Some turntable measurements
Goofball_star_dot_etal
http://www.wareing77.plus.com/Valle-TT-Polar.jpg
http://www.wareing77.plus.com/Linn-1k-1-Polar.jpg
The .wav files from which these are obtained are those used by
Stereophile, kindly supplied by JA .The data are shown with JA's and
Carl Valle's approval.
The scales are different for the two with a frequency span of 0.5% for
the Linn and 1.5% for Valle. This is because Carl's LP was not so well
centred and this accounts for much of the increased width of the
published spectrum. The average frequency is much closer to the ideal
1kHz for "Valle" (direct drive) than "Linn". More types of plots are
available, if anyone is interested.
Barry Mann
Are the Valle traces made without lifting the arm? It seems like
successive revolutions are not as well registered as with the Linn.
Did you use a high pass filter?
Have you done any spectral analysis? It seems almost like there are
side tones. Is the arm being excited?
The peak to peak variation seems lower on the Linn. Would you attribute
this to less bearing problems or a flatter record?
-----------------------------------------------------------
spam: u...@ftc.gov
wordgame:123(abc):<14 9 20 5 2 9 18 4 at 22 15 9 3 5 14 5 20 dot 3 15
13> (Barry Mann)
[sorry about the puzzle, spammers are ruining my mailbox]
-----------------------------------------------------------
Goofball_star_dot_etal
>In <3ffb302e...@usenet.plus.net>, on 01/06/04
> at 10:03 PM, w...@needs.email.anyhow (Goofball_star_dot_etal) said:
>
>>Here are some plots:
>>http://www.wareing77.plus.com/Valle-TT-Polar.jpg
>>http://www.wareing77.plus.com/Linn-1k-1-Polar.jpg
>>The .wav files from which these are obtained are those used by
>>Stereophile, kindly supplied by JA .The data are shown with JA's and
>>Carl Valle's approval.
>>The scales are different for the two with a frequency span of 0.5% for
>>the Linn and 1.5% for Valle. This is because Carl's LP was not so well
>>centred and this accounts for much of the increased width of the
>>published spectrum. The average frequency is much closer to the ideal
>>1kHz for "Valle" (direct drive) than "Linn". More types of plots are
>>available, if anyone is interested.
>
>Are the Valle traces made without lifting the arm?
The various (other) traces seem continuous so I very much doubt that
the arm was lifted.
>It seems like
>successive revolutions are not as well registered as with the Linn.
>
The Linn plot reminds me of a squashed fly perhaps it is something to
do woth squashed plastic that the "Valle" does not resolve. The Linn
has flutter mainly at 50Hz and 100Hz and a visible component at around
tone arm resonance. The direct drive machine "Valle" seem to have some
control system wobble when I expected to see distinct poles.
>Did you use a high pass filter?
>
It is low pass filtered with a brick wall, set in this case at 200Hz.
The displayed frequency is the sum of the local oscillator frequency
used ( which may be manually or automatically set) and the rate of
change of the phase difference between the signal frequency and the
oscillator. Since this involves differentiation of the sampled
detected signals it needs to be filtered to get rid of high frequency
noise.
>Have you done any spectral analysis?
Yes, see for yourself:
http://www.wareing77.plus.com/turntable-web.htm
> It seems almost like there are
>side tones. Is the arm being excited?
>
More in the case of "Linn", I think.
>The peak to peak variation seems lower on the Linn. Would you attribute
>this to less bearing problems or a flatter record?
>
The peak to peak variation is more on "Valle" but if it can be heard
at all I would guess the higher frequency error on "Linn" might be
more objectionable.
I am reluctant to interpret the results too much. People here can
probably do that better for thenselves.
Arny Krueger
news:3ffbba49...@usenet.plus.net
>
>> Have you done any spectral analysis?
>
> Yes, see for yourself:
> http://www.wareing77.plus.com/turntable-web.htm
>
Very worthwhile analysis.
Very praiseworthy.
Are you going to distribute the analysis tool?
Should shed some light on why turntables may sound different from each
other, all other things being equal.
Goofball_star_dot_etal
wrote:
>"Goofball_star_dot_etal" <w...@needs.email.anyhow> wrote in message
>news:3ffbba49...@usenet.plus.net
>>
>>> Have you done any spectral analysis?
>>
>> Yes, see for yourself:
>> http://www.wareing77.plus.com/turntable-web.htm
>>
>
>Very worthwhile analysis.
>
>Very praiseworthy.
>
Thanks!
>Are you going to distribute the analysis tool?
>
Not unless I clearly understand what the license says. . . Small
risk, no gain at all does not appeal to me.
>Should shed some light on why turntables may sound different from each
>other, all other things being equal.
>
>
I do wonder whether having a turntable that performs better than
squashed plastic is a significant gain but so far they have not been
quite at that level.
cwvalle
"Goofball_star_dot_etal" <w...@needs.email.anyhow> wrote in message
I want to thank you for taking the time to work on these plots.
I agree that turntable accuracy is at a certain level, a moot point.
It is interesting though I think.
Again thank you for your time and energy.
Carl
Goofball_star_dot_etal
wrote:
No problem. Get the grease gun out for that old tape recorder of
yours. Or the cotton wool buds if you are a poofy tweako-freako, note.
;-)
Thomas A
I made som plots of my Linn Axis II, Shure V15VxMR, and a very warped
Cardas record, 1 kHz.
http://hem.bredband.net/b113928/vinylspektr_1kHz.htm
Statistics are:
Minimum 997,53003
Maximum 1000,19
Sum 24970,45
Points 25
Mean 998,818
Median 998,98999
RMS 998,81849
Std Deviation 1,0057601
Variance 1,0115533
Std Error 0,20115201
Blackman window was used for. Size/resolution of FFT 32768. Speed used
to be too fast at 33 RPM, but adjusted it using the pots inside the
player. However, I did not bother to set it perfectly, since I would
need a perfect record. 50 Hz side bands can be observed, at -88 dB,
while 100 Hz are not there on the Axis. Top peak at -20.7 dB.
Other vinyl measure perhaps of interest that I've made:
http://hem.bredband.net/b113928/Vinylmeasure.htm
http://hem.bredband.net/b113928/Dustcover.htm
T
Goofball_star_dot_etal
wrote:
>w...@needs.email.anyhow (Goofball_star_dot_etal) wrote in message news:<3ffb302e...@usenet.plus.net>...
>> Here are some plots:
>> http://www.wareing77.plus.com/Valle-TT-Polar.jpg
>> http://www.wareing77.plus.com/Linn-1k-1-Polar.jpg
>> The .wav files from which these are obtained are those used by
>> Stereophile, kindly supplied by JA .The data are shown with JA's and
>> Carl Valle's approval.
>> The scales are different for the two with a frequency span of 0.5% for
>> the Linn and 1.5% for Valle. This is because Carl's LP was not so well
>> centred and this accounts for much of the increased width of the
>> published spectrum. The average frequency is much closer to the ideal
>> 1kHz for "Valle" (direct drive) than "Linn". More types of plots are
>> available, if anyone is interested.
>
>I made som plots of my Linn Axis II, Shure V15VxMR, and a very warped
>Cardas record, 1 kHz.
>
>http://hem.bredband.net/b113928/vinylspektr_1kHz.htm
>
It is reassuring to find a number of people getting similar results.
What is the difference between spectrum 1 and 2?
How did you manage to get frequency out? I would hate to have to count
cycles. .although a computer would not mind.
>Statistics are:
>
>Minimum 997,53003
>Maximum 1000,19
>Sum 24970,45
>Points 25
>Mean 998,818
>Median 998,98999
>RMS 998,81849
>Std Deviation 1,0057601
>Variance 1,0115533
>Std Error 0,20115201
Do you happen to know how the normal wow and flutter figure is
calculated? Something weighted RMS % I think.
>
>
>Blackman window was used for. Size/resolution of FFT 32768. Speed used
>to be too fast at 33 RPM, but adjusted it using the pots inside the
>player. However, I did not bother to set it perfectly, since I would
>need a perfect record. 50 Hz side bands can be observed, at -88 dB,
>while 100 Hz are not there on the Axis. Top peak at -20.7 dB.
>
>Other vinyl measure perhaps of interest that I've made:
>
>http://hem.bredband.net/b113928/Vinylmeasure.htm
>http://hem.bredband.net/b113928/Dustcover.htm
>
>T
The aliasing shound not be there by rights. I have seen that before
with undithered signals as the distortion is then generated after the
filtering.
If you want me to run the 1kHz file through my "software radio" just
ask.
Thomas A
> On 8 Jan 2004 17:54:52 -0800, Thomas_...@hotmail.com (Thomas A)
> wrote:
>
> >w...@needs.email.anyhow (Goofball_star_dot_etal) wrote in message news:<3ffb302e...@usenet.plus.net>...
> >> Here are some plots:
> >> http://www.wareing77.plus.com/Valle-TT-Polar.jpg
> >> http://www.wareing77.plus.com/Linn-1k-1-Polar.jpg
> >> The .wav files from which these are obtained are those used by
> >> Stereophile, kindly supplied by JA .The data are shown with JA's and
> >> Carl Valle's approval.
> >> The scales are different for the two with a frequency span of 0.5% for
> >> the Linn and 1.5% for Valle. This is because Carl's LP was not so well
> >> centred and this accounts for much of the increased width of the
> >> published spectrum. The average frequency is much closer to the ideal
> >> 1kHz for "Valle" (direct drive) than "Linn". More types of plots are
> >> available, if anyone is interested.
> >
> >I made som plots of my Linn Axis II, Shure V15VxMR, and a very warped
> >Cardas record, 1 kHz.
> >
> >http://hem.bredband.net/b113928/vinylspektr_1kHz.htm
> >
>
> It is reassuring to find a number of people getting similar results.
> What is the difference between spectrum 1 and 2?
It's just two arbitrarily chosen parts of the 1 kHz signal.
> How did you manage to get frequency out? I would hate to have to count
> cycles. .although a computer would not mind.
In the simple first graph I just measured the f at every 0.2 s using
the software and plotted them. Manually.
>
> >Statistics are:
> >
> >Minimum 997,53003
> >Maximum 1000,19
> >Sum 24970,45
> >Points 25
> >Mean 998,818
> >Median 998,98999
> >RMS 998,81849
> >Std Deviation 1,0057601
> >Variance 1,0115533
> >Std Error 0,20115201
>
> Do you happen to know how the normal wow and flutter figure is
> calculated? Something weighted RMS % I think.
Is it the CV (standard deviation/mean)? I have no idea.
>
> >
> >
> >Blackman window was used for. Size/resolution of FFT 32768. Speed used
> >to be too fast at 33 RPM, but adjusted it using the pots inside the
> >player. However, I did not bother to set it perfectly, since I would
> >need a perfect record. 50 Hz side bands can be observed, at -88 dB,
> >while 100 Hz are not there on the Axis. Top peak at -20.7 dB.
> >
> >Other vinyl measure perhaps of interest that I've made:
> >
> >http://hem.bredband.net/b113928/Vinylmeasure.htm
> >http://hem.bredband.net/b113928/Dustcover.htm
> >
> >T
> The aliasing shound not be there by rights. I have seen that before
> with undithered signals as the distortion is then generated after the
> filtering.
I should say that the recording is made by a Macintosh iBook and a
simple iMic A/D. It is quite noisy.
>
> If you want me to run the 1kHz file through my "software radio" just
> ask.
It would be interesting. The A/D + computer is however quite noisy in
the 50 Hz-600 Hz region (peaks) and the record is "warped".
The file can be downloaded (wav) at
http://hem.bredband.net/b113928/1_kHz_Axis_V15.wav
It is about 13 seconds, taken from the left channel and made into
mono.
Regards,
T
Goofball_star_dot_etal
wrote:
>The file can be downloaded (wav) at
>
>http://hem.bredband.net/b113928/1_kHz_Axis_V15.wav
>
>It is about 13 seconds, taken from the left channel and made into
>mono.
>
First two revs:
http://www.wareing77.plus.com/1_kHz_Axis_V15.zip
>Regards,
>
>T
Thomas A
Thanks for the analysis. I realized that the first file I gave you was
only around 5-6 seconds (the file became truncated due to overloaded
server). Now the file is 13 s, but I guess it does not matter for the
results.
I don't really see the 50 Hz sidebands in your analysis, which I see
more of using Hamming window, less with Blackman window.
Secondly, the warped record is quite obvious on the polar plot, and
that speed is a little slow. Perhaps I open the player one day and
adjust the speed (again).
Thanks again. Very interesting analysis!
(If you wish, you could keep the file for any type of publishing or
creating a library. I can also get other turntable data using the same
record. It would be nice to see a library of various TTs published
with these kinds of plots (perhaps JA is interested too?). A player
which adjusts bad centered records (Nakaminchi?) and warped records
with vacuum (Luxman, Basis) would be nice to see!)
T
Goofball_star_dot_etal
wrote:
>w...@needs.email.anyhow (Goofball_star_dot_etal) wrote in message news:<3ffffcd3....@usenet.plus.net>...
>> On 10 Jan 2004 03:29:33 -0800, Thomas_...@hotmail.com (Thomas A)
>> wrote:
>>
>> >The file can be downloaded (wav) at
>> >
>> >http://hem.bredband.net/b113928/1_kHz_Axis_V15.wav
>> >
>> >It is about 13 seconds, taken from the left channel and made into
>> >mono.
>> >
>>
>> First two revs:
>> http://www.wareing77.plus.com/1_kHz_Axis_V15.zip
>>
>>
>> >Regards,
>> >
>> >T
>
>Thanks for the analysis. I realized that the first file I gave you was
>only around 5-6 seconds (the file became truncated due to overloaded
>server). Now the file is 13 s, but I guess it does not matter for the
>results.
>
I longer file would give greater frequency resolution. I may try
again.
I did not mind too much. :-) My program is getting so bloated it slows
down dramatically as files get bigger
>I don't really see the 50 Hz sidebands in your analysis, which I see
>more of using Hamming window, less with Blackman window.
>
I have tested my program quite well with various things such as FM on
one channel and AM on the other, generating the waveforms from
formulae. Perhaps it just needs more resolution to see the 50Hz or
perhaps the 50 Hz may just not be there as FM or AM.
>Secondly, the warped record is quite obvious on the polar plot, and
>that speed is a little slow. Perhaps I open the player one day and
>adjust the speed (again).
>
Any idea of the source of the 5.5 Hz? I doubt whether you have ten or
so waves in your record.
>Thanks again. Very interesting analysis!
>
>(If you wish, you could keep the file for any type of publishing or
>creating a library. I can also get other turntable data using the same
>record. It would be nice to see a library of various TTs published
>with these kinds of plots
I assume that if the source file is made public that the owner does
not mind if the analysis is made public. If you want it confidential
E-mail it while I still have an address. I don't mind if anyone keeps
or distributes any of the plots that I have made public. As I don't
intend to keep them on the web, anyone who wants them should save them
and do what they like with them.
> (perhaps JA is interested too?).
He might be . . .
I am happy do do some more but I should modify the program soon. It
was made to work with stereo files but to run on mono I just parallel
the mono up which wastes a lot of speed.
> A player
>which adjusts bad centered records (Nakaminchi?) and warped records
>with vacuum (Luxman, Basis) would be nice to see!)
Yes. And a fused silica test record. .
>
>T
Goofball_star_dot_etal
operating the computer remotely)
There is some 50Hz FM but it does not seem particularly pure or
significant.
Thomas A
I should say that the resonance frequency of the arm (medium mass) and
cartridge (Shure V15VxMR) is rather low, around 6-7 Hz. I do have the
damping device on the cart but I have no clue how it efficient it is.
My guess is that the vertical and horisontal resonance is slightly
different due to the damping device, and that the warped record
excites these resonances to a certain extent. The two peaks could be
those resonances. I have however no problems playing records, even
quite warped ones. For highest quality i would probably change
tone-arm to a light one, but the critical errors are still much higher
in vinyl reproduction than this.
>
> >Thanks again. Very interesting analysis!
> >
> >(If you wish, you could keep the file for any type of publishing or
> >creating a library. I can also get other turntable data using the same
> >record. It would be nice to see a library of various TTs published
> >with these kinds of plots
>
> I assume that if the source file is made public that the owner does
> not mind if the analysis is made public. If you want it confidential
> E-mail it while I still have an address.
Ok. No problem. Information is good if it can be distributed to
others.
> I don't mind if anyone keeps
> or distributes any of the plots that I have made public. As I don't
> intend to keep them on the web, anyone who wants them should save them
> and do what they like with them.
Ok. So is it ok if your plots are presented on another web page (like
mine)? It would be sad if they would disappear from web space. :( (It
would be nice to have identical scales on the polar plots, since
different ones can fool the eye.)
>
>
> > (perhaps JA is interested too?).
>
> He might be . . .
>
> I am happy do do some more but I should modify the program soon. It
> was made to work with stereo files but to run on mono I just parallel
> the mono up which wastes a lot of speed.
So what was the software intended for originally?
>
> > A player
> >which adjusts bad centered records (Nakaminchi?) and warped records
> >with vacuum (Luxman, Basis) would be nice to see!)
>
> Yes. And a fused silica test record. .
Or the laser player....:)
Thanks again for you kind help!
>
> >
> >T
Goofball_star_dot_etal
wrote:
Snips in progress.
>w...@needs.email.anyhow (Goofball_star_dot_etal) wrote in message news:<40004551....@usenet.plus.net>...
>> Any idea of the source of the 5.5 Hz? I doubt whether you have ten or
>> so waves in your record.
>
>I should say that the resonance frequency of the arm (medium mass) and
>cartridge (Shure V15VxMR) is rather low, around 6-7 Hz. I do have the
>damping device on the cart but I have no clue how it efficient it is.
>My guess is that the vertical and horisontal resonance is slightly
>different due to the damping device, and that the warped record
>excites these resonances to a certain extent. The two peaks could be
>those resonances. I have however no problems playing records, even
>quite warped ones. For highest quality i would probably change
>tone-arm to a light one, but the critical errors are still much higher
>in vinyl reproduction than this.
To genereate FM there must be a component of movement along the
groove. I'm not sure what generates AM but often they seem to be quite
different, which is a bit of a puzzle since changing the speed should
change the output amplitude. Something to do with equalization, I
guess, must reduce the AM due to speed.
>>
>> >Thanks again. Very interesting analysis!
>> >
>> >(If you wish, you could keep the file for any type of publishing or
>> >creating a library. I can also get other turntable data using the same
>> >record. It would be nice to see a library of various TTs published
>> >with these kinds of plots
>>
>> I assume that if the source file is made public that the owner does
>> not mind if the analysis is made public. If you want it confidential
>> E-mail it while I still have an address.
>
>Ok. No problem. Information is good if it can be distributed to
>others.
>
>> I don't mind if anyone keeps
>> or distributes any of the plots that I have made public. As I don't
>> intend to keep them on the web, anyone who wants them should save them
>> and do what they like with them.
>
>Ok. So is it ok if your plots are presented on another web page (like
>mine)?
Yes.
If you need to drag my name up it is "Goofball"
> It would be sad if they would disappear from web space. :( (It
>would be nice to have identical scales on the polar plots, since
>different ones can fool the eye.)
Yes, it is a problem. I think the span of your polar plot is the same
as "Valle" but of course some turntables need a bigger span than
others and are off in frequency. If you gave them all +/- 10% you
would not see much. Some scales may be set to "Autoscale" either by
choice or by mistake. If anyone wants to come up with a "Standard" for
the future that is OK.
>>
>>
>> > (perhaps JA is interested too?).
>>
>> He might be . . .
>>
>> I am happy do do some more but I should modify the program soon. It
>> was made to work with stereo files but to run on mono I just parallel
>> the mono up which wastes a lot of speed.
>
>So what was the software intended for originally?
>
I fancied making a radio and this came up:
Re: Appeal of Vinyl (any comment on my speed stability test?)
http://www.google.com/groups?safe=images&ie=UTF-8&oe=UTF-8&as_ugroup=rec.audio.opinion&as_umsgid=%3C6xE0a.848%24uP7.6...@newssvr15.news.prodigy.com%3E&lr=&num=100&hl=en
I don't use vinyl much myself and have no test record but once upon a
time I thought I might make a turntable and the thought has occured
again. Now that I think I might have the world's best wow and flutter
meter it might be fun to make the world's best turntable, just for
fun. It won't be flash or expensive and I am quite good with a lathe,
note.
>
>Thanks again for you kind help!
>>
John's dad say: "Meter on its own useful like bra with three cups."
Goofball_star_dot_etal
Now overwritten with six rev. data.
Thomas A
Thanks.
It is obvious that a flat record would improve the results. I am
working on a vaccum solution on this.
The 4 Hz and 5.5 Hz peaks; I have no idea what they are (about half
the tonearm/cart resonance, I checked it). The remaining peaks are
quite low in level, including the 50 Hz.
There is a damper on the Shure cartridge which affects the fundamental
resonance. Perhaps this one has something to do with the 4 and 5.5 Hz
peaks?
T
Thomas A
Addendum:
From Klaus excellent summaries of publications of tonearm and
cartridge resonances at the Vinyl Asylum it can be read that there is
a statistical maximum of warps around 6 Hz (Happ and Karlov). The
distribution of vibrations are:
record eccentricities : 0.55 Hz
warp : 0.55 - 10 Hz
The tonearm/cartrdige resonance in my case is 7.13 Hz with an addition
of 2 Hz due to the damper (found by Nakai), thus 9.13 Hz.
My belief is that the warped record induces these resonances at 4.1,
4.4 and 5.5 Hz (of which 4.4 and 5.5 is the 7th and 9th overtone of
the record eccentricity and very close to the statistical maximum of
the warps)
A perfectly flat record would probably reduce these resonances
significantly.
T
Goofball_star_dot_etal
wrote:
It makes sense. This problem would be less if the arm was pivotted in
the plane of the record. It is probably not practical do the same for
the stylus cantelever.. . Unless it has a wishbone type suspension
with a virtual pivot. . .
You have another test record?
Thomas A
Yes,
but it is the Hifi & News test record with a 1 kHz signal and a
sweeping 20 Hz-5 Hz tone to localise the resonance. So it would not be
good to use in this particular test.
I don't know how long you can wait, but I may be ready with the vaccum
method to flatten the record in a couple of weeks. This will be the
ultimate test since it would make use of the same record and player
with or without warps.
T
T
Goofball_star_dot_etal
wrote:
>w...@needs.email.anyhow (Goofball_star_dot_etal) wrote in message news:<400282f3...@usenet.plus.net>...
>> On 12 Jan 2004 02:50:21 -0800, Thomas_...@hotmail.com (Thomas A)
>> wrote:
>>
>>
>> You have another test record?
>
>Yes,
>
>but it is the Hifi & News test record
Is that the one with a 300Hz tone I used for "Moi" and "Nick"? If so
300Hz is OK and is known to be quite accurate by the "Stanley Knife"
test. I would not mess with your speed until you know your sound card
and test tone is accurate enough for the precision you want.
>with a 1 kHz signal and a
>sweeping 20 Hz-5 Hz tone to localise the resonance. So it would not be
>good to use in this particular test.
>
>I don't know how long you can wait, but I may be ready with the vaccum
>method to flatten the record in a couple of weeks. This will be the
>ultimate test since it would make use of the same record and player
>with or without warps.
>
>T
>
We can do that as well. We only need to look at the FMFFT.jpg to check
that the problem was warp.
Thomas A
I have now experimentet with a DIY vacuum method to flatten out the
record. I did not manage to flatten the record out completely,
although it should be flatter than previously. During the experiments
I had an accident with the test record, so it could be a faint
repeated "tick sound" in the file. Looking at the FFT I am almost
certain that the 6/8 Hz peaks are still there, but the peak look a
little more symmetrical. It would be nice if I could get the same
(complete) analysis as before, including the polar plot, if you have
time to do it. Another difference is that I used the stereo track
before the "left channel track" since the left channel track was even
more damaged by the accident. The left channel was taken and made into
mono.
Here is the file:
http://hem.bredband.net/b113928/1kHz_flat_mono.wav
T
Goofball_star_dot_etal
wrote:
>I have now experimentet with a DIY vacuum method to flatten out the
>record. I did not manage to flatten the record out completely,
>although it should be flatter than previously. During the experiments
>I had an accident with the test record, so it could be a faint
>repeated "tick sound" in the file. Looking at the FFT I am almost
>certain that the 6/8 Hz peaks are still there, but the peak look a
>little more symmetrical. It would be nice if I could get the same
>(complete) analysis as before, including the polar plot, if you have
>time to do it. Another difference is that I used the stereo track
>before the "left channel track" since the left channel track was even
>more damaged by the accident. The left channel was taken and made into
>mono.
>
>Here is the file:
>
>http://hem.bredband.net/b113928/1kHz_flat_mono.wav
>
>T
No obvious improvement. I will put the 'before and after' mono files
into a stereo file and try to get the results out soon.
You speeded the turntable up a bit. I think you need to come down.
Latest results seem to indicate that the test record (JA) used is
about 998.5 Hz but I don't know what the error bars are, perhaps as
much as +/- 1 Hz. I looked at a blip on the polar frequency plot and
adjusted the "reference frequency" (which affects the angular scale)
until the blip was stationary from rev to rev. Using an
autocorrelation function on the "FM v T" gave the same result but it
is not very "sharp". I might get something from your scratch. :-)
Thomas A
I also noticed an increase in speed this time, however, I did nothing
to change speed in the player. I checked the strob and it looked ok.
The only thing I did was to change track (which is the track just
before the track used previously) and applied a vaccum method to
secure the record to the platter. Could the vaccum method have changed
the speed? Previously I just used the felt mat on top of the platter,
which itself is not secured to the platter. Stylus drag???
T
Goofball_star_dot_etal
wrote:
>
>I also noticed an increase in speed this time, however, I did nothing
>to change speed in the player. I checked the strob and it looked ok.
>
>The only thing I did was to change track (which is the track just
>before the track used previously)
Ho hum, I think the turntable speed is about the same, perhaps the
tracks *are* different.
> and applied a vaccum method to
>secure the record to the platter. Could the vaccum method have changed
>the speed? Previously I just used the felt mat on top of the platter,
>which itself is not secured to the platter. Stylus drag???
>
>T
http://www.wareing77.plus.com/1_kHz_Axis_V15c.zip
Some notes:
Both the old and new "flat" tracks are analysed at the same time.
The "flat LP" track is the left channel coloured green.
The old file was amplified by 7.4 dB so that the two tracks would be
about the same amplitude.
The "flat" file had about 1/2 rev's worth lopped of the start so that
the FM would be roughly in phase.
Page 1: The detected frequency and THD are wrong because at 1000Hz the
modulation index is such that the carrier is down compared to trhe
sidebands and at this resolution the wrong peak is selected, I think.
Page3: The reference frequency (LO) was chosen manually so that the
phase error over 5 revs is minimised on the "flat" track. You can see
the phase error of the other track increasing in magnitude. The
average frequency can best be calculated by adding the (change in
phase over 5revs) / (5* 2PI*1.8) cycles/second to the reference
frequency. (subtract this number in the case of the old ( right )
channel since the slope is negative)
Page8: Notice that the peak near 4.1Hz is not a harmonic of the
rotational speed. Also most of the peaks above that match nearly
perfectly when I zoom in on them (not shown)
Notice that the peaks below 4Hz have changed and are harmonics of the
rotational speed and that the AM and FM seem to behave quite
independently.
Thomas A
Thanks,
I guess another test record is the best way to check the results.
Having 2001.2 Hz and 998.7 Hz from the tracks just next to each other
may indicate a poor record.
T