You could replace your DD with a digital front end. Some of them have very low jitter now. Like DD's they are only a little bit out all of the time.
Showing 50 responses by dover
Lewm - he didn't say elastic belt drive, he said belt drive. My silk thread "belt" does not stretch, the ac motor doesn't cog, and I dont get belt creep whatever that is. According to both the Timeline and KAB the speed on my deck is rock steady. I have dispensed with both a 301/401 so I know what those idlers sound like - musical but far far away from state of the art in resolution & transparency. I agree with you they are all flawed, but which has the least flaws. |
The belt creep that Mark Kelly expounds assumes the belt stretches and thins slightly, ergo if the belt doesn't stretch then there is no belt creep. Japan. Big and round. There are lots of ways to reducing or eradicating cogging, varying the ratio of slots to poles, angled slots or stators, variable drive applied to the motor windings, running the motor at high speed etc In the Final Audio it uses an AC synchronous motor with precisely controlled regenerated sine/cosine waves for the motor and variable voltage regulation to optimise the torque applied to the moving high mass/high inertia platter. I can actually dial in cogging via the torque controller if someone likes to listen to that. |
Mosin - please re-read my post. I did not say that elasticity causes the belt creep. I said it is an assumption and that if the belt doesn't stretch then there is no creep. Mark Kelly's calculation of belt creep is : Creep = T/r x A/E If the belt has no elasticity, then the Elastic Modulus (E in the above calculation ) is ∞ ( infinity ), and Belt Creep will be effectovely 0 ( zero ) exactly as I said. To give some examples of tensile strength : Rubber 15mpa Human Hair 380mpa Silk 1000mpa Aramid fiber 2557mpa These are single fibres only. Here are the calculations using Mark's example, and assuming the cross sectional area is arbitrarily 1/10th the size of the rubber belt. T R A E Creep 1 5 10 50 0.04000 1 5 1 380 0.00053 1 5 1 1000 0.00020 1 5 1 2257 0.00009 On the subject of thread drive TT's, they do have to be designed properly, as in the Final Audio. The Final Audio uses an AC synchronous motor with precisely controlled regenerated sine/cosine waves for the motor and variable voltage regulation to optimise the torque applied to the moving high mass/high inertia platter. In addition to this the pulley profile must be designed for a thread rather than a belt. If I recall correctly the pulley should present a concave hemisphere to the thread. |
Mosin - no worries. The silk thread I use only seems to slip fractionally on start up, but the 20kg platter can be up to speed in less than 1 rotation, depending on how I tension it, which I think is pretty good considering the pulley is only about 10mm versus the 20kg platter. This suggests the silk has very good grip. With the high mass platter it does drive the motor pulley as well. If I turn the motor off, the platter/thread will drive the motor for a while. The designers original intent was that the silk should be "chalked", I presume for a controlled slip, but I prefer direct. The silk lasts a long time as in can go for a year or two. The only breakage I've had in 20yrs is when I've moved the deck and once when I went on holiday for 6 weeks and left the TT running ( I think high humidity weakened the silk fibre ). Timeline speed checked on playing records inside/out etc and it's rock solid. These things all come down to the quality of the implementation, whatever direction is chosen. |
Spirit of music Where are the specifications that suggest the Transfi has high torque. On the contrary, from their website - "Note1: The 9kg platter has immense inertia & will take a few seconds to respond to the control knob. Only very slight tweaks are needed. Note2: You will find speed drift may require frequent minor adjustments until the bearing, motor & electronics are burned in. Typically, starting from cold the system will run slightly fast. As operating temperature is reached it will slow. A suggested method of operation is to let the system warm up then set the speed. Play several records & fine tune until the speed remains constant.....then leave the control alone. When you switch the system off, the motor does not need to be disengaged from the traction belt. The next time you start the system from cold, it will run fast. If you try to adjust the speed at this stage it will slow as the system warms, so best to let it run fast until it settles to the correct speed. IMPORTANT The whole system of speed control is finely tuned, & any variation in consistency may require adjustment, from VTF of the cartridge to the weight of record clamp used. Sophisticated motor speed control methods have been ditched in the case of Salvation as these all proved detrimental to the sound using this direct rim drive method. In practice, speed drift on Salvation is not an issue providing you are not varying the way you play your records from record to record or are a stickler for speed accuracy! Please consider this before ordering." If you have a temperature controlled room, the angle of your dangle is stable, and you are not a stickler for speed accuracy, then this is the turntable for you - this is their own words. The Michell Gyrodeck is not a paragon of TT speed accuracy that should be used as a standard either - tiny motor, stretchy flimsy elastic belt, floppy suspension and all. |
Richardkrebs - Interesting post. Could you please elaborate on a couple of points. Contrary to some opinions expressed in these forums, the servo does not intervene unless an external load causes a speed change. ( this assumes that the TT has been put together as the designer intended). the motor does slip back in phase due to stylus drag...This can be seen by scoping its power supply. And watching the current draw. It was actually possible to recognize the music being played at the time on the scope. The speed sensor was measuring these tiny errors and correcting. This could be observed by watching the motor draw current in sync with the music. An amazing observation. The conclusion is that some form of speed correction is required if we are to seek good dynamic speed stability. So as I read it Para 1 - The servo only cuts in when speed changes Para 2 - Motor slip is occurring as a result of varying stylus drag to a degree that you can follow the music by scoping the current draw Para 3 - Your conclusion is speed correction is required for dynamic speed stability based on your observations. These seem contradictory. Are you saying that most direct drives dont speed correct all the time, but that they should be correcting more frequently for variable stylus drag ? Isn't speed correction for stylus drag shutting the door after the horse has bolted ? If the current draw is up AND down with the music, when do you propose to correct the speed UP and DOWN before, during or after the event ? Doesn't this just suggest that for whatever drive system/platter/arm/cartridge you have it must have enough energy storage or overkill in terms of motor drive & inertia to ensure the variable stylus drag does not impact speed at a micro level. Are you not presenting a strong argument that all TT's should have a very high mass platter ? ( unless of course you can scan the grooves and preprogram the required changes to the motor controller to preserve micro dynamics ). The other question I have for you is torque ripple, or cogging. As I understand it the torque ripple or cogging torque will vary because a Direct Drive motor is running much slower than a small pulley/platter. For example DD - 33.33rpm x 20 poles = 20 poles per revolution of record. AC/belt - 1800rpm x 4 poles = 216 poles per revolution of record. Disclaimer here - I know you also have to factor in the number of slots, which could increase or reduce the differential in this example, but doesn't eliminate it. Cogging torque results in torque as well as speed ripple; however, at high speed the motor moment of inertia filters out the effect of cogging torque. Presumably a high mass platter, provided it is direct coupled to the pulley with an inelastic belt, can assist in smoothing out any torque ripple if it exists in both instances. Can you explain why I can hear such a big difference in speed stability between the Technics SP10mk2 and SP10mk3 ? Do you think it is the higher torque, superior motor controller, higher mass platter of the mk3 or all of the above ? Cheers, Dover |
Mosin - Absolutely right, I used mass loosely. I read somewhere the EMT idler has a platter of 5.5kg ( same weight as the copper mat on my deck ) but due to its weight distribution ( 16 & 1/2" platter ) is the equivalent of a 50kg solid platter. Although not as radical my TT has an inverted bearing with a substantial subplatter which is an inverted truncated cone (frustum), so most of the mass of the 20kg platter is between the inner and outer grooves & the thickness increases inside out ( and below the bearing point ). So yeah the effective mass of my TT platter is well in excess of 20kg. |
Richardkrebs Yes I agree on moving away from which is best to what do we need. Mosins previous posts on motors used in various idler drives for example would indicate to me that you cant even put Garrard 301/401's in the same camp as say the EMT 927 - I'm not saying one is better, more that the motor drives are completely different in motor characteristics, use of motor flywheels/eddy brakes, distribution of mass in the platter etc. So the fact that they are both idlers is all they have in common. With regard to my stylus drag testing I wanted to test two things - 1. Does my TT deal effectively with stylus drag 2. How accurate is my KAB speed checker ( assuming the Timeline is more accurate ). I have always set the speed with a record playing using the KAB speedstrobe. Some folk on the forum reckoned they couldn't see stylus drag with the KAB. So my test procedure was : 1. Set the speed with the KAB with no record playing. 2. Use the Timeline to validate the speed at both inner at outer grooves. 3. Reset the speed again with the KAB with the record playing tracking at 2g 4. Use the Timeline to validate the speed at both inner at outer grooves. In both instances the TT passed the Timeline test and also confirmed that my KAB was accurate. The Timeline indicated no errors when using the KAB. As for consistancy of stylus drag - presumably stylus drag has record groove modulation, stylus pressure, antiskate forces all contributing. All I can say is with my TT the inner/outer grooves made no difference as verified with the Timeline and with the KAB, running the record as far as it would go, the KAB remain rock steady. I suppose the next step might be to set the speed using a flute concerto and check whether you get through a Mahler Symphony or Reference Recordings Dafos on time without speed correction to quantity that differential. The point I take from your Goldmund example is that you are suggesting that mass alone will not provide stability with a motor that simply doesn't have enough torque to start with. |
Halcro - I think your test methodology would prove that only turntables with error speed correction built in will pass. We know that stylus drag exists and should always set speed with the stylus playing. The question then becomes how significant is stylus drag variation, should we speed correct, and if so how. Remember also that the timeline is only a sampling of 1 per revolution, what is happening in between that interval is unknown. You could have Turntable A that has a much bigger variation when dropping the stylus than turntable B but has a better instantaneous speed correction. Which is better for sound reproduction ? We also know that the natural recovery of AC motors and Brushless DC motors if the motor "slips" due to back emf and goes out of phase slightly that the recovery is different - some say the AC motor recovers sinusoidally whereas a DC motor tends to recover trapezoidally. Can you measure this ? What difference does it make to the sound ? Lots of unknowns here. |
Halcro/Tony I've just done the Halcro drop test of dropping the stylus on and off without adjusting the speed. The fall off in speed is minute. I would estimate it would take a couple of hours at least for the error to translate to 1 complete revolution. So according to my rudimentary maths for my TT 1.8seconds per revolution/7200 seconds ( time to complete 1 rotation error ) = a speed variation of 0.025%. Even an hour would only be 0.05%. ( Tony correct me if I'm wrong on the maths, my brain is disfunctional at the moment ). With the speed adjusted for the tracking weight there is 0 drift. To me the big question is how consistant is the impact of the stylus drag on speed and if error correction is used, is the bark worse than the bite. This will no doubt come back to the quality of the particular solution offered. |
Richardkrebs Can you please review the video again. From 0-24 seconds you can read the 33 quite clearly, but the 33 is moving forward and back, it is not stable. From about 57 seconds on - he now has a constant load on the platter. 33 is unreadable, but it looks to me as though the surging has ceased. The conclusions I see from the video are : Unloaded, it is sort of speed stable but "hunting". Can you not see this ? Loaded - appears to me to suggest this deck is not accurate with a load ( blurred 33 ) but ironically the long term speed variation seems to have improved with the finger load on it ( less drift ). To be fair to the deck, the load on it probably takes the deck outside of its error correction operating parameters. |
OK just for fun here's something to try on a Linn/Oracle/Clearaudio...take your pick http://www.youtube.com/watch?v=lwugFlbCOww I've recorded the KAB speed strobe on my Final Audio Parthenon VTT1 thread drive TT while I bash the record with my knuckle. Anyone else want to have a go - I used a Carol Kidd record.... from Linn. |
We all know the TT in Halcro's video was the Transrotor Fatbob. I dont think it is representative of a true "high mass" or "high inertia" TT for the following reasons - Platter is only 11kg - even Brinkmann suggest minimum 15kg Drive belt is elastic rubber - should be thread or inelastic material Most high intertia TT's use a small pulley and drive the platter at the perimeter. Because of this gearing ratio it enables you to run the motor much faster and reduce the cogging effect by increasing the number of poles per revolution. eg HM 1800rpm/4poles = 216 poles/revolution vs DD 33.333rpm/20poles = 20poles per revolution. The Fatbob is driven by small pulley, small subplatter which means the poles per revolution will be significantly less than the 1800rpm scenario, but more than the DD. So the Fatbob to me is an inbetween deck - mid mass platter , low speed motor, elastic drive and no speed correction. What is really disturbing is that despite what was demonstrated the magazine reviews describe its superior ability on timing rhythm and pace. When I cued up the same track on the Fat Bob, the recording became even more believable. The timing, rhythm, and pace were improved over the Leonardos.Obviously the reviewers system went from really bad to just bad. |
Halcro - Breaking newsflash - This Fatbob could have the magnetic drive, where the belt drives a subplatter and it relies on magnets on the platter to "lock" to the attracting magnets on the subplatter. They say it reduces wow and flutter by a factor of 10. Personally I would call this floppy drive - maybe thats the issue, certainly not as direct as a DD, idler, or thread drive.. http://www.positive-feedback.com/Issue50/transrotor.htm The main difference is the heavy platter in comparison to the Linn/Rega/VPI. The VPI's are driven around the perimeter of the platter, so they run higher speed motors. As far as motor for the Transrotor goes, I couldn't find any specs on the Transrotor whether its ac or dc. I assume no error correction but you could measure the back emf in the motor and use that to control speed, but there is nothing measuring the platter. As Richardkrebs says you have to engineer the motor/drive system/platter as a group. From what I hear in the video, my old Roksan, Sota & Townsend Rock TT's, which use rubber belts, would be more speed stable than the Fatbob. |
Lharasim - I did read your response and had another look at the video. The "lurching" or "hunting" that I see occurs in the 1st 19 seconds before you put your finger on the platter. Did you have another finger on the platter out of sight ? From the lurching and hunting could I guess the deck might be an old Technics DD - maybe an SL1200 or early SP10 ? |
Lharasim - My suggestion then, based on what I see in the video, and given the age of the deck, would be to get the power supply checked over and ensure it is in spec, check the electrolytic caps which can dry out. In my experience decent power supplies contribute hugely to the sound of a turntable, not just speed accuracy. For a few dollars you may get quite an all round improvement. |
Richardkrebs, A simple dynamic load test you might want to try if you havent already done so, is to hold your finger against the rotating platter while playing music.I'm not sure how relevent this is - the force reative to the incremental change in stylus drag has not been quantified. In my earlier post I quantified the impact on speed due to the stylus drag differential between stylus not playing versus playing, with my high mass thread drive. The Final TT has a 20kg platter with offset weight distribution, it is around 3920kg/cm vs your 1100kg/cm of the SP10mk3 (10kg platter). Most DD's would have substantially less than your SP10mk3. The speed change with and without the stylus playing was approx 0.008%, so presumably if there were variations in the drag due to musical peaks and troughs then the variation it would be a very small small proportion of this. You would have to agree this is tiny. So the issue becomes what is happening at a micro level. The thumb test is not measuring this micro behaviour. To me the nub of the issue is what is happening momentarily. If the direct drive has less inertia than say a high mass thread drive, then it may have a bigger momentary drop in speed, but its recovery might be quicker if properly designed. In this scenario we clearly have big speed drop/quick recovery versus small speed drop/slow recovery. Now an interesting comparison here is your old Goldmund Studio versus the Kenwood L07D - both DD with speed correction, but with different feedback behaviour. Did your old Goldmund Studio pass the thumb test. I doubt it very much. You could hear the poles ( 8 pole motor from memory ) as they slewed past the start line. The Kenwood L07D blitzed it in timing and microdynamics. Ironically the Goldmund would probably pass the Timeline test as would the Kenwood due to the speed correction. As far as music reproduction goes they were on different planets. Interestingly if I recall correctly the Kenwood trumpeted "Transient Load Fluctuation" in their manual as opposed to "wow and flutter". Again the method of drive/ control comes into play here and we each need to decide which does the least harm.Here you seem to have shifted your stance from an earlier post of championing good design to one of "choosing the method of drive that does the least harm". I would argue that it is the quality of the design irrespective of drive as in the Goldmund/Kenwood example above. There are other factors in speed. I tend to judge a system by its speed and coherency. Many systems are not that quick or coherent and the speed and coherency of the front end is masked somewhat. As Tonywinsc has quantified eccentric records can have a significant impact on speed and timing. Cantilever design now comes into play. I have a new Koetsu Black Goldline ( shelved ) and a friend has a new Koetsu Rosewood - both of them slew around on eccentric records. Having had many Koetsu's over 30 years I suspect they have screwed up the cantilever design integrity somewhere along the line. Mismatched arms/cartridges in terms of compliance mismatches may well skew speed and timing. Interestingly Bruce Thigpen argues that his Eminent Technology produces less distortion on eccentric records than a conventional pivoted arm. Stylus shape may well have an impact on speed and coherence. There is an argument that a spherical stylus will have less phase errors than a fine profile tip. Phase errors will skew timing. Tangential tracking versus pivoted arms - again the tracking error of pivoted arms may skew phase & timing. An example of this was my old Sota Star/ET2/Denon 103 Garrott - very very good speed and coherence despite the rubber band drive. Yes the motor regulation had been worked over to reduce cogging somewhat. But I am resonably sure that that front end had much better timing than say any Technics SP10mk2 or Goldmund Studio I've heard and I believe that the sum of the parts including TT, tonearm and cartridge all come into play if we want to talk about musical timing and coherence. Let's face it no one wants DD done badly, Belt Drive done badly or Idler done badly and unfortunately most of them are, irrespective of cost. To use a hackneyed phrase - there are no free lunches when it comes to good turntable design. |
Richardkrebs, Just a little plug here. Customers of both MK2 and MK3 upgrades have universally noted a marked improvement in percieved speed stability. How is this possible ? For the last 5 days you have argued that the SP10mk3 is impervious to stylus drag and to substantiate this you quoted the manufacturers specifications. Do you think the specifications are erroneous ? Are there other speed stability issues, not related to stylus drag that you are aware of ? |
Lewm - same experience, but from the motor & power supply mods we did, the cogging in the early motors on the SOTA was significant. According to our tech the "cogging" was induced by shortcomings with the power supply regulation on the papst motor board, not the traditional cogging as in poles/slot rotation of the motor itself. That being said, the motor/platter should be rigidly coupled for stability, and obviously SOTA addressed this issue later by mounting the motor on the same chassis as the platter & arm. After the SOTA I temporarily used a Townsend Rock with the ET2 and the stability of the fixed plinth design was apparent, despite the sloppy build quality of the early Townsend. |
Richardkrebs "The perfect TT has absolute speed stability under all load conditions" I would draw your attention to the following : Newtons Second Law of Motion - Force = mass x acceleration or Acceleration = Force/mass In other words deceleration due to an external force is inversely proportional to mass. The moment of inertia for the following TT's are - Final Audio high inertia TT - 3920kg/cm2 Kenwood L07D - 1025kg/cm2 Technics SP10mk3 - 1100kg/cm2 This means that the high inertia Final Audio TT requires a 300% increase in stylus drag variation compared to the L07D and SP10mk3 to achieve the same loss of speed. In other words for a given additional force due to stylus drag variation the Technics SP10mk3 and Kenwood L07D will have a deceleration 4 times higher than the Final Audio high inertia TT. Now if we look at the design intent of both the Technnics and Kenwood we find some interesting facts : Kenwood L07D Manual Clearly the Kenwood designers believed that the magical figure for moment of inertia of 1000kg/cm2 was sufficient to overcome stylus drag fluctuations. Now let's look at the Technics SP10mk3 Sales Brochure Technics SP10mk3 Brochure So the Technics designers have a foot in both camps. On the one hand they believe that a moment of inertia of 1100kg/cm2 will support 1000 tonearms tracking at 2gm but then they go on to say that the fast error correction helps to maintain speed. Furthermore the Technics SP10mk3 Brochure also explains that purpose of the high torque motor is to ensure for that the platter can be brought up to speed within 0.25seconds and reverse torque can stop the platter in 0.3seconds to meet the requirements of Radio Stations. With regard to your comments on the Techdas, I would draw your attention to the design. Although the Techdas has a 30kg platter, the motor drive is low torque. They deliberately drop the torque as low as possible just to keep the platter spinning accurately and no more. They have done this to minimise noise and vibration. This is explained on their website. To reiterate my earlier posts on cogging - The torque ripple or cogging torque will be lower on a high inertia thread drive because a Direct Drive motor is running much slower. For example DD - 33.33rpm x 20 poles = 20 poles per revolution of record. AC/thread drive - 1800rpm x 4 poles = 216 poles per revolution of record. In summary then, the benefits of a high inertia thread drive over direct drive are : Final Audio High inertia TT has a resistance to stylus drag variation 4 times higher than the SP10mk3 & Kenwood L07D Final Audio High inertia TT has less cogging effect ( and lower amplitude ) due to the higher motor speed. Speed correction is required on Direct Drive TT's due to insufficient moment of inertia which is required to resist to transient load fluctuations. |
Richard 1) The 0.0008% speed impact from stylus drag I measured on the Final is insignificant in the context of the other issues raised such as eccentric records etc - the ones you believe are spurious. The measured speed change is less than the wow and flutter quoted for both the SP10 and L07D with their speed correction engaged. 2) I recall you said it was a myth that the servos are operating frequently, and I assume they only occur after the drop off in speed, which will initially be higher as you acknowledge. 3) Cant comment on cogging as I dont have enough knowledge other than I can hear it on most DD's. You have acknowledged that whilst they may occur at different frequencies the magnitude is lower with the higher speed motor. 4) Pass 5) I dont believe much of what I read. Having studied engineering at the University of Auckland it is easy to identify the contradictions within the brochures. For example, as you have alluded to, there is no way the Technics would support the load of 1000 tonearms without the speed correction being employed. 6) No doubt there are many paths to enlightenment, but you haven't convinced me that speed correction is required on all turntables, which was your original assertion, based on your observations of the performance of a Goldmund Studio which I believe is a very poorly designed turntable. Yes, thank you, I have enjoyed the Final for over 20 years. It has not required a new plinth, redesigned power supply, additional mats & clamps , or any further expenditure to address inherent design issues. It continues to be state of the art some 40 years after production commenced. |
Lewm - in my posts I have tried to relate what I hear with the design etc. Some folk misread stuff. I am not anti direct drive or any drive. What I have heard in very good systems is that there are very few DD's that dont sound awful, by which I mean grainy, grey, music dissembled, fragmented, a lack of dynamics and transparency. The only ones I've heard that dont sound awful are the SP10mk3 and L07D. The Goldmund I heard had a rebuilt and upgraded power supply, many other upgrades and was in very good state of tune. The standard L07D demolished it - demolished it. Demolished it in terms of speed and timing. The Kenwood sounded more lucid, crisper and cleaner. I use this example to illustrate that in my view if you are running direct drive it needs to be at the level of the SP10mk3/L07D at least or you may be better off with an idler or belt drive. Ironically the Goldmund would probably pass the Timeline test, so the Timeline test is very valuable to give you assurance that the basic speed is accurate and maintained, but then we get into the dirty unknown. What happens in between at the micro level. I have made a statement supported by physics that the high inertia of the Final will result in deceleration due to stylus lag variation of 1/4 in magnitude of that of the Technics SP10mk3 and Kenwood L07D. Furthermore results that quantified the stylus drag with and without the stylus playing are at a level less than the wow and flutter levels of the Technics SP10mk3 and Kenwood L07D. A counter argument has been proposed that properly designed servo speed correction is a fundamental requirement for any TT. No numbers or empirical evidence have been put forward. This is the nub - both the Final and the SP10mk3 are 2 of the best TT's made, they are at or near the top of the tree - so the question is what if you cant afford or obtain the top of the tree? What compromises should you make if you are going to make any ? Most think that what they have is the best. My nature is to try and understand why things sound the way they do. I have been able to improve my system by doing this - understanding what I have, its design construct and selecting products that achieve the goals which for me are important, speed, timing, coherency, harmonic structure. I personally dont have perfect pitch, but I'm extemely sensitive to timing - if the timing is off, I cant stand the music, fall asleep, lose interest. The posts here that have references to gear I might have heard, hard facts are valuable and enlightening. Recent examples are Halcro's speed testing and Mosins' debate on belt creep which led me to study up previous posts and discussions on vintage idler motors, which was informative. Folk shouldn't get offended by posts - just get pissed off enough to go away and do some research. We all learn from this process. |
Richardkrebs : Let's assume we are tracking at 2.0g. Now let's break down the stylus drag components into 2 parts - one being stylus drag due to the tracking force ( A ) and the other being the variation in stylus drag due to the variations in the music ( B ). I measured A, not B. A - Stylus Drag Due to Tracking Force ( assume constant ). When you drop the stylus into the groove the high inertia TT drops 0.008%. ( By the way - this was a very conservative estimate to avoid debate, it's actually less ) Now I believe the Technics will drop by 4 times this if the servos are not on due to the lower inertia. The Technics maintains speed because the servo kicks in. This is no different to me adjusting the speed via the controller to account for the constant drag...except for the following - When I bring the speed up to adjust for the 2.0g tracking force I bring it up manually until it the speed is stable. The Technics servo will react by increasing the speed. The catch here is that the servos only react to errors or changes in load and they dont know the magnitude of whats coming, so they ramp up the response until such time as they detect the speed is too fast - they overshoot. Then they have to correct for the overshoot. This becomes a constant cycle of overshoot and undershoot. You might not be able to measure it, but it is there. B - Stylus Drag due to Variations in Music I cannot measure any stylus drag with the high inertia TT. I checked for variations on both inner and outer grooves. I assume if it exists it is too small to measure. Same argument exists as in A above, if there is any variation then your DD servo will kick in according to some preprogrammed parameters, and once the servo kicks in, again you end up with micro overshoot and undershoot - too slow, servo, too fast, servo, too slow etc Now we can argue until the cows come home a to which is better or worse, and which cow comes home first will vary depending on the design and quality of the individual design. Wow and flutter versus Stylus Drag I would give equal weight to these. If I am listening to a piano concerto, I want to hear the attack, intensity and decay of each note as well as the ebb and flow of the performance. I would also suggest that to reproduce the attack, intensity and decay of the note, then the record/platter interface and platter/bearing/plinth design and how it deals with excess energy and maintaing a rigid loop will have a major bearing as well as stylus drag. Certainly the designers of the Final believe there is significantly more energy generated than that required to move the stylus which needs to be dealt with effectively, whilst maintaining a rigid loop between cartridge and record to measure the groove accurately. Removal of this excess energy in my experience is analogous to lowering the noise floor - increased resolution, less smearing of notes. Re the cogging : I cannot be sure that the instability I heard is cogging. I have described what I heard a the response to Lewm which was posted prior to your last response. It may be the servos, lack of inertia or other issues, but I would add to your comments that cogging can be induced by poor power supply design if the current waveform driving the motor is not maintained accurately as well as the motor itself. I have seen a number of TT's where the power supply regulation is poorly designed and literally turns on and off whilst playing. Cheers. |
Richardkrebs I seemed to have missed the olive branch. Was it before or after "end of story"and "for the last time". I have continued the discussion as you have misconstrued some of my comments and not fully addressed some of my queries. I cannot understand how the servo/speed correction system knows whether you are going to play Mahlers 2nd Symphony or a Beethoven Sonata. You seem to be certain that there is no measurable stylus drag on the DD by measuring the power supply. This simply doesn't quantify the the stylus lag in real time that is occurring at the stylus tip. The comments on loop rigidity and energy dissipation were put forward in the context of maintaining the attack, intensity and decay of each note. You have overlooked the fact that even if you had perfect speed stability, the attack, intensity and decay of each note can be distorted by an inadequate plinth that is not rigid and doesn't deal with the energy reflected into the platter. It is of no consequence to me really, but you assumed I measured the variation in stylus lag using the Timeline. That assumption is not correct. Finally, I am trying to understand how Direct Drives address these issues, but you have offered no explanation for the differences I heard and described between the Goldmund and the Kenwood L07D, particularly in speed, timing and coherency. |
Halcro - I'm a bit like you - the Raven doesn't look that bad. When you study their website, the motor supposedly has some accuracy, and the belt is elastic, so presumably the steepness of the initial drop will be due to the elasticity. I have listened to the Raven - but not played with it, I gather some have said that the speed issue is more around that the adjustment is stepped, and you cant get it spot on ?? Techdas on your shopping list for this year ? Retro EMT idler will go nicely with the classic Porsche. You do need more TT's for all those arms dont you. |
How about this ... very retro....very rare Fidelity Research idler....as Syntax would say the quiet rumble of the London Underground on every record can be comforting... http://www.hifido.co.jp/KW/G0301/E/450-10/C12-65456-39066-00/ |
You can fix the off centre hole by checking the run out and carefully open up the hole with a needle file to centre it. Then you simply mark with a pencil on the label which part of the elongated hole butts up to the spindle for next time. Option 2 is to drill/mill the hole oversize. Then centre the record to minimise runout. You can make it easy by making up a washer with a spindle hole, and once you are happy that you have found the true centre, then push the washer over the spindle and glue it to the record. Now you have a record with a true centre. PS As Tony has pointed out, dont assume the record is round, it isn't. You must play the record and work out how far it's out by measuring the lateral movement of the stylus. PS2 : As this advice is free, there is no responsibility on my part if you screw up. |
Halcro - exactly that. If the hole isn't in the centre, then drill a big one. Position the record so that the arm is not moving in and out from the centre. Definition of run-out - oxford dictionary - noun slight error in a rotating tool, machine component, etc. such as being off-centre or not exactly round: Have you been watching too much cricket ? |
Lohanimal, It is more a function of the error correction systems used in direct drive TT's. Like digital, they have inherent jitter that destroys the time and phase relationships of the music signal. This results in soundstage aberrations as well as a lack of resolution due to smearing. The designers of the SP10mk3 & Kenwood L07D tried to minimise this by running higher mass platters, but it is still there. |
In_shore - The centre disc method you outlined is flawed. It assumes the record is round, which often is not true. Therefore the diameters are not consistant. The groove is what is round, not the edge of the record. The file method I suggested in the speed accuracy thread was trialled by Tonywinsc. 11-26-11: Tonywinsc |
Richardkrebs Re Your post on the 24/1 I have alluded to another problem with speed stability in DD TTs that is unrelated to stylus drag or cogging... Thanks for your illuminating post. The jitter you are referring to is one of the fundamental shortcomings of the Technics SP10 that other manufacturers have taken pains to design out. It is all too obvious for those of us with a quality turntable that the Technics possesses jitter that would make its purchase unacceptable. I am all too aware of the Technics jitter because my Final simply does not have it. Your attempts to remove the jitter from the Technics is laudable. |
Richardkrebs The jitter artefact, as Mosin quite correctly points out, is not limited to the Technics line, he hears it in DDs in general. Actually my testing and listening has shown it to be present in any motor that uses feedback be it local feedback ( self correction) as in free running motors like those in most BDs / idler drives or global feedback as in DDs. That is it occurs in all motors. Since all TTs have motors ...... This statement is an oversimplification and a little disingenuous. You are conflating apples and oranges. There are 2 areas of instability 1. Any slip in the motor. 2. Any instability induced by speed correction and servo action. You are grouping 2 issues together, which have quite different outcomes in terms of how they impact a turntables speed stability and sound. Jitter is associated more with instability due to speed correction and servo action. Motor slip and its self-correction is much softer and more benign than jitter induced by speed correction and servo action. 1. Motor Slip Local feedback is softer and slower to respond to demand, it creates a slurring effect. The music is slightly blurred like a soft focus lens. Drive and dynamics are diminished. I would agree with this statement, but it should be qualified in that the amount of slip and its impact on the sound will depend on the quality and type of motor, its power supply and the inertia of the platter. One of the fundamental differences between AC synchronous motors and Brushless DC motors as used in the Technics SP10 is how they lock into speed. AC synchronous motors are dependent on the frequency of the drive current. In the Final Audio TT sine and cosine waves are generated from a motor controller and are not dependent on the stability of the mains supply. In Brushless DC motors, the stator windings are energized in a sequence. The sequence and timing is determined by the rotor position. Therefore accurate measurement of the rotor position is required for speed accuracy. In a perfect world these would sound the same, but the world is not perfect and they dont. My research into motors has shown that this can be reduced only slightly. . This is a generalization - how the motor self corrects will vary considerably depending on the motor design and how it is used. For example AC Synchronous motors and DC Brushless motors can behave quite differently. AC synchronous motors tend to be sinusoidal; DC motors tend to be trapezoidal. Trapezoidal motors give a back EMF in trapezoidal fashion and the sinusoidal motor‟s back EMF is sinusoidal. In addition to the back EMF, the phase current also has trapezoidal and sinusoidal variations in the respective types of motor. This makes the torque output by a sinusoidal motor smoother than that of a trapezoidal motor. Similarly the self-correction in a sinusoidal motor will be smoother than a trapezoidal motor. The method of coupling the motor to the platter is immaterial, as this effect is built into the motor itself. You are wrong. If a platter is direct coupled to a motor then the platter will act as a flywheel, with inertia. This will increase the resistance to drag induced motor slip and the behavior of the self-correction by the motor to recover will also be variable, for example higher inertia slower recovery. Factors that will influence the motor slip will be the design and implementation quality of the motor design, power supply, inertia of the system etc. You have not provided any comment on the magnitude of any instability due to motor slip you measured. How many motors did you measure? What types of motor did you measure? What parameters did you measure? How did you measure them? Did you use your thumb and pencil method you recommended in earlier posts? 2. Any instability due to speed correction and servo action. This is what I hear. I agree with this, it is, I assume, what I have heard from most DDs I have heard, including the SP10. It is less apparent in the L07Ds that I have heard. The music does not flow and does not properly connect with the listener. It is not servo overshoot, hunting or cogging. This sound is what the BD people talk about. This problem, in my opinion, can be largely removed. I agree with the hunting and cogging these issues have quite a different impact on the sound, musical timing and instability, particularly noticeable in the bass range, for example the Goldmund Studio. With regard to loop control feedback systems as used in most DDs you seem to be in denial of system overshoot. As far as I am aware most feedback system will have input parameters based on some sort of modeling. Usually algorithms are used requiring control inputs based on measuring past errors, the present error and a prediction of future error based on the current rate of change. Past errors (Integral) are included to accelerate the process and remove any steady state errors in the present error (or proportional if you want to get technical). Since the integral term responds to an accumulation of errors from the past, it can cause the present value to overshoot the set point value (where you want to get to). In terms of correcting the present error, if the gain is too high you get instability, too low and you get a less responsive controller that may not deal with fluctuations. The prediction of future errors (called the Derivative) is required to minimize overshoot. So, summarizing this then these inputs have to be balanced to maximize the correction and minimize instability and overshoot. There are always tradeoffs. For any given Speed Correction Feedback Loop you should be able to measure or calculate the following parameters: Rise Time Overshoot Settling Time Steady state error Stability Since the Technics was built there have been significant advances in technology. We now have computer simulation software to model motor behavior and generate more accurate algorithms, from which the input parameters are derived. There are now faster chips and amplifiers for driving the motor PWMs are switching at a trillionth of a second today. The problem is the Technics is full of obsolete chips and trying to optimize the feedback loop would be like trying to tune a car that has a very basic engine management system. You would probably be better to throw the boards away and start afresh. There have also been newer methodologies to apply global feedback, which are the subject of patents. I think from memory Fujitsu has some. For the DIY person because the whole feedback loop performance is based on accurately measuring the rotor, eliminating any flex or instability in the physical motor could in theory help to reduce the issues endemic in feedback loops. You have claimed that you have managed to reduce these issues related to speed correction/servos. It would be helpful to see some hard data measurements to quantify the level of improvement. I would also be interested to know if there are any turntables available commercially that do not have speed correction/servo induced issues. |
The myth of Direct Drives and Lathes Direct Drive devotees often refer to cutting lathes and make claims about the suitability of DD's for replay based on their use in record cutting lathes. I would like to point out the following : The Technics SP02 is considered one of the best DD lathe motors. Yes it is Direct Drive, but in most cutting lathes the platter is 70lbs or so ( Neumann for example ) and designed to act as a flywheel - more weight on the outer. So in other words cutting lathes are usually using very high inertia platters to provide stability in conjunction with whatever drive system is chosen. Most of the best classical music in my view was recorded before the mid sixties, when multi miking became popular and the engineers took over the asylum. Most music was recorded on shaft driven or belt driven ( some Scully lathes for example ) lathes with very high inertia flywheel/platters. This to me this is a compelling argument for using high inertia, irrespective of the drive methodology chosen. |
Lewm, Lawrence Could I ask you respectfully to reread my post. My post addresses some points made in earlier posts about "motor slip" and "speed correction systems/servos", both of which can be present in any type of turntable or motor. AC motors have been used in Belt Drive,Direct Drive & Idler DC motors have been used in Belt Drive,Direct Drive & Idler Furthermore there are many variants of both AC & DC motors. Speed correction systems have been used in both Belt Drive and Direct Drive. Speed correction systems have been used in both AC & DC motor applications. I use those particular TT's mentioned as examples because I know them and know what type of motor they use. Ct0517 wanted to know more about speed controllers & servos. My post tries to explain that most speed correction feedback loops would include the following parameters - Rise Time, Overshoot, Settling Time, Steady state error, Stability - and any solution has to be a compromise between these parameters. The question is " Do the issues inherent in speed correction/servo loops outweigh the benefits ? " The correct answer is "We dont know". Why ? Because nobody has quantified the error versus the errors generated in correcting the error. |
What about the latest megabuck DD turntables? I think the one from the Northwest (NVS) does not use servo correction. Is it free of the sort of coloration we are talking about? Rumor has it that it does not pass the Timeline test.From the Wave Kinetics website Speed Control Capable of both 33 1/3 and 45 RPM Utilizes a laboratory grade motor control system High inertia platter Speed accuracy to 1 part per 1000000 Direct drive system Tuning optimized using a 17 degree of freedom tuning model In linear regression the degrees of freedom is the number of estimated predictors. Degrees of freedom is often used to quantify the model complexity of a statistical modeling procedure. This tells me that they have sophisticated speed control programmed into their speed controller based on the prediction of errors ( see my post of 02-02-13 where I explained how the algorithms are derived ). 17 degrees of freedom does not mean 17 parameters either - I can send you the linear regression mathematics behind this statement if you like. It tends to suggest that their programs are using statistical analysis, and a continuous shrinkage technique to improve the prediction of error on the fly. In other words, the input parameters for the speed correction/servos are not set in concrete, as in the old days, they are calculated on the fly. As suggested in my earlier comments on the age of the Technics SP10 circuitry, these folk are using much greater computer power and new thinking on statistical prediction that is now available. Caveat : Just so everyone is clear - I make no comment on how this turntable might sound and I am ambivalent as to the drive mechanism used. |
Lewm cc Richardkrebs I dont think the Goldmund numbers quoted are indicative of the sampling rate. I dont think you can count the number of wavy lines on a pcb and equate this to sampling rates. The Goldmund uses a coreless JVC motor. It uses a quartz lock servo. It is more likely that the wavy lines on the circuit board are used to generate a phase signal that is compared to a reference in a feedback circuit. The loop bandwidth of the feedback circuit determines the speed of the feedback loop. It would be more illuminating if Richardkrebs actually explained what type of motor the Goldmund uses, how its speed correction works and why he chose to abandon the TT self build project with this motor. |