Origami Idler Drive?

[EC8010]

Hello mosin,

the Lenco idler needs to be narrow because if you imagine an idler of finite thickness as being two separated idlers of infinitesimal thickness separated by a finite thickness, then one idler will be driven at one speed, and the other, driven by a different diameter of the conical motor shaft, will rotate at a slightly different speed. But those two hypothetical idlers are locked together, so one or other must slip, and that must make noise. This is fundamental to the tapered motor shaft idea and one of its weaknesses.

Because you need a very narrow cross-section tyre touching the motor taper, the rubber must be hard to avoid deformation with contact pressure. That reduces the filtering effect of the tyre. Conversely, a wide idler tyre like the 301 can get away with a much lower pressure because it has a larger contact patch. That lower pressure means softer rubber can be used. You can check this by comparing the two.

Your bearing and rigid support certainly sounds like an improvement on the original Lenco scheme and should significantly reduce noise.

Personally, I think that idler drive has more flaws than belt drive, but I'm always susceptible to a reasoned argument. In the end, it probably comes down more to the quality of the engineering...

 

[mosin]

Hello mosin,

the Lenco idler needs to be narrow because if you imagine an idler of finite thickness as being two separated idlers of infinitesimal thickness separated by a finite thickness, then one idler will be driven at one speed, and the other, driven by a different diameter of the conical motor shaft, will rotate at a slightly different speed. But those two hypothetical idlers are locked together, so one or other must slip, and that must make noise. This is fundamental to the tapered motor shaft idea and one of its weaknesses.

Because you need a very narrow cross-section tyre touching the motor taper, the rubber must be hard to avoid deformation with contact pressure. That reduces the filtering effect of the tyre. Conversely, a wide idler tyre like the 301 can get away with a much lower pressure because it has a larger contact patch. That lower pressure means softer rubber can be used. You can check this by comparing the two.

Your bearing and rigid support certainly sounds like an improvement on the original Lenco scheme and should significantly reduce noise.

Personally, I think that idler drive has more flaws than belt drive, but I'm always susceptible to a reasoned argument. In the end, it probably comes down more to the quality of the engineering...

The wild card here is the shape of the rubber, and the footprint of the contact area. These variables alter any steadfast conclusions significantly.

 

[Mark Kelly]

one idler will be driven at one speed, and the other, driven by a different diameter of the conical motor shaft, will rotate at a slightly different speed. But those two hypothetical idlers are locked together, so one or other must slip, and that must make noise.

Another effect is that the contact patch with the platter itself must have finite extension in both the radial and tangential directions and for a constant angular velocity the linear velocities must vary. I think this effect is smaller than that due to the tapered shaft. The 301 and other idlers with edge contact have a different problem which is that if both the idler and the platter have constant angular velocity but with different radii they must have different linear velocities through the contact patch so the contact must deform and slip to make up the difference. As you note, this creates noise.

I think that idler drive has more flaws than belt drive

I think that the problems are of a different natures, relative severity comes down to tolerance of these different problems. Idler problems all seem to centre around transmission of motor and bearing noise, whereas belt drive problems seem mostly to do with speed variation due to the loose coupling (which helps them avoid the problems of noise). In both cases better quieter motors must help.

I know you and I disagree over the relative importance of distortion, quadrature error and torque cogging on the performance of synch motors but I think we are both working in similar directions.

 

[mosin]

...Idler problems all seem to centre around transmission of motor and bearing noise, whereas belt drive problems seem mostly to do with speed variation due to the loose coupling (which helps them avoid the problems of noise)...

Mark,

I agree that the two types of drive have dissimilar problems, but I am most interested in what you say about the idler's problems centering around the transmission of noise from the motor and bearing. You have seen a photograph of my first generation idler, and in my post on the previous page, I outlined the makeup of the second generation one. I went to a lot of trouble in developing the new idler because I agree with you regarding the transmission of motor noise, and feel that I have successfully addressed the idler's problems in that regard with a given motor, as least as far as it can practically be done. I did manage to reduce motor noise to a level that is imperceptible, and the tapered shaft problem is adequately minimized due to the shape and stiffness of my assembly, and also due to the fact that my particular turntable is fixed at a single speed which also lessens wear on the associated rubber parts. That said, it would be nice to be rid of the tapered shaft entirely.

You also mentioned the contact area of the idler rubber with the platter. The primary reason that I selected the Lenco is because I felt this was less of an issue than the one both you and EC8010 so elegantly describe in regard to the other idler drive type. That said, the skew still exists where the wheel contacts the underside of the platter, and it is the reasoning behind my decision to use not only a thin material in width, but one in height as well. Hopefully, I have minimized it adequately enough to gain additional reliability, in addition to an improvement in the sonic performance of the turntable. The idea was to tighten the interface without compromising the grip of the wheel.

This leaves the question of bearing noise. I assume you mean platter bearing noise, and if so, wouldn't the belt drive's noise and the Garrard type idler's bearing noise be essentially the same due to outward pressure? If that is the case, wouldn't driving the idler on the underside of the platter make more sense?

In any event, I believe that I have done pretty much all that I can do with specific regard to the idler wheel itself in a standard Lenco style implementation. As far as performance goes, I don't know if I agree with EC8010 based on the reasons he states when comparing idler and belt types of drive. However, my next project will be a belt driven type, and that exercise is to satisfy my curiosity of belt drive issues. I have to see exactly what they are for myself.

A thought occurred to me after building this first turntable, and that is one of idler turntable motor isolation. It might be feasible to build an idler using the same techniques that Roskan uses in their turntables, and that is one of separated plinths with the motor being mounted on a lower plinth that is critically placed, but with the only point of contact between the two plinths being the idler wheel itself.

Whatever the case, building an idler is probably a more difficult undertaking than building a belt drive. I'll find out soon enough.

-mosin

 

[mosin]

Two points because I edit eternally...

I) I understand the differences between the words elegantly and eloquently. You gents stated your points both ways!

2) I want to clarify my remarks about Roksan's method of motor isolation. First, they do not make idler drive turntables, and the plinths do contact at small points. Still, I feel the idea is worth a closer look.

-mosin

 

[Zero One]

Here is a thought on limiting motor noise, have a look at this thread I posted on vinyl circle last night, the ideas are applied to a belt drive TT, but it might work for an idler drive too.

I can absolutely say it is a ripper mod on the belt drive I am using and it has a large 240v motor much like those found in idler drives.

http://www.audiocircle.com/circles/index.php?topic=38056.0

The bearing mods is very worthwhile too is nicely reduces surface noise.

Zero One

PS:
Just another thought but probably of no use, would gluing a thin layer of rubber under the lencos platter for the idler to run against help, it would not effect speed as it would on a rim drive.

 

[mosin]

Originally posted by Zero One Just another thought but probably of no use, would gluing a thin layer of rubber under the lencos platter for the idler to run against help, it would not effect speed as it would on a rim drive.

Zero One,

I do not know, but the idea of running the wheel on an entirely different platter material, like Delrin, has crossed my mind.

-mosin

 

[Mark Kelly]

I assume you mean platter bearing noise, and if so, wouldn't the belt drive's noise and the Garrard type idler's bearing noise be essentially the same due to outward pressure? If that is the case, wouldn't driving the idler on the underside of the platter make more sense?

I was actually referring to idler and motor noise. In both the Lenco and the Garrard the original idler bearings are poorly designed and become noisy with time. Shaded pole motors tend to be quite noisy.

Whatever the case, building an idler is probably a more difficult undertaking than building a belt drive. I'll find out soon enough.

Couldn't agree more. The development cost of just doing upgrade parts for Garrards has gone beyond my (very limited) means. I'm also playing with a belt drive at the moment because it's so much easier.

 

[EC8010]

Whatever the case, building an idler is probably a more difficult undertaking than building a belt drive. I'll find out soon enough.

Couldn't agree more. The development cost of just doing upgrade parts for Garrards has gone beyond my (very limited) means. I'm also playing with a belt drive at the moment because it's so much easier.

This is something we can all agree on. It's much harder to make a good idler drive than a good belt drive.

 

[graeme uk]

'A thought occurred to me after building this first turntable, and that is one of idler turntable motor isolation. It might be feasible to build an idler using the same techniques that Roskan uses in their turntables, and that is one of separated plinths with the motor being mounted on a lower plinth that is critically placed, but with the only point of contact between the two plinths being the idler wheel itself.'

This is exactly what ive done, the two plinths will be seperated by a sandbox. The only pint of contact is the idler wheel.

As for running the idler on a different surface, mine runs on glass as i use a rega platter. There is less noise from the idler to platter contact than there was with the lenco platter.

 

[mosin]

In both the Lenco and the Garrard the original idler bearings are poorly designed and become noisy with time.

Gents,

Picture in your mind a Lenco motor with its armature shaft fixed by bushings on each end with the drive area between them, or picture a motor like the Garrard's with a yoke attached to the outside body, so that it goes over the shaft on both ends. Wouldn't it be feasible to dispense with the bronze bushings entirely by fixing a single ball to each end of the yoke, like I did with my idler wheel? It would be a rather simple modification with the Lenco motor because it already has a nearly perfect housing for the modification. A small recess could be machined on each end of the shaft, so the balls would stay positioned. Then a couple of small pieces could be made to hold them in place with one providing tension adjustment. I believe that it will work, and it should go a long way to quieten the motor, especially if the balls run on a structural engineering plastic that is suitable for the application. I have already taken my Lenco motor halfway there, and it works.

I have a spare Lenco motor, so I'll try to modify it, and report the results.
The Lenco Gyro mod.

-mosin

 

[Mark Kelly]

A small recess could be machined on each end of the shaft, so the balls would stay positioned. Then a couple of small pieces could be made to hold them in place with one providing tension adjustment. I believe that it will work, and it should go a long way to quieten the motor, especially if the balls run on a structural engineering plastic that is suitable for the application.

I can't see what part of this arrangement takes the lateral load on the shaft.

 

[mosin]

Originally posted by Mark Kelly I can't see what part of this arrangement takes the lateral load on the shaft.

Mark,

I envision it as being the same as my idler wheel with the load being shared between the ball and the race it runs in. In the case of my idler, the Si3N4 ball doesn't spin at all, but the Torlon hub does. The ball is extremely hard and precise, so any wear that may occur will be on the Torlon, and it will be evenly distributed. By the way, my assembly does exert significantly more pressure to the hub area than in a typical idler wheel application, in order to keep it in place. So far, I have seen no wear, however. The ball rests only in a shallow divot, but that particular plastic is so incredibly rugged that it seems impervious to the load even with such a very small recess. It is so small that it is barely visible, yet it is sufficient enough to keep the ball in place.

It might be prudent to design the motor modification, so that the Torlon spins, rather than the ball. (That would be a change from my first idea in the post above.)

Lack of wear was not the case with the first generation assembly because the Si3N4 ball caused the bronze hub to wear out very quickly, so I feel that the high tech material is key to the success of the idea. Heat shouldn't be an issue with Torlon, either. It withstands a fairly high sustained temperature. If that does become a problem however, there are plastics that have extrusion temperatures that are much higher. Maybe one of them would be more suitable.

With the platter removed, I can press the idler assembly against the motor's shaft with significant pressure, and the assembly continues to spin without any sign of strain, and it is getting a lot more lateral stress than under normal conditions. I cannot see why the same sort of modification made to the motor's shaft wouldn't behave in the same manner. If I am right, and it is aligned properly, I don't see why it won't work.

-mosin

 

[Mark Kelly]

I think you will have a problem with wear, it will just take longer to become evident.

Without going into a long tutorial on bearing design, I think you need to do an analysis of reaction forces on the shaft. You will find that the lateral reaction forces through the bearing act at an angle which is dependent on the penetration of the ball into the surface.

With a hard surface (brass) the limited penetration means the contact angle is very acute and the stress in the bearing is high.

With a softer material (torlon) the degree of penetration is higher and you have effectively created a wedged journal bearing - the stress is lower but it's not the point contact bearing you thought you had.

 

[EC8010]

I'm not clear about what you're trying to achieve. The rationale for replacing/redesigning the idler bearing was obvious, but the motor bearings? Are they really that bad? My experience with even cheap shaded pole motors is that the bearings are remarkably good. Also, as Mark has said, those bearings were designed for the load they had to withstand. Your proposal will wear quickly on the motor. You get away with it on the idler because the forces are pretty well balanced.

 

[Mark Kelly]

Back to the original question regrading idler mechanics, I posted some information today on Vinyl Asylum which may be germane to the discussion. I don't know what the moderators here think of links to other fora so at the risk of repeating myself, here's the gist of it:

I thought it might be useful to compare available drives using a different metric, and I came up with the speed change produced by a notional 10 mN force applied at the platter rim. I used a nice round number right in the middle of the probably range of forces produced by modulation drag. If this worries you, just use the table as relative data - any other force can be substituted, the numbers all change by the same amount. The other advantage in doing it this way is the figures can be added - if you have a speed change of x for one motor and y for one drive mechanism, the total speed change will be x + y.

I decided to express the speed change produced in parts per million as the range is quite high. As an "anchor", piano tuners work in cents, one cent being 1/100th of a semitone, roughly 600 ppm. This is probably close to the smallest pitch change that can reliably be detected. I am sure some of the golden ears out there will claim to be able to hear much smaller changes....

So, to the figures. First, drive mechanics (eg belt creep)

Direct Drive 0

Belt drive (rim, 300 rpm motor) 750

Belt drive (rim, 1500 rpm motor) ~ 4000

Belt drive (subplatter, 300 rpm motor) 1500

Idler drive (Garrard Type) 150

Idler drive (Lenco Type) 50

It's also worth noting that any drive which uses platter derived feedback will have much improved figures, depending on the configuration of the feedback loop. A competently designed PLL will give speed drift of zero.

Next, motors

Synch motor (any) 0

Induction motor (Garrard / Lenco type) 600

Induction Motor (Papst type) 300

DC motor (uncompensated) 3000

DC motor (compensated) 30


Remember, to get a usable figure add the motor and the drive together. Interestingly, synch motor plus standard belt drive = 750, Garrard motor plus Garrard idler = 750.

This begs the question: so why do we hear "better" speed stability with the idler type drives? Well it's perfectly possible we are imagining it but I think there is more going on so my next couple of installments will be about torque spring effect and flywheel effect.

The promised extra analysis of flywheel and torque spring may be a while, once I get the info together on a web page I'll post a link.

Mark

 

[EC8010]

Back to the original question regrading idler mechanics, I posted some information today on Vinyl Asylum which may be germane to the discussion. I don't know what the moderators here think of links to other fora so at the risk of repeating myself, here's the gist of it:

In the end, it comes down to "the good of the membership". Hopping backwards and forwards between forums is irritating - especially if you only have a peripheral interest in the subject matter and are not a member of the other forum. Thus, your action was the logical one and we wouldn't have minded in the least if you had added a link to allow reading the more complete post.

I'm looking forward to seeing the next instalment.

 

[cubastreet]

Mark, seems I was too quick to slate dc motors.
I was generalising from the audioorigami one which iirc runs off a couple of aa batteries and is pretty low torque.
glad that someone`s doing it though.

 

[Mark Kelly]

Further calculations

I have psoted another in a series on drive mechanics at Vinyl Asylum. I won't place the whole post here, the gist:

Calculated moments of inertia (Icm) of several platters: all figures are in g m^2 or 10^-3 kg m^2 (the SI base unit).

Direct Drive: 10
Medium mass belt drive (40mm acrylic) 37
Medium mass idler (Garrard 301) 38
High Mass (say 100mm solid aluminium) 215

Next some motors:

DC motor 1 x 10^-3 (0.001)
Synch motor 0.01
Induction motor (Garrard Type) 0.03
Induction motor (Papst type) 1
No figures available for direct drive due to the integration of the rotor.

The effective moment of inertia of the motor referenced to the platter:

600 rpm DC motor 0.3
300 rpm synch motor 0.6
1350rpm Garrard motor 50
1470rpm Papst motor 1950

You can see why the Papst motor is known as a flywheel motor!

And some times constants for combinations of drives and platter inertias;

Direct Drive 0
Medium Mass belt drive 0.075s
Typical idler 0.075 s
High Mass belt drive 0.375s

Consider that one beat in the bar will usually be between 0.5 and 1 second we see that the moment of inertia of even the heaviest platter is not sufficient to govern speed stability at the time scales of interest in music.

The post is at http://www.audioasylum.com/audio/vinyl/messages/638888.html