Turntable Wow and Flutter "Polar" Plot Script

[n3mmr]

The nominal f being bang-on is rather trusting as well.

Exactly, but the nominal frequency will tell you what the effect of a deviation will be on the rev time.

But if the nominal freq is X Hz, and the real frequency measured is Y Hz, then the one-rev time for 33⅓rpm nominal is (X/Y)*(60/33⅓) seconds

 

[JP]

My point is that it doesn't help if the f that's on the test record isn't what it's advertised to be, which is often the case. Within reason this is never an issue unless someone what's to plot a very high number of revolutions which seems a bit pointless.

 

[n3mmr]

My point is that it doesn't help if the f that's on the test record isn't what it's advertised to be, which is often the case. Within reason this is never an issue unless someone what's to plot a very high number of revolutions which seems a bit pointless.

Is it really that common for the nominal f, if stated as 3150, not to be 3150 but some significantly different f??
How large a deviation should one expect?

P. S. It might not be pointless to plot two revolutions from widely separated times.

 

[morillon]

the good old stroboscope, even a 300hz one, is good for adjusting the speed when it is possible ;-)

 

[n3mmr]

the good old stroboscope, even a 300hz one, is good for adjusting the speed when it is possible ;-)

I've tested the Stroboscope Engineer android app.
Seems to work ok, and seems to indicate my Dual Golden 1 runs true.
It has a "calibrate" facility that checks for "latency".

 

[n3mmr]

I've tested the Stroboscope Engineer android app.
Seems to work ok, and seems to indicate my Dual Golden 1 runs true.
It has a "calibrate" facility that checks for "latency".

The Strobily app seems even better. And indicates my TT runs at a very stable and accurate 33⅓ RPM

 

[JP]

Is it really that common for the nominal f, if stated as 3150, not to be 3150 but some significantly different f??
How large a deviation should one expect?

P. S. It might not be pointless to plot two revolutions from widely separated times.

As the software doesn't support doing elapsed-time groups, you'd be looking at a manual process to try to align those plots regardless.

My best test record is 4Hz off, and the speed accuracy of my 'table is quite high being PLL with a modern clock source. You can't infer both, so you need to assume one is correct. The is just shifting the error from one cause to another.

 

[n3mmr]

As the software doesn't support doing elapsed-time groups, you'd be looking at a manual process to try to align those plots regardless.

My best test record is 4Hz off, and the speed accuracy of my 'table is quite high being PLL with a modern clock source. You can't infer both, so you need to assume one is correct. The is just shifting the error from one cause to another.

Just musing: 4 Hz is about 1.3 ‰ at 3khz, which should be about the same as a drift of 3-4 strobe lines per minute, at 50hz.

 

[mfalcon]

I've been realigning my cartridge and playing with the Clearaudio test records this morning. I've run 3 test records now and keep getting some bad W&F plots. My only other way to really test is the KAB strobe and the built in strobe to the table. Those look fine.
So my question is should I just chalk this up to three bad 3150 W&F test tracks or is something wrong. The Technics 1200 G should have great W&F and I'm driving myself crazy thinking I'm measuring wrong. Thanks

 

[JP]

Center the record and do 2-4 revolutions. Doing that many just obscures what can be seen.

 

[mfalcon]

I’ve got to reem it out the center. It’s has 0 play in it.

 

[SeriousSeri]

I am playing around with this Script to classify Test Records, i love the idea!

I am getting spikes though which render the plot useless. I can sometimes find a "sweet spot" (by offsetting and limiting revs for the measurement) with no spikes, but on some records there are just too many.
I will try to implement a filter that will mark, but cut-off those, so it does not affect your plot to the point you can't read it anymore. Or fix the frequency axis, that may be better...
Example attached. Keep up the good work <3

 

[JP]

I’d suggest finding the root cause. The method has rather good noise immunity, and in the hundreds of measurements I’ve done and thousands including others, I’ve not seen problems like that in normal use.

 

[Balle Clorin]

I see those strange spikes too, I wonder if it is pops or too high wow or drift that cause it

 

[Balle Clorin]

I’ve got to reem it out the center. It’s has 0 play in it.

I use NAK T-100 software to help me center record before making a polar plot. Easy to see immediate effect of different centering

 

[morillon]

wfgui , for windows can help too

 

[Balle Clorin]

Too jumpy for me…

 

[mfalcon]

I use NAK T-100 software to help me center record before making a polar plot. Easy to see immediate effect of different centering

What is your process for this? Trial & Error? I messed around trying to center a test record last weekend without much success.

 

[Balle Clorin]

Trial and error yes. I start with the label/text pointing from 9->3 o clock. I the push the record slightly towards center in one direction say from 12 o clock, and measure, then in the opposite direction and compare to see what is best. Then I do the same for 3 and 9, then I have a clue where the record should be pushed, and can try 1 and 2 etc. then ican optimise by varying say a full push or a halv/light nudge . This way I have reduced the wow peak . from 0.15 to 0.03. see red arrow it is the WOW at 0.55Hz you are adjusting, nothing else

The adustmentr range I have on my best record is ca 0.5mm certainly less than 1mm

 

[JP]

Why not just watch a locked groove?