The Korf Blog

The inside story: our research,
development and opinions

10 February 2021
Any Point in Wow & Flutter? Part III
Since we have published our two previous posts (Part I, Part II) on wow and flutter, we had a privilege of talking to a distinguished guest, an AES author Mr. Ray Kilmanas.

Mr. Kilmanas did some very interesting work back in 1982, showing that not only the turntable drive, but the tonearm configuration too can have an impact on wow and flutter. As the tonearm traces the warps of the LP, it is "scrubbing" — frequency modulating the signal that's being picked up. We have touched on the subject of scrubbing earlier. This should have exactly the same effect as the turntable drive errors.

But does the scrubbing exist "in the wild"? It is quite easy to provoke with a specialized test LP, but the real world LPs have no content below 20 Hz or so. We have prevoiusly shown that the only possible source of such low frequency disturbances are vertical warps. Unless the LP is completely out of shape, their influence is quite small. Is it enough to induce scrubbing?

How can we test the scrubbing hypothesis? Linear tracking arms, having no headshell offset angle, should be more resistant to scrubbing. What if we could measure two turntables with the same (or very similar) drives but different types of tonearms — one with pivoting, and one with a linear tracker? Will it show the wow and flutter advantage of the latter?

The Turntables
As luck would have it, I have two turntables with theoretically similar direct drives. One is a beautiful Technics SL-7 that we have fully restored a decade ago. And the other is slightly tatty, but functionally perfect, Grundig PS 3500.
But, you would say, what can possibly a European Grundig have in common with Japanese Matsushita?

A lot, actually. It was built by them and differs from the Technics SL-D2 mainly in cosmetics (and in having a garden hose for an arm tube). Its drive is based on the ubiquitious Panasonic AN630 IC.

The SL-7 is a bit more advanced, with a later AN6635 IC driving the coils. But both are 3-phase Technics designs of similar configuration.

SL-7's drive has an advantage, but it is largely in the absence of higher frequency fluctuations. Its low-frequency behaviour, where the scrubbing effects might be, should be broadly similar.

The Measurements
First, the speed deviation charts. They are familiar to most from Dr Feickert's own sadly retired Platterspeed app that is still being used by the Stereophile magazine. We are doing ours in Virtins Multi Instrument. Technics SL-7 is on the left, Grundig PS 3500 on the right. This is the raw FM-demodulated data:
Unfiltered data gives about 0.37% unweighted wow and flutter for Technics, 0.7% for Grundig. Ouch. Twice as bad. It looks like Grundig's drive is much worse than SL-7's.

And this is the lowpass-filtered data:

The lowpass-filtered data, where any tonearm scrubbing effect should be, is a bit closer: 0.24% wow & flutter for Technics, 0.33% for Grundig. Ok, but it's still 1/3rd off. So much for our "similar drives" experiment.

Let's look closer at the low modulating frequencies. Maybe we'll be able to separate the possible tonearm scrubbing influence there?

After digging through the vast Virtins Multi Instrument manual I have finally figured out how to increase the FFT resolution up to about a tenth of a Hertz in the low frequencies. Virtins is an amazing tool that would have really benefited from a command line interface, instead of a zillion options randomly spread around hundreds of dialogue boxes.

Here's the spectrum of both speed deviations. X axis is frequency in Hertz, Y is speed deviation in %. If low-frequency induced scrubbing is a thing, we should see some extra peaks from the Grundig data that are absent in the SL-7 data.

There are more very low frequency peaks in the SL-7 data, and they are more pronounced. Grundig only has one smallish peak at about 13.5 Hz that isn't present in the SL-7 data. There isn't enough information here to make any conclusions.

So What Have We Discovered?
The low frequency noise by itself is not strong enough to induce scrubbing
Today's experiment was not a smashing success. The drives of two turntables turned out to be quite dissimilar even after applying the lowpass filter. After FM demodulation, we could not find any pronounced peaks that could be attributable to the scrubbing effect of the pivoting tonearm.

It is perhaps not very surprising, given what we have found out about the low frequency behaviour of tonearms. The low frequency content of the LP and its warps is by itself not strong enough to induce scrubbing.

But this doesn't mean that scrubbing is not a thing. There are other forces acting on a cantilever that can provoke scrubbing. The most obvious one is the variation in the loudness of the material being played back. The resulting change in stylus drag is well-known (and measureable). Obviously, it would not show up in any experiment done with a stable 3150 Hz signal.

Currently I can not think of any straightforward way to measure such scrubbing. If you, dear readers, have any ideas — please share them in comments and I'll try to implement the best ones.

Please subscribe to receive blog updates in your inbox!
comments powered by HyperComments