Fun with vinyl measurements

[Balle Clorin]

setting anstiskate by intermodulation distortion testing. More antiskating tetsing using AnalogMagik V1 45 rpm record- 45/45 stereo modulated VTA 60:7000 Hz IM test..'' First track on record

Setting Antiskate AS= VTF value maintains consistent IMD values

But changing anti-skate for will possibly skew the stylus/cantilever... here is the effect of 1.75g and antiskate of 0-1-2-3 Left channel,only, right behaves the same...

Then lowering the VTF give even higher % AS side force... more push on cantilever angles

So high side force 20% of VTF( VTF 2.5 AS 3) gives lowest %IMD, but 20% side force is unreasonable high, so i think my stylus should be tilted a bit to match the tilt it has at high AS force.. This tests could be done on Tacet and AP Ultimate test record too, but I am too lazy now

Any comments to which I should adjust first, Azimuth or horizonal angle/zenith?

Script to make plots are in the attachment, rename txt to py

EDIT. AP ULTIMATE TEST RECORD , same improvement by higher Aantiskate. BUT NO RESPONSE TO VTA CHANGE( PLATETR MATS ADDED)

 

[narud]

so i recorded the shure trackability tests from the era v record which is a combined tone of 200hz,2.1k and 17k at 6 different levels using the carts i have currently mounted on tables and asked chatgpt to analyze them for distortion. this is what it gave me. im not sure if im prompting it correctly or if this track is useful for this analysis. am i on the right track? should i revise the prompt? the carts marked new are really just under 50 hours or so and still show a rounded ridge from the front view with out obvious wear flats. the 745xml is really a 45xml mounted on the body from a 510xcb so its the same generator. the 350 hour shure is an estimate. i bought it used with light wear. it looked broken in under the scope but still looked fairly new. id guess a couple hundred hours and ive put 150 on it since putting it in use.

 

[Curupira]

so i recorded the shure trackability tests from the era v record which is a combined tone of 200hz,2.1k and 17k at 6 different levels using the carts i have currently mounted on tables and asked chatgpt to analyze them for distortion. this is what it gave me. im not sure if im prompting it correctly or if this track is useful for this analysis. am i on the right track? should i revise the prompt? the carts marked new are really just under 50 hours or so and still show a rounded ridge from the front view with out obvious wear flats. the 745xml is really a 45xml mounted on the body from a 510xcb so its the same generator. the 350 hour shure is an estimate. i bought it used with light wear. it looked broken in under the scope but still looked fairly new. id guess a couple hundred hours and ive put 150 on it since putting it in use.
View attachment 502235View attachment 502236

Nagaoka has seen better days, at this point I'd question if it's damaging the grooves.

 

[Jimi Floyd]

so i recorded the shure trackability tests from the era v record which is a combined tone of 200hz,2.1k and 17k at 6 different levels using the carts i have currently mounted on tables and asked chatgpt to analyze them for distortion. this is what it gave me. im not sure if im prompting it correctly or if this track is useful for this analysis. am i on the right track? should i revise the prompt? the carts marked new are really just under 50 hours or so and still show a rounded ridge from the front view with out obvious wear flats. the 745xml is really a 45xml mounted on the body from a 510xcb so its the same generator. the 350 hour shure is an estimate. i bought it used with light wear. it looked broken in under the scope but still looked fairly new. id guess a couple hundred hours and ive put 150 on it since putting it in use.
View attachment 502235View attachment 502236

It definitely seems to make sense. I am curious, did you give ChatGPT just the raw files + your question?

 

[narud]

It definitely seems to make sense. I am curious, did you give ChatGPT just the raw files + your question?

yes. i just upload the wav files and ask it to analyze.

 

[Balle Clorin]

Vertical tracking Angle and Imd, page 21

https://www.worldradiohistory.com/Archive-All-Audio/Archive-Audio/60s/Audio-1962-Nov.pdf

 

[Balle Clorin]

so i recorded the shure trackability tests from the era v record which is a combined tone of 200hz,2.1k and 17k at 6 different levels using the carts i have currently mounted on tables and asked chatgpt to analyze them for distortion. this is what it gave me. im not sure if im prompting it correctly or if this track is useful for this analysis. am i on the right track? should i revise the prompt? the carts marked new are really just under 50 hours or so and still show a rounded ridge from the front view with out obvious wear flats. the 745xml is really a 45xml mounted on the body from a 510xcb so its the same generator. the 350 hour shure is an estimate. i bought it used with light wear. it looked broken in under the scope but still looked fairly new. id guess a couple hundred hours and ive put 150 on it since putting it in use.
View attachment 502235View attachment 502236

Very interesting, did not know that chatgpt could analyse wav files directly ! ,
Can you try my Python script and see What that gives.? Or share the files with me?
Nagagoka seem to be with very high VTA , can that be confimed optically?
What you have plotted is What some Audio articles I Linked to recently advise to do.IMD is proportional to level and VTA error…..

 

[narud]

Very interesting, did not know that chatgpt could analyse wav files directly ! ,
Can you try my Python script and see What that gives.? Or share the files with me?
Nagagoka seem to be with very high VTA , can that be confimed optically?
What you have plotted is What some Audio articles I Linked to recently advise to do.IMD is proportional to level and VTA error…..

i barely know what im doing with any of this. i was surprised i ever got jps script to work in the first place! im game to try it though. wheres your script? ill put the files in a dropbox though in case i cant get it going.

the vta is actually tail down if using the top of the headshell as a reference. i spent an embarrassingly long time taking fr measurements trying to find the right vta that the previous owner was using. if not lined up correctly, id get a big high end drop in the right channel. i could even it out by shimming one side to change the azimuth, but it was a pretty extreme tilt. i can see where the idea that shibatas are finicky to set up came from because the at155lc i had that was used also needed the perfect arm height to not get a drop in one channel. none of the used microridge carts ive picked up have had this issue.

i started going down this rabbit hole because of some high end smearing im noticing on the nag 500 with two copies of the stylus and an old mp-30. i found some old reviews of the mp30 and mp50 from hi-fi choice where they complain of high distortion.

 

[Balle Clorin]

Post 2421 has script attached,

 

[JP]

Personally I find ChatGPT to be a pain to interact with so I use Claude. It's more natural language, and if you swear at it, it'll swear back.

This entire analysis and visuals were done by Claude. I uploaded the SJPlot CSV files, framed what they were and told it what I wanted to know. From there it was a "conversation" reminiscent going through data sets with my old analytics team from my corporate days. As it uses Python natively for data analysis, at the end I had it give me the scripts it was using with some modifications to make them easier to use.

The SJPlot web front-end was written by AI, as was this DAC/ADC generation test orchestration utility. On the latter I haven't looked at a single line of the code. The array transposition bug that was causing the fundamental and crosstalk to flip in SJPlot was diagnosed by AI. I gave it the script and the audio file and told it what was happening and to figure out why.

@narud once you have a working Python environment with the usual libraries for this kind of work you should be good to go with other scripts. Regardless, if you give AI the script and the audio/data files it should run them for you.

Not to say that AI isn't extremely frustrating at times - you have to pay attention what it's doing. For me, worth it.

 

[Balle Clorin]

Not sure if I'm on the right track since I've not seen much discussed in the past regarding VTA/VMA and crosstalk effects. I've can only recall the Yosh pages regarding this matter.

 

[narud]

i put the script into chatgpt and had it run the files. heres two outputs for a vn5mr. low+mid pair and mid+high pair

 

[Thomas_A]

View attachment 502406
View attachment 502407

I saw some notes about azimuth and crosstalk and that test records vary with left and right symmetry. There was a formulate there with inclination angle

Fig 7 shows a drawing of the stereo groove. is the vertical tracking angle and is the cutting angle. is 45 deg., the half angle of the angular distance between the two minimum points for the two curves for crosstalk as a function of the angle variation a.. If we assume that the cutting direction is 45/45 in the cutting plane, there is the following relation between the three angles:

tan()=cos()/cos (ß)

If we know and measure we can thus find .

But I think Yosh excel macro shows the essentials (section J for crosstalk):

https://www7a.biglobe.ne.jp/~yosh/distortions.xls

 

[Thomas_A]

@Balle Clorin

have you ever tested the phase of the AP record 1 kHz signal, the one on first track first side, using your script? It is only around 120 s but it could be of help to check alignment/zenith. Does it look similar in phase difference as your 3150 Hz test the first 120 s?

 

[Balle Clorin]

@Balle Clorin

have you ever tested the phase of the AP record 1 kHz signal, the one on first track first side, using your script? It is only around 120 s but it could be of help to check alignment/zenith. Does it look similar in phase difference as your 3150 Hz test the first 120 s?

Analog Productions Ultimate analog test record

Clearaudio trackability record, 3150 track

The Sperling TLP-1 that Korf used for distortion across the record is on the way..

 

[Thomas_A]

Thanks! Do you have the link to your script? I tried to look at the expected phase diff vs tracking error but got different answers. It is both frequency and level-dependent(?)

 

[Balle Clorin]

here is the script.
it is sensitive to frequency and wow and flutter.. The phase difference will vary with frequency since a given time difference between channels gives a larger L-R phase difference for higher frequencies.

Script for intermodulation distortion testing of vinyl playback, setting VTA ,Zenith etc by test record

Thanks. I have the STR-112, it has duotones with 4kHz and 400Hz lateral and vertical, and 200 Hz lateral and vertical, all with different levels of the intermodulation. I think the results from such a test can only be compared to tests with the same record. Anyway, the vertical intermodulations...

www.audiosciencereview.com

Fun with vinyl measurements

Horizontal Trackng Angle, or Zenith , the thing that protractors help with Thank you... I'm not used to this abbreviation... (What are the current zero points—the compromise points of your tonearm—at what distances?)

www.audiosciencereview.com

Post in thread 'Script for intermodulation distortion testing of vinyl playback, setting VTA ,Zenith etc by test record'
https://www.audiosciencereview.com/...-zenith-etc-by-test-record.66496/post-2446294

play with these factors

 

[Thomas_A]

Thanks. I think I might have something to work on regarding zenith. I looked at what you calculated before regarding tracking error vs. phase error, but I am not sure what the value will be. AI have given me conflicting results where inter channel phase difference is dependent on frequency and level, some answers give me around 4° in the beginning and end of record (approximately) for a Loefgren setup with 230 mm tonearm. Other tries have given lower values, so I do not know what gives here. My analysis gave just below 5.5° for the stretch which is ≈146-135 mm. I think the level of the AP record is 7 cm/s.

AI

Bottom line​

For a 230 mm tonearm with Löfgren alignment playing a 1 kHz, 7 cm/s signal:

• Inter-channel phase difference ≈ 2 × tracking error (in radians)
• Zero at null points
• Typically ≤ 4–5° maximum across the record

Groove Radius (mm) Inter-channel Phase Diff (deg)
-----------------------------------------------
55.0 ~+4.0
66.0 ~0.0
80.0 ~-3.0
95.0 ~-3.0
110.0 ~-1.5
120.9 ~0.0
130.0 ~+1.5
145.0 ~+4.0

 

[Thomas_A]

A very small, slight turn of cartridge in headshell. Decrease to ≈4.5°. But it does not change very much from beginning to end of track.

 

[Balle Clorin]

Inter-channel phase difference ≈ 2 × tracking error (in radians)

Did AI come up with this? It cannot be that simple.. ? The phase diff will vary with frequency. , tracking error does not