DIY linear tonearm

Hi Bill.

The chatter I was referring to is not caused by rotation. Within the bearings the radial load is supported by only the balls at the bottom of the race. If axial load is low the rest of the components in the bearing will only be in light contact with each other. During record replay vibration from the stylus passes via the carriage to the bearings causing the loose components to rattle against each other producing further vibration which in turn is passed back to the cartridge. The effect occurs at discrete frequencies where the components resonate against each other. The effect is audible as a harshness associated with certain notes.

This is probably a moot point as cantus, clearaudio and the contributors to this thread all use unlubricated ball races and don't seem to suffer chatter. I have heard an old sme with shot bearings that suffered chatter.

Pin bearings are definitely worth investigation. They're relatively easy to make and cost a lot less than premium ball races.

Niffy

Niffy, Paul. et al (and especially et and al)/

The following is a quote from Bo Hansson and I wonder how this plays out in this discussion of rattles and resonances:
"The vertical movement from the record, shall be damped. In Cantus it's solved by using ballbearing with a little sideway play. This means the contact point in the bearing is changed when the tonearm moves up and down. As the contact point is changed there is automaticly a damping, because it's not a resonant point."

BillG
 
Miniature Ball Bearings in Clocks by Rex Swensen - Sydney

Hi,

Interesting thoughts on bearings - so as usual I asked my friend Google and found this article:


where the author concludes:
"From these test results it would appear reasonable to assume a µ of 0.003 for the 2 mm ball races, if well soaked out in solvent. For plain pivots and jewelled pivots, a value of 0.15 - 0.18 is representative. This means that the plain pivot has about 60 times as much friction as the ball race. To achieve the same friction in a plain pivot it would have to have a diameter of 0.017 mm or 0.0007.

I haven't critically read the article to see whether he is comparing apples with apples - but interesting non the less.

What did interest me was the assortment of part on this site that the author referred to:

www.smallparts.com.au

A lot of good stuff to enable building a smart looking arm.

So .... I better get building :eek:
 
Bill

Interesting quote.
Not too sure about its validity in this case. I believe Bo is talking about damping the vertical motion of the arm not the vibrations that may cause chatter. In the arm design as it stands at the moment this vertical motion of the arm is damped by the sliding friction between the bearings outer race and the glass rods.

Bob

Good links.
The jewelled pivot discussed in the Rex Swensen artical is a totally different beast to the vee cup bearings I am using. He is testing a plain pivot of 1mm diameter. This is a 1mm diameter parrellel sided steel pivot sitting within what is basically a brass or ruby tube.
A vee bearing has a steel pin with a spherical tip that sits inside of a spherical cup, in my design made of teflon but more traditionally ruby.

If we assume a 30g carriage with the load distributed evenly beteen the 4 bearings (2 per axil) each bearing will be supporting 7.5g. Using hertzian contact equations I calculated the required pin and cup diameters required for both teflon/steel and ruby/steel. The limiting factor in this discission is the maximum pressure the different materials can safely and reliable take . The teflon used a 0.5mm pin tip in a 0.65mm cup. The ruby a 0.12mm pin tip in a 0.18mm cup. (Ruby can take much higher pressure than teflon)
Used in this application the contact diameter for the teflon bearing is 0.09mm and for the ruby only 0.008mm. At this point the ruby looks much better (and only half the diameter Rex Swensen suggested to match the miniture ball race)
The teflon makes a comeback when you consider the contact pressure and the coefficient of friction between pin and cup. The contact point in the ruby bearing has pressure over 120 times as great as in the teflon and the coefficient of friction 3 times as great. The calculated figures suggest the friction within the two types of bearing are very similar. I chose teflon for its "soft" contact and theoretical advantage when chatter was concidered.
The calculated cofficient of friction for the teflon bearing, when converted to the set of conditions stated by Rex, would be a u of 0.0022. slightly lower than for the miniture ball race.

Niffy
 
Forgot this bit.

Check out Boca Bearings. They have 3x10x4mm bearings. 440 stainless races with silicon nitride balls, NO LUBRICATION so no tedious cleaning and easily removed rubber seals. Not the cheapest but I challange you to find better ball races.
Oh. and smallparts used to sell ruby vees and accociated hardware, cant find them in their web site. hope they still do them.

Niffy
 
As to the bearing rattle issue, a couple of points:

1_ if you get any resonance or rattle it will be from the non-loaded balls especially if the races are cleaned with solvent. It will be roughly equivalent to setting free balls in a pan on top of the pivot point. You can purchase balls from most hardware stores or online by the way. The question though, is even if they rattled and you hear them when you are near to the tonearm, is if the resulting 'rattle' would get coupled back thru the cartridge. You would need some time of generator as in 2, below, to test it.

2_I was wondering about how you would find the resonant frequency and relative amplitude(s) of rattles or problem areas. I was thinking if you took the tonearm with a sacrificial cartridge (?) and place it on the cone of a small speaker you would be able to run various tones thru the speaker to stimulate vibrations. Probably don't need to couple it thru the cartridge, just set the tonearm on a stick or something with a mass analog of the cartridge. Tonearm manufacturers must have some type of test setup. Maybe if you used a speaker as a generator and put an accelerometer on the tonearm somehow you would be able to track the resonances. You can buy accelerometers and hook them to an arduino board to map the vibration energy. So a testbed might be a horizontally mounted speaker with the tonearm sitting on it and then use an accelerometer to measure the resonance peaks??

I am working on a linear tonearm with an air bearing and have been kicking around testbeds.

Good luck with the project. You guys have handled issues in very creative ways sometimes coming-up with solutions by using very crude tools or construction methods. I don't mean to be critical here, quite the contrary. I think back to when I was working with RF and most of the times 'manhattan style' or 'dead bug' construction performed better than custom PCBs by factors.

I still think you need to mount the glass bar with jewel bearings on each end to overcome the stiction in the vertical plane. Though it would add to the cost, if you also use jewel bearings on the track wheels you probably would have lower friction and cut down on any bearing rattle. The jewel bearings, mounts, etc aren't that expensive.

Jerry
 
Hi All ,

Find it encouraging to see people manufacturing their own bearing versions !
I must realy read this all through in prior of making comments . but I'm glad
to read opinions on this subject .

As for a testbed for tonearm resonance ... I have been thinking about using
transducers , such as ultra sonic transducers . variable freq. amplifier etc.
but not sure about their bandwidth .
But than again , can't we just use a testrecord , track various frequencies and
check the output signal for harmonics ? shape ?

THX,
Paul
 
Last edited:
This is some really interesting theory and a lot of good points are brought up. Should t we easily be able to test bearing chatter with a frequency sweep?, since we are talking about resonantly induced chatter in theory. I guess too much of this would also depend on bearing size and tolerances too. I had once owned a cheapie s arm that exhibited this bearing chatter that was excited by a few different frequency nodes, could easily detect this anomaly through headphones.

I have found with this Linear arm that just the loading of the carriage itself and the way the mass is distributed is enough to load the bearings to eliminate chatter. A simple test of possibility of chatter is to take the carriage with bearings installed and shake it like a piece of fruit :), sloppy bearings will sound more like quiet shakers :D. Good bearings will have a lot less of this slop, I have to put my ear close to hear any slight ticking shaking the bearings I'm using which with a little load bias will be next to nil. If there is bearing chatter, it is out of the audio band and not seemingly excited by any frequency in the audio band.


Colin
 
Bearings can chatter or resonate within the audio band not unlike someone shaking a can of rocks in time to a beat. I read an article where a engineer went into the whole pivot bearing slop issue, I didn't agree with all of it as you don't need measurable slop for movement, but just about any tonearm or mechanism with a high degree of freedom will have some 'slop' but maybe not measurable.

The test record idea has problems in application as the tracks don't allow you the time to tune the frequency and you might pass thru the resonance before it has a chance to take off. Using a variable transducer of some type, or my speaker idea, would be best in that you can dial it around to isolate the problem, if any.

I don't want to get into a debate on why the cantus needs stiction or friction in the vertical plane but it seems contrary to all the best practices in tonearm design. If it does, then I suspect it is mitigating some other design compromise. I find it hard to see how an arm with any stiction or measurable friction in that plane will have consistent tracking force, something that has a marked measurement in response. I read somewhere that the stiction required the arm to drop opposed to lightly setting down to maintain tracking force. From everything I have read and experienced over the years a tonearm with a high degree of freedom in all directions that relies on the mass of the cartridge and arm is preferred. I would think a free floating mass will give the system consistency whereas any stiction or friction will be a variable and almost impossible to control. The mass of the arm and cartridge become the stable platform, within limits. But then again, I am not overly familiar with the cantus arm.

Another approach to linear tracking that I haven't seen would be to just use free balls running in a track that would allow almost complete freedom of movement in two planes. Using mating groves that trap the balls in a linear line might work and I've been kicking that around. There are problems with this design in that you have to account for the balls congregating at one end or the other.

For all these reasons and more I decided to go the air bearing route like the Kuzma though I originally started with something more along the lines of longer curved air bearing with only one upper side of the horizontal cylinder under pressure. Both of these designs have challenges as well. That's why I continue to read this thread to see how good the ball bearings on glass tube will work out. Both my turntables (TNT MK IV and a Micro Seiki) have traditionally pivoted arms and sound great but could be better.

Try the speaker idea to modulate the tonearm and I'm sure it will give usable data.

Jerry
 
From everything I have read and experienced over the years a tonearm with a high degree of freedom in all directions that relies on the mass of the cartridge and arm is preferred. I would think a free floating mass will give the system consistency whereas any stiction or friction will be a variable and almost impossible to control. The mass of the arm and cartridge become the stable platform, within limits. But then again, I am not overly familiar with the cantus arm.

I don't want to misread Cantus' concept but I think this idea of "freedom of movement" in conventional tonearm is exactly what Mr. Hansson arguing that it has too much freedom and it's harder to control when the energy of the retrieved information by the cartridge is wasted into the oscillation of the vertical movement. That's my interpretation anyway. I think Moray James probably has more to add to this. It's something I never thought of and just like many people I used to think friction free is the ultimate goal in bearing. I at least can appreciate Mr. Hansson's outside of the box thinking.

Another approach to linear tracking that I haven't seen would be to just use free balls running in a track that would allow almost complete freedom of movement in two planes. Using mating groves that trap the balls in a linear line might work and I've been kicking that around. There are problems with this design in that you have to account for the balls congregating at one end or the other.

Isn't that a ball bearing is doing anyway, except it's running a series of balls in circular form?
 
I don't want to misread Cantus' concept but I think this idea of "freedom of movement" in conventional tonearm is exactly what Mr. Hansson arguing that it has too much freedom and it's harder to control when the energy of the retrieved information by the cartridge is wasted into the oscillation of the vertical movement. That's my interpretation anyway. I think Moray James probably has more to add to this. It's something I never thought of and just like many people I used to think friction free is the ultimate goal in bearing. I at least can appreciate Mr. Hansson's outside of the box thinking.



Isn't that a ball bearing is doing anyway, except it's running a series of balls in circular form?

If you take the shields off the ball bearings I think you will find that in most if not all of them there is a pressed metal separator ring that traps the individual balls in their own little pockets. No possibility of them bunching up or bumping into each other much less anything else. It will probably do a decent job of dissipating most or all of this rattling energy that is so confounding us.

BillG
 
I still think that what may be confounding us in the bearing chatter/rattling issue is that we're chasing a problem that isn't a problem. No has reported the harsh sound associated with severe chatter. As Frank mentioned, when referring to less severe chatter, on page 160 "still very common are compromised detail retrieval and/or colorations." The current carriage designs being used, especially Colin's, have moved away from the cantus style and are now becoming ever increasingly clearaudio like. Read any review of a clearaudio tt tonearm and detail retrieval and low coloration are at the top of its list of attributes. Doesn't sound like low level chatter is a problem here. The bearings used by clearaudio are unlubricated ball races.
Adding mechanisms to deal with something that is not an issue will probably compromise the arm in other areas.

Niffy
 
I don't want to misread Cantus' concept but I think this idea of "freedom of movement" in conventional tonearm is exactly what Mr. Hansson arguing that it has too much freedom and it's harder to control when the energy of the retrieved information by the cartridge is wasted into the oscillation of the vertical movement. That's my interpretation anyway. I think Moray James probably has more to add to this. It's something I never thought of and just like many people I used to think friction free is the ultimate goal in bearing. I at least can appreciate Mr. Hansson's outside of the box thinking.


I'll have a go at explaining this one.

Conventional wisdom says you want the cartridge to be constrained from moving only by the effective mass of the arm. In other words a zero friction bearing. ARA of sme said he tried to design his arms to act as a "pure mass" but even he had to add a silicone fluid damping trough to get the V to perform at its best. What is actually required is mass with a small amount of damping.

In a mechanical linear tracker the damping is applied via a small amount of friction. With a knife edge bearing there is virtually no friction in the vertical plane and too much in the horizontal creating an imbalance. What we are trying to achieve is a reduction in the horizontal plane friction and an increase in the vertical plane to bring the system into balance.

Niffy
 
In one of Bo's letters to me he responded to my struggle with having increased friction in the vertical plane when playing warped records. Bo told me it was a matter of controlling resonances. He said that the only think you really wanted to have moving was the cantilever and that if the arm was free to move then you would loos energy in the cartridge. Since you don't want skipping you allow the arm to let the record drag the assembly up the inside of the tube to track the warped record (which he said you should not be playing anyway) and then gravity would bring the carriage back down for you after the warp had passed. The slack in the bearing assembly of the Cantus had enough available motion to cove normal arm motion with non warped records. To the added friction is a kind of resonance damping. You arm has enough freedom of motion to play a normal record (or even a warped one) but the built in friction stops system resonances from stealing energy and dynamics from the system.

Bo's first version of the Cantus was what I would call a rolling knife edge the arm could rock back and forth like a balance scale. The next and improved version had four points of contact inside the glass tube and the only free motion was due to the slack in the bearing assembly itself, past that to play a warp the carriage would have to be drawn up the inside of the glass wall as previously mentioned. The new (and current) version is a significant step up and forward in performance. So the new version is free to roll in the horizontal plane and has progressively damped (friction) motion in the vertical plane. I apologize if I am doing a poor job of describing this but I am not an engineer.

It is to me rather ironic that Bo's arm would be copied by a German company who would sell a sophisticated version for near to $15,000.00. That copy was of the first version although the execution is visually a a little different. I asked Bo about this he laughed and said that they could not get it right. Personally I expect that it was a conscious decision because that version permits you to play outrageously warped records and when a "customer" pays that much they want to be able to play what ever kind of warped record they may want to, even if it results in a reduction in performance. Best regards Moray James.

PS: keep in mind that Bo was the original importer of the Sowther arm into Europe and that Bo and Lou Sowther were good friends. Bo's Cantus design was intended to deal with the weaknesses that Bo saw in the Sowther arm. So it is easy now to clearly see the German audio connection as that is where the Sowther designs were later sold to.
 
Hi Moray.

Thanks for your in-depth description. I think Bo description might have come at the problem from a different angle than I did but arrived at the same conclusion.
Many years ago I read an article by a designer who like Bo advocated not playing warped records/dealing with warps at the turntable not the arm. This designer said that when playing non-warped records the best location for the vertical pivot was on the line that passes through the stylus and the centre of the cartridges generator, rather than being level with the record surface. For most cartridges this would be at an angle of 20degrees from the stylus. As wow, vta etc are not issues, when raising the pivot, with flat records he may have had a point.
Do you know if this designer was actually Bo? This idea seems to fit nicely with the discrption you just posted. I'm not entirely convinced by this idea but I can see positive merits to it.

Niffy
 
I still think that what may be confounding us in the bearing chatter/rattling issue is that we're chasing a problem that isn't a problem. No has reported the harsh sound associated with severe chatter. As Frank mentioned, when referring to less severe chatter, on page 160 "still very common are compromised detail retrieval and/or colorations." The current carriage designs being used, especially Colin's, have moved away from the cantus style and are now becoming ever increasingly clearaudio like. Read any review of a clearaudio tt tonearm and detail retrieval and low coloration are at the top of its list of attributes. Doesn't sound like low level chatter is a problem here. The bearings used by clearaudio are unlubricated ball races.
Adding mechanisms to deal with something that is not an issue will probably compromise the arm in other areas.

Niffy

Niffy, you said "I still think that what may be confounding us in the bearing chatter/rattling issue is that we're chasing a problem that isn't a problem." copied from quote above. I really wonder that nobody has tumbled to the irony in my postings and my tone of voice or are they all being polite? That is what I have been talking about. The next few posts (thank you Moray) also are germain. At any rate we are all learning a bit more about ball bearings. Let's all go listen to some relatively flat records and enjoy the great performance these arms are giving.

BillG
 
Hi Bill.

Thanks for the heads-up on the magnet wire.
My previous wiring was a litz multi-strand which was a too heavy and was giving me no end of trouble. I tried simply twisting the wires as you suggested but was getting way to much noise. I've just made a new set of wires using a single strand of 0.112mm magnet wire with 7 strands of 0.08mm magnet wire wrapped around it giving full shielding. Dead fiddly to make but killed the noise. Much cleaner extension into the treble than the litz. Thanks once again for your invaluable suggestion.

Niffy
 
Hello everyone
It seems I'm one of the few people who have actually owned and used both a Cantus and a Clearaudio. I have over 2500 lp's and have maybe a few (and I mean a few) that suffer warps-there are more like undulations rather than great cavernous valleys-I stick by the tenet 'if the warp is bad, don't bother playing it' and buy a flat one.

I'm beginning to think that the obsession with warps is becoming slightly anal in it's nature. The only times I've had tracking issues have been when I hadn't dressed the wires effectively and they were pulling the carriage one way or another. Not sure I want to learn much more about bearings-after all they are just a load of balls if you get my drift.

I agree with Bill-start listening to your records-if they're warped you are probably storing them badly! IF I like what I'm hearing, experience tells me things are working. Let's have a bit more about actual building experiences, the results and how good the music can sound. I'm currently listening to a Cannonball Adderley and the sax is the best I've heard!
Chris
 
Thanks Chris and Bill,


It would be nice to see more designs and people actually building the arm. As is expected there are always the few that want to shoot holes into a theory or a build like this that works very well. Compounding this is that we are not tonearm designers professionally and not all will happily take us lowly diy guys coming up with something that has any chance of besting a commercial product, such is the nature of this. Though it would be nice to see some more effort put into confirming the results rather than to theoretically trying to find issue that may not exist:).

Colin