Discrete Opamp Open Design

diyAudio Member RIP
Joined 2005
I found one more interesting discrete opamp. Looks like an instrumental amp loaded on output transformer. V/I converters drive controlled by current common cathode stages with deep feedback by voltage. However it is power amp, but can be used as line level output stage with smaller tubes. Thanks to transformer output can be balanced, for professional studio usage.

Actually getting back to the thread topic! Give that man a coupon.
 
The multitone tests most closely simulates a music signal source and produces both Im and harmonics. In checking the available published test data when using such, the 'noise' or grass between the tones grows by 20db when 15 tones are used. Rises another 20db when 30 tones are used (40db rise). I assume this increase in noise/unwanted freqs continues with ever more applied tones. So, we can assume that the level will contribute to the coloration of sound heard if the amplifier (opamp) does not have an extreamly low distortion to begin with. Thx - RNm

Reference? On the surface these numbers don't make physical sense. 2X the tones at the same total energy yields 10X the "noise" power. There are a lot of details missing here, crest factors, levels, 15 tones is 20db over what?
 
Disabled Account
Joined 2012
Yes. And iirc you can drive a ribbon microphone to low but audible levels, though you wouldn't want to. Reminds me of the time I drove a tuning fork crystal a little too hard and heard it break.

Speaking of bad news for cartridges, one could put d.c. current through one you weren't too fond of :eek: and listen to the changes to the frequency response.

Or put the same make and model cartridge in the feedback loop.
 
Or put the same make and model cartridge in the feedback loop.

...on the same disk, in the same point. :)


You can't compensate resonances by filters. The only way to minimize resonances is to suck an energy from them. For example, loading the cartridge on zero dynamic resistance (or on negative one) is the best that you can do, but such a way you can damp only resonances that are immediately and linearly translated into electrical output.
 
Disabled Account
Joined 2012
...on the same disk, in the same point. :)


You can't compensate resonances by filters. The only way to minimize resonances is to suck an energy from them. For example, loading the cartridge on zero dynamic resistance (or on negative one) is the best that you can do, but such a way you can damp only resonances that are immediately and linearly translated into electrical output.

You might also take out nonlinearitties caused by the cart (magnetics). That is what I was thinking. I agree re. resonances. I published a circuit that had very low Zin specifically for MC cart (TAA 1/82). Been there, done that. :)
 
Last edited:
Its difficult to get a copy as it was filed in australia, patent AU1986058306, anyway it has expired many years ago.

It looks like US4644517 is it and it has nothing to do with mechanical damping or resonances. He proposes a transformer coupled virtual ground load to extend low frequency distortion of the magnetics to a lower frequency. Is this a problem in reality?

John would appriciate that he has far more patents on advanced fuel injection for cars.
 
Last edited:
Member
Joined 2002
Paid Member
No, that's not minimum phase and it's also not single-valued.


We are making progress. :up:
The two necessary conditions for a mechanical or acoustical resonance to be electrically equalizable is that the physical system that resonates should be minimum phase and expressed mathematically by a single-valued function . Is this correct?

For minimum phase, you will ask for linearity and time invariance. Say you have it.
Then causality. You have it too.
Next comes stability. Thank you, that’s a valuable input. But oops! Problem. Is there a maximum Q allowed?

Single-valued function. :bomb:
My understanding is that a function is a single-valued one when and only when it’s graph contains no closed loops. Please correct me if I am wrong.


George
 
It seems we have gone way off topic here but let me throw it off a little bit more here. When speaking of a loudspeaker and resonance what do you do about sympathetic resonances that are not caused by any electrical impulse stimulation? Harmonic stimulation's caused by acoustical excitations? How so do you counteract that, and wouldn't this same phenomena happen with the cantilever, headshell, tonearm and gymbal mechanism?
 
diyAudio Member RIP
Joined 2005
The example of a Q =100 resonance is a bit far-fetched, although perhaps useful to capture the concept. Even with a Q = 10, which is rarely encountered, feeding the output of such a resonator to a matched notch filter and switching on a sine wave at resonance produces a burst that peaks after abut 3ms and is 20dB down from the peak after 13ms.

Another observation: the presence of frequency response variations doesn't have to be due to ringing of a resonator as we tend to think of them. For loudspeaker baffles there is a transition in the radiation pattern with frequency that manifests as a gentle boost with quite low "Q". It is entirely appropriate to compensate for this with a simple bridged-T dip in the electronics.
 
Last edited:
Member
Joined 2002
Paid Member
No, that's not minimum phase and it's also not single-valued.


We are making progress. :headbash:

The two necessary conditions for a mechanical or acoustical resonance to be electrically equalizable is that the physical system that resonates should be minimum phase and expressed mathematically by a single-valued function . Is this correct?

For minimum phase, you will ask for linearity and time invariance. Say you have it.
Then causality. You have it too.
Next comes stability. Thank you, that scoops many cases out. But oops! Problem. Is there a maximum Q allowed for the stable ones?

Single-valued function. :bomb:
My understanding is that a function is a single-valued one when and only when it’s graph contains no closed loops. Please correct me if I am wrong.


George
 
The example of a Q =100 resonance is a bit far-fetched, although perhaps useful to capture the concept. Even with a Q = 10, which is rarely encountered, feeding the output of such a resonator to a matched notch filter and switching on a sine wave at resonance produces a burst that peaks after abut 3ms and is 20dB down from the peak after 13ms.

Another observation: the presence of frequency response variations doesn't have to be due to ringing of a resonator as we tend to think of them. For loudspeaker baffles there is a transition in the radiation pattern with frequency that manifests as a gentle boost with quite low "Q". It is entirely appropriate to compensate for this with a simple bridged-T dip in the electronics.

You are correct, the Q of 100 was used only for concept capture.

My point was that for a chain of sequential elements from input to output, when the transfer of energy from one element to the next relies upon a resonance to build up the amplitude to a sufficient value that it achieves I/O equilibrium, that kills transient response.

I wasn't actually thinking about baffle or cone decoupling issues (but agree), but more of a reflex or 4th order bandpass kinda thing.
j
 
Disabled Account
Joined 2012
I thought we were talking MM, we have discussed virtual ground input MC before. I think John said it makes things sound like "under water".
Our discussion about spherical stylii and the use of pre-distortion on many LP's was a real eye-opener for me.

OK I thought we were talking abpout MC.... typical load was 100 Ohms. I can see JC comment working for the higher Z MM types which used 47K-ish. One thing all this brings up for ME is why I junked the whole thing relating to phonographs and LP's. Just is a big headache that never really gets cured. Esp the distortion levels and on and on. Lets just do the Line level opamp before I get sick and expel my lunch. :)
 
Last edited:
Member
Joined 2002
Paid Member
and wouldn't this same phenomena happen with the cantilever, headshell, tonearm and gymbal mechanism?

Oh man! I had kept these same questions for my 120th post to SY.

Another observation: the presence of frequency response variations doesn't have to be due to ringing of a resonator as we tend to think of them

Brad
You are close to my asking, especially with your last sentence.
When I frequency sweep for to check the impedance of a circuit or a loudspeaker in a Bode plot, it is difficult to see if some small aberrations are indications of small resonances. I then plot the impedance in a Nyquist diagram (real part of the impedance on the X axis. The imaginary part on the Y axis. There, if there is a resonance, the plot makes a closed circle. The area of the circle is indicative of the energy stored there.

George
 
I found one more interesting discrete opamp. Looks like an instrumental amp loaded on output transformer. V/I converters drive controlled by current common cathode stages with deep feedback by voltage. However it is power amp, but can be used as line level output stage with smaller tubes. Thanks to transformer output can be balanced, for professional studio usage.

303328d1348674903t-discrete-opamp-open-design-p-p-amp.png


This was published in Electronics World, author Jeff Macaulay, if I remember well.
 
Member
Joined 2002
Paid Member
Is 20 khz triangular realistic? I would have expected sine for this derivation. Are the peak G's representing slew max or where the stylus hits the vinyl wall?

jn

I was expecting that objection. ;)
Jneutron, I had to do it in my primitive way for to be able to understand it. I used only elementary geometry
http://www.diyaudio.com/forums/analogue-source/209205-grooves.html#post2952803
Acceleration is on part two that is to be published sooon together with an Excel spreadsheet trimmed for SnMs :D

Of course, there are readily available equations for calculating velocity and acceleration for sines, but these are for normal people. (this is that :eek: Jan has not understood about me and asked for new glasses)

Comparing my triangular peak values with the sinus values gives the nice number 1.41… :)

George

>Edit: Here is a nice method to actually measure acceleration : http://www.luckydog.demon.co.uk/images/Baxandall.pdf
 

Attachments

  • sin.jpg
    sin.jpg
    127.1 KB · Views: 395
Last edited:
Ah, so you are using the zero crossing straight line portion for the calculation?

The first derivative (velocity) of the triangular wave I was thinking of is a square wave, the second (acceleration) is infinite at the transitions.

I thought you might be calculating based on the stylus mass,contact area, vinyl modulus, and vinyl density at the abrupt transition in direction.

j
 
For a tip with an effective mass of 0.27mg (Denon DL-103)following a theoretical groove with an 20KHz triangular 5cm/sec modulation, the resulting peak acceleration is 10,000m/sec^2 (1020G).
The inertial force due to this acceleration btn tip side and wall land would be 0.0027N.

Increasing by 20db the modulation velocity to 50cm/sec, these numbers are multiplied by 10.

In a more common case (1KHz at 5cm/sec) acceleration drops down to 500m/sec^2 (51G) and force is at 0.0000135N

I am sure that calculations based on tip dynamics will give wild peak temperatures but wouldn’t published images prove any thermal induced evidence at the groove lands, if that high temperatures were really occurring?

Wow. Thanks, George. That's enough to get into the right order of magnitude, for sure. From here, I'd need to know whether the vinyl acts like a liquid or not, rate of heat transfer, temperature rise due to pressure, effective contact area, and (I'm sure) a dozen other things. It's just too far over my head, and I'm not sure I'd believe anything on these tiny size scales without a measurement to keep me honest. Here's a few references if anybody's interested:

Turntable Forum • Frictional heating of vinyl revisited

but their arguments are not my own. They're interested in tip temperature and friction and I'm interested in vinyl temperature and pressure.

Anyway, I've failed, and need to retract my claim of 400 F instantaneous temperature, pending a measurement.

Thanks,
Chris

ps: Thanks for the great link. I have some photos made for Stanton Corp back in the 70's, but they're matted and framed, so difficult to scan to share. I really need to do it anyway.