View Full Version : Turntable speed analysis
Some of you guys may not be aware, but there is a very interesting (well, I think it is) thread on pfm (pink fish media) regarding the speed analysis of turntables. This requires a test tone to be digitally recorded (3150hz is common) and a number of very interesting graphs can then be generated (particularly one called a polar plot).
A number of people have already provided recordings for evaluation, but so far no Technics 1200/1210 (a Technics SP10 has been submitted though). As I know that a number of AoS members own modified 1200's, it would be very interesting to compare them.
The thread can be found here: http://www.pinkfishmedia.net/forum/showthread.php?t=70027
sq225917
22-10-2011, 08:39
You assume that they are interested in attempting to measurably verify their mods as being a genuine improvement. A lot of mods succeed solely because no one ever does that and people just assume, different, is better.
Of course some mods are so obvious they need no verification.
You assume that they are interested in attempting to measurably verify their mods as being a genuine improvement.
Yes, you are right, that is what I assume: in fact, I assume this to be true of anyone genuinely interested in furthering turntable design (or at least interested in understanding it).
I'm not saying that measurement is essential, but if a measurement can be made that correlates to what we hear, then it has to be worth pursuing. In this case I think that Paul has come up with something that does help unravel the mystery between the performance differences between direct drive, idler and belt drive (samples of all three technologies have been submitted).
Well, after 177 views it would appear that nobody is interested - how 'disappointing'!
Stratmangler
26-10-2011, 15:28
Well, after 177 views it would appear that nobody is interested - how 'disappointing'!
More like follow the link to a multiple page thread and losing the will to live.
How about giving us a précis here of the procedure?
sq225917
26-10-2011, 23:33
http://www.kilmory.demon.co.uk/pics/af_polar.png
This should help explain it. The graph is drawn as a 'circle' where one rotation of the platter = a full circle. 4 or 5 total revolutions are shown in each graph. A turntable that ran at a perfect theoretical speed would give a perfect circle, the further away from perfection you get the less it looks like a perfect circle.
Because they show speed error amplitude and rate of change you can discern a great deal about the type of speed error from the shape of the polar. Wow displays as eleiptical deformation while high sped flutter would turn the line into a series of steep, but small, peaks and troughs. Lumps and bumps on idler wheels show up as dips and troughs and wobbly suspension shows up as well. once you get your head around them they are a great tool.
And the plot above. Two well maintained decks, one a belt driven suspended sub-chassis deck and the other a 1210. Now who'd like to take a guess which is which?
What is unusual is that it shows a graphic presentation of speed modulation in action, rather than an averaged + and - figure which only tells a very limited story. In particular, it is capable of showing very high frequency anomalies, above what would often be called flutter/ Indeed, the terms wow and flutter are themselves rather over simplistic and there area number of different types of each.
It is easy to jump to conclusions regarding drive types, but this data suggests that they do not behave in the manner often suggested (both positive and negative).
WOStantonCS100
27-10-2011, 04:05
As an owner of several MK2 Technics models (but not a "DD is the only way" type person), I do find this quite interesting; not disheartening at all. If anything, I think it points out more clearly the objectives of the engineers at Technics at the time.
Very simplistically speaking, if the goal was to get that polar graph looking more like a circle than an oval and the trade off was a bit of smoothness, then they achieved their goal. Perhaps the converse could be said for the other. Which set of trade-offs is more tolerable than the others will forever remain a matter of personal preference.
Has anyone performed this test with an idler?
It would also be more telling, I think, if these measurements could somehow be taken without the use of a vinyl pressing as that introduces many variables that ultimately influence the outcome.
I will have to read that thread in detail to make sure I'm not missing the essence or important details. Just quick thoughts here.
EDIT: I can't help but think, it would have to be one helluva drive system to actually produce a perfect circle.
It would have to be a hell of a test record too.
Very interesting, Mark. I shall have a look at the procedure and see if I can measure my 1210 with Mike New bearing & platter.
Ah, jolly good, that would be very interesting indeed :).
In amongst all that material on PFM, can you point me to the software used to create the polar response plots?
Here are two plots taken from my own turntable. The green plot is after a modification to the drive system (it is lower down the page because the speed is slightly different between the two samples). The lines represent five rotations of the platter. These plots would suggest that the rotation speed per revolution is relatively good (how much like a circle it is). But the red one has a lot of very high frequency noise (the VVVV). The green one is significantly better in this regard, but there is a bit of speed drift per revolution (the circles being a different size to each other).
http://www.kilmory.demon.co.uk/pics/YNWOAN_34_polar.png
In amongst all that material on PFM, can you point me to the software used to create the polar response plots?
You would have to contact Paul R who has said he will supply it to any that would like it - it is not commercially available.
Here is a link to his pfm PM:
http://www.pinkfishmedia.net/forum/private.php?do=newpm&u=44
sq225917
27-10-2011, 10:46
Your green plot is very circular Mark, compare that to the 1210 and it's much more regular and has equally good high frequency tracking as well. It shows how elliptical the 1210 actually is.
sharif did mention that his test record may not be in the best condition.
Here are all four plots with true circles imposed over the top:
http://i62.photobucket.com/albums/h90/markemark_2006/Turntablepolarplots.jpg
How does one interpret the 'circles'? And what do the figures displayed at the bottom left mean (e.g. 1.12856 and 1.60949)? (Incidently, isn't the high level of resolution somewhat academic?)
I assume the 'spikey' circle implies a high level of flutter, whereas a more gentle 'ripple' implies the presence of wow.
What does it mean if the mean circle is not centred on (0,0)?
sq225917
27-10-2011, 15:52
just means it's running fast or slow, or maybe both fast and slow.
How does one interpret the 'circles'?
I tried to do that a bit in post #14.
And what do the figures displayed at the bottom left mean (e.g. 1.12856 and 1.60949)?
They aren't directly related to comparative analysis - ignore them.
(Incidently. isn't the high level of resolution somewhat academic?)
I don't understand why you say that? There isn't much point in having a low level of resolution.
I assume the 'spikey' circle implies a high level of flutter, whereas a more gentle 'ripple' implies the presence of wow.
To a point, but the the VVV quality equates to very fast flutter indeed. As I said earlier, the terms wow and flutter are very generic. None of these decks show traditional wow as such, the plots would have to look much more evenly oval, or egg shaped, to do that.
What does it mean if the mean circle is not centred on (0,0)?
Nothing very important; as sq states, it just means that the recorded frequency was slightly different.
Reid Malenfant
27-10-2011, 18:25
Interesting thread :) No-one appears to be mentioning what is each system being measured, however I'd find it hard to believe that some system that generates a good deal of flutter (I think that would be the red traces in the pictures above) would be from a high mass platter. With high mass it should act more like a flywheel & I'd expect less flutter, the drive unless ridiculously powerful should have less of an effect on the platter once up to speed.
Just my way of thinking, I'm not a mechanical engineer.
With a high mass platter I'd expect wow to be more prevalent depending on the drive system...
I'm not sure exactly what the graphs are saying, but I'd hazard a guess & say direct drive with a high mass platter should give the least wow & flutter :scratch:
:cool:
Interesting thread :) No-one appears to be mentioning what is each system being measured,
Well it's all in the pfm thread. The first graph that sq posted shows a Phonosophie turntable in red (belt drive, flywheel platter, suspended chassis, stepper motor) and a Technics 1200/1210 (direct drive, low mass platter, non suspended, feedback loop)
Both of my plots are from my own belt drive, flywheel platter, suspended design, AC synchronous motor design turntable.
however I'd find it hard to believe that some system that generates a good deal of flutter (I think that would be the red traces in the pictures above) would be from a high mass platter. With high mass it should act more like a flywheel & I'd expect less flutter, the drive unless ridiculously powerful should have less of an effect on the platter once up to speed.
Well I can see why you might think that but the evidence does not support it - in fact, the opposite appears to be true. Or rather, it depends at what frequency you think flutter is defined. The sharp VVV pattern is extremely rapid speed variations over a very short period. The smoother ununu pattern is more like actual flutter and this pattern does seem to be absent from the higher mass platters. It might also be tempting to blame the VVV pattern on the turntable suspension but the SME10 shows a similar quality but has no suspension and uses a feedback equipped DC motor.
With a high mass platter I'd expect wow to be more prevalent depending on the drive system...
I wouldn't particularly. There isn't really any evidence in the plots to support such an assumption either.
I'm not sure exactly what the graphs are saying, but I'd hazard a guess & say direct drive with a high mass platter should give the least wow & flutter :scratch:
It will be interesting to see how the high mass platter interacts with both the torque of the motor and the feedback circuit - I feel sure that the bearing will also play a significant role. At present the absolute speed stability of the DD designs has not showed itself to be particularly strong. They are largely free from very rapid fluctuation though and speed up and slow down by very similar amounts per rotation (though in both these regards the belt driven Phonosophie is better and mine varies less per revolution).
They aren't directly related to comparative analysis - ignore them.
If the figures are to be ignored, what is the point in quoting them to six decimal places? That was my point about resolution (vs. accuracy).
It seems a useful tool to assess performance, but how do you interpret the deviations from a perfect circle: what do the horizontal and vertical scales mean?
Sorry to be a bit thick over this, but I come from a scientific background (30 years working as as a Research Physicist in the research laboratory of a major electronics company), so am sceptical of data presented graphically without well defined scales.
sq225917
28-10-2011, 00:45
forget the horizontal and vertical scales, they are the same, consider the scale being a line running from the crossover point of the axis out at 90 degrees. The slower the speed at any point the closer that point is to the crossover point of the axis, and vice versa.
The graph shows speed of rotation at every given point per rotation of the platter. The actual scales are the same for each polar plot. I don't know what the distance from centre to edge of graph correlates to, it's pretty small though more like 0.1rpm than 1.0rpm
At present the absolute speed stability of the DD designs has not showed itself to be particularly strong.
....and yet I've read so many times that the superb speed stability of direct drive designs is a strong contributor to the way they sound. Yet none of this seems true based on these graphs.
Well, not entirely, I would say that they are strong in some respects; there is very little high frequency 'super flutter'.
If you listen to my recordings made when the deck exhibits this 'super' flutter and when it is suppressed, the suppressed recordings do have the kind of stability often attributed to DD designs.
The plots of the DD do not show wow either (not in the traditionally accepted definition of wow), it's more like an overlayed beat.
So I guess that the thing demonstrated by your recordings of your own turntable is direct drive is inherently no more speed stable than belt drive.
Yes, I think that's right - I think the recoding of the Phonosophie P3 (which is also a belt drive suspended design, but using a stepper motor) also supports it. I think DD can be more stable than belt drive, but I don't think it is inherently better.
But don't get me wrong, I'm not saying that subjective adherents of DD decks are wrong, or making it up, because I don't think they are. What I do think is wrong is to associate the positive qualities they hear so explicitly with the drive method used. I think it's possible to make different drive systems behave like each other, not that they necessarily do.
What I think is that a turntable is a superficially simple combination of components that actually interact in a surprisingly complex manner. It is tempting to apply definitive performance descriptors to the individual components, motor, bearing, platter, chassis etc. but the truth is that the influence and performance of these components is not a fixed point, but is actually quite fluid and varies depending on the specific manner in which they are mixed together.
So I guess that the thing demonstrated by your recordings of your own turntable is direct drive is inherently no more speed stable than belt drive.
There are two things going on there: absolute speed stability depends largely on whether a crystal reference is used and can apply to any of the drive methods; speed variation around one revolution is where I think the jury is out pending a lot more measurements from people.
I will do my own soon and hope to get a response from the PFM people.
absolute speed stability depends largely on whether a crystal reference is used and can apply to any of the drive methods..
To be honest, I think there is considerably more to it than that, or rather I think there a number of other elements that seek to work against such perfection. Whilst the crystal may oscillate in a very accurate manner, actually applying that information in a manner that leaves it uncorrupted is an entirely different matter.
There are five revolutions represented by each plot.
There are five revolutions represented by each plot.
Yep, I appreciate that, and the eye is visually averaging them. There seems to be remarkable consistency to most of the plots.
As a matter of interest, are you recording WAVs to your laptop at standard 16/44 resolution or something else?
Ah yes, I appreciate that - It would be easier to follow each trace if they were colour coded - I agree that the eye tends to average them.
Yes, I recorded mine at CD standard 16/44 - phonostage plugged into external USB A2D into my laptop using Audacity (WAV - 14 seconds long).
I am very much looking forward to your plots - I do expect an improvement.
Powered by vBulletin® Version 4.2.3 Copyright © 2024 vBulletin Solutions, Inc. All rights reserved.