Turntable speed stabilty

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2014
Paid Member
This is a new thread that has been spun off from here http://www.diyaudio.com/forums/analogue-source/303389-mechanical-resonance-mms-32.html#post5115798 to discuss the measurement, analysis and interpretation of turntable platter speed and stability, with the end goal of understanding it more and the fixes for it when needed. The Dr Feickert test records and apps have made measurements a lot more accessible, but the presentation and analysis part has been missing. For example Fremer always posts Feickert output on his reviews (e.g. https://www.analogplanet.com/content/technics-direct-drive-sl-1200g-turntable) but you can see that they heavily filter the measurement, throwing away a lot of useful information (and the reviewer wouldn't know what to do with that information anyway).

Member Luckythedog has taken this up a couple of notches by plotting the speed on a polar chart, so you can see speed variation on a rotation by rotation basis. This immediately gives useful additional information you can use to work out where the speed changes might be coming from.

I should cover on this post why this all matters, after all we are talking subsonic effects in the most case. The reason it matters, along with record eccentricity and cart-arm resonances is that you are frequency modulating the wanted signal. Unless you are chasing a 'phat' analog sound this is something that it's worth trying to minimise.

Floor open for discussion
 
Compare two test records

Interesting subject I did a compare between two testrecords Adjust and a DIN W@F testrecord both 3.15khz center freq. span 400 hz on the same turntable Linn LP12 with SME 5 tonearm. Both after carefull centering the record.

Linn LP 12 -SME-V Speed spectrum Adjust testrecord    .png

Linn LP 12 -SME-V  Speed spectrum DIN 45545 Testrecord   .png
 
I have a polar plot for an LP12, attached below. It's from a 3rd party who provided a short 3kHz or so test file. I made a habit of collecting these test files for 'TTs of some reputation', so chances are I have examples of most of them in my archive and can reel them out as we like.

It seems a good place as any to explain the polar plot, which presents how frequency of a 3kHz or so test tone varies once around a platter revolution. So once around the plot is one platter revolution, and radius is 'instantaneous frequency' in Hz. Each radial line represents pitch variation of 0.1%, said to be about the audible threshold of pitch variation for most people. Inset is spectrum of whatever modulates the test tone.

This LP12 example is actually quite good as these things go. One can readily see mostly there's a once per platter revolution 'eccentricity', which might well be record mis-centering. Otherwise, one can see contributions around the 8-12HHz region which might well be cart-arm resonance - often this can dominate stability in a good TT, IME, as it appears here.

So in this case although the underlying TT seems good, the cart-arm resonance might well be dominant in determining pitch stability, or there might be something else TT drive-wise that contributes the 8-12Hz instability.

Chances are this TT was running slow and the tone was intended to be 3kHz I suspect.

Interesting way of looking at things?

LD
 

Attachments

  • LP12.JPG
    LP12.JPG
    325.1 KB · Views: 2,188
I have a polar plot for an LP12, attached below. It's from a 3rd party who provided a short 3kHz or so test file. I made a habit of collecting these test files for 'TTs of some reputation', so chances are I have examples of most of them in my archive and can reel them out as we like.

It seems a good place as any to explain the polar plot, which presents how frequency of a 3kHz or so test tone varies once around a platter revolution. So once around the plot is one platter revolution, and radius is 'instantaneous frequency' in Hz. Each radial line represents pitch variation of 0.1%, said to be about the audible threshold of pitch variation for most people. Inset is spectrum of whatever modulates the test tone.

This LP12 example is actually quite good as these things go. One can readily see mostly there's a once per platter revolution 'eccentricity', which might well be record mis-centering. Otherwise, one can see contributions around the 8-12HHz region which might well be cart-arm resonance - often this can dominate stability in a good TT, IME, as it appears here.

So in this case although the underlying TT seems good, the cart-arm resonance might well be dominant in determining pitch stability, or there might be something else TT drive-wise that contributes the 8-12Hz instability.

Chances are this TT was running slow and the tone was intended to be 3kHz I suspect.

Interesting way of looking at things?

LD


Yes it is which testrecord was used for this measurement ?
 
Yes it is which testrecord was used for this measurement ?
It works with any test record that provides a continuous sine tone at about 3kHz or 3.15kHz, and just needs a short recording of 10s or so in any lossless format at any reasonable level.

So, over a few years, I collected together recordings of test sample tones that people have posted on forums, but usually I don't know and it doesn't matter what the actual test record was.

I don't mind running a few examples here, if anyone likes?

The software is a set of VB modules I wrote a few years back for myself to diagnose and improve pitch stability on my rigs. I never got round to making a proper user interface, so I still have to run them them manually.

Meantime, here's another good example from the archive. This is from a Micro Seiki SX-777. Red and Blue traces are for 2 consecutive revolutions.

LD
 

Attachments

  • SX-777.jpg
    SX-777.jpg
    338.6 KB · Views: 2,149
Interesting subject I did a compare between two testrecords Adjust and a DIN W@F testrecord both 3.15khz center freq. span 400 hz on the same turntable Linn LP12 with SME 5 tonearm. Both after carefull centering the record.

View attachment 622759

View attachment 622760
Thanks, Volken - that's interesting.

I venture the difference might be due to different warp/vertical run out, or to different stylus-groove friction between the test records?

Stylus-groove friction isn't steady, it varies with time in a 1/f flicker noise like profile about an average value IME. This variation causes variation in skate force in an offset arm, and in downforce due to VTA on all arms. This is a stimulus that can cause the headshell to move which is a source of FM or pitch variation. What happens next depends on the cart-arm resonant system, and how much damping there is around.

Could you post short recordings, of the test tones please - about 10s or so ?

LD
 
Last edited:
Thanks, Volken - that's interesting.

I venture the difference might be due to different warp/vertical run out, or to different stylus-groove friction between the test records?

Stylus-groove friction isn't steady, it varies with time in a 1/f flicker noise like profile about an average value IME. This variation causes variation in skate force in an offset arm, and in downforce due to VTA on all arms. This is a stimulus that can cause the headshell to move which is a source of FM or pitch variation. What happens next depends on the cart-arm resonant system, and how much damping there is around.

Could you post short recordings, of the test tones please - about 10s or so ?

LD

These two speed spectrum measurements shows that the test record is very important , and indeed it shows the tonearm/cart freq. and the 100hz motor vibration .

Pitty enough I have no records from these I use the spectrum measurements for revision work before and after .

Here a EMT 930


EMT 930 st ser.8828 Speed spectrum after revision 2.png
 
Hi Volken, I have an archived 3rd party 3150Hz test recording from an EMT 930, it's polar plot is attached below.

I think one can see that for this example of a EMT930, there was something big going on at a few Hz, but the 25Hz drive component was there but not prominent. It does show up in the amplitude spectrum, but not so much in the FM.

Also inset in this is the conventional spectrum, which also only hints at a 25Hz component.

I don't know what cartridge was fitted, unfortunately, but the cart-arm resonance isn't prominent either in this case.

LD
 

Attachments

  • EMT 930 polar plot.JPG
    EMT 930 polar plot.JPG
    233.2 KB · Views: 927
Any other method other than test records ? Since offcenter and warps can not be avoided any modern method to measure speed stability ?
A Postscript printer can print precise dots (or markings) we can make perfectly center thick white sheet with such markings and optically measure the speed variations. Right ?
How do classic tables like Thorens, Garrard, Lencos and Technics compare with regards to speed stability. Any one has measured them and have data on them ?
Regards.
 
....if table is direct drive can we measure current required in motor in speed control circuits and ***** its speed stability ? Heavy platters with tight tolerance highly polished/lapped bearings will have comparatively stable speed isn't it ?
Regards.
 
Member
Joined 2014
Paid Member
Hiten: You certainly can use a non-record method to measure things. The question then becomes what you gain and what you lose from this. My Kenwood has a 300 slot optical code wheel on the spindle that is used to provide the correction signal to the electronics. So that would give a 540Hz square wave if you probed the opto connectors.

AEDB-9140-A13 | Broadcom Incremental Encoder -0.5 ? 7 V dc | Broadcom is an example of a 500 slot code wheel with detector that is fairly cheap. You'd need to work out how to fix it in place. If you want more resolution then HEDS-6140#B09 | Broadcom 1000 Pulse Code Wheel with a 0.5 in, Through Hole | Broadcom but the cost nearly doubles as you have to buy wheel and sensor seperately. You could also on tables with a strobe marking on the platter pickup off that.

I think all of these are work investigating and comparing to the test record method.

Ref your direct drive question, the answer is yes and no. Yes you can measure current and back EMF to do magical things in DSP, but the motor is turning too slowly to really make use of it. But this is an area I would love to have time to explore further (and would require another thread).
 
That's not even vaguely round, is the 930 idler driven?

I have the hifi news test record, badly centered and poor pitch on the 3150 cut. Acquiring a good din record of an optical disc as per regas dev unit is on the cards.
Yes the EMT 930 is idler drive. It's very well respected and was part of the EMT studio family that has serious heritage. I also have a test tone recording from an EMT938 in the archive that I can dig out..............

How round the polar plot appears depends upon the radial scale, and I set it to deliberately see the variations. I always use 3Hz as the spacing, because that is said to be the threshold of most people's ability to hear the difference. I try to keep the plots consistent, so there are 18 radial lines and the peak excursion clips the outside.

What we generally find, IME, is that TTs with excellent sonic reputation don't necessarily have good pitch stability when looked at this way - often far worse than those considered good or mediocre..... which is interesting and surprising at the same time perhaps?

Also, focus on accurate motor speed control is typically misplaced, IMO, because that's seldom a dominant factor in overall pitch stability, it seems.

LD
 
Any other method other than test records ?
I like the idea of displaying a servo signal where there is one, if nothing else, for a health check......!

But pitch stability is only partly about platter speed stability. Typically factors like spindle centering, platter vertical run-out, warp, and cart-arm stability contribute far more in an otherwise healthy TT. So this is where bang-for-the-money is, and is well within most DIYers capabilities methinks.

LD
 
Luckythedog,

your polar diagrams, translated into conventional xy graphs, represent the rotation angle by the x and frequency by the y axis, respectively. How did you record the angle/x axis, in order to get the x/y relationship in your diagrams?

@Hiten: Maybe you've frogotten Dual in your list of classic TT manufacturers?

Best regards!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.