Discrete Opamp Open Design

Member
Joined 2002
Paid Member
George,

Your diagram shows explicit filter elements or intrinsic resistances and inductances?

It showed intrinsic resistances and inductances.

A bit of this depends on exactly which, since serial filtering *might* yield vanishingly low noise and hum at the end of the chain.
_-_-bear

This is correct bear. It comprises a cascaded RC or LC low pass filter.
In terms of noise and hum reduction is very good and is an economic solution. In theory and more so in practice though has one drawback: Each succeeding stage's DC, rides on the modulations of the previous. If you probe there with an oscilloscope, you will see low frequency modulation. This, together with the voltage drops at the common return line which are different at each stage, cause the signal input from stage to stage to low frequency wander up and down. Visualize a coupling capacitor, the two ends of which are on different, -unequally moving- DC potentials. This, in severe cases may give rise to low freq. oscillation (motorboating).
All these do not occur with parallel connecting the amplifying stages to the PSU

For clarity, I have attached a new sketch, this time with explicit filter elements.
Circuit B is no good at all. It is not a parallel connection (common return) and it lacks the cascaded low pass filtering of the serial circuit.

a good way is to determine the stage with lowest PSRR and put the psu closest to that.
jan
With series connection, the circuit closest to the PSU has to be the one with the higher current demands. The furthest away has to be the least hungry for current.
And, if you're serious, use remote sensing for the regulator at that stage.
jan

Do me a favor. Have a look at my avatar please :p



I know that phono cartridges with -20dB crosstalk can just sound fine. Some bleeding in the other channel may even enhance the perception of spaciousness…I rather prefer that not be a characteristic of the electronics.

Good stereo separation in the electronics is VERY IMPORTANT. Don't be fooled by phono cartridges Joachim.

Joachim
This time :D Mr. Curl is correct :cheers:
The crosstalk within the cartridge is relative “clean”.
The crosstalk between L/R channels in electronic circuits happens mostly through the common PSU lines and less so through spacial multiple capacitive/inductive couplings. All these produce “dirty” crosstalk signals (undefined amplitude phase/delay relationship).

George
 

Attachments

  • PSU 2.jpg
    PSU 2.jpg
    57.1 KB · Views: 397
George,

There is a 4th possibility. That would be the first example, but with the grounds of the third.

Also, if the L used is sufficiently high, and the L/C combination "tuned/scaled" well it seems unlikely that there would be much VLF passing. I suppose that if a prior stage was a big current hog that had wild current draw swings, perhaps. But this is not likely for class A stages certainly, class AB stages that are output stages might be another story. There, if we're talking a situation like this thread's design goals, I'd guess that the way to handle it would be along the lines of what Jan suggested, separate regulator/psu.

Where the "ground" is physically/electrically is another issue. Jneutron might be the one to have something on that.

Hey, kids! Get your all NEW and IMPROVED do-it-yourself Op-Amp design kit today! It's easy, It's fun, It's ON SALE NOW!!


_-_-bear
 
Last edited:
Disabled Account
Joined 2012
Blend, is flat over frequency, Xtalk is NOT!

Does it matter? As any frequencies tend towards mono the effect at those freqs is to cancel and loose information. That it may happen more at certain freqs (highs usually) isnt the point.

If you want to focus on phono cartridges for a source.... they tend towards mono at the high freqs. But the same thing happens due to crosstalk for a number of reasons and mechanisms. Thx - -RNM
 
Last edited:
Disabled Account
Joined 2012
In addition to a losing details, when the channel seperation become less, the imaging shifts towards the middle/center.

So to have max accuracy in detail info reproduction and to keep accurate imaging, you need to have much better isolation between channels than was recorded.
-RNM
 
Last edited:
diyAudio Member RIP
Joined 2005
Another item of interest: based on level differences only, apparently the threshold for perceiving "all" of the signal coming from one loudspeaker of a conventional stereo pair, in the absence of time delay differences: 13dB, in the most sensitive spectral region (!). The notion that crosstalk reduction into the highest quoted extremes (like the dual mono pursuit) is audible is not supported by data.

However, if the "crosstalk" is really some bleedthrough of severe distortion associated but not identical to signals on the other channel, which can be observed in poorly designed amplifiers, especially low-cost chip amps, this can be quite audible at lower levels. Whether it's appropriate to call it crosstalk or not is another issue.

Also, it's often stated that standard same-signal crosstalk that gets worse (i.e. larger) at higher frequencies is due to capacitative coupling. Although this can be a mechanism, it doesn't have to be the only one.
 
I wonder, how can the red curves be so different from the blue ones?
Very large difference between left and right channels...


/S

Here is a comparison of measurements. One is the Kleos that has both coils wound on the same square former in 90° and the Atlas that has a cross with 4 arms. 2 coils for each channel.
The improvement in crosstalk is obvious. Still this would be very poor for an amp. Yes, i know, Vinyl has no right to sound as good as it does.
 
staccatiss, Stereoplay put the cartridges in a standard head shell when they measure.
Adjusting the azimuth the crosstalk could have being more symmetrical. How to get best azimuth would require a separate thread because there are several philosophies about what is best to do.
Stereoplay uses the biggest Pass Labs Phono as a reference but they also have other stages like Naim and Lyra. I could ask how they loaded the cartridges for the measurements they took.
When i do my phono stages double mono incl. PSU they sound better.
The peak can be damped in the phono stage, sure. Bob Cordell has shown a way to EQ a
Shure V15/3. OK, this is not an MC cartridge but the same method can be used for MC too. I asked Lyra why there is a peak and it has to do with damping the cantilever and with more damping to get away the peak they do not track as well so this is a compromise.
 
AX tech editor
Joined 2002
Paid Member
staccatiss, Stereoplay put the cartridges in a standard head shell when they measure.
Adjusting the azimuth the crosstalk could have being more symmetrical. How to get best azimuth would require a separate thread because there are several philosophies about what is best to do.
Stereoplay uses the biggest Pass Labs Phono as a reference but they also have other stages like Naim and Lyra. I could ask how they loaded the cartridges for the measurements they took.
When i do my phono stages double mono incl. PSU they sound better.
The peak can be damped in the phono stage, sure. Bob Cordell has shown a way to EQ a
Shure V15/3. OK, this is not an MC cartridge but the same method can be used for MC too. I asked Lyra why there is a peak and it has to do with damping the cantilever and with more damping to get away the peak they do not track as well so this is a compromise.

Joachim,

I don't think it has to do with tracking, unless you think that the coil dynamically damps the mechanical movements of the cantilever.

When you damp away the peak and ringing your hf output decreases and as Bob has shown, you can compensate for it in the pre.
What Bob (and others) have shown is how to damp the peak and ringing and place the zero precisely at the RIAA turnover. Then you delete the RIAA zero and presto! a perfectly flat response with no ringing and lower parts cost.
Neat, really.

jan
 
AX tech editor
Joined 2002
Paid Member
jan
With series connection, the circuit closest to the PSU has to be the one with the higher current demands. The furthest away has to be the least hungry for current.

Don't agree. The idea is to give the stage with the worst PSRR the best supply, so you place that close to the reg and/or use remote sensing. This *can* be the one that's most power hungry, but is not so necessarily.

Do me a favor. Have a look at my avatar please :p
George

OK, I just did. I still feel the same. Do I miss something? New glasses? :eek:

jan
 
AX tech editor
Joined 2002
Paid Member
I was not talking about the coil damping the cantilever. The cantilever is suspended on a wire that goes into the cartridge body. That gives the cantilever the opportunity to move
and gives a restoring force. That suspension is damped by a rubber ring.

OK yes I know I probably misunderstood "with more damping to get away the peak they do not track as well" to mean mechanical damping.

jan
 
AX tech editor
Joined 2002
Paid Member
You know, I'm not really a phono guy but reading all these phono pre articles in Linear Audio makes me start to think about it. ;)

What I wonder: assuming that the hf ringing comes from mechanical ringing of the cantilever system. If we then damp the coil output to get rid of the ringing/overshoot, that's electrical damping, correct?
Meaning, the cantilever system merrily rings along, except we now don't see it in the electrical output? Or is my assumption incorrect?

jan