Calling Adjust+ experts...

bazelio

Well-Known Member
Sep 26, 2016
2,493
1,745
345
California
PS - I sent Chris Feickert a few questions about what the software Adjust+ is actually doing. The 180° phase window and it's relationship to azimuth is really not intuitive the more I've thought about it. I will post anything I learn if he responds.
 
  • Like
Reactions: MPS

ianm0

Well-Known Member
Feb 22, 2018
36
16
108
PS - I sent Chris Feickert a few questions about what the software Adjust+ is actually doing. The 180° phase window and it's relationship to azimuth is really not intuitive the more I've thought about it. I will post anything I learn if he responds.
I think I might have an explanation for tima’s azimuth phase measurements in his OP:


I’ll skip the mathematical details, but briefly discuss the physics involved.

First a perfectly manufactured MC should have the two coils exactly at 90 deg to one another, then the crosstalk should be very close to, but never equal to, zero. Let me explain how crosstalk arises. Let’s say signal is present only in coil A but not in the other coil B. When A moves through a magnetic field, an emf is induced – that’s the measured signal strength. This emf drives a current around Coil A, but the current lags the emf by a phase angle ? given by tan ? = ?L/R. L and R are respectively the self inductance and DC resistance R of A. This will induce a magnetic field and consequently an emf in Coil B, both of which are in phase with the aforementioned induced current in A. The latter emf is the crosstalk signal. If the angle between A and B is smaller than 90 deg, this phase difference is simply ?<90 deg . If the angle between the coils is larger than 90 deg,the phase difference is (180 +?). (Note: I skip the physics explanation.)

Let me illustrate with some numbers. If ? = 30 deg, the phase angle could be 30 o or 210 deg, depending on the angle between the coils. But 210 deg is equivalent to -150 deg.

Adjust+ only shows angles between 0 deg and 180 deg. I suspect in this numerical example, 210 (-150) deg is recorded as 150 deg. I stand to be corrected on this conjecture.

Of course, one remaining question must be answered. Why should the two crosstalks be so different (like 30 deg and 150 deg). Shouldn't they always be very close since they only depend on L and R of the coils? Again that’s theoretically true if the coils are rigidly fixed relative to the stylus, that is, no relative motion at play. So what could possibly happen? Again my guess with a specific example:


Let’s say the R channel coil is a bit loose and L coil rigidly attached to the stylus. Let me for augment sake assume that the angle between the coils is less than 90 deg initially. When only L channel signal is present, R channel coil feels no mechanical stimulation. So the phase ? is given by tan ? = ?L/R, and ? < 90 deg. However, when R signal is played, the R coil experiences a larger mechanical stimulus and the angle between the coils could possibly become larger than 90 deg. The phase is then closer to 180 deg than 0 deg.

Hope this explains the observations in the OP of tima. I hope I have made it clear. I tried.

Another point on tima's measurements. The two crosstalks seemed to be quite different. Dr Feikert said that incorrect VTA setting could cause disparity between crosstalks even the phase relations remain pretty much the same with different arm heights.
 
Last edited:

bazelio

Well-Known Member
Sep 26, 2016
2,493
1,745
345
California
I was chatting offline about this, and sent some questions to Chris Feickert in order to confirm what Adjust+ is actually reporting. Take a look at the following picture to understand the signal phases coming off cartridge coils (picture source vinylengine.com)

stereo cartridge.PNG

The right and left coils are reversed in polarity. Why? The lathe cutter heads are reversed and L/R channels 180 degrees out of phase and then the cartridge "re-inverts" the right channel. So, referencing this, then we know the following:

1. Upward vertical movement of the stylus creates a positive signal in the L channel, negative signal in the R channel.
2. Downward vertical movement creates a negative L signal, positive R signal.
3. Horizontal rightward movement creates positive R signal, positive L signal.
4. Horizontal leftward movement creates negative R, negative L signals.

Another way to visualize cart movement and resulting L/R signals (picture source https://www.analogplanet.com/content/crazy-little-thing-called-azimuth-part-1)

Screen Shot 2020-05-10 at 9.40.23 AM.png

Now in the case of the test record, we have single-channel test tracks and either of the top two images apply. So if the vertical movement results in 180 degree out of phase signals and the horizontal movement creates in phase signals, then splitting the difference (i.e. 45 degree motion of the stylus) should result in 90 degree signal phase shift.

I guess, then, the "ideal" crosstalk measurement would show the signals 90 degrees out of phase, for both L->R and R->L. Further, when the phase plots cross somewhere other than at 90, and if the stylus is perpendicular to the groove at that point, then I interpret that to mean the stylus and the coils themselves are misaligned. However, other factors may still be contributing, e.g anti-skate and VTA. I'm asking Dr. Feickert to confirm all of the above.

Lastly, it is interesting to note that Richard Mak (AnalogMagik creator) does *not* seem to agree that phase alignment should trump lowest crosstalk:

You can adjust azimuth until the phase shift number is as close together as possible. However, this is not a very common occurrence and Phase shifts bear a weaker correlation and it is sometimes not repeatable unless if we incorporate a lot of averaging to stabilize the number. In our opinion, Crosstalk/%THD, measurement is usually sufficient.
From: https://www.analogmagik.com/azimuth (A good read). He is taking a more holistic view, and emphasizing that we shouldn't focus solely on a single parameter, but rather spend the time to remeasure and minimize all distortion artifacts. I'm tempted to say don't settle for crosstalk minimization, phase alignment at some odd point. But try to adjust all parameters in order to see the effect they have on bringing crosstalk phase closer to the ideal. Then, how to prioritize phase alignment vs lowest crosstalk ... Still debating that.
 
Last edited:
  • Like
Reactions: MPS and tima

ianm0

Well-Known Member
Feb 22, 2018
36
16
108
I am in complete agreement with your last paragraph. Just let your ears be YOUR final arbiter. I have had the experience that often I still have to fine tune the azimuth after I think I have the best setting from measurements. After fine tuning, I would get smoother vocals, be that male or female. Same for strings. To be truthful, I set my azimuth by ears. My theoretical pursuit is merely to satisfy my academic interest to understand the physics behind it.
 

ianm0

Well-Known Member
Feb 22, 2018
36
16
108
I made crosstalks measurements on two cartridges using Visual Analyzer software, which can double up as an oscilloscope and a spectrum analyzer. The latter can display frequency spectrum, as well as phase spectrum.

First, the cartridge with quite normal phase relationship shown in the bottom plots.

1589257444331.png
Now another cartridge:
1589257497033.png
Signal/crosstalk almost in phase in one channel, but almost out of phase in the other – similar to tima’s plot which was reproduced in post #62. Also note jumpy phase/frequency relationship in bottom right plot when R channel signal is played. This is what led me to conjecture the right channel signal coil is not rigidly attached to the stylus. Please refer to my explanation in post #62.

I am not claiming my conclusion is the final say. Just a thought. Ironically the second cartridge is better sounding.
 
Last edited:
  • Like
Reactions: Calle_jr

bazelio

Well-Known Member
Sep 26, 2016
2,493
1,745
345
California
Did you intend to post pictures? I don't see any.

I'd bet the coil is fine. Richard Mak alludes to an issue in the quote I posted above: "Phase shifts bear a weaker correlation and it is sometimes not repeatable unless if we incorporate a lot of averaging to stabilize the number.". So it becomes a software DSP problem. I've seen phase jump around on my cheap digital oscilloscope too.
 
Last edited:
  • Like
Reactions: Calle_jr

bazelio

Well-Known Member
Sep 26, 2016
2,493
1,745
345
California
current lags the emf by a phase angle ? given by tan ? = ?L/R. L and R are respectively the self inductance and DC resistance R of A.

Phase angles in MC carts look small. One example, using specs from the new Lyra Atlas Lambda MC:

4.2ohm self-impedance, 11uH inductance
Code:
Frequency (Hz)    Phase Angle (degrees)
--------------    ---------------------
20                0.019
1000              0.92
10000             9.3

I'm personally not ready to say we should hang our hats on loose coil theories, though.
 

ianm0

Well-Known Member
Feb 22, 2018
36
16
108
Did you intend to post pictures? I don't see any.

I'd bet the coil is fine. Richard Mak alludes to an issue in the quote I posted above: "Phase shifts bear a weaker correlation and it is sometimes not repeatable unless if we incorporate a lot of averaging to stabilize the number.". So it becomes a software DSP problem. I've seen phase jump around on my cheap digital oscilloscope too.

Perhaps we should clarify whether we are talking about the same phase angles. The numerical values of phase angles in my VA plots are of no significance. It is the difference between those of the two channels that matters, and it is given by tan ? = ?L/R. Your are absolutely correct that this difference should be quite small for nearly all MC cartridges. Hence the phase spectra for both signal and crosstalk should be almost on top of one another. That should also apply for both channels. The small difference can be attributed to slight mismatches of L and R of their coils. All that is theory, of course.

I don't have a good digital oscilloscope to check out. John Elison did this:
https://www.audioasylum.com/messages/vinyl/919668/an-oscilloscope-is-simply-a-measurement-tool
His plots follow:
1589268388660.png
The crosstalks are almost out of phase with signals for both channels. You are correct the phase difference between signal and crosstalk should be quite small if L~ uH. This is borne out by Elison's plots. Also the out-of-phase consistency between the two channels in Elison's plots means the coils are at an angle > 90 deg with each other.

I wish the plots for my first cartridge would have been like Elison's, ie, both channels are either in phase or out of phase. Their not to be made me suspect non rigid right channel coil. As I said before, just a guess.

In his talk with Fremer, Dr Feickert showed phase difference between signal and crosstalk in the range 60-80 deg - a pretty large angle. If you find out from him what Adjust+ measures, please share.
 
Last edited:
  • Like
Reactions: Calle_jr

ianm0

Well-Known Member
Feb 22, 2018
36
16
108
The way that phases in crosstalks are measured motivated me to dwell further into the issue. I'll use my experience using Visual Analyzer (VA) as an illustration of what I found.

VA provides two ways to do it: from captured scope or captured spectrum. For the former, read zero crossings of rising (or falling) edge for both channels to obtain the phase difference between signal and crosstalk. Numerical value of phase measured with VA available from the captured spectrum. Difference from both methods could amount to a few degrees. In the diagram below, the plots on top are from the captured scope and the bottom plots from captured spectrum using 1000 Hz test signal:

1590287054463.png

From the captured scope plots, the phase difference between signal/crosstalk is close to zero for both channels. From the captured spectrum, the same difference amounts to a few degrees, which are displayed numerically by VA (but regrettably not shown -my oversight when snapping these shots before I had those numbers displayed)

Conclusion: Use the captured scope to determine phases since how VA designed their software for measuring phases is not known. Besides latter results are not quite reproducible.

Question: How does Adjust+ measure phase? I suspect quite similar to the aforementioned captured spectrum method of VA. Unless we know how the software is designed, we have no way to pass judgement on its accuracy. One point continues to irk me is that Adjust+ only provides a range from 0-180 degrees.

Footnote: In my earlier post, I mentioned: tan ? = ?L/R. L and R are respectively the self inductance and DC resistance R of coil A . Actually R should include the load seen by the cartridge. That will make the phase smaller still. The two top plots in this post confirm that the phase difference between signal/crosstalk is indeed close to zero. The cartridge is a well built one: it's a Fidelity Research FR-1 MkIIIF I have had for overly thirty years and recently rebuilt by Expert Stylus.
 

J.R. Boisclair

Well-Known Member
Jun 30, 2020
182
321
135
Hi guys,

i noticed in this thread some of you are tilting your azimuth setting quite a bit past 2 degrees. HERE is a calculator we built to determine how much you need to adjust your SRA by to compensate for the impact on it caused by the extreme azimuth tilt. Applies to “Tyoe I” arms only. Those are tonearms where the azimuth mechanism is at the pivot housing or on the arm wand.

Tonearms with azimuth adjusted on a plane perpendicular to the horizontal alignment of the cantilever do NOT affect SRA when azimuth is changed. (Think most removable headshells - Kuzma and SAT arms excepting)
 
  • Like
Reactions: uber2 and tima

ianm0

Well-Known Member
Feb 22, 2018
36
16
108
I think it was AnalogMagic who made the remark that phase in crosstalks measurements is meaningless, but without elaboration. So I decided to do some further investigation by measuring crosstalks traces at different times. The outcome is summarized in the attached photo. What are your thoughts on the phase measurement in Adjust+?
 

Attachments

  • Shelter 901 Crosstalks Phase Investigation.jpg
    Shelter 901 Crosstalks Phase Investigation.jpg
    1.2 MB · Views: 23

MPS

Well-Known Member
Jun 20, 2016
112
84
160
Finland
I think it was AnalogMagic who made the remark that phase in crosstalks measurements is meaningless, but without elaboration. So I decided to do some further investigation by measuring crosstalks traces at different times. The outcome is summarized in the attached photo. What are your thoughts on the phase measurement in Adjust+?
I think measurement requires averaging over time to minimize random variance.
 
  • Like
Reactions: bazelio

ianm0

Well-Known Member
Feb 22, 2018
36
16
108
That's what I thought too. An important statistical question:. Are enough samples taken to measure the average of the random variable? I doubt it. If not, the measurement is meaningless.
 

tony22

Well-Known Member
Nov 4, 2019
557
239
133
62
I know I’m late to the thread, but do we deduce from these last few posts that the Adjust+ method for azimuth adjustment can lead to an erroneous adjustment?
 

Calle_jr

Well-Known Member
May 24, 2015
160
14
250
Sweden
www.euphonia-audioforum.se
I know I’m late to the thread, but do we deduce from these last few posts that the Adjust+ method for azimuth adjustment can lead to an erroneous adjustment?
I don’t.
It’s unusually unclear what the software is actually doing. Even to talk about phase for azimuth adjustment is a bit weird. But when you realise that;
- Green curve shows magnitude and phase in the right channel relative to 1kHz sine in the left channel and
- Red curve shows magnitude and phase in the left channel relative to 1kHz sine in the right channel,
it becomes rather obvious.

Both magnitude and phase would be affected by build tolerances in stylus, cantilever, suspension, dampers, coils and magnets.

Phase should be close to 90o in this test. Is it realistic that it varies by some +/-70 to 80o due to small changes in azimuth for a 1kHz sine?
Even if it is; will phase deviation in the opposite channel due to crosstalk even matter?
Does a few dBs worse crosstalk even matter?

In my opinion phase deviation have greater impact than magnitude for crosstalk, because phase deviation mess up the tone. Magnitude in crosstalk only causes a small mono effect.
 
Last edited:

NekoAudio

Member
Oct 28, 2023
44
34
20
San Jose, California
www.nekoaudio.com
In my opinion phase deviation have greater impact than magnitude for crosstalk, because phase deviation mess up the tone. Magnitude in crosstalk only causes a small mono effect.
That has been my experience. I use Adjust+ for cartridge calibration. When the phase difference of the two channels is large, the sound becomes really messed up. Vocals become harder to understand, and sounds get smeared. A couple dB worse of crosstalk is barely noticeable, in relative terms.
 

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing