I recently purchased a mint condition 1961 Empire 208 turntable. Everything was in excellent condition, sans the belt and rubber mounts. I replaced the belt and rubber mounts. I also re-lubed the motor thrust plate, turned the thrust plate over to the fresh side and inserted a new precision ceramic ball in the motor thrust plate. The platter spindle has a steel ball bearing which needs to be replaced, but I have not done so yet.
Using the iphone RPM app to record speed and wow/flutter (I can not vouche for the accuracy of the RPM app, but its the best I have), I recorded several measurements for speed accuracy and wow/flutter. FYI: I use a Long Dog Audio LDA Quartz Regenerator power supply which keeps a steady voltage. The Long Dog Audio unit allows the platter speed to be adjusted easily with 0.01 increments. So, I was able to lock into 33.33 RPMs and it did not fluctuate during use.
Using the stock platter (aluminum which rings like a church bell), the iphone RPM app recorded a wow & flutter measurement of 0.16 - 0.18%. This range was consistent across 3 measurements.
Then, I added a 60 pound platter from a prior DIY project to the top of the stock platter. The 60 pound platter was centered within 0.001" TIR. With the added mass, the wow & flutter measured 0.06 - 0.08% (again measured 3 times). The only variable between the measurements was the platter weight.
Also of note, with the stock platter and motor off, I spun the platter and it continues to spin for a few minutes. I used a stethoscope to listen for bearing noise. None detectable. In all my years of using a stethoscope, I always detected bearing noise until today. Quite a nice feat for a stock 1961 turntable.
Then, with the 60 pound platter added to the stock platter, I once again spun the platters and used the stethoscope. Bearing noise was clearly audible. The bearing noise was cyclical which indicates that the sleeve bearing is where the noise is occurring, not the steel thrust plate. The sleeve bearing noise was likely caused by a slight imbalance in the platters weight distribution and thus would shift the bearing laterally and create the bearing noise.
So, where does this leave us:
1. 60 lbs incremental platter mass reduced wow/flutter by 50-66%.
2. Heavy platter needs to be appropriately balanced to no longer cause the sleeve bearing noise.
The 60 lbs platter is not suitable for the current setup (see photos), as it was taken from a 17" OD Gates CB500 transcription table. So, I am considering a new high mass platter and bearing.
Meanwhile, the stock papst motor purrs like a kitten, albeit with the fan noise that is inaudible at the listening position.
More to come.
Pat
Using the iphone RPM app to record speed and wow/flutter (I can not vouche for the accuracy of the RPM app, but its the best I have), I recorded several measurements for speed accuracy and wow/flutter. FYI: I use a Long Dog Audio LDA Quartz Regenerator power supply which keeps a steady voltage. The Long Dog Audio unit allows the platter speed to be adjusted easily with 0.01 increments. So, I was able to lock into 33.33 RPMs and it did not fluctuate during use.
Using the stock platter (aluminum which rings like a church bell), the iphone RPM app recorded a wow & flutter measurement of 0.16 - 0.18%. This range was consistent across 3 measurements.
Then, I added a 60 pound platter from a prior DIY project to the top of the stock platter. The 60 pound platter was centered within 0.001" TIR. With the added mass, the wow & flutter measured 0.06 - 0.08% (again measured 3 times). The only variable between the measurements was the platter weight.
Also of note, with the stock platter and motor off, I spun the platter and it continues to spin for a few minutes. I used a stethoscope to listen for bearing noise. None detectable. In all my years of using a stethoscope, I always detected bearing noise until today. Quite a nice feat for a stock 1961 turntable.
Then, with the 60 pound platter added to the stock platter, I once again spun the platters and used the stethoscope. Bearing noise was clearly audible. The bearing noise was cyclical which indicates that the sleeve bearing is where the noise is occurring, not the steel thrust plate. The sleeve bearing noise was likely caused by a slight imbalance in the platters weight distribution and thus would shift the bearing laterally and create the bearing noise.
So, where does this leave us:
1. 60 lbs incremental platter mass reduced wow/flutter by 50-66%.
2. Heavy platter needs to be appropriately balanced to no longer cause the sleeve bearing noise.
The 60 lbs platter is not suitable for the current setup (see photos), as it was taken from a 17" OD Gates CB500 transcription table. So, I am considering a new high mass platter and bearing.
Meanwhile, the stock papst motor purrs like a kitten, albeit with the fan noise that is inaudible at the listening position.
More to come.
Pat
Attachments
I'm guessing from the pictures that the heavy patter is larger diameter, the higher rim speed would mean much more flywheel effect for the same weight.
Even a ball supported spindle might cause periodic noise if the ball is not perfectly centered. A tiny fraction of a millimeter is enough to make it wandering around. It could not be your case, though.
Your measured W+F with the standard platter is much higher than I would expect possible, especially given the iPhone App's optimistic measurement. In comparison measurements I have made, the iPhone RPM App gives a consistently low W+F figure. I assume this is because: 1. it integrates (averages) some of the HF flutter; and 2. it does not include vertical runout in the wow component. I do not use this App because I find it too inaccurate to be useful.
For the benefit of anyone reading this who isn't familiar with the terms, wow is heard as a sort of doppler effect, a 'coming and going' of pitch whereas flutter is audible as a sort of tremolo effect. W+F is created by 'noise' in the motor torque, inconsistency in the belt thickness, roughness on the surfaces the belt runs one (spindle and rim), bearing noise, and eccentricities (both vertical and horizontal) of the spindle and platter, amongst other things. IME with reasonable quality turntables the biggest contributor to W+F is the drive belt, because most belts are not accurately ground and have excessive surface roughness and inconsistency in thickness.
If you run a belt between your fingers any surface roughness you can feel will produce flutter, and any variation in thickness (which is hard to feel but can be measured with a micrometer) will produce wow. It is not uncommon to find 0.05mm of thickness variation along the length of a belt which causes serious amounts of wow. An extremely heavy platter is needed to smooth out W+F introduced by the belt.
Wow that is of slightly longer period than the revolution of the platter is an indication of belt thickness variation. Many people (and most mechanical engineering text books) will tell you that the thickness of the belt can be disregarded in rotation speed calculations, but in fact that is an inaccurate simplification of the physics. The relevant diameters of the spindle and drum surface for calculating the gear ratio are the effective diameters, which is the diameter of the contact surface plus very slightly less than half of the belt thickness (slightly less than half because of the elongation of the belt in use and compression of the rubber as it runs around bends, particularly the spindle).
Synchronous motor belt drive Japanese turntables from the '70s with replaced belts consistently run slow because in my experience replacement belts are typically 0.5 and 0.6mm think, whereas the original belts were often 0.8mm thick.
Belts also introduce a flutter noise component caused by stiction of the inner belt surface as it compresses at the contact point where it bends. The solution to this is to lubricate the belt, traditionally with talcum powder as recommended by Linn, Thorens, et al.
A few years ago I lashed out and purchased about 30 different expensive and inexpensive 'upgrade' and standard replacement drive belts for Rega turntables and measured and graded them for speed, and wow and flutter. With just one exception, irrespective of brand, price or belt rubber material the belts were all pretty bad and easily the main contributor to W+F. (Obviously the surface smoothness, eccentricity and flattness of the sub-platter is critical as well, so in the case of Rega and similar turntables, replacing the plastic sub-platter with a TangoSpinner or similar is mandatory.) The one belt was head and shoulders above the others was the blue coloured Little Belter from Edwards Audio. I bought three and all were very good, but I may have fluked a good batch - I just don't know. All of the other 'known' quality brands, even Rega's own expensive upgrade belts, were quite poor on W+F. With the Little Belter W+F was well under 0.05%, most belts were well above 0.1%. (There is residual W+F on all test records. Of the ½ dozen or so that I own, ClearAudio's has the lowest W+F, but I am not able to measure the residual W+F and therefore not able to know with certainty what the turntable is contributing and what is residual from the record.)
I have not repeated that controlled experiment with flat drive belts for either sub-platter or rim drive turntables so I cannot objectively recommend a particular brand of replacement belt. For what it's worth I usually buy Thacker belts from Germany which have given OK results in my experience. I currently have an LP-12 on the bench supplied by the client with an original Linn replacement belt. Although I haven't measured the W+F figure yet, the Thacker belt I have put on it is audibly much better.
Once the drive system is sorted out, in my experience eccentricity of the platter is the next main contributor to wow. Wow that has the same period as the rotation of the platter is an indication of platter runout. This is usually only vertical runout but in extreme cases there may be a component of horizontal runout causing wobble. I use a dial gauge to measure the runout components. With turntables that use a platter / sub-platter assembly it is often possible to find a significant null point where the eccentricities of the two parts nearly or completely cancel out. Where this is possible I mark the relative positions of the platter / sub-platter so that they can be reassembled consistently. In a couple of uncommon instances I have successfully corrected spindle to platter non-perpendicularity by judicious tapping of the shaft with a rubber mallet.
For the benefit of anyone reading this who isn't familiar with the terms, wow is heard as a sort of doppler effect, a 'coming and going' of pitch whereas flutter is audible as a sort of tremolo effect. W+F is created by 'noise' in the motor torque, inconsistency in the belt thickness, roughness on the surfaces the belt runs one (spindle and rim), bearing noise, and eccentricities (both vertical and horizontal) of the spindle and platter, amongst other things. IME with reasonable quality turntables the biggest contributor to W+F is the drive belt, because most belts are not accurately ground and have excessive surface roughness and inconsistency in thickness.
If you run a belt between your fingers any surface roughness you can feel will produce flutter, and any variation in thickness (which is hard to feel but can be measured with a micrometer) will produce wow. It is not uncommon to find 0.05mm of thickness variation along the length of a belt which causes serious amounts of wow. An extremely heavy platter is needed to smooth out W+F introduced by the belt.
Wow that is of slightly longer period than the revolution of the platter is an indication of belt thickness variation. Many people (and most mechanical engineering text books) will tell you that the thickness of the belt can be disregarded in rotation speed calculations, but in fact that is an inaccurate simplification of the physics. The relevant diameters of the spindle and drum surface for calculating the gear ratio are the effective diameters, which is the diameter of the contact surface plus very slightly less than half of the belt thickness (slightly less than half because of the elongation of the belt in use and compression of the rubber as it runs around bends, particularly the spindle).
Synchronous motor belt drive Japanese turntables from the '70s with replaced belts consistently run slow because in my experience replacement belts are typically 0.5 and 0.6mm think, whereas the original belts were often 0.8mm thick.
Belts also introduce a flutter noise component caused by stiction of the inner belt surface as it compresses at the contact point where it bends. The solution to this is to lubricate the belt, traditionally with talcum powder as recommended by Linn, Thorens, et al.
A few years ago I lashed out and purchased about 30 different expensive and inexpensive 'upgrade' and standard replacement drive belts for Rega turntables and measured and graded them for speed, and wow and flutter. With just one exception, irrespective of brand, price or belt rubber material the belts were all pretty bad and easily the main contributor to W+F. (Obviously the surface smoothness, eccentricity and flattness of the sub-platter is critical as well, so in the case of Rega and similar turntables, replacing the plastic sub-platter with a TangoSpinner or similar is mandatory.) The one belt was head and shoulders above the others was the blue coloured Little Belter from Edwards Audio. I bought three and all were very good, but I may have fluked a good batch - I just don't know. All of the other 'known' quality brands, even Rega's own expensive upgrade belts, were quite poor on W+F. With the Little Belter W+F was well under 0.05%, most belts were well above 0.1%. (There is residual W+F on all test records. Of the ½ dozen or so that I own, ClearAudio's has the lowest W+F, but I am not able to measure the residual W+F and therefore not able to know with certainty what the turntable is contributing and what is residual from the record.)
I have not repeated that controlled experiment with flat drive belts for either sub-platter or rim drive turntables so I cannot objectively recommend a particular brand of replacement belt. For what it's worth I usually buy Thacker belts from Germany which have given OK results in my experience. I currently have an LP-12 on the bench supplied by the client with an original Linn replacement belt. Although I haven't measured the W+F figure yet, the Thacker belt I have put on it is audibly much better.
Once the drive system is sorted out, in my experience eccentricity of the platter is the next main contributor to wow. Wow that has the same period as the rotation of the platter is an indication of platter runout. This is usually only vertical runout but in extreme cases there may be a component of horizontal runout causing wobble. I use a dial gauge to measure the runout components. With turntables that use a platter / sub-platter assembly it is often possible to find a significant null point where the eccentricities of the two parts nearly or completely cancel out. Where this is possible I mark the relative positions of the platter / sub-platter so that they can be reassembled consistently. In a couple of uncommon instances I have successfully corrected spindle to platter non-perpendicularity by judicious tapping of the shaft with a rubber mallet.
Last edited:
Praise and request
Excellent post @johnmath !
If you had the time or inclination it would be awsome if you started your own thread on the ins and outs of turntable set up and tuning from start to finish.
Would make an awsome sticky.
Sorry for the off topic comment OP
Dennis h
Excellent post @johnmath !
If you had the time or inclination it would be awsome if you started your own thread on the ins and outs of turntable set up and tuning from start to finish.
Would make an awsome sticky.
Sorry for the off topic comment OP
Dennis h
Johnmath - great stuff. Agree with Dennis. Have you tried any custom belts that have ultra tight tolerances? Also, I was thinking of trying a stiffer belt like Kuzma is using on their new tables. Have you tried anything with stiffer / less compliance?
Thank you banpuku and Dennis for your kind words, however there are many gaps in my knowledge, which is mainly practical experience dosed with a smattering of 1st year mechanical engineering studied 50 years ago.
I haven't been setting up as many turntables over the past few years since I have moved to a wilderness location on an island. I'm currently working on old Rega Planar 3 and Linn LP-12 turntables, which are not really high end decks.
When I was measuring belts some years ago I purchased as wide a selection of branded belts suitable for Rega turntables that I could find at the time on the internet. Rega use a round section O-ring belt which is a limiting factor on achievable W+F. An O-ring belt can run in a V notch which increases the amount of torque that can be delivered by a given belt, but also increases the sensitivity of the drive to belt irregularities. Unfortunately I don't think Rega's small motor pulley and sub-platter diameters are suitable for a flat drive belt conversion.
In most cases resilience is required in a drive belt to filter torque variations from the motor. Synchronous motors have a torque profile that 'cogs' at a rate dependant on the number of motor poles. Most clock type motors will have 16 or 24 poles, the higher number being preferable as it shifts the cogging frequency upwards whilst reducing the amplitude of the torque fluctuations, making it easier to filter them out with resilience in the belt and mass in the platter, which together act as a quasi second order low pass filter.
If the motor has a constant torque, the belt can be stiffer. For example, a properly designed 3-phase AC motor will in theory have no torque fluctuations if fed with 3-phase sine wave power accurate in wave-shape, amplitude and phase.
I have made some DC motor conversions at clients' requests, but to be honest I have not thoroughly investigated the torque profiles of those motors and the subject is outside of my knowledge base. In those instances I have replaced the belt if necessary to achieve an acceptable level of W+F as real-world non-uniformity of the belt creates W+F even with a perfect drive motor. There seems to be some controversy over DC motor drives in the Linn / Thorens et al camps.
This is a common misconception. A 12 pole and a 24 pole motor will have the same cogging signature; while the latter has twice as many poles, it is moving at half the speed so the cogging frequency will be the same. If the motors have the same power rating, they should have nearly identical vibration.
This was explored and measured in this thread:
Hurst Motors: 300 RPM vs 600 RPM--Upgrade or Myth?
The cogging is the result of the variable reluctance between the PM rotor and the steel core and air gaps in the stator. As the magnetic pole approaches the steel core, it accelerates and as it rotates past the steel core, the speed is retarded, causing the rotor speed to oscillate at 2x the driving frequency. This occurs with any PM motor with magnetic stator cores regardless of the number of poles, windings or drive method (AC or DC).
With a decent rubber slipmat the platter will not ring at all. It makes an incredible difference. You can damp it underneath with rubber too it you like.
I think you've measured the precision of iphone's app there, the heavy platter is probably way below that in reality.
That's clearly mashing up the bearing, its not designed for such weight.
The best mat I've come across over the past few years is Herbie's Audio Labs Way Excellent II:
Way Excellent II Turntable Mat
Way Excellent II Turntable Mat
