Cartridge Alignment: Setting Azimuth with a PC and Soundcard

Ugh, sorry, totally misread the question - thought we were back on azimuth and that the answer would be the same as for zenith - which is: tracking distortion turns out to be inversely proportional to linear speed, so is more in the inner grooves than the outer,as stated in the 2nd paragraph of P2 of Serious's link and as is evident on the graphs here:

https://db.audioasylum.com/mhtml/m.html?forum=vinyl&n=595959

But searching just now, even with the aid of beer I couldn't find where Elison (and everybody else) got that relationship from. Maybe Loefgren's original paper.

Edit: There's a summary on pp. 1-5 to 1-7 of Dennes's review of all the previous work:

http://brujic.gradjevinans.net/HiFi/Graeme Dennes -Tonearm Geometry Analysis.pdf

He describes how Loefgren showed that tracking distortion is not only proportional to tracking error but also inversely proportional to radius - which came from an earlier proposal that distortion should be related to the ratio of amplitude of groove modulation / wavelength of groove modulation (wavelength at a given frequency being proportional to groove velocity, which is faster at the outside of the record). Hence the use by Loefgren and everyone subsequently of weighted tracking error (angular tracking error / radius) as a proxy for tracking distortion.

And then he goes through all the equations

 

There are also a few 'live' oscilloscope programs around that'll show the waveforms as you play the test track or recording of it in real time (plus the latency of your system). A simple free Windows one is SoundScope, available here (Microsoft alert):

https://www.microsoft.com/en-us/p/soundscope-free/9n19c75120t2?activetab=pivot:overviewtab



This allows interactive scaling of the X-axis and doesn't require any voltage or SPL calibration for the Y-axis (but neither does it allow that, meaning you only get relative amplitude measures on the XY crosshairs readouts). Hitting the pause button (bottom right) will freeze it at any desired point.

While this is convenient for seeing how things are changing through the whole test track, I'd still use something like REW or ARTA for comparison of different trials as well as to measure crosstalk (getting back to the acutal topic of this thread).

 

If the azimuth is adjusted by rotating an arm wand of diameter 13 mm, a rotation over 0.1 mm changes the azimuth by 0.88 degrees. It's so easy to overstep and miss the optimum azimuth. Bigger changes if wand is thinner.
This chase could turn into a hit and miss exercise unless the tonearm comes with a reproducible fine adjustment device - a micrometer of sort comes to mind. Audiomods arms have this for VTA adjustments and I love it.

 

Interesting take on this process using the Parks Audio Puffin phono-preamp:



This feature is in beta and not available to the general public yet, though they do have VU meters in the current firmware which can accomplish pretty much the same thing. At around $400 used this is a pricey option but if you already have that preamp around...

UPDATE: Shannon at Parks Audio will put up a link to the beta firmware shortly, I have it in house now and be will testing it out.

 

Interesting differences between test records, evidently this is widespread (@Serious ). Of the two I tried, the 1967 Popular Mechanics one (results on p2 of this thread) gave about 25 dB whereas the Feickert 7" gave about 34. The Feickert is a useful record, but too eccentric for good speed stability / wow & flutter results.

Ordered the Tacet. Looks useful at a decent price.

 

I did some initial testing using the Puffin process in the video above. First thing is not to set azimuth of a cartridge while there's an off-center level sitting on top of it.

While the results sound pretty good I feel I'm missing something about the process. Starting at 4:30 in the video he explains he's trying to get to the max amount of 'separation' in each channel, and he doesn't specifically want to get to equal amounts, just the maximums. Can anyone tell me exactly what this is he's computing?

 

Steve, I was just emailing with Shannon, and he says the Azimuth check is measuring crosstalk. High signal minus low signal--using two successive tracks that have signal just in the left and then just in the right channels.

I think that makes sense to me. It's getting things aligned electrically more than physically.

I think the idea is that this is better than just trying to get each channel's output level the same, as there can be a channel imbalance in the cart itself. (And that can be corrected at the regular preamp.) I believe the Fozgometer also measures crosstalk.

I'm wagering you know more about this than me though.

 

LOL, certainly if your arm is a unipivot!

The method in the OP here is to measure absolute amplitude in each signal and non-signal channel, then take the difference between each pair to calculate the two crosstalk results. As @wbass says the Puffin is evidently doing it the same way, but it only seems to provide the results rather than the inputs - so if there's a channel imbalance inherent in the cart (a difference between signal-channel amplitudes, presuming the L and R tracks on the record are at the same level), it won't be reflected in the outputs.

I think what he's trying to get at in that part of the video is that it's best to have both separation figures as high as you can get them (= minimum crosstalk in each channel) rather than striving to make them exactly equal - because if they were equal but lower, it could be because of a combination of one channel being quieter than the other and azimuth being out.

 

Just reading the last two posts but I figured out the equation in my head about an hour back. Simple and it works, it just makes the adjustments themselves a little trickier since the numbers don't indicate which way to turn the cart. Pure trial and error but its a fast process.

There's another setting in the Puffin that lets you compensate for balance so there might be some functionality there for measuring imbalance, I'll look at it again tomorrow.

 

I've been trying out an expensive way to do this measurement, using the AnalogMagik software. I did a write up of my results in this thread.

 

Here's the record I mentioned:

(This is really all that's needed for vinyl set up, really. My custom protractor has zero points at 67.5mm and 115.0mm. Getting the alignement right isn't easy without a proper way to measure it, but so far no such software and test record exist. However that doesn't mean there never will be one.)

I haven't found another record which has broadband pink noise like this one with lateral, vertical and each channel seperately in quick succession like that. The best part is that with multi input capture in REW Pro we can immediately see from the FFT results how close we are and adjust accordingly. Saved me so much time, highly recommended!


With the Skoll bringing my focus and most of my listening time to vinyl I of course set up the phono carts again:
My worn Lyra Delos on the VPI Classic and my father's Dynavector DV20X2L and Denon DL103 with a Shibata stylus extracted from a dead AT33SA (shit happens) on his DIY TTs.


(lower volume on phono pre, hence lower level)





The Delos took some time and quite a bit of cleaning (not just the stylus) until I got it to track properly again at its 1.75g tracking force, which also removed a slight warm tilt to the FR it had by now. It still doesn't sound like new again, but it does seem like a replacement can still wait a bit



The record seems to have a bit more ultrasonic extension in the left channel. Also that step after 20kHz doesn't inspire confidence. But look at the Delos go! 80kHz in the left channel with still significant output.

You can also see 50Hz and 100Hz hum. I mentionened how I got hum on the XLR inputs on the Skoll. Well, this is that.
The 150-300Hz things could be tonearm or even cartridge resonances, I'm not sure.

I want to get a record with good channel separation and I have my eyes on a couple. Clearaudio also sells a modern copy of the JVC TRS-1007. I'm likely going to get a different one, but the Clearaudio one is a modern production no-fuss option.

 

None of your recommended test records are available from US vendors it seems