advertisement


Turntable speed analysis part II

I'm still interested, but having moved back to the UK I'm not set back up yet to take measurements or run numbers. Hopefully I will be soon.
 
I completely missed your earlier post. The software has been developed for other reasons and is now better. So tomorrow afternoon I'll see what transpires.

(I've been using it to demodulate jitter from tones captured from digital sources. Quite interesting.)

Paul
 
At first glance the Teflon washer looks like a big improvement, and the foil washer a small backward step.

The pictures aren't right though, I've got to check out some calibration stuff that I've obviously messed up.

Paul
 
Hi Paul, have you had a chance to work on the plots?
Im really curious how the simple teflon washer trick worked.
Thanks,
phi
 
I used your latest files as a reference phix to test my current turntable.

Here was your standard:
phixstandard_zps3d8228e0.jpg


Teflon washer no foil:
phix-teflon-nof_zpsd124433b.jpg


Teflon washer with foil:
phix-teflon-f_zps2223d590.jpg


And finally my Bix, which uses a premotec dc motor with a basic speed controller:
bix_zps042d1023.jpg


I need to go back through some earlier files to make sure its working, spectrum lab has changed the fm demodulator behavior.
 
Isn't the Premotec is usually an AC Motor ?

That last result looks quite remarkable, what test record are you using?
 
sq225917 said:
Isn't the Premotec is usually an AC Motor ?

That last result looks quite remarkable, what test record are you using?
Click to expand...

The picture seems to be having issues on photobucket, I will fix it when I have a chance.

It is a small brushed DC motor, when I get the chance I will get the part number off it. The test record is a hifi sound test record, HFS 74. Which is 3000hz not 3150, but I adjusted for that in the demodulation and plot. The result was so smooth when I first went through the process I assumed something was wrong, which is what led me to plot phix's results using the same scale.

I hope to verify them better. When I get the chance I will upload the file and Paul can plot it.
 
Edd, thank you!

Wow, what a difference! The idler must have been working against the notorious
fiber washer (the one with the pin), producing the fast tik-tik-tik sound I could hear with
a more advanced arm-cart than the stock EMT.

Edd, do you also have the demodulated spectra? Would be interesting to see how much is the 2.72Hz peak attenauted by the teflon washer.

BTW, the difference between my files 1) and 2) lies not in any foil--both 1) and 2) were taken with the same teflon foil washer in place. The only difference is a nearby fridge beeing switched on (track 1) and off (2)...:)) The deck was out of its console so it was sensitive to all vibrations and I did not remember if when taking the standard 3) the fridge was on or off :)...but the difference is so obvious that it doesn't matter.

Your file looks very smooth=absence of flutter? But the shape is more oval, could it be the test LP eccentricity?

Good job guys and thanks lencoheaven Dimitri for the tip!

PS Edd, I presume your scale is different than Paul's as my standard polar plot did not look that ugly (red trace):

http://www.pinkfishmedia.net/forum/showpost.php?p=1788378&postcount=341

or something else has happened?
 
I will have a look at trying to post the spectrum. That is something I have not managed as well as Paul.

The scale wont be the same. I use spectrum lab to demodulate the signal. Demodulation makes a waveform from the deviation from the frequency. If you set the centre to 3150, as the speed goes up, say to 3160 that gives a positive peak, as it goes under you get a negative peak. How big those resulting peaks are in the plot depends on the gain of the demodulator in spectrum lab, and of course the scale in gnuplot. I just try to be consistent, this isn't about absolute numbers, its about better or worse.

On the bix, I think its at the point where I'm not sure, I could be eccentricity, It could be the recording on the lp itself, it could be minor speed variations from the bearing or belt. I haven't had it set up long.
 
edd9000 said:
The scale wont be the same. I use spectrum lab to demodulate the signal. Demodulation makes a waveform from the deviation from the frequency. If you set the centre to 3150, as the speed goes up, say to 3160 that gives a positive peak, as it goes under you get a negative peak. How big those resulting peaks are in the plot depends on the gain of the demodulator in spectrum lab, and of course the scale in gnuplot. I just try to be consistent, this isn't about absolute numbers, its about better or worse.
Click to expand...

This is how I understand it too. Your scale seems to be more revealing
and less forgiving :)) Would be nice to compare it to Paul's plots.

edd9000 said:
On the bix, I think its at the point where I'm not sure, I could be eccentricity, It could be the recording on the lp itself, it could be minor speed variations from the bearing or belt. I haven't had it set up long.
Click to expand...

The spectrum plot should give a clue. When you attempt it, please try to make it in two versions: a larger picture up to, say, 100Hz and a more specific one up to 25Hz. I use a tube phono, so my flutter could be partly due to tubes' thermal noise.

BTW, is this the soft you're using:

http://www.qsl.net/dl4yhf/spectra1.html ?
 
Ed the eccentricity looks to be at multiples of platter revolution . I'd say its record eccentricity rather then related to deck mechanics.
 
Ok here are the spectrums, the main difference with Pauls are mine are a log scale on the verticle which leads to less specific peaks.

0-10hz Standard
phix-standard0-10_zpse45755da.png


0-10hz Teflon
phix-teflon0-10_zpsf7c8a4c1.png


0-50hz Standard
phix-standard0-50_zps8ccf4d7a.png


0-50hz Teflon
phix-teflon0-50_zpsbdb5e3ca.png


You can see the huge drop in the the idler frquency of 5.55hz.
 
And my bix, bare in mind the verticle scales, I had to adjust the verticle scale to fit the plots on.

0-10hz
box0-10_zpsa4dda253.png


0-50hz
bix-0-50_zps09879585.png


All of these were done on the raw unfiltered output from spectrum lab. I only trimmed the start and end of the files. On the teflon samples from phix, they were at 22100 so had to be resampled to 44100, which should have little effect.
 
sq225917 said:
Ed the eccentricity looks to be at multiples of platter revolution . I'd say its record eccentricity rather then related to deck mechanics.
Click to expand...

While its playing I cant see the arm wandering back and forth. But it could easily be the pressing or the cutting, it is old, and could be warped. But for £7 on ebay I figured a test record was better than no record.
 
Thanks Ed! I'm quite puzzled.
This 5.5Hz peak (2nd harm. of the idler fundamental) is a novum at that magnitude (higher than 2.72). The washer did remove it, leaving 2.72Hz untouched--this is probably idler surface irregularity (a bump?). It also removed some higher order components,leaving clearly the first two motor related peaks: fundamental at 12.5Hz and it's 2nd harm...but I'm a bit perplexed.

The standard picture is uglier than it used to be. There are 2 factors: the deck has been mover recently but with all the care: platter aside, idle protected etc; the deck was nacked, no console, resting on it's frame. I don't rememeber how it was last time (post #334) but I'm 99% sure it was in the console. Could it be that the console damps some of the idler related vibration (it doesn't reach the arm via the chassis)? I hope to make your soft procedures work and will check it myself (SoX won't run ok under my win7 :-((
 
You must log in or register to reply here.


advertisement


Back
Top