Don't listen to records on FM!

[diyrayk]

I don’t mean FM radio, I mean FM as in Frequency Modulation Distortion (FMD, or pitch instability). FM occurs as a result of the stylus tip scrubbing along the groove during tonearm mass/compliance resonance or transient events. The amount of scrubbing depends on the amplitude of the resonance or transient event but I want to show that it also depends on arm geometry, and that there are arm geometries that can minimize it. I’ll present graphics, computer animations, and video sound clips to demonstrate that arm geometry can affect apparent turntable pitch stability which, in turn, can degrade the measured wow & flutter figure of an otherwise excellent turntable. This is not another alignment thread.

I’m starting this new thread as a spinoff or extension of topics discussed in the diyaudio “Turntable speed stability” thread:
https://www.diyaudio.com/forums/analogue-source/309349-turntable-speed-stabilty.html
…and there are some roots in the “Test LP group buy” thread:
https://www.diyaudio.com/forums/analogue-source/313335-test-lp-buy.html
Much has already been written and discussed about the AM effects of resonance and ways of dealing with it, including various implementations of tonearm damping and commercial/diy low-pass electronic filters. However, the FM effects of stylus scrubbing cannot be removed electronically downstream by low-pass filtering.

If there’s a mass/compliance resonance event, the stylus scrubs. If the stylus scrubs, there is FM.

Any arm motion that causes cantilever deflection will also result in stylus scrubbing fore & aft along the groove. Stylus scrubbing changes the relative groove velocity, which causes a pitch variation or speed instability that can sometimes be audible, depending on program material and individual listener pitch sensitivity. The mitigation of scrubbing arising from horizontal versus vertical displacements involves different arm design tactics. Here is a familiar graphic from Shure tech notes which shows how horizontal mode scrubbing is affected by the offset angle of a typical pivoted arm:

​

[01] Shure Figure 2 Simplified Lateral Scrubbing​

Shure’s graphic shows a simplified relationship between horizontal motion and scrubbing along the groove and a simple formula that describes it. Notice that the scrubbing motion is asymmetrical WRT headshell/cantilever displacement. For a symmetrical sideways displacement of the headshell, the stylus scrubs farther aft than it does forward. More on that later. The motion of the stylus point in the groove can be calculated if the groove radius, the stylus cantilever length, the arm pivot to spindle distance, and the distance from the arm pivot to the cantilever pivot/suspension is modeled as a four-bar mechanical linkage. This model works for all types of arms, whether S shape or J shape with an offset bend in the arm tube, or straight type with an offset angle at the headshell. It also works for most servo linear arms, which are generally just motorized pivoted arms with no offset angle. Air bearing and passive/mechanical linears can also be analyzed but within a certain range of assumptions.

​

[02] Four-Bar Linkage Model​

It’s easy to demonstrate the scrubbing motion yourself by making a crude working mockup of this four-bar linkage system with a plastic coffee stirrer. Cut a flexible joint notch in the stirrer at about ¼ of the length. Bend it 90 degrees at the notch. Anchor the long end of the stirrer to the turntable spindle (this becomes the groove radius ‘link 1’), then slip the short end over the cartridge stylus (this becomes the stylus cantilever ‘link 2’, and you’ll want to use a junk cartridge for this). The line from the arm pivot to the spindle becomes ‘link 3’, and the line from the arm pivot to the stylus cantilever pivot (aka cantilever suspension) becomes ‘link 4’. Note: ‘link 4’ as defined is not the same as the arm’s “effective length”! For this mockup demo I used a cheap ceramic cartridge and then rocked the tonearm sideways left/right to see what happens:

[03] Tonearm/Stylus Scrubbing at Mass/Compliance Resonance - YouTube
​

[03] Tonearm/Stylus Scrubbing​

More on the way…

Ray K

 

[diyrayk]

For sideways headshell displacement on a generic 9” pivoted arm with an offset angle of around 24 degrees, Shure’s simple graphic formula Vsinα shows that roughly 40% of the sideways cantilever displacement gets translated into stylus tip scrubbing motion along the groove. The amount and character of the scrubbing depends on the magnitude of the horizontal displacement, and the interaction of the links in the four-bar linkage model. Before you get ahead of me, yes, a 12” arm is better in this regard than a 9” but more on that later.

Stylus scrubbing along the groove causes a pitch fluctuation (FMD) of the audio signal in the groove and occurs at the same rate as the mass/compliance resonance frequency of the arm/cartridge combination. It is the same audible flutter effect as the cyclic or transient speed variations caused by an eccentric capstan on a tape deck, or a defective drive component in a movie film projector. The audible effect of flutter in these types of mechanical reproducers has already been extensively researched and documented in technical papers and journals. The British Broadcasting Company (BBC) published an excellent technical paper wherein their researchers found that listeners were most sensitive to FM flutter distortion when the audio was frequency modulated in the range of about 5Hz to 10Hz. This is in the same ballpark as commonly encountered mass/compliance resonance frequencies of typical arm/cartridge combinations.

Note that as the stylus scrubs forward the pitch and effective groove velocity decreases, whereas as the stylus scrubs backward the pitch and effective groove velocity increases. Most of us use magnetic cartridges whose output is velocity sensitive, so the cartridge output of the mid and high frequency program material will alternately experience some asymmetrical AM in sync with the FM distortion when that occurs. I’m still debating in my head the net effect of this velocity dependent AM interacting with the slope of the RIAA eq curve. Maybe a topic for discussion?

I used a freeware mechanical linkage motion program called “Linkage2” (thank you Carlo for alerting me to this) to model typical arm/cartridge/cantilever/stylus mechanical systems and display stylus tip scrubbing in action. The black line in the model image is the radius from the spindle to the stylus tip (model Link 1). The blue line is the stylus cantilever (model Link 2). The green line is the arm and cartridge body from the tonearm pivot to the cantilever suspension ‘bearing’ (model Link 4). The gray line is the arm pivot to spindle (model Link 3). Here’s a Linkage2 model of what I will call “Arm 1”. It is a typical 9 inch pivoted arm with a nominal 24 degree offset angle, positioned at a 60mm radius null point (okay, it’s a SME-3009 S2 Improved). The model only needs to be concerned with the geometry of the actual moving parts, and doesn’t show the cosmetic superstructure of the arm. In other words, it doesn’t matter whether the arm is “S” shape, “J” shape, straight, or a pretzel.

[04] Arm 1 SME-3009 Linkage2 Model
​

Here’s a close-up Linkage2 video of the simulated mechanical motion of the stylus and cantilever during one cycle of horizontal resonance. Notice the asymmetry of the scrubbing. As the arm moves outward from the spindle, the stylus scrubs backward farther than it does forward when the arm moves inward to the spindle by the same amount:

[05] Arm 1 - Linkage Cantilever View - YouTube

[05] ARM 1 – Linkage Cantilever View

​

Here’s a video of a test track of music from the JVC Electro-Dynamic Servo Tone Arm Test Record. It’s a piano piece with a superimposed horizontal LF signal swept from 4Hz – 9Hz for 60 seconds. The amplitude of the LF tone really stresses the cantilever:

[06a] Arm 1 - Piano - YouTube

[06a] Arm 1 – Piano​

Here’s a video of a test track from a vintage Ortofon 0002 test record. It’s a tone complex of 2349Hz and 2960Hz with announced LF spot frequencies from 25Hz down to 4Hz. It’s less severe than the JVC disc:

[06b] Arm 1 - Test Tone - YouTube

[06b] Arm 1 – Test Tone​

“Arm 1” for this test is a SME 3009-S2 Improved, aligned per SME template.

The cartridge is a Stanton 681 with genuine EEE stylus, brush down and corrected for 1.5g VTF. AS string is set at the 1.25g notch. The test record is a vintage pressing of “Ortofon Pickup Test Record 0002”. Ortofon’s vintage 0001 test record also has these test signals but be aware that Ortofon’s vintage 0003 test record and Ortofon’s new currently released test record DO NOT have these test signals!

For comparison, here are video/sound clips of another arm setup, which I’ll call “Arm 2”. It is playing the same two test tracks, but it has a different configuration and geometry than SME 3009-S2 “Arm 1”. The cartridge is the same Stanton 681EEE as used in the “Arm 1” test, same 1.5g VTF, with brush, same turntable:

[07a] Arm 2 - Piano - YouTube

[07a] Arm 2 – Piano

[07b] Arm 2 - Test Tone - YouTube

[07b] Arm 2 – Test Tone

​

Well, wuh d’ya think?

More on the way…

Ray K

​

 

[billshurv]

I have to say, on the arm1 test tone it's mesmerising just watching the groves. What interests me is how damping affects this. My naive view is that damping wont remove the scrubbing, it will just stop the cantilever getting into a total tank slapper.

 

[diyrayk]

Yes. Any comments on arm 1 versus arm 2?

Ray K

 

[billshurv]

Arm 1 seems to have a higher Q resonance but beyond that hard to call.

 

[Hiten]

Arm 2 sounds clean.
On other hand Piano strikes of Arm 1 sounds wobbling/oscillating (lack of better word). My hearing is not good though. 😱

Thanks for doing this diyrayk.
Best regards.

 

[diyrayk]

Hiten,

Thanks for listening. You are on the right track. I think it is easier if one listens with ears instead of eyes. Let's see if anyone else is interested in this.

Ray K

 

[Tubelab_com]

Anyone else old enough to remember the Garrard Zero-100. This thread reminds me of the Garrard sales literature. I got one from the Garrard rep in 1971. Neat idea, but it took constant maintenance to keep it working, and I was a Garrard authorized repair tech. It got swapped for a basic Technics Direct Drive in the mid 70's and I still have it.

 

[The Peasant]

Arm 1 appears to have a slightly lower resonant frequency and is more easily excited by the LF signal. The FM is clearly audible in this case. Arm 2 appears more stable and much less affected by the LF, even when it is visually oscillating. Audible FM is much reduced.

Take care,
Doug

 

[sgrossklass]

On the test tone in particular, arm 1 is dramatically more affected than arm 2.

It's interesting that resonance frequency for both does not appear to be too different - about 9 Hz for arm 1 and 10 Hz for arm 2 by the looks of it? What happens if you add the right amount of mass to arm 2 to make them equal?

I imported both into Audacity to have a look at spectra, which didn't show the dramatic differences I had hoped for. In a direct comparison, the setup with arm 2 was revealed to have quite a bit more background noise though (not a function of the arm but rather the platter driver, I guess - noisy bearing or a rim drive?).

 

[Mark Tillotson]

Hiten,

Thanks for listening. You are on the right track. I think it is easier if one listens with ears instead of eyes. Let's see if anyone else is interested in this.

Ray K

Listen with eyes on the spectrum analyzer screen if you want to measure FM artifacts, surely? Very distinct signature.


The term flutter is used for FM artifacts above 4Hz in audio reproduction.

 

[billshurv]

I think you mean 'was used'. Unless you have a citation for that from the 21st century?

 

[Hiten]

Hiten,

Thanks for listening. You are on the right track. I think it is easier if one listens with ears instead of eyes. Let's see if anyone else is interested in this.

Ray K

May be wrong words chosen to describe; but actually I did listened to it properly with limited resources/methods I have. (because I like to see mechanical working of vinyls/TTs.)

I listened to it on my PC.
Creative Speaker.
Clicked on both youtube links.
Only one speaker.
Low Volume.
Close listening about 1.5 feet.
Choose short section at around 0.10 sec in video.
Few seconds of repeated alternate listening for about 7/8 times.
Then moved on to listen to the remaining part of music.
Again by alternatively listening to short pieces by switching between tabs of the youtube links.
Regards.

 

[diyrayk]

Several of you heard a difference, so I’m going to take that as a validation of my hypothesis.

Here’s what you were listening to:

​

[08] Arm 1 Setup​

​

[09a] Arm 2 Setup

​

[09b] Arm 2 Front View

​

[09c] Arm 2 Cartridge​

“Arm 2” is the same SME 3009-S2 Improved as “Arm 1”, except that the cartridge was re-aligned so that it is in the same vertical plane as the arm pivot, such that the offset angle has been effectively changed to zero. Effectively, I turned the 3009 into a straight DJ arm. The arm was then repositioned with some underhang to have a null point occurring in the center of each test track.

Mathematical analysis of 9” pivoted Arm 1:

​

[10] Arm 1 Pivoted Scrub Motion for One Cycle of Cantilever Deflection​

This graph shows stylus tip motion along the groove through one complete cycle of resonance for peak stylus deflections of 0.381 mm (0.015 inch) and 0.127 mm (0.005 inch). The FM scrubbing associated with a 0.381 mm horizontal displacement of the cartridge is equivalent to several % peak flutter at this inner groove radius. 0.3 mm is the thickness of a typical heavy-paper business card, so even barely visible sideways ‘bobbling’ of the headshell can introduce significant amounts of scrubbing and flutter into the audio signal. This FM will occur even if the turntable is a mega-expensive aerospace machined marvel with % flutter spec’d in parts per million. The pitch variation of this FM has nothing to do with the pitch stability of the turntable platter or drive system, but the wow & flutter measurements I’ve seen don’t separate out the contribution to measured pitch instability that comes from arm/cartridge scrubbing.

12” pivoted arm with reduced offset angle:

Given that the scrubbing in a 9” arm is influenced by the offset angle, then what happens when you have a 12” arm and reduce the offset angle to around 18 degrees? As expected, for the same resonance amplitude, the amount of stylus scrubbing is lower (better) than the 9” arm, all other things being equal. Of course, all other things are never equal, but at least in terms of FM scrubbing, the typical 12” arm geometry has a leg up on a 9” arm.

Mathematical analysis of 12” version of pivoted Arm 1:

​

[11] Arm 1 (12” version) Scrub Motion for One Cycle of Cantilever Deflection​

Sorry, I don’t have a 12” arm to demonstrate comparison sound clips.

Arm 2 analysis is on the way…

Ray K

 

[diyrayk]

For some reason the photos don't seem to be posting correctly, let me try again:

Ray K

 

[sgrossklass]

Good ol' Dual rim drive. Still not sure why noise level would have been that much different between the two geometries (might have been caused by something unrelated).

So basically, FM intermodulation distortion is the price we pay for the two nulls in tracking angle of the conventional geometry? Bummer. The question is, is there enough subsonic energy on non-pathological recordings to make this a problem?

 

[diyrayk]

Arm 2 - Zero Offset Angle Arms and Linear Trackers

So, if less is more, then what happens if we go with a geometry that has zero offset angle, as in “Arm 2”? Such would be the case for linear tracking arms, including servo-driven “pivoted” arms and air-bearing and mechanical bearing arms. For such arms with zero offset angle, the scrubbing vs. cantilever deflection looks like this:

​

[12] Arm 2 Simplified Lateral Scrubbing​

​

For the same sideways displacement of the headshell as was the case for 9” and 12” pivoted arms, the stylus scrubbing is greatly reduced when the offset angle is eliminated.

Here’s the Linkage2 model of “Arm 2” - a zero offset angle arm:

​

[13] Arm 2 Linkage2 Model​

Here’s a close-up of the cartridge end of the arm:

​

[14] Arm 2 Linkage2 Model Cantilever View​

Here’s a Linkage2 video of the simulated mechanical motion during horizontal resonance:

[15] Arm 2 - Linkage Cantilever View - YouTube
​

[15] ARM 2 – Linkage Cantilever View Video

​

The scrubbing in “Arm 2” is so much less than “Arm 1” that it’s difficult to see in the same scale. Here’s a close-up stylus view of the same video. Note that the magnitude of the scrubbing has dropped to ~0.05mm.

[16] Arm 2 - Linkage Stylus View - YouTube
​

[16] ARM 2 – Linkage Stylus ViewVideo​

Mathematical analysis of zero offset angle Arm 2:

​

[17] 9” Arm 2 Scrub Motion for One Cycle of Cantilever Deflection​

“Arm 2” is the same 9” pivoted “Arm 1” that was used in the first test, except that it was reconfigured as a straight zero offset angle arm for playing the test band. It is equivalent to a 9" servo-linear arm. For the same peak horizontal deflection as in the 12” pivoted arm, the calculated scrubbing in this zero offset angle “Arm 2” (and associated FMD) is much less than of the 12” arm.

It’s the same arm, same cartridge, same stylus, same compliance, same effective mass, same turntable & mat, same test record, same resonance, same damping, etc. What’s different is only the geometry. A ViV Lab “Rigid Float” or RS Lab “RS-A1” or a Tru-Glider should perform comparably for this test, but all “other things” then wouldn’t be the same and the validity of the test comparison would be murky.

​

[17b] VivLabs RigidFloat
​

More on the way...

Ray K

 

[diyrayk]

DOUBLE YOUR FUN

If you compare the scrub motion graph for the linear “Arm 2” with the scrub motion graphs for the 9” “Arm 1” and the 12” "Arm 1a", you can see that for the 9” and 12” pivoted/offset arms the scrubbing motion is asymmetrical. For a symmetrical excursion of the headshell/cartridge horizontally, the stylus scrubs farther aft than it does forward. In fact, relative to a symmetrical excursion of the headshell/cartridge horizontally, the mechanical action of the four-bar linkage introduces a “second harmonic” of the mass/compliance resonance scrubbing frequency which also frequency modulates the audio signal in the groove. As the offset angle is reduced to zero (as would be the case for a linear tracker), the modulation frequency and scrubbing becomes just the second harmonic of the mass/compliance resonance frequency. If you view the “Arm 2” linkage video carefully you will see that for every cycle of horizontal resonance there are two cycles of FM stylus scrubbing. For example, if the cart/arm resonance is 8Hz, the AM rise is still at 8Hz but the frequency modulation due to the scrubbing in fact occurs at 16Hz. The origin of the second harmonic effect is a purely mechanical consequence of the geometry and linkage motion.

Normally we associate the phrase “second harmonic” with distortion and something that has negative connotations. Counterintuitively, this second harmonic frequency doubling effect of the FM in zero offset angle arms and linear arms is beneficial. Remember that the BBC research team found that listeners were most sensitive to flutter/FMD when the audio frequency modulation occurred in the range of about 5Hz to 10Hz. In other words, there is a threshold curve for sensitivity to FMD that varies with frequency, and the maximum sensitivity region (the dip) in that curve is in the range of 5Hz to 10Hz.

[18] FM Threshold Curve
​

​

Let’s say we have a zero offset angle arm or linear tracker such as a Rabco, Goldmund T3/T5, Yamaha PX-1/2/3, Mitsubishi LT-20/30, Pioneer PL-L1000 (you get the idea), and the arm was set up for mass/compliance Fres of 10 Hz. On a horizontal low frequency sweep the arm will exhibit an amplitude resonance peak at 10Hz, the same as it would for a pivoted arm set up for Fres of 10Hz. However, the horizontal mode stylus scrubbing on the linear tracker will be frequency modulating the audio at 20Hz instead of 10Hz. The reason that this modulating frequency doubling effect is beneficial is that the FMD sensitivity curve rises to a slightly higher threshold level at 20Hz, so we are less sensitive to flutter at the higher frequency and it becomes subjectively less audible.

Comparison of 9" Pivoted Arm 1, 12" Pivoted Arm 1a, and 9” Zero Offset Arm 2

It is easier to compare the differences in horizontal FM susceptibility of the three arms if the stylus displacement scrubbing motion is plotted as a function of cantilever deflection. In this graph the ideal would be that, as the cantilever gets deflected left and right, there should be no scrubbing at all. In other words, the stylus displacement scrubbing ‘curve’ would ideally be a straight line on the horizontal axis. The closer the ‘scrub’ line is to the horizontal axis, the lower the FMD. The ‘scrub’ line approaches that ideal as you reduce the offset angle to zero. As you approach zero offset angle, the remaining scrubbing motion is mostly affected by stylus cantilever length.

[19] Arm 1, 1a, & 2 Comparison​

What about high mass air bearing and mechanical linear trackers that have very low mass/compliance resonance frequencies? Well, these will still exhibit less FM scrubbing than pivoted arms because of the elimination of the offset angle and they, too, will experience the FM scrubbing frequency doubling effect but an arm with, say, 2-3Hz Fres could see the resulting scrubbing FMD flutter frequency land near the beginning of the 5HZ to 10Hz high sensitivity zone.

I’m suggesting that the more compelling explanation for afficionados of linear trackers describing their tonearms with subjective terms such as “solid”, “focused”, “clear transients”, blah, blah, has more to do with reduction of the time-smear effects of FMD (pitch instability heard as flutter) rather than the reduction of tracking error distortion. I’ve seen reviews of the “Rigid Float” and “RS-A1” use those same glowing descriptive terms, despite the fact that those arms introduce boatloads of tracking error. More on the RS-A1 later.

More on the way…

Ray K

 

[dennis h]

thanks

As one considering an airbearing arm your methods have got my attention,thank you for your efforts and teaching.
(and yes I could hear it thru my cheapie ancient HP computer speaker and even more so with cheapie headphones thru same putor)
dennis h


















a

 

[diyrayk]

(and yes I could hear it thru my cheapie ancient HP computer speaker and even more so with cheapie headphones thru same putor)

Thank you for your comment. I decided to do the videos with simple (some might say 'low') technology for two reasons, partly out of laziness spurred by wanting to get this idea out and posted, but also to demonstrate that FMD can be audible on the most modest of systems, even computer speakers and cheapie headphones. I can't stand it when someone plays the card "well, if you had a good enough system...", yada, yada.

I did the simulations on a freeware program called 'Linkage' which you can download here:

Linkage Mechanism Designer and Simulator | Dave's Blog

I attached the 'linkage2' model files for Arm 1 and Arm 2 so anyone can view them in far better resolution than the YouTube videos. Feel free to play with, modify, or pass them around.

Ray K