It has been more than month that I have really good TT rotation (form 0.02 to 0.04% wow). I should have being playing with cartridge and preamp by now. But I got this obsessive compulsive disorder to try to get it as perfect as possible (in given setup). In the meantime I re-polished the drive pulley to get everything where it was before etching, and on a picture below you can see my challenge. Following is brain stream:
This is one of worst measurements, but highlights the issue nicely: biggest dips are 1.8 second apart:
Mostly I get better than this, but than it is more difficult to spot 1.8 second issue... So, 1.8 second is 1 rotation of the platter, this is obviously not motor related and if I manage to fix this I can as minimum half the speed osculations....
1st , must be the driven pulley, or platter in my case. Installed vernier caliper on the plinth, established where the metal is touching the platter rim (where belt works) and found biggest gap on the other side of the platter. Measured that gap with feeler gauges... The difference is about 0.06mm (feeler gauge goes in tight but smooth) and 0.08mm (feeler gauge goes in but hard). Worst case 0.08 mm over 300mm platter gives 0.00267 % eccentricity... This is not it! Platter is good. I could improve it by same method used before for drive pulley, but this is not it.
2nd, must be the bearing... Awe, my bearing is closed at bottom end, machinist never could go with honer through it... Must be it. Now I did something with great potential to damage bearing. Still have finest grinding paste for finishing engine valves. I put small amount of grinding paste on bearing instead of oil, slowly pressed platter in by rotating, and rotated, rotated and rotated until it was smooth (warning, don't try this at home unless you know exactly what to do).... Ok, if there was any regularity inside this would fixed it. Washed all through, re oiled and run, result to my disappointment exactly the same! (with more luck than brains, bearing behaves the same as before, grinding paste did not take much away)
3rd, Even many years ago I established the best lubricant, started again; from copper grease to lightest PTFE oil and anything in between. Washing, cleaning every time between changes.
Well established (for me) motorbike chain lube just proved to be by far the best. Thicker lubricants (greases) just slowed platter (more slip on drive pulley) and thinner ones did not. But wow was worst on all . Most importantly, 1.8 second pattern did not change at all regardless lubricant.
4th , what is left? The belt of course.... Earlier I reported that every change in belt makes huge difference. Now I examined my best resulting belt and found visible irregularity at one spot. Belt in my set up is just little bit longer than platter circumference, easy to confuse 1.8 second with slightly more belt lenght.
Than tried 2 looser belts one on top of other (making one thicker belt with hope that belt mistakes will equalize over 2)) and result was bad wow, but 1.8 second issue reduced a lot, ok, first clue. Run to 2 shops where I was buying belts earlier, both have no stock of my size, bummer!
Finally today I did something just to prove the concept. Used sewing elastic thread and made 2 belts , each with knots. I installed them with knots on opposite sides. In theory I should get 0.9 second issue now as there are 2 bumps in the belt instead of one. Here is the measurement of this, set up and picture of elastic band:
The speed run down because these cords are not belts, but not the point. Point is that 1.8 second issue is now about 0.9 - 1 second issue.... 2 knots!
I think this is the clue... Need to get more right sized belts to confirm this!
PS, funny setup with 2 elastic cords for laundry knotted in the belt gives 0.12% wow, acceptable in most cases 🙂
This is one of worst measurements, but highlights the issue nicely: biggest dips are 1.8 second apart:
Mostly I get better than this, but than it is more difficult to spot 1.8 second issue... So, 1.8 second is 1 rotation of the platter, this is obviously not motor related and if I manage to fix this I can as minimum half the speed osculations....
1st , must be the driven pulley, or platter in my case. Installed vernier caliper on the plinth, established where the metal is touching the platter rim (where belt works) and found biggest gap on the other side of the platter. Measured that gap with feeler gauges... The difference is about 0.06mm (feeler gauge goes in tight but smooth) and 0.08mm (feeler gauge goes in but hard). Worst case 0.08 mm over 300mm platter gives 0.00267 % eccentricity... This is not it! Platter is good. I could improve it by same method used before for drive pulley, but this is not it.
2nd, must be the bearing... Awe, my bearing is closed at bottom end, machinist never could go with honer through it... Must be it. Now I did something with great potential to damage bearing. Still have finest grinding paste for finishing engine valves. I put small amount of grinding paste on bearing instead of oil, slowly pressed platter in by rotating, and rotated, rotated and rotated until it was smooth (warning, don't try this at home unless you know exactly what to do).... Ok, if there was any regularity inside this would fixed it. Washed all through, re oiled and run, result to my disappointment exactly the same! (with more luck than brains, bearing behaves the same as before, grinding paste did not take much away)
3rd, Even many years ago I established the best lubricant, started again; from copper grease to lightest PTFE oil and anything in between. Washing, cleaning every time between changes.
Well established (for me) motorbike chain lube just proved to be by far the best. Thicker lubricants (greases) just slowed platter (more slip on drive pulley) and thinner ones did not. But wow was worst on all . Most importantly, 1.8 second pattern did not change at all regardless lubricant.
4th , what is left? The belt of course.... Earlier I reported that every change in belt makes huge difference. Now I examined my best resulting belt and found visible irregularity at one spot. Belt in my set up is just little bit longer than platter circumference, easy to confuse 1.8 second with slightly more belt lenght.
Than tried 2 looser belts one on top of other (making one thicker belt with hope that belt mistakes will equalize over 2)) and result was bad wow, but 1.8 second issue reduced a lot, ok, first clue. Run to 2 shops where I was buying belts earlier, both have no stock of my size, bummer!
Finally today I did something just to prove the concept. Used sewing elastic thread and made 2 belts , each with knots. I installed them with knots on opposite sides. In theory I should get 0.9 second issue now as there are 2 bumps in the belt instead of one. Here is the measurement of this, set up and picture of elastic band:
The speed run down because these cords are not belts, but not the point. Point is that 1.8 second issue is now about 0.9 - 1 second issue.... 2 knots!
I think this is the clue... Need to get more right sized belts to confirm this!
PS, funny setup with 2 elastic cords for laundry knotted in the belt gives 0.12% wow, acceptable in most cases 🙂
Last edited:
Dražen, in your place I'd try something non elastic, like dental floss. Don't worry about the knot. The ultimate test would be a flat non elastic belt, like a magnetic tape or video tape with 45° splicing. But you would need a cylindrical pulley instead of a barrel shape, and a second spring loaded pulley (on the opposite side of the platter?) for keep the tape tight.
Szia Laszlo,
Ok, I can run floss, even I don't believe in knots on the belt, I tried fishing line and dental floss.
Cassette tape I can try with different motor and set up with external motor.... Just doubting if it is worth. If you have that prepared you can show result?
Fishing line, not bad but still twice as much wow than common belt.
Dental floss was at 0.33% , so first try renders it useless.
Ok, I can run floss, even I don't believe in knots on the belt, I tried fishing line and dental floss.
Cassette tape I can try with different motor and set up with external motor.... Just doubting if it is worth. If you have that prepared you can show result?
Fishing line, not bad but still twice as much wow than common belt.
Dental floss was at 0.33% , so first try renders it useless.
Recently I got my TT measurement by test LP where I received great and instrumental help from Mr Hans Polak who analyzed my recorded files by his software, it is incredibly detailed.
I recorded 1kHz test tone from Ortofon test LP, once at 33.33 RPM and once at 45 RPM and Hans analyzed them.
This is the picture of 33.33 RPM, so 1kHz. 3 traces represent 3 revolutions :
Wow for IEC386 is 0.056%
With an additional notch filter at 0.55Hz Wow becomes 0.047%
This the same track but recorded at 45 RPM, now center frequency is 1,35 kHz. Here the vertical axis is showing mV, but 1mV actually means 0.1% speed deviation.
Now Wow IEC386 is 0.062%
Wow IEC386 plus 0.75Hz notch filter is 0,049%
These 2 pictures present the same (just one is concave and other convex, it just depends where measurement started) and show that almost all deviation is caused by platter or LP eccentricity, or belt. I measured platter, looks excellent, but I will revisit it, as well as belt and get another test LP..
Nevertheless, results as is are already very good. They also complement phone measurements with app (that are 0.03 - 0.04% ) but this time LP arm and cart are added to the picture. Phone app also shows biggest deviations at 0.55 Hz so eccentricity is not only LP. I can again confirm that this phone app is sensitive with right phone, if it doesn't show absolute correct numbers, it records every change and its graph points to right issues, and so easy and quick.
What is very nice is higher frequency deviations, they are very small and obviously repeatable, If I count them it is close to 1 motor revolution, so I could polish drive pulley a bit more (or improve motor bushings) if it makes sense at all.
There is no sign of motor torque ripple (it would be 8 times shorter than short deviations present), seems Supaspin and tape recorder capstan motor really do the job.
I will play with this forever, but motor and drive story is finished for this thread. Soon I will make summary in post #1 and from then on just here and there report any upgrades.
Big thanks to Hans Polak for this!
PS, goal was to do this within one year, now it is 1 year and 10 days. All of future upgrades I can do without taking TT apart, so it goes back to listening room finally!
I recorded 1kHz test tone from Ortofon test LP, once at 33.33 RPM and once at 45 RPM and Hans analyzed them.
This is the picture of 33.33 RPM, so 1kHz. 3 traces represent 3 revolutions :
Wow for IEC386 is 0.056%
With an additional notch filter at 0.55Hz Wow becomes 0.047%
This the same track but recorded at 45 RPM, now center frequency is 1,35 kHz. Here the vertical axis is showing mV, but 1mV actually means 0.1% speed deviation.
Now Wow IEC386 is 0.062%
Wow IEC386 plus 0.75Hz notch filter is 0,049%
These 2 pictures present the same (just one is concave and other convex, it just depends where measurement started) and show that almost all deviation is caused by platter or LP eccentricity, or belt. I measured platter, looks excellent, but I will revisit it, as well as belt and get another test LP..
Nevertheless, results as is are already very good. They also complement phone measurements with app (that are 0.03 - 0.04% ) but this time LP arm and cart are added to the picture. Phone app also shows biggest deviations at 0.55 Hz so eccentricity is not only LP. I can again confirm that this phone app is sensitive with right phone, if it doesn't show absolute correct numbers, it records every change and its graph points to right issues, and so easy and quick.
What is very nice is higher frequency deviations, they are very small and obviously repeatable, If I count them it is close to 1 motor revolution, so I could polish drive pulley a bit more (or improve motor bushings) if it makes sense at all.
There is no sign of motor torque ripple (it would be 8 times shorter than short deviations present), seems Supaspin and tape recorder capstan motor really do the job.
I will play with this forever, but motor and drive story is finished for this thread. Soon I will make summary in post #1 and from then on just here and there report any upgrades.
Big thanks to Hans Polak for this!
PS, goal was to do this within one year, now it is 1 year and 10 days. All of future upgrades I can do without taking TT apart, so it goes back to listening room finally!
Still needed to try cassette tape belt as it was strongly suggested, installed pitch roller on opposite side than motor to tension it and made splice with superglue (very thin) over approx 5mm of overlapping tape. Like this it runs nice on right place of crown pulley like rubber belt.
The result is nevertheless a disaster, not only that splice is sharply visible, but it also has sharp harmonics kind off. Don't think this is the way to go really...... I ordered some new belts 🙂
The result is nevertheless a disaster, not only that splice is sharply visible, but it also has sharp harmonics kind off. Don't think this is the way to go really...... I ordered some new belts 🙂