Hi Peter,
One thing I noticed on a couple of your videos, as you had mentioned, the red line is sometimes shorter. Do you see this when watching it live? I suspect that the occasional blip in the red line on the video is due to the shutter speed of the video camera. Do you know if you are recording at 24 fps, 30, fps or 60 fps? If you can, try changing the shutter speed and see if the issue is gone. |
Well, the Timeline is a strobe; but I do not say that with disdain. It is just a fact. It is a very precise strobe with reported 0.0002% accuracy. As with any measurement device it must be used properly. The results must be interpreted properly as well. I know it is too late in some cases; but I propose that when using the timeline, 9 minutes minimum should be the measurement cycle. That is just short of 1000 rotations of the platter. If turntable speed accuracy is in the 0.003% range like Peter's SME deck, then that will show an error of 0.03 rotations. For example, 0.03 rotations is 10.8 degrees. If the laser line is projected out to a wall 18" away, then the drift after 1000 rotations will be 3.4". Measuring 3.4" vs. less than 0.5" will reduce the affect of measurement error and give more accurate speed measurement results. |
A DC Motor is a simple device. The speed is proportional to voltage input and load. If the load increases then the speed will drop and current will increase. My Sota tt has a simple power supply and DC motor. It has a trim pot for fine tuning speed; but speed will vary as load changes. (A tt with closed loop controller for a DC motor would vary voltage to maintain speed under varying load conditions.) Using the iPhone app I can adjust the speed to within 0.02%- about an order of magnitude off compared to Peter's SME deck. Also, speed drifts on my tt as it warms up so I must let it run for 15-20 minutes before fine tuning the speed. I know the higher end Sota tt's have much more sophisticated motor controllers. I would like to know how they perform in comparison.
Synchronous AC Motors are based on frequency. The rotor of an synchronous AC motor is always following the rotating magnetic field. That is phase lag or slip as Richard mentioned. As the load on an synchronous AC motor increases, phase lag increases. Current draw in the motor will increase which increases torque in order to bring the phase lag back to near zero. The frequency source for a synchronous AC motor can come from the 60Hz AC line or from a seperate frequency generator. Voltage will not change speed on a synchronous AC motor but it will change the amount of torque available to minimize phase lag. (More voltage means more current draw is available.) A closed loop synchronous AC motor controller must vary frequency to adjust speed. |
Halcro, My test record that I use for the iPhone app is also flawed. The record hole is off-center so I filed the hole open a bit and I must center it as best as I can on the platter. I get differing Wow&Flutter measurements (between 0.06% down to 002% filtered readings) depending on how well I center the record. My record also has a slight warp and at around 12 seconds into the track the frequency shifts due to this warp. So my overall measurements look better when I start the recording past that warp. I did screen saves on sevaral measurement trials using the iPhone app and finally saw this pattern of speed or frequency shift at the same point every time. |
I graduated from college and started working in Aerospace Engineering just as the WWII generation was starting to retire. I had the privilege of being mentored by engineers that started working just before and after WWII. This is the generation that designed and developed moon rockets, supersonic jet fighters, television and hifi. They did it all using slide rules and look up tables. It was an age of soulful design. On paper, vinyl playback looks bad. On paper, vacuum tubes look bad; but separately and even better together they make music magic. There is a lot to be said for Computer Aided Design and increased engineering productivity; but there is something about the Art of Engineering in vinyl records and the whole recording chain of the past, something remarkable that still amazes us in the 21st Century. |
Rather than hijack this thread some more (sorry Peter:-) I have done a write up on my turntable mod at my system page. |
I think the confusion lies in trying to use the Timeline, which shows average speed over any period of time, to evaluate momentary, micro-variations in speed due to stylus drag. The timeline can only infer that speed is varying once the stylus is in the groove if the laser mark drifts. The problem is you can't know if the drift is due to speed being something other than 33 1/3 exactly or if speed is actually changing during play. The only sure way to measure the effects of stylus drag on speed for a particular tt is with a very fine tachometer.
Think of driving your car. You time yourself from point A to B and knowing the distance determine your speed. That is average speed. If you want your speed to be exactly 30 mph then you must leave and arrive within a specific time interval. The thing is, you don't know what your speed variation is between points A and B. In order to know your speed variation you must watch your speedometer. The speedometer is giving you instantaneous speed. Showing your stopwatch to someone as proof that you drove exactly 30 mph between A and B doesn't prove that your speed was a constant 30 mph. Maybe you went up a hill and dropped to 25 and then down a hill and got up to 35 for a moment averaging out to 30.
The timeline is like a stopwatch giving you average speed. |
There are two subjects here: 1) How well do the differing turntables hold speed to exactly 33 1/3 rpm? 2) What is the effect on platter speed during heavy groove modulation?
Subject 1 is being answered with the Timeline device testing. And the Timeline device is being used appropriately for these measurements. This is a perfectly valid and interesting exercise.
Subject 2 is also a valid and interesting point. The Timeline device, however is not the appropriate tool in this case. A fine tachometer is required. I find this subject interesting and especially if it were possible to compare the results of speed control with a heavily modulated record on all three basic drive types- belt, direct and idler. |
Hi Lewm,
How can micro variations in speed due to groove modulations be measureed with a laser light that flashes just once every 1.8 seconds? The answer is that it cannot. The Timline gives you average speed- that's it. Don't try to read into it anymore than that. Sure, you can see the difference in average speed for an unloaded platter vs. a platter with stylus load; but you cannot measure the micro variations in speed with an averaging device. It must be an instantaneous device like a tachometer. The Timeline is a nice little package that has laboratory grade accuracy for a realatively low cost. A Tachometer with equivalent precision will cost significantly more, I'm sure. The danger is mis-using the data from an averaging device. For exmaple, the speed error for variable stylus drag is cumulative. So the longer you play the record and measure speed, the more the average, in percent, will deviate from 33 1/3. (If the stylus drag is constant, then the error will not accumulate- ie. the average speed will remain a constant delta in percent from 33 1/3). See, things can get murky real fast when working with averages. You can infer stylus drag is causing variations in platter speed with the Timeline, but actual measurements are not possible.
Back to my car analogy: Try driving and maintaining speed with the Average Speed readout on your car's trip computer. Once you have accumulated a couple of miles on the average speed computer, the speed reading will not change much even if you go very slow or 120 mph. The police officer is using an instantaneous speed readout device. Good luck convincing him not to ticket you because your average speed was below the limit. (If that does work, let me know how you did it!) |
I'm not sure it has been confirmed that the newer Timeline devices have 6 flashes per rev. At least the ones everyone is using seems to have just one flash per rev. Regarding tachometers: most today and even the one in your car uses either optical or hall effect sensing devices. They are not mechanically linked. The accuracy is dependent upon the precision of the encoder ring. Good data capture for micro variations in speed would need to be in the 40kHz range (two times max frequency response of a record), at least. That is several orders of magnitude beyond even 0.3 seconds. |
Hi Halcro, let me explain something about motors here without the added detail of charts and graphs. A motor runs at a constant speed to the casual observer; but if you take a much closer look you would see that the motor armature actually speeds up and slows down as it passes by the individual poles. Cogging is an accentuated aspect of this motor property. A flywheel is added to the armature to smooth out this periodic motion. This is a property of any motor type. Now, with the exception of DD turntables, the motor is just one part of the platter drivetrain. If the pulley on the motor shaft, or the idler as the case may be, is not perfectly centered then the pulley will add a periodic motion. The roundness of the pulley and of the platter where the belt or idler rides as well as the eccentricity of the platter contributes to this periodic motion. So just like a record that has an off center hole moves back and forth during play, the motor pulley and platter are doing this too; but hopefully at an almost undetectable amount. Therefore, the motion of the platter- any platter is not perfectly constant during a single rotation. Being periodic, meaning that it slows down and speeds up, it can still hit the 1.8 second mark every time because the average speed is 33 1/3. You can see an eccentric record moving back and forth during rotation causing the tangential speed at the stylus to vary. The record is still spinning at exactly 33 1/3; but if it is off by a good amount then the WOW can be heard at every peak of rotation. That goes back to my motor analogy- you can speed up and slow down your car between mile markers and still have the correct average speed. |
One other thing: Look at the specs of most turntables. The WOW & Flutter measurements of most high end tables all seem to be around +/-0.02% to +/-0.04%. Most of these very nice turntables here are showing excellent speed settings to within 0.00X%. It's clear that many companies have come very close to perfection with regard to speed control. WOW & Flutter numbers do not seem to vary much among competitive tables which is I believe due to limits of machining tolerances. WOW & Flutter being the periodic motion of the platter that I described. |
Maybe I am pointing out the obvious here. Our ears don't care about average speed. Our ears hear the minute variations in speed. |
It is a Series III from the 90s. I think that is good advice. I will check with Sota. I had the impression that only the Cosmos features a sub-chassis mounted motor with the platter and tone arm. I also have some ideas that would be reversible, to try out on my tt to fix the horizontal axis.
This thread has been a good learning experience. I bet a number of people suspected suspended tables didn't hold speed as consistently but didn't really understand why. I knew before that any movement of the turntable due to footfalls or just pulling on the cue lever affected speed; but I didn't realize til now how much the variable drag by the stylus would affect belt tension and ultimately speed. Plus, the improvement in detail and distortion that I experienced with better speed control is an eye opener. |
Hi Harold,
Cool video. Very cool turntable and tonearm. I watched the video on my iPad and I could hear the WOW. I was supposed to be able to hear it, right? I think it was most pronounced when the stylus tracked over the wave in the record rather than the eccentric motion of the record; but hearing it in person would certainly be better. I think we all own a few records like that, unfortunately. Ironically, my test record is one that has an off center hole. So when I play the pure test tones, the WOW is very evident. The WOW caused by eccentric records is not so obvious to me when playing music- as long as the record is not too bad.
And thank you Halcro for that link. Much of what he wrote I already understood; but he gave me some insight into my Sota turntable. I can see his point about how the horizontal axis being unconstrained is a bad thing on a floating sub chassis. I can see how variations in stylus drag will change the belt load/tension which will cause the sub chassis to move in the horizontal direction. That movement, no matter how slight will alter the belt tension some more. Any change in belt tension will affect platter speed. It gave me some ideas. First, I'm going to fix the sub chassis with some rubber shims and see if I can hear a difference in the rhythm and pace. If I can, then I have an idea how I might constrain the horizontal axis without fixing the vertical axis. |
I had a chance to experiment this evening. I used some rubber shims to constrain the sub chassis on my turntable in the horizontal axis. My Sota has the motor fixed on the main chassis while the tonearm and platter float on the sub chassis. First, I put on my test record to check speed with the iPhone app. Speed was 5 Hz high. So belt tension must have changed. I tried to keep the sub chassis in a neutral, level position. I adjusted speed and checked WoW & Flutter with the app. I saw right away that the FFT waveform is a clean sawtooth pattern now. Before the sawtooth pattern had a lot of hitches in it. I think this is already an improvement. Speed is within 0.025%. That is an order of magnitude higher than Peter's turntable; but pretty good I think, for an open loop motor. The timeline laser mark would move around 10X faster on my table than on some others. WoW & Flutter measured 0.02%/0.03%; typical.
Listening: First, I tried "For Duke". This is a technically near perfect direct to disc record and good music too. Rhythm and Pace seemed about the same to me. The bass seems cleaner and sharper. Should I expect that to be the case?
Next, I put on Beethoven's Appassionata. This is another direct to disc cut at 45 rpm. Sounded fantastic as always. Very powerful. I was hearing some micro detail that I had not heard from this record in years. eg. I could hear the pianist take in a breath just before hitting the keys. Would you expect to hear more detail? Again, I think the lower registers of the piano had more power.
I think overall it is an improvement. It was clearly seen on the iPhone app. So next step for me is to come up with a way to constrain the horizontal axis and leave the vertical axis free. |
I just played "Muddy Waters" Original Master Recording. He convinced me. The rhythm and pace was obviously better to me playing this record with the horizontal axis constrained on my turntable. Plus, his voice on the peaks did not break up. I would have never attributed that to speed variation; but I just heard it with my own ears. I would have posted this sooner but when Muddy Waters is playing, I'm glued to my chair until both sides of the record have played out. |
Yes, my speed was within 1.5Hz of 3150Hz before. One thing I can do is recheck speed after I remove the shims and see if it shifts down. The other thing that I found last year with the iPhone app was that the tt must run for a couple of minutes for speed to settle in. I don't know if the motor must warm up, the electronics or the platter bearing lube. |
I had my turntable dismantled the last couple of days. I found some bushings at the hardware store that slip over the threaded studs supporting the main chassis. The sub chassis has holes for the studs to go through. These bushings that I found were 1mm smaller in diameter than the holes in the sub chassis. This made them a perfect fit. Where the sub chassis used to cycle up and down several times if pushed (underdamped) it now moves back to position in one cycle (critically damped). I was concerned that the sticktion was too much but using a level on the platter; the sub chassis seems to move back to the correct position every time. I have very good isolation of the whole turntable. I had developed the isolation last year which made a considerable improvement. That surprised me since the turntable is a suspended design and I thought was already well isolated.
I put my turntable back together and set it up to play this evening. After dialing in speed I started listening to some records. The iPhone app showed speed was off by 0.1Hz out of 3150Hz. WoW&Flutter at +/-0.02%. After dialing in 33 and 45 rpm I started playing some music. Right away I was bothered by the difference in the highs. I realized that they are cleaner now, which makes the highs like cymbals more apparent. It took me a couple of records to get used to them. The whole character of my turntable has changed. Nothing has been lost as far as I can tell. The soundstage is more airy. The detail and clarity of the highs are startling. Imaging is more focused. I can't say I hear a difference in rhythm and pace. I don't think speed was off by that much before. The improved speed control seems have affected more subtle things about the music. I have now experienced things that several people have said would happen with improved speed control. Constant belt tension even under micro conditions such as stylus drag is critical. I am a believer now. |
I forgot to mention that nearly all play in the horizontal plane has been eliminated. The added bushings took up that horizontal play. Corresponding to this is increased damping of motion in the vertical direction. Like I was saying- before if I pushed on the platter, it would bounce up and down for several cycles. Now it moves back to position in one cycle. |
It is the Dr. Feickert iPhone app. It requires a test record with a 3150Hz test tone, or now the app will also recognize a 1000Hz test tone. I happened to have a test record with the 3150Hz test tone; but Dr. Feickert's test record can be purchased at their online shop.
I think this app is excellent. Just start playing the test tone and the app will react to the tone and ask to start recording. It plots frequency versus time and has a bar at the top that helps you to dial in speed. Once you save a plot the app analyzes the data and gives you average speed, raw WoW&Flutter as well as filtered Wow&Flutter. It is easy to filter out the record's contribution to WoW&Flutter due to eccentricity of the center hole since the period is 1.8 seconds. But the app does all that for you. The filtered plot shows you raw and filtered frequency over time. For example, my record gives a frequency plot with the peaks going from 3160Hz down to 3140Hz (my test record center hole is not very well placed). The filtered line or the actual tt performance shows shows frequency right on 3150Hz with slight waves about that line. That is the +/-0.02% or -0.6HZ/-0.7Hz variation that the chart shows me.
I tried this app with my CD player too. I had a test disc with a 1000Hz tone. The app showed a perfect flat line at 1000Hz with WoW&Flutter at 0%. |
Just for the sake of argument about calibration of the iPhone's microphone. I downloaded an FFT app a few years ago. FFT is a method of analyzing a noise spectrum. It plots frequency vs. amplitude. I was at an automaker's NVH lab one day a few years back. This is a Hemianechoic chamber for measuring noise levels and frequency on a car. (If you are amazed at how quiet your car is, then I may have had a little something to do with that on one little component. If you are not too pleased with the sounds your car makes, well then someone else must have worked on that model). They had their B&K calibrated test tone source used for verifying microphone amplitude calibration sitting there. It generates a 1000Hz tone. So I turned it on and checked my iPhone's calibration (just for frequency, not amplitude). The app was dead on. These apps utilize the iPhone's internal oscillator/clock for calibration. Frequency accuracy should not be a problem for an iPhone app; but amplitude (SPL) is a different matter. I would not count on the iPhone's mic for exact amplitude measurements. |
I did lose something it turns out. After dark when the ambient noise level was at its lowest, I could hear some rumble. It is faint; but there. So the bushings are transmitting some mechanical vibration from the chasis to the sub chassis. I have some ideas to try out. I can't go back to an unconstrained horizontal axis... |
I overreacted. The last record that I played the other night must have had noisy vinyl. I listened last night and heard no rumble. I cranked up "For Duke" again. This record has dead quiet vinyl. Sounded great- saxophones and trumpets all over the room. |
