From an audio forum... FR4 question

[John Larkin]

That story is some kind of deceitful ruse, no such thing occurred...

Hey, Fred, see pics in abse.

Quoting Johnny Depp in the Willy Wonka movie,

"You are REALLY weird."

John

 

[mw]

I submit that if you open up your amp and start tapping parts with a
pencil, the loudest noise you'll hear is the sound of parts being
tapped with a pencil. To solve this particular problem, quit hitting
parts with pencils.

What if the "tapping" is caused by environmental factors, such as road
vibration, acoustic waves from loudspeakers? Do you just stop the
vehicle by the side of the road? Tell the musicians using your
amplifier at the concert to stop playing so loudly?

My NMR gradient drivers have a couple of PPM noise, dc to 50 KHz, and
use regular surface-mount parts on FR-4, with lots of noisy fans.
Aside from pure semiconductor and resistor noise, the next biggest
hazard is magnetic loop pickup from fields leaking out of transformers
and fans.

Sounds like your NMR machine must be a large heavy device, mounted in a
quiet building somewhere. No worries, then.

How will a microphonic cap make hissy, buzzy audio?

Maybe you missed my earlier post:

Ref: Linear Tech AN83, page 14:

"A piezoelectric device generates voltage
across its terminals due to mechanical stress, similar to
the way a piezoelectric accelerometer or microphone
works. For a ceramic capacitor the stress can be induced
by vibrations in the system or thermal transients. The
resulting voltages produced can cause appreciable amounts
of noise, especially when a ceramic capacitor is used for
noise bypassing. A ceramic capacitor produced FigureÊ B4’s
trace in response to light tapping from a pencil. Similar
vibration induced behavior can masquerade as increased
output voltage noise."

I'm trying to be helpful.

mw

 

[John Larkin]

What if the "tapping" is caused by environmental factors, such as road
vibration, acoustic waves from loudspeakers? Do you just stop the
vehicle by the side of the road? Tell the musicians using your
amplifier at the concert to stop playing so loudly?

Consider that some external source is producing vibrations in the
audible range. Assume the level in the listening room is X dB. How
much additional sound level will be added to the X dB by the gain
associated with a microphonic capacitor and the following
amps/speakers? In other words, the truck itself will make a hell of a
lot more noise than what the truck induces into the cap.

Ever hear a system howl from microphonic caps?

Sounds like your NMR machine must be a large heavy device, mounted in a
quiet building somewhere. No worries, then.

The superconductive magnets themselves are supported by pneumatic
vibration isolators; my gear is not. Again, if your listening room is
next to a railroad track, it's going to be noisy now and then, and
microphonic caps won't add measurably to the din.

John

 

[Boris Mohar]

Consider that some external source is producing vibrations in the
audible range. Assume the level in the listening room is X dB. How
much additional sound level will be added to the X dB by the gain
associated with a microphonic capacitor and the following
amps/speakers? In other words, the truck itself will make a hell of a
lot more noise than what the truck induces into the cap.

Ever hear a system howl from microphonic caps?


John

I wonder how much distortion do ceramic caps produce by reacting to the
stress produced by their own internal electrical fields that are set up by
the signal itself.

 

[Ken Smith]

I wonder how much distortion do ceramic caps produce by reacting to the
stress produced by their own internal electrical fields that are set up by
the signal itself.

The ceramic caps that have a microphonics problem are already non-linear
for other reasons. The high K ceramics have both issues. NPO/COG don't.
If a signal voltage is to be developed, it is best to use NPO capacitors
in that part of the design.

 

[Pooh Bear]

I wonder how much distortion do ceramic caps produce by reacting to the
stress produced by their own internal electrical fields that are set up by
the signal itself.

No decent audio designer wil use ceramic caps in signal paths other than
NPO/COG types.

Medium and Hi-K dielectrics are well known to distort the signal.

Graham

 

[John Larkin]

No decent audio designer wil use ceramic caps in signal paths other than
NPO/COG types.

Medium and Hi-K dielectrics are well known to distort the signal.

Graham

How can they do that? A coupling cap almost by definition is a low
impedance at signal frequencies, so the voltage drop across it is
small. And if it's inside an overall feedback loop, any nonlinearity
is further reduced.

The only interesting case is a cap that itself defines a low-frequency
rolloff point. If used at high signal levels, that one could in theory
generate harmonics. In practise, I suspect the effects are generally
small.

John

 

[Fred Bartoli]

How can they do that? A coupling cap almost by definition is a low
impedance at signal frequencies, so the voltage drop across it is
small. And if it's inside an overall feedback loop, any nonlinearity
is further reduced.

The only interesting case is a cap that itself defines a low-frequency
rolloff point. If used at high signal levels, that one could in theory
generate harmonics. In practise, I suspect the effects are generally
small.

How do you make filters with zero AC voltage across caps?

 

[[email protected]]

I had a pll running at 155.52 MHz with a narrowband (2 khz) loop, on a
VME module. Whenever one pulled an SMB test cable out of its
connector, the loop briefly lost lock. After some tapping and bending,
we found it was the expensive Vectron crystal oscillator, not the
caps. We designed some tiny springs to isolate the oscillator can,
made it wobble like one of those gooney-head dolls you see in the
backs of cars.

Holy crap! I had exactly the same problem with a Vectron VS700
155.52MHz VCXO. Only I used an SMA connector for the test points
Stress on the board would pull the oscillator frequency.

My fix was to ignore the problem, as that level of stress only happened
during testing.

Regards,
Allan

 

[Pooh Bear]

How can they do that? A coupling cap almost by definition is a low
impedance at signal frequencies, so the voltage drop across it is
small. And if it's inside an overall feedback loop, any nonlinearity
is further reduced.

You're kidding me right ?

Non-linearities in feedback loops can be amplified !

A least, that certainly applies to noise. Cue an unforgettable repair moment when
a noisy transitor amp stage was discovered to be due to a defective C-E
capacitor.

Anyway, the reason that med and hi-K caps distort is due to a non-linear
coefficient of capacitance with applied voltage. And ceramics aren't normally
used for coupling. I'm thinking of EQ applications for example.

The only interesting case is a cap that itself defines a low-frequency
rolloff point. If used at high signal levels, that one could in theory
generate harmonics. In practise, I suspect the effects are generally
small.

Oh, I've measured that. Electrolytics used for coupling that have greater than ~
100 mV A.C. across them generate some weird products. It's not especially
low-level either. The trick is to *never* have that much voltage there. I.e. use
big caps. Using the 'bipolar' versions doesn't help either despite much common
belief to the contrary.


Graham

 

[Fred Bloggs]

Hey, Fred, see pics in abse.

Quoting Johnny Depp in the Willy Wonka movie,

"You are REALLY weird."

John

The lengths some people will go to perpetuate their lies...is this some
kind of stock option scam on Vectron?

 

[Ken Smith]

Oh, I've measured that. Electrolytics used for coupling that have greater than ~
100 mV A.C. across them generate some weird products. It's not especially
low-level either. The trick is to *never* have that much voltage there. I.e. use
big caps. Using the 'bipolar' versions doesn't help either despite much common
belief to the contrary.

The bipolar ones are often the worst way to go. Using two polar ones and
applying a large pre-bias works better.

 

[John Larkin]

How do you make filters with zero AC voltage across caps?

Digital!

But seriously, some ceramic caps have a nonlinear voltage/capacitance
curve, so if you use them in filters where harmonic generation might
be an issue, just keep the signal level low relative to the cap rated
voltage. Or test the damned things.

High-voltage nonlinear transmission lines (shock/soliton lines) have
been made with ceramic caps as the nonlinear elements.

John

 

[John Larkin]

You're kidding me right ?

Non-linearities in feedback loops can be amplified !

You're kidding me, right?

John

 

[John Larkin]

The lengths some people will go to perpetuate their lies...is this some
kind of stock option scam on Vectron?

You don't believe that crystal oscillators are shock/acceleration
sensitive? You think we designed and fabbed these springs to trash
Vectron stock?

OK, I admit it all.

John

 

[John Larkin]

Holy crap! I had exactly the same problem with a Vectron VS700
155.52MHz VCXO. Only I used an SMA connector for the test points
Stress on the board would pull the oscillator frequency.

My fix was to ignore the problem, as that level of stress only happened
during testing.

Regards,
Allan

Geez, don't tell Fred!

I don't think it's Vectron especially; all XOs are shock/acceleration
sensitive to various extents. SC-cuts are supposed to be better, at
roughly 3-5x the price.

After we spring-mounted the cans, you could whack the board all you
wanted with no effect; I recall that the spring-mass mechanical
resonance was in the 15 Hz range maybe, enough to lowpass filter the
shocks pretty well, enough for the pll to track easily. But just the
delicate touch of a small screwdriver against the can would make us
lose lock for some milliseconds.

This is a very dumb pll, using a single EclipsLite flipflop as a
bang-bang phase detector, so it wasn't the best tracker in the world.
Once we had a solid lock, the uP reduced the loop bandwidth from
roughly 5 to about 1 KHz to reduce jitter, which made it even easier
to kick out of lock. This is for the NIF laser, running megabucks per
shot, so I didn't want to ignore the problem.

John

 

[Steve Sousa]

All these "audiophooles" keep going on about OFC etc.
Then I try to imagine (badly) what the hell is going on at the
molecular level inside a decoupling electrolytic. It must surely
outweigh FR4/Teflon problems

Is there any truth behind the argument that OFC doesn't oxidise as much
as normal copper, so it gives less trouble at the speaker and amp
connectors?

 

[keith]

All electrical-grade copper is OFC.

OMG, are the audiophools going be pissed at you!

 

[John Larkin]

Is there any truth behind the argument that OFC doesn't oxidise as much
as normal copper, so it gives less trouble at the speaker and amp
connectors?

All electrical-grade copper is OFC.

John

 

[Ted Edwards]

You don't test audio equipment like this. You need meters for....

Musicality
Etch
Resolution
Glare
Detail
Tightness
Speed
Depth
Size
Location
Pace
Brightness
Balance
Flavour
Bloat
Cleanliness
Extension
Energy
Warmth
Fullness
Weight
Smoothness
Liveliness
Sparkle
Power
Sweetness
Correctness
Naturalness

and that is just for cables.

Especially line cords.

Ted