Phantom Power Capacitors

[gyraf]

..it's a 1uF, loaded 1K Ohm

F(c) then at 159Hz

so off course you get that - I'm not sure what else you'd expect?

Rule-of-thumb always was: Set your F(c) way out of audio band, do not use as crossover

 

[abbey road d enfer]

Making something that makes sense.

Because you're the only one here that makes sense...

But you define the problem very poorly and you are deliberately creating an unsolvable task by unnecessarily denying available resources to the solution.

Not really. I'm putting reasonable constraints.

+/-18V and +48V regulated and > 54V unregulated under all operating conditions

Unreg voltage is not always available, for example in the case one uses an smps.

Is there any problem? This 2024, not 1980. It should be trivial. Nu?

Wayne Kirkland proposes a solution. It is probably different than what you would do, but it certainly puts at work the same principles.
It took him a few years to arrive at a satisfying solution.
Atfer all, CRT oscilloscopes presented the same kind of issues, that were solved about 70 years ago.

Now, you may think you're much smarter than him.
I don't know one single mixer manufacturer that uses a similar approach. It's a matter of return on investment and perceived value.
And again, in practice, it's a solution in search of a problem.

 

[merlin]

Make a Microphone Preamplifier that requires no coupling capacitors in to out, including any such coupling capacitors inside feedback loops,

Would this include no cap to keep DC off the gain pot?

 

[thor.zmt]

Would this include no cap to keep DC off the gain pot?

Of course, I mean this 2024, what kind of person would still put a cap there?

Thor

 

[JohnRoberts]

Back when I was still thinking about this, one idea I floated during my late night beer soaked scribbles, was floating an analog mic preamp up to phantom voltage (more like 24V than 48V and also floating a good quality A/D up there with it). The digital output could be optically coupled out. Look ma... no capacitors in the audio path. The last time I was chewing on this (decades ago) I could not settle on a no-compromise A/D to use for a presumably exotic high end design.

Perhaps high end A/D sellers could incorporate some version of this into their SKUs as a marketing hook?

JR

 

[abbey road d enfer]

Even a regular electro under the worst conditions will produce distortion well below 0.001% at mic levels.

I don't know where this comes from, but it seems to show a singularity at about 4.5V, where THD is minimal. I'm not sure all electrolytic caps show the same behaviour.

 

[Mr. Moscode]

Did anyone mention dissipation factor, dielectric absorption, stability vs time or vs temp or vs F? 1 test of an electrolytic at 1 frequency at 1 moment in time for 1 parameter at 1 age does not a study make.

This is an interesting paper on the subject: https://vbn.aau.dk/ws/portalfiles/portal/484307748/FINAL_VERSION.pdf

 

[abbey road d enfer]

Did anyone mention dissipation factor, dielectric absorption, stability vs time or vs temp or vs F? 1 test of an electrolytic at 1 frequency at 1 moment in time for 1 parameter at 1 age does not a study make.

This is an interesting paper on the subject: https://vbn.aau.dk/ws/portalfiles/portal/484307748/FINAL_VERSION.pdf

Indeed it's interesting, but I have never seen a distinct correlation between DF and distortion.
DF is a very important factor in reliability, though.

 

[JohnRoberts]

Why would you do this?

Do you know how (for example) how a high side current sensor IC is build, that senses current on +72V rail and has a ground referenced output?



Way too slow. But modern 100Mbps+ digital isolators now have kV isolation rating and single digit dollar pricing.



I think a ground referenced normal audio level out is easier and less headaches for system integration.

It is always best if a "new thing" is an improved drop in for the "old thing".

If we incorporate A/D, why not use the same style as a digital out MEMS microphone and use the capsule as modulation element for a a 24.576MHz PWM modulator, say 5th or 7th order with physiological noiseshaping?

No more Anna Log anything.

Perfect sound forever.

Thor

IMO this DC coupled stuff is all mostly marketing hyperbole.... I know I can't hear 0.0000x % THD

I recall talking with a fellow console designer (now RIP) about doing this (DC coupled) inside a console and he was not convinced that there was any merchantable (sellable) benefit.

Another quasi marketing suggestion is to have the phantom power switch short across the phantom capacitors when phantom voltage is not needed... I suggested he could market this DC coupled mic preamp to the phools making passive summing boxes to use for the needed make-up gain.

JR

 

[abbey road d enfer]

Of course, I mean this 2024, what kind of person would still put a cap there?

That said, as I normally shoot for MCU control of gain via SPI it idles not really meaningful anyway.

But no caps in series with switchable gain setting resistors (you could use a pot if you want to be boring and conventional - in Thailand we smoke pot, not put it in the signal path).

One secret is that where you see the need for a "big ole evil polar cap" (this is a formal trademark of Hamas used with kind permission) I see the need of a servo with correct pull range, using 0.1uF C0G 0805 SMD capacitors.

Unmicrophonic, undistorting, unexpesive, unlarge etc... Essentially use an "electronic capacitor" over a real one.

There is a reason I specified a Quad OPA (at 40 cent USD pef quad in 1ku) in my problem definition, where normally two will do the job perfectly well.

The two spares are used as servos to eliminate any coupling caps.

And at the core it's still a Cohen meets Birt mic-pre dating back to 70's actually and trivial in 2024 - even as IC integrating the 5V input PSU on chip.

The biggest challenge for an IC is the need for a bimos process with nice PNP's which for example TI and AD have these days, TSMC too.

Incidentally, done properly the Ein can be < -142dBu at max gain (and no, I do not consider this exceptional or an achievement, it would have been in 1980).

So asking -130dBu I was giving out freebies.

Thor

So, what do you propose in practice?

 

[JohnRoberts]

Indeed it's interesting, but I have never seen a distinct correlation between DF and distortion.
DF is a very important factor in reliability, though.

There was a great deal of drama in the audiophool community back in the 1970s about DA, DF and other obscure capacitor performance parameters with claims of audibility in typical line level audio applications (perhaps in a passive loudspeaker crossover).

JR

 

[jp8]

I concur with john12ax7 that there are some exceptions.
In order to pinpoint more accurately, it would be interesting to make the same tests with separate capacitors.
It is quite possible that some effects in one type compensate those in the other type, so it makes it very difficult to analyze.

Agree. Even order harmonics might well be partially canceled out by the test setup that I used. But I don't want to temporarily modify the preamp of my Audio Analyzer just for the purpose of this test. Maybe we'll get a better picture of whether THD will rise or fall under certain conditions, but the outcome will still be that we're measuring THD levels that are orders of magnitude below what anyone will be able to hear. Maybe something that I'll do when I'm retired...

Jan

 

[jp8]

So, what do you propose in practice?

Maybe something as described in this thread?

 

[ruffrecords]

Incidentally, done properly the Ein can be < -142dBu at max gain (and no, I do not consider this exceptional or an achievement, it would have been in 1980).

So asking -130dBu I was giving out freebies.

Thor

What source impedance and bandwidth does this -142dBu EIN assume?

Cheers

Ian

 

[thor.zmt]

What source impedance and bandwidth does this -142dBu EIN assume?

Input short, 20Hz-20kHz.

Thor

 

[abbey road d enfer]

Maybe something as described in this thread?

No. Thor has clearly expressed his disdain for Wayne's approach.

 

[ruffrecords]

Input short, 20Hz-20kHz.

Thor

On I see, a marketing spec not a real world one.

Cheers

Ian

 

[abbey road d enfer]

Using one's brain. Even if it is as poor as my own.

And defining/framing the problem correctly and then looking for the solution, as opposed to determine the solution and then trying to make it somehow fit the problem.

And no, I'm not posting more details, on the principle that all these minor elaborations of the basic principles are obvious to those skilled in the arts, as it commonly says in patents.

All that is needed was in thread above.

Thor

Sounds like an Agatha Christie mystery, in which you are Poirot.
I know, it's Colonel Vertdegris, in the slaughterhouse, with an enema.

 

[thor.zmt]

On I see, a marketing spec not a real world one.

Ein. Equivalent Input Noise.

If you want to measure resistor noise, as opposed to the circuits self noise, be my guest.

The source impedance is not standardised. Yes, 150 Ohm is often used.

Let's use a Schoeps circuit with 4mA in each follower. Zsrc is 13 Ohm.

Mind you the microphone noise is likely more than 150R equivalent.

So we might as well use Transformers and Tubes and accept extra noise that is still less than gaussian noise of air molecules bouncing off the diaphragm.

;-)

Thor

 

[thor.zmt]

Sounds like an Agatha Christie mystery, in which you are Poirot.

Nah, more Inspector Gadget.

I know, it's Colonel Vertdegris, in the slaughterhouse, with an enema.

Wrong.

It was the Butler. The killer is always the Butler, except if it's the gardener.

Or the German Sniper.

Or some guy with a knife shouting in Arabic.

Thor