Measuring for rumble

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
How many times have you made a new plinth, or "upgraded" your turntable with parts and components that promise to improve the sonic qualities of your turntable? What if there is a simple method to determine by repeatable methods that can put a number of that performance?

I suspect that there is and that it may be very simple to do this in the relative comfort of your own listening environment.

In this case I am measuring for rumble on my Thorens TD124. SN 2729.
The TD124 has been refurbished by yours truly some years previously and listening sessions told me that I had greatly improved the player in comparison to what it had been when I initially bought it. Now, however it seems appropriate to put a number on that performance.

Rumble is an important figure because it largely determines how much detail the player will be able to reproduce from the record grooves without being masked by the mechanical noises of the player. The quieter it is mechanically, the more pure clean music you hear.

With this in mind I considered a simple way to come up with a valid measurement. Here below is what I think works.

DSC_0188.jpg


Above photo: Tracing the "blank" between tracks of HFN001 test record.
It seems important to choose a "blank" space that does not have a lead-in groove that will guide the stylus between tracks; like on a regular record. That is why I chose this test record. by recording this blank I can be sure not to be measuring any rumble that may have been imparted onto the test record by the cutting lathe that mastered this record.

The recorder:
DSC_0187.jpg

I used my 2nd hand Masterlink 9600 to record a 24/96 file of the "blank" space. And I had it burn a CDR of this file.

2729rumble2.jpg


Taking the CDR over to my PC, I used Audacity to analyze this sound file.
2729rumble.jpg

2729rumble3.jpg

And it was the Spectrum Plot that gave me the above images. Frequency and Amplitude of the noise. The number (in negative dB) shown is closely in agreement with published rumble figures for the TD124. Probably closer to what a TD124 mkII will do, thanks to its improved motor mounting scheme. the algorithm shown was Spectrum. Whatever that means.:D

Disclosure time:
DSC_0186.jpg

I used a Minus-K platform under the Thorens for this test. But it is how I listen to the player all the time. It is quite possible that the Minus-K will improve the noise floor of any turntable standing on it. I suppose in order to measure the contribution of the Minus-K into this equation, I would need to repeat the test but with the Thorens standing on something else. I won't be doing that today. But I imagine it is a factor. How much is unknown. but my listening sessions tell me that it is a small but identifiable difference.

Any comments on this methodology? Is it really this simple? for the moment, I think that it is.

-Steve
 
Last edited:
The first thought is gee, is the record blank truly flat? Just looking at the top photo I see that images in the blank regions of the disk do not reflect background as a truly flat surface would.

My first search for "turntable rumble measurement" brought me to this link, which on page four brings up very same topic of "subsonic surface irregularities".

They go on to mention that these are typically much greater signal source than motor and bearings, and that they excite resonances of tonearm system.

So, perhaps instead of record, optically flat surface is needed. Apparently when you are riding the grooveless blank region the anti skating mechanism is set such that stylus remains in smooth grooveless region.

Rumble in real recordings is dependent on smoothness of master cut by lathe, smoothness of lathe travel, and then further by smoothness achievable in pressed records.

In your frequency analysis picture "size 1024" refers to number of samples used for FFT used in transforming time domain into frequency domain. This only has marginal resolution for the frequencies of interest.

You should compare your result with using 1kHz test tone.

Otherwise, yes, your test methodology is apparently very similar to that used in the industry, at least back in the 70's.
 
Ohhh I remember the thread regarding that antivibration device.
1) it's based on a series of leverages and weights, 100% it wasn't designed
for audio purposes
2) If it was designed for that, it's also a really bad implementation (100 %)
It would need an anti-seismic base, and not a wooden stand.
3) Not to talk about the machinery placed on it : whazzat? Graphite ? and the Thorens is...plastic ? Viewing the whole thing with X-ray ( eyes ) it's just a bunch of moving metal parts soaked in some moisty fluid.
4) so the whole column is totally wrong; it's called 'miracles on vinyl' and I'm with you on that side. It's a very difficult matter and needs to be studied very
scientifically

The minus-K... K is the symbol of elasticity coefficient or something.
 
Last edited:
Picow
(1) What? It was designed for anything placed on top to reduce movement to the lowest frequencys possible. That includes lasers, optical, turntables, etc.
Herzan is another active device that attempts the same results rather than passive here, seems like a serious attempt IMO
(2) No
(3) Yes
(4) ?

Regards
David
 
Hi user510. Thanks for pushing the measurement side of things. All too often the proof of concept is "it sounded better". I do think the record and arm contribute to your total reading but for changes you make to your own setup you have found a way to view changes. very cool. Might have been nice to have this kind of view before you modded your table just to see what actually changed. Could Audacity (or something similar) not have been used to capture the data instead of the Masterlink 9600? Not too many people have one of those :)
 
Your recorder must have a flat response down to 6 or 7Hz before you can make accurate measurements.


Hi pass filtering of phono preamp and rest of recording chain are easily assessed. A reasonable correction curve is made, and noise gain is kept in mind.

tabarddn;

In your spectrum picture is 50Hz signal part of test track or due to 50Hz mains voltage?

How is rumble signal referenced to signals in groove?
 
4) so the whole column is totally wrong; it's called 'miracles on vinyl' and I'm with you on that side. It's a very difficult matter and needs to be studied very
scientifically

Each piece has its modes of vibration, and they get altered and mixed ( add resonances ) when joining; then there are the forces and the masses; so each piece needs to be modelized - the tonearm itself has different forces involved in the process, that don't match with the requisites of the platter; at least, a uniform material ( e.g. aluminum) that has the same response should be used, so the reactions are more predictable when modelizing.
Mixing & matching of different forms-functions- materials leads to a random - more difficult & pain - result
 
I didn't see where anybody measured a known signal so the measurements can be referenced to anything. Audacity tells you dB in terms of full signal, but you need to tie that to some known displacement or the number is meaningless. (though I could have missed it!) Don't forget to apply weighting if comparing to commercial weighted numbers. (one of the best sites is http://www.diracdelta.co.uk/science/source/a/w/aweighting/source.html)

I've looked at the Thorens rumble coupler measurement device and in all honesty I don't understand how the thing can exceed the performance of a well designed and fabricated bearing. IMO, making a near-perfect bearing isn't that difficult. Eliminating motor coupling is a whole 'nuther story, and that signal should probably be described as something other than rumble. A lot of modern motors generate stuff in the hundreds of Hz.
 
Last edited:
thanks for all of the replies...

To respond to some:
1) Method: Yes, this process records the subsonic signals generated by the less than perfectly smooth uncut vinyl. But keeping in mind that under the best of conditions, this vinyl can be as flat and will tend to mirror the surface roughness of the stamper die that pressed it. With that considered the HFN001 copy I have appears to be a very flat record,..... as far as records go.

2) Right. This method can not be compared directly to any published "weighted" values that turntable manufacturers used to reference back in the day when rumble measurements were actually taken. As I see it, its' only repeatable reference will be within the scope of my own listening room while using the same group of equipment. Example, using the same test conditions while recording from a different turntable. then comparing the spectrum plot generated between them. Or, better yet, same turntable after refurbishment has been carried out.

I have to admit that I thought I'd see a "noisier" decible figure from this TD124. something around -40 db weighted would have been appropriate. And my spectrum plot is, I think, unweighted raw data. Even that is useful provided the same conditions of the test are adhered to from one player to the next.

3) The Masterlink. It would be possible to feed the line level signal coming out of the phono preamp directly into the sound card of a suitable PC. Thus eliminating the need for the ML9600. However, not all PC sound cards are created equal. Some are more agreeable with audiophile pursuits than are others. In my case, I just have standard issue sound cards on the computers in my care. The Masterlink, on the other hand, was designed to record high quality sound. So, on that general assumption, I presume it to be as good a bet for getting the "unpolluted data" into Audacity as any other available to me. also, this particular model is long out of production. It was expensive when new. but now, 2nd hand samples may be found and purchased for around a few hundred usd. And that is how I got mine.

4) known signal. The test records I have are rather limited. Two of these offer tracks with unmodulated grooves. The HFN001 record calls this track system residual noise. An older test record I have, Allied Radio, calls their unmodulated groove track a rumble test. but when you look at that track in good lighting it can be seen that there is some very slight modulation within those grooves. Rumble from the mastering lathe. But the HFN001 record does not show any visible evidence of modulation within its residual noise track.

It has been suggested that I use this method to analyze a 1000hz signal. Unfortunately none of the test records I have offer this. Although the HFN001 does offer a 300 hz signal at various different amplitudes. I suspect that a steady tone signal could be used to examine more than the mechanical noise of the turntable. But it could be used to determine performance in the domain of time. Like wow and flutter. So it does seem like a good idea to get a test record that does offer a 1000hz track. Fwiw, the Allied Radio test record (actually 3 records) does offer some wow/flutter tracks. These consist of a single piano note being sustained. This is probably useful for computer analysis. I'm just not certain what software would be the best choice for this test. (wow/flutter)

re: some references available on the web.
http://www.theanalogdept.com/images/spp6_pics/.../MechanicalResonances.pdf
this is the Bruel and Kjaer document that is widely available on the web. Afaik, this company still carries out audio testing and offers its equipment to do so. But none of this is available to me at a price that would make sense.

-Steve
 
Hi Steve

The following measurements are done with the Thorens messkoppler on a special antivibration table .
All the ,,rumble ,, is between the 0,55 -100 Hz as you can see the motor has two vibration components its own resonance from 23 hz at 50 hz power frequence and the magnetic vibration at 100hz .
Others are from the tonearm, steppulley, idler and its harmonic components.
The first picture is with the normal rubber motor decoupling 100Hz at -65db the second with spring decoupling 100 hz at - 72db . same results at 23 hz motor frequence.

Volken
 

Attachments

  • IMG_7754.jpg
    IMG_7754.jpg
    426.2 KB · Views: 1,170
  • Rumble Thorens TD124.2 Messkoppler rubber decoupling Meting_0000.bmp
    104.1 KB · Views: 208
  • Rumble Thorens TD124.2 Messkoppler Spring motor decoupling .bmp
    104.1 KB · Views: 130
375375d1381176991-measuring-rumble-img_7754.jpg


Good image. and while we have it why not discuss the merits of the rumplemesskoppler. The rumble coupler was designed by Thorens engineers back when it became clear that record players were capable of producing less rumble than many of the mastering lathes still being used to produce records. And so a test record with an un-modulated track was no longer of any use.

But now some have suggested that this device, the "rumplemesskoppler", by itself, does produce some bearing noise. So the question becomes; is the reading taken 'a measurement of the record player' or of the instrument itself? My guess would be that it depends upon the record player being measured. If its' rumble levels are greater than those of the rumplemesskoppler, then those readings should show up on the graph but in addition to the noise produced by the instrument.

Then compare readings taken using this instrument versus the readings taken using a blank flat bit of vinyl while the stylus skates across it. I see two very different reads. And it seems that the frequencies recorded by myself using the later method did generate some very low frequency sound waves. down to below 20 hz, while the rumplemesskoppler did not go that low. But with the Thorens coupler instrument we see noise frequencies consistent with motor vibrations and spinning pulleys and driving idler wheels, etc.

Anyway, just brainstorming. I probably missed an important point or two.

Fwiw could you let us know more about your methodology? What is the signal chain in use to acquire the two charts you show?

Interesting stuff. And that is a great photo of a Rumplemesskoppler in use.

-Steve
 
[
Fwiw could you let us know more about your methodology? What is the signal chain in use to acquire the two charts you show?

Interesting stuff. And that is a great photo of a Rumplemesskoppler in use.

-Steve[/QUOTE]


Steve the messkoppler is connected to a B@K Response test unit and after that to the HP 35665A Dynamic Signal Analyzer .
I did also measurements with a Denon test record posting follows.

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