedance of the passive in ohms and C is the cable capacitance (plus 100pf, its a guess, but a good one) in Farads. Pico farads are ten to the minus 12th power.... hope you like math.

Here's a typical example worked out. Say you have a 25K pot. The worst case output impedance (assuming a low Z source like a CD player) is about 6 Kohms. and let's say the cable is 500 pf (which is kinda high, but some are) you will be 3 db down at 53 KHz and next to nothing at 20 KHz.

Another way to use this formula is to decide where you are willing to be 3 dB down and see how much capacitance you can tollerate. Choose about 40KHz so you won't be more that a dB down at 20KHz. Then C=1/6.3xRxF. (simple algebraic rearrangement solving for C).

I've used pots as high as 100K with a few hundred PF and it's just fine.

Roger A. Modjeski
RAM LABS/Music Reference

[amplifierguru]

Hi Roger,

Sorry to throw one from left field but source impedance as high as even 1K can seriously degrade both the PSRR of the amplifier and it's THD.

I recommend using max 10K stepped attenuator/pot and finding the typical user setting , say 800 ohms, subtracting that from the 1K input filter on my amps which are balanced with the feedback R. This ensures the PSRR of the amplifier is maintained (so you are not listening to PS artefacts -ugh!) and THD is as spec'd, both in magnitude and makeup.

Cheers,
greg

[Roger A. Modjeski]

greg,

That may be true of your amplifiers, but not of my amplifiers. They are unaffected by any reasonable source impedance. Frankly if you are talking Power Supply Rejection Ratio (PSRR) how are you gonna deal with any reasonable source impedance variations from one brand to another.  Many tube preamps have a few Kohms output impedance. How are you going to handle that? What kind of input circuit do you have?

[Terry Demol]

Hi Roger,

Sorry to throw one from left field but source impedance as high as even 1K can seriously degrade both the PSRR of the amplifier and it's THD.

I recommend using max 10K stepped attenuator/pot and finding the typical user setting , say 800 ohms, subtracting that from the 1K input filter on my amps which are balanced with the feedback R. This ensures the PSRR of the amplifier is maintained (so you are not listening to PS artefacts -ugh!) and THD is as spec'd, both in magnitude and makeup.

Cheers,
greg

Greg,

Are you referring to CM distortion of IP pair due to mismatch of
impedances between IP side and FB networkside? Ideally these
impedances should be identical but that is hard to get in the real world,
as Roger stated.

Do you have a series R that feeds the RF shunt / slew limit cap to gnd
on your amps?

How is PSRR affected by the IP Z (within reasonable limits). Is his
a side effect of the impednace mismatch as stated above?

Cheers,

Terry

[95bcwh]

Barry,
Now the math....    The 3db down point will be F=1/6.3xRxC, where R is the worst case outp ...

Roger,
 Thanks! I do have some engineering background (not electrical though) so I do like math

  But then, when I plug in your numbers into the equation I don't get the 54kHz that you mentioned:

  F = 1/ (6.3 * 25e3 * 100e-12) = 1.27e4 Hz or 12.4kHz ?!!

 Perhaps I need to start reading up some electrical text books related to audio frequency..

  Actually I have just sent my system specs to Luminous audio and ask them to do a calculation on what impedence should the passive attenuator have:
----------------------------------------------------------------------------------
Speaker efficiency at 1w/1m = 85dB, 8 Ohm

Input imp  of  power amp = 50K Ohm

Power amp has to level of gains:
(1)1.8Vrms in, 29 dB gives 300W into 8 Ohm
(2)3.6Vrms in, 23 dB gives 300W into 8 Ohm

Output voltage of SB3 = Max 2.1 Vrms, output impedence ~230 Ohm.
-----------------------------------------------------------------------------------

 They have responded back and said I need only 1.2k Ohm  impedence!!    Anyone can guess what formula are they using?

Rgds
barry

[amplifierguru]

Hi Roger,

I was commenting in a generic sense for amplifiers with a typical differential (or dual differential as mine) input stage.

Hi Terry,

Yes CM performance is degraded by unequal impedances on the input and FB sides. I use 1K FB R and !K input filter R - both very common values. If you sim for PSRR (as I've hilighted many times in DIYA) and experiment with source R's you will find both THD (probably CM related) and PSRR deteriorate quickly outside balance.

My solution , for those with passive vol feeding directly is to find their typical setting, measure the Rout of it and reduce the input 1Kby this amount thus re-establishing approximate balance.

It's a rough and ready solution, a buffer being best.  Most amps are designed for low Z drive and most pre's have < 100R out. Inthe case of tube amp outputs of 5K, you must like the consequences or voltage divide the output down to a 1K. It WILL degrade PSRR on most amps - mine start off a high base, so there's somewhat less effect.

cheers,
greg

[Roger A. Modjeski]

Barry,

Nice math, but you used the pot's resistance value in the equation. Go back and note what I said about the impedance to use in the math. A pot driven by a low Z source sees ground at the bottom and another zero on the top. At mid electrical rotation there is half the resistance in parallel with the other half so the impedance at the pot is 1/4... cool huh. Ask a dozen people who think they know ohm's law this question and see how few get it right. Its one of my favorites.

I have no idea what the 1.2 k refers to without more info.

roger