It will show that the designer used a motor with sufficient torque to maintain a stable speed. |
It is actually 1.8 sec/rev. 33 and 1/3 rev/min divided by 60 sec/min yields 0.5556 rev/sec (0.555555555 repeating). Your tt makes 100 revolutions every 3 minutes or 100 revolutions every 180 seconds. So if you divide 180 seconds by 100 revolutions, you get 1.8 sec/rev. |
Halcro, another way to do it is with a stopwatch if you are fast enough. The motor should be able to accelerate the platter from 0 to 33.33 rpm in less than 3.6 seconds. So you could start the laser then start the motor and stopwatch and time it until the laser spot stops moving. |
I have read this post with fascination and it has lead me to do some measurements and calculations. I am an engineer. This sort of thing is what I live for. I have an expensive belt drive tt and therefore I became very interested in determining if I need to invest $4k+ into a tt upgrade because I am missing out on some musical pleasure from my vinyl.
My first measurement was to pull out my strobe disc and fluorescent light. (Ugh I can hear some of you saying, but just listen a moment) I fixated on one black mark and in 10+ revolutions, that mark did not drift as best my eyes could tell. Next, I dropped the needle on the record while remaining fixated on that black mark. Again, no drift at all as best my eyes could tell. I must comment that considering the cost of most of our turntables these days, they sure better be able to hold speed with or without stylus drag. It would be a pretty poor turntable that could not. After all, this is the primary function of the tt- spin the platter.
So at this point, I'm pretty convinced that my tt holds rotational speed very accurately even considering the apparent disadvantage of being a belt drive. Next I put on my Stereo Test Record and played a 1000 Hz test tone. I can hear some Wow in the tone. The "A" tone, 440 Hz makes the Wow much more apparent. So where is the Wow coming from? The specs on my tt state a Wow and Flutter measurement below 0.03%. I believe it. That's part of what I paid for in this very expensive tt. I look at the record, and this is supposed to be a test record, and I see some serious runout. I can hear the Wow precisely when the tonearm is rocking outboard as the record is spinning. As with many records, the tonearm is swaying back and forth due to the runout- ie. the center hole is off position relative to the center point of the grooves. At this point, I would like to borrow a dial indicator from work to measure this runout. (I would also like to see just exactly how much runout is in my platter.) Since I cannot do that today, I measured the distance of the center hole of the record to the outer edge. I found the center hole of this record to be off center by 0.8mm or 0.0315 in. This would yield a runout measurement of 0.016 in. After searching the internet a while I found a specification for records that says the runout tolerance of the hole can be +/-0.015in. Allowing for some measurement error on my part, this test record has runout at the maximum industry tolerance. What does that mean from a sound standpoint?
Well, I measured the radius from the center of the record to the grooves with the "A"/440 Hz. That measured 3.5 in. This puts the groove velocity at that point at about 12.22in/s. Now factor in the runout of 0.016 in and the speed change is 0.05497 in/s. That causes a frequency shift of 2Hz. It is actually +/-2 Hz. This is the theoretical calculation. Now to measure the actual shift.
I happen to have a FFT analyzer handy. (Engineers, sheesh!) So I measured the frequency of the 440 Hz tone being played on my tt and coming out of the speakers. Guess what! I see the periodic Wow in the trace and I also see the frequency varying from 338 Hz to 442Hz! How about that?
I conclude that my tt maintains speed at least an order of magnitude better than a record with production tolerances. If the runout is 0.016in, then the Wow will be 0.45%. That is over 10 times the spec for my tt.
Okay, what does all this mean subjectively? I think that rhythm and pace is definitely a variable among turntables. I have heard it myself on different systems. I don't understand how it differs from one to another. Maybe no one does completely which explains the myriad of solutions in the marketplace for spinning a platter. Maybe someone could come up with a platter design that can true the record to the center of rotation. |
I found an article online about the Nak. Very interesting tt. The author had an excerpt from an interview with the designer. He substantiates my point exactly. He says vinyl record runout is the elephant in the room that tt designers ignore. As for inertia: Torque= J*omega, which is the angular term for F=ma. The tt motor provides the torque and the platter bearing and stylus apply a counter torque. The inertia of the platter determines the rate of change in speed (deceleration). Say for example the motor is uncoupled from the platter. The platter is spinning at 33.33 rpm. (ignore bearing friction for a moment) Now drop the stylus onto the record. A 20 kg platter is going to decelerate at a lower rate, for example, than a 2 kg platter.
Now let's hook the motor back up to our platter. The motor is either clocked to the 60Hz line frequency or is feedback servo controlled. So it holds the platter at 33.33 rpm. Any perturbation in the platter speed causes the torque output of the motor to change in order to restore 33.33 rpm. The motor could do it's job regardless of the amount of inertia in the platter. The stability of the platter speed is based on the control loop and torque of the motor combined with the system inertia. That means the designer has to couple a motor and platter as a system. The platter is designed to be a mass damper. We use mass dampers in dynamic systems. We use mass to tune System Natural Frequencies and keep them out of certain operating ranges. A bigger platter requires a higher torque motor in order to be stable. Perhaps the youtube example is a tt design with an undersized motor. I would say as a rule of thumb, the motor in a tt should be able to accelerate the platter up to speed within one rotation. To me that would indicate that the motor has sufficient torque to maintain a stable speed. btw- I just checked my tt and it is up to speed within one rotation. |
Halcro, can you turn on the laser with the platter stopped and then start the platter? See if the speed is stable within one rotation. It is a bit difficult. The best way to see it would be with instrumentation and a plotter. |
Keep in mind that the effect of the vinyl record runout is worse at the inner groove. As the needle is further out from the spindle the radius of the runout is proportionally smaller so Wow will be less at the outer groove of the record. So effectively another problematic issue when playing the inner grooves of records. Does anyone notice a rhythm/timing difference in the music from the outer to inner tracks? |
I found an article online about the Nak. Very interesting tt. The author had an excerpt from an interview with the designer. He substantiates my point exactly. He says vinyl record runout is the elephant in the room that tt designers ignore. As for inertia: Torque= J*omega, which is the angular term for F=ma. The tt motor provides the torque and the platter bearing and stylus apply a counter torque. The inertia of the platter determines the rate of change in speed (deceleration). Say for example the motor is uncoupled from the platter. The platter is spinning at 33.33 rpm. (ignore bearing friction for a moment) Now drop the stylus onto the record. A 20 kg platter is going to decelerate at a lower rate, for example, than a 2 kg platter.
Now let's hook the motor back up to our platter. The motor is either clocked to the 60Hz line frequency or is feedback servo controlled. So it holds the platter at 33.33 rpm. Any perturbation in the platter speed causes the torque output of the motor to change in order to restore 33.33 rpm. The motor could do it's job regardless of the amount of inertia in the platter. The stability of the platter speed is based on the control loop and torque of the motor combined with the system inertia. That means the designer has to couple a motor and platter as a system. The platter is designed to be a mass damper. We use mass dampers in dynamic systems. We use mass to tune System Natural Frequencies and keep them out of certain operating ranges. A bigger platter requires a higher torque motor in order to be stable. Perhaps the youtube example is a tt design with an undersized motor. I would say as a rule of thumb, the motor in a tt should be able to accelerate the platter up to speed within one rotation. To me that would indicate that the motor has sufficient torque to maintain a stable speed. btw- I just checked my tt and it is up to speed within one rotation. |
The mass of the platter has nothing to do with speed control other than it is the "m" in the equation, F=ma or T=J(Omega). The motor controls the speed. Its that simple. As shown by the equation, more mass means the motor must generate more torque to provide the same acceleration. How much acceleration is enough? Well, it depends on how well the designer wants to hold the exact rotational speed under dynamic conditions. I stated as a rule of thumb based on my knowledge and experience designing dynamic systems, (not turntables) that a tt motor should be able to accelerate the platter from 0 rev/s to 0.5556 rev/s within the first rotation. Some motors may be able to do better than that, but if it takes a motor more than one rotation to get the platter up to speed, then I think it might lack sufficient torque to maintain a stable speed. I could be wrong here, I've never designed a tt myself. But to carry it further, the motor is designed to spin at a certain speed. That control system for the motor can be as simple as being clocked to the 60Hz sine wave of the AC power to a very sophisticated electronic servo feedback control system. However it is done, the resolution and rate of response of this control system results in the level of stability of the platter speed. The system design dictates how much speed error can occur before it corrects it and how fast it can correct the speed. At this point I could get into the impact of an underdamped or over damped system, but the point is the tt designer must come up with a system that includes the motor torque, speed, resolution of the control loop, platter inertia.
I might be wrong here, but I have observed that DD turntables tend to have lower mass platters versus later BD tables. I'm guessing that the torque available from DD motors is less than for BD considering BD motors get a mechanical advantage from the pulleys. Higher mass platters offer a nice stable sink for the vinyl record to minimize vibrations. The added inertia of high mass platters is probably more of a problem for tt designers in terms of speed control- just because it requires more torque from the motor.
I'm starting to think the rhythm and pace might be linked more to the tonearm and stylus. The runout of the vinyl record, as I showed before, has more impact on Wow than does the tt. But the swaying of the tonearm, which is part of the Wow problem must also influence the tempo of the music I think. Couldn't the side to side movement of the stylus in the groove alter the tempo? Is that what tonearm damping is all about? I don't have damping. Does anyone have experience that says tonearm damping improves the tempo of the music? |
H Lewm, yes, you are correct. That is called cogging. I'm sure BD motors do it too if they are unloaded. You just can't see it with your eyes because they spin too fast. That is what I was saying before that designers use mass as a damper- mass damping. All motors have some level of cogging, or ripple. Even your car engine has it as each piston fires. The flywheel smoothes out those ripples just as the tt platter smoothes out the ripples from the motor. The typical tt specs showing Wow&Flutter below 0.03% shows how well these massive platters are doing their job. btw, the more poles that a motor has, the smaller the ripples. I notice some of these expensive tables have quite sophisticated motors. Caveat Emptor always applies, but I think you are getting what you pay for in some of these expensive motor designs. |
I downloaded the iPhone app and did some testing. The same test record that I used before also has a track with 3150Hz. As I mentioned before, I measured considerable runout in my test record which by itself causes some audible Wow when playing constant frequency test tones. The center hole of this test record has a slight amount of play on my tt spindle. After much patient work to minimize the record runout I got the following measurements: Mean Freq: 3151.3, Raw Frequency: -0.22%/+0.24% relative, -7.0Hz/+7.4Hz absolute. Lowpass Filtered Frequency: -0.02%/+0.01% relative, -0.5Hz/+0.5Hz absolute. I'm guessing that the filtered values subtract out the record runout. The reason I conclude that is as I meticulously adjusted the record on the platter to minimize runout, I had raw numbers as high as -9.8/+10.7Hz while the filtered values never exceeded 0.03%.
Another observation: The original Mean Freq was 3155.6Hz based on my setting using the strobe disc. I adjusted speed down and can certainly work some more to dial it in at precisely 3150Hz. The other observation is that holding the iPhone while taking the measurement adds more error. It is sensitive enough to pick up hand movements. I set the phone on a table about 1 meter away from the front of one speaker.
So how does my humble BD table compare to some others?
btw- I have a record clamp. So I adjusted the record relative to the spindle and then clamped the record down. |
I found the owners manual online for Adjust+ and some reviews. As I surmised the low pass filter is set for 0.56Hz to eliminate record runout. I was considering drilling the center hole out a few tenths of a millimeter on my test record, but I began imagining what a catastrophe that would be me loose with a hand drill with one of my precious records. Fortunately this iPhone app cancels out the record runout for you. Pretty cool app.
Regarding the tt bearing: I do my own maintenance on my tt. Periodically I breakdown the tt and clean the main bearing. I re-lubed last time with a synthetic grease. I think it made the tt sound better. Anyone else had that experience? I have never touched the motor. |
Dover, good suggestion. I have a small, round bastard that should do the trick. I might give that a try. |
I filed the hole in my stereo test record. I improved the runout a bit. According to the iPhone app, I now have the raw Wow down to +/-0.16%. The filtered didn't change, -0.01%/+0.02%. The total spread within the published specs for my tt. With a little more filing of the center hole and more work/patient effort to center the record on the platter, I could improve the raw values even more. I have a feeling that I will be bringing a dial indicator home from the machine shop at work to measure record runout with extreme precision. I will likely be filing the holes on my records now to perfectly center them. My neurosis is ratcheting up a level, I think.
Anyway, this app is nice, you can play the test records 3150Hz track at 45rpm and this app will detect the higher frequency and tracks the speed at 45rpm automatically. I must agree having seen it live with this app. Having the motor via a belt separate from the floating chassis/platter is a deficit. The slightest vibration/movement of the floating chassis causes a speed error. It is a small error, but still there. |
Runout is the measurement of the eccentric motion in a rotating body. What I have been talking about in terms of a record is that the center axis of the record grooves are not matched precisely to the center axis of the platter. (Shaft runout would normally be measured in V-blocks in a lab and would consist of a combination of roundness as well as straightness of the shaft.) Additionally, the platter can have some runout- hopefully that is nearly unmeasurable considering the cost of some of these platters. The best way to measure the runout of a record would be to measure the swinging motion of the tonearm when it is playing a round groove at the inner diameter of the record. It would be difficult to measure while playing in the music groove since the tone arm is continuously moving inward towards the center of the record. The OD of the record is likely not that round and not necessarily tied dimensionally back to the music groove. I would say that if you can discern a swaying motion in the tonearm with your eyes as it is running in the music groove, then the runout is probably high. This was the case with my test record. But even then, I could hear the Wow only when playing pure test tones. |
Your timeline device is averaging speed. It is not giving you an instantaneous reading. The laser pulses every 1.8 seconds. As long as the turntable achieves one revolution in 1.8 seconds, the laser spot remains stationary. Your turntable could be speeding up 25% and slowing down 25% in one revolution and the timeline will happily show a stationary spot on the wall as long as the platter rotation averages 1.8 seconds per turn. The best way to determine how your turntable is behaving is to look at speed real time. One way is some type of encoder with the output to an X-Y plotter. Another way is to play a record with a fixed frequency such as 3150 Hz. The inherent errors are 1) accuracy of the recorded frequency and 2) record runout. Using an FFT analyzer you can see any speed drift realtime. A good substitute is the iPhone app. It seems very good to me. It will show you your turntable wow and flutter real time and filters out the 0.5556Hz frequency caused by any record offset to the platter.
I would like to know if anyone's turntable wow and flutter measures less than 0.01% regardless of drive type. |
I'll ask this question again since the other post was deleted. What does everyone expect from their turntables when using the Timline device? Absolute zero drift of the laser mark? Here is another way to look at it. Leave the Timeline device on for 30 minutes. If you are comfortable cueing up the same record side twice, do that. So in 30 minutes the platter will rotate 1000 times. Since most tt's are spec'd to have a speed accuracy of around 0.02%, then you can expect to see the laser drift about 72 degrees in 30 minutes. That is 0.2 rotations out of 1000. Is that good or bad? That would seem very good to me. btw- the Timeline maker advertises an accuracy of 2ppm. I take that to mean the Timeline device is about 2 orders of magnitude better than a typical tt. |
Hi Halcro, I saw your video. It is impressive how well your tt holds speed, but that was a relatively short period of time. Can you let it run for 30 minutes to see the results over the long term? I am very curious to see those results. If the laser mark has no drift after 30 minutes that would mean your tt has speed accuracy on the order of 0.00002%. That would be far beyond expectations in my mind. Some suggest that the speed accuracy fluctuates, but I believe that speed error is cumulative when using a device such as the timeline. That means tha the longer you run your tt with the timeline the more error, or drift you should see. I don't think that cueing and dropping the tonearm will make the line move the other way due to the motor controller. It reacts to the torque changes. |
Let me clarify my one statement a little better. Cueing up the tonearm is a one time event that may or may not cause the laser mark to move. It is not part of the cumulative error. So over the long term it has little or no impact on the results. Unless something funny is going on with the motor controller, the laser mark should drift slowly in one direction over the 30 minutes. That is the cumulative error. |
Hi Halcro, That is excellent news. So is speed accuracy something that was once mastered and now lost on more recent tt's? How do we get it back? Do you feel you lose something with the thread drive since it drifts a bit? I would think not, but your opinion/experience here would mean something. |
I said it once before long ago- any tt motor that brings the platter up to speed in about a half of a revolution (less than one revolution) has plenty of torque to hold speed stable.
Also, I wanted to be the 300th poster on this thread. Yay!! |
The Timeline is giving you one data point per rotation. So what is happening in between those 1.8 second pulses? Your ears just told you something more is happening than just precise timing of the platter. It is like peeking into the room and everytime you see the cat sleeping, but still the canary disappears. The timeline is giving only one dimension of speed control- ie. timing of the rotation period. What is happening during the rotation? The platter could be cogging, it could be wobbling or even have Wow due to an out of round pulley or out of round platter OD.
Beyond the speed accuracy measured by the timeline, as someone else mentioned, speed precision is also important. That means how well does the platter hold 33 1/3 during the entire rotation. This has to be measured either with a test record or with a precision tachometer. The best instrument might just be our ears. Test records have limitations- the accuracy of the center hole in records causes them to be at least one order of magnitude worse than most turntables. (Two order of magnitudes worse than Halcro's DD turntable.) |
The comment about using viscous drag or friction to dampen platter oscillations on bd tt's struck me. I do my own maintenance on my Sota tt. It is easy enough to disassemble the platter, clean and relube the spindle and bearing. I let it go for a few years and then about two years ago, I got around to pulling it apart and relubing it. I noted then that the tt sounded better, but I didn't really understand why. So now I clean and relube it every year. Some other Sota owners have also mentioned they found some synthetic lubes make it sound better than other lubes. I tried some synthetic lube last Fall and didn't notice a big difference in sound myself. The synthetic lube is very tacky and definitely added drag to the spindle compared to the lithium grease that I was using. Since my speed is pretty rock solid, I guess the drag is not too high. |
I have to say, this thread has been a good learning experience for me. Other than changing my phono cartridge now and then and doing basic maintainence on my tt, such as belt change and bearing lube, I have just listened to my records on occasion. Over the years I set my platter speed by the supplied strobodisc and flourescent light and called it good. But over this past month here is what I have learned. First, I learned that my speed was low by a couple of percent. I was able to dial in my speed exact by using the 3150Hz test tone on an old test record that I had and the iPhone app. That app is a very useful tool. Interestingly, with the speed dialed in dead on now, the strobodisc still shows a stable pattern. So the strobodisc/flourescent light method is not very accurate, ie. >2%. Second, not only has the rhythm and timing of the music changed, but the bass is tighter, the attack stronger. I noticed this playing a couple of old records that I have listened to many times over the years. The timing was so different from what I was used to that I had to go back and double check the speed setting again. Now I have to listen to all of my records all over again at the correct speed. Good thing I am still young. I conclude based on what I have learned that speed setting is as at least as important as speed control. In fact, I am wondering now if the differences heard with respect to rhythm and timing are really based on the speed setting of a tt more than Wow and Flutter- ie. if Wow and Flutter are less than 0.03% or so. |
Hi Peter,
Excellent. First, that is a really long tonearm. I'm not envious, just impressed. I noticed just a slight amount of drift over the 5.2 minutes of the song. It appears the speed is just a little fast by about 1.6%. I estimated that the line drifted about 1 inch over the 5.2 minutes and I also estimated the radius from the timeline center to the album cover to be 10 inches. So, if I did the math right, the angle changed 5.7 degrees over 5.2 minutes. If you have exact numbers then the calculations would be more meaningful. That means the number of record rotations was off just 2.8 rotations out of 173.33 rotations. Are you able to adjust the speed a bit? At the end it seemed like the line did not shift when you lifted the tonearm. Is that what you saw? Is your speed controller closed loop? When you dropped the tonearm onto the record did you see any shift in the line? |
Hi Peter,
I updated my calculations and also found an error. My previous calcs were off by 100 because I had calculated a % and then used that for an absolute. The correct numbers now with your measurements are speed is fast by 0.003%. The 5.2 minutes of playing time were off by 0.0059 rotations out of 173.33. (That 100 makes a big diff in the numbers). Drift to the left means slow? Ok, I see that now too. |
Hi Peter,
Sorry for the confusion this morning. I was rushing because we were preparing to go to a wedding. I was excited to see your post and wanted to reply before we left and my wife was bugging me to get off the computer and get ready to go.
The drift is 0.4 degrees/minute or 0.001133 revs/min. So after 5.2 minutes, the total error was 0.00589 revolutions out of 173.33 total revolutions. That is a speed error of just 0.003%. It may not be 0 as some DD turntables claim to be; but it is pretty close.
Your turntable is absolutely perfect. Its speed control is on the order of any type of drive system, ie. direct drive, rim drive, etc. in terms of accuracy. The only other check would be with a scope to see the constancy of speed over a single revolution (Wow and Flutter); but I'm sure it is laboratory grade as well. I wonder why the Timeline device does not flash more often. If it did, it would give some indication of speed variation during one revolution.
Thanks for sharing. |
