[ Home | Staff & Contacts | HiFi Playground | Listening tests | DIY & Tweakings | Music & Books ]

Arms design

How many ways can you skin a cat...?

[Italian version]


Recently I was in conversation with a manufacturer of turntables and arms. He said that in his opinion, given a basic level of competence in the turntable, the arm was responsible for 80% of the sound quality produced.
During my series of turntable and arm reviews I've started to come to see his point, though I'd put the split nearer 60/40 in favour of the arm. Whether you agree or not, it's not in dispute that the tonearm has a fundamental effect on the sound produced by a vinyl front end.

Over the last four years I've had nine arms here on a long term basis (1 month+), six of them in the last six months with more to come. They came in all shapes and sizes and each sounded different so I thought it was time for me to find out what makes arms tick and pass on that knowledge to the readers of TNT-Audio.

Before I start a few words of warning - what follows is a basic overview of tonearms, what they do and why they are designed the way they are. Each design has it's proponents and camps are deeply entrenched over what is best, if in my ignorance I offend such groups then I apologise, all I can say in my defence is having played with a lot of designs I know what I like...

The problem

On a record the music is encoded as wiggles in a piece of vinyl. In fact to get the idea imagine a piece of corrugated iron, fold it down the middle across the corrugations to make an angle of 90 degrees. OK - with me? One side of the resulting 'V' is the left-hand channel, the other side the right. The stylus (imagine a traffic cone) moves along the 'V' and is wiggled back and forth, these vibrations go up the cantilever and are turned into an electrical signal by (generally) two sets of magnets and coils also set at 90 degrees to each other - each perpendicular to the groove wall it is tracking - this gives a stereo signal for the preamp. It's delightfully simple...

So going back down to the microscopic dimensions that actually exist, it is the arm that has the job of holding the cartridge whilst the 'V' shaped groove hurtles past.

So what is the arm up against? Well the basic problem is this - If the record surface was perfectly flat and ran under the stylus in a straight line like tape over a tape-head, the arm could be a block of rigid concrete holding the cartridge absolutely immovable. This way when the stylus moves these movements become the music signal - no energy is lost moving the cartridge back and forth in sympathy with the music. But records, NO record, is absolutely flat, and rather than travelling in a straight line the groove is a spiral so the arm must allow movements both up and down and sideways across the disc.

So the perfect arm must allow the stylus to move up and down for warps and sideways to take account of the spiral BUT at all music frequencies must hold the cartridge absolutely solid as if mounted in that block of concrete.

Tough call...

The only answer is of course to use bearings of one sort or another to allow the arm to pivot in both vertical and horizontal plane, they must do this with as near zero friction as possible but not allow any movement at music frequencies including twisting. It must also press the stylus into the groove at a constant downforce so as to resist a modulated groove 'spitting' the stylus upwards and out of intimate contact with the groove walls where the information is stored, and in order to keep the cartridge generator aligned.

To answer this impossible brief three main schools of arm design have evolved. The gimballed arm, the unipivot and the parallel tracker. Each has its advantages as we'll see, but they all share one design problem. If you take a rod of any kind and hold it rigidly at one end it will 'ring' at certain frequencies rather like a plucked string. The stiffer the arm the higher up the frequency of the ringing (like a tighter string). Real 'stiffies' like the big SME's and Roksan's Artemiz have main ringing modes at 1 kHz or more where it is less noticeable than lower down.
Clever shaping or damping can spread or reduce these resonances, but tuning them to be as unobtrusive as possible is one of the 'black arts' of arm design. To add to this every part and particularly joint of an arm will produce its own sonic signature, which is why we have so many one-piece arms since SME, Mission and Rega led the way.

The gimballed arm

This is the most common design and probably the most obvious. Just put two sets of bearings at 90 degrees to each other at the balance point of the arm. Assuming decent bearings these will hold the arm rigidly, stopping any twist whilst allowing the arm to move freely up and across. If the centre of gravity is placed level with the stylus the downforce will remain constant as warps force the arm up and down - this is good. Likewise if the horizontal bearing is at the level of the stylus then as the stylus rises due to a warp it will not tend to move back and forth as with a bearing above the stylus.

The snag... These bearings are ball bearing races (or variations). In order to have low friction they must have some free play. So as the stylus tries to twist the arm the bearings will allow slight movement and perhaps more importantly tend to rattle against each other - this is called 'bearing chatter'. As you tighten the bearings to take up this slack the friction will increase thus compromising the arms performance. So a gimballed arm is a compromise.
Have the bearings tight enough to reduce (you cannot eliminate) chatter but loose enough to allow minimal friction. This is why top arms are so expensive, they require the bearings to be perfectly aligned and for the compromise to be exactly where it will be most beneficial - this takes time. Talk of expensive bearings is generally misleading, ABEC3 bearings cost pence but the time and precision needed to make them work costs...

There are two other snags with such a design. One - the bearings each act as a boundary where resonances can be reflected or produced, so though an SME has a one-piece armtube, it has a bunch of other boundaries in its bearings. Two - over time these bearings will wear and go out of adjustment, either needing replacement or adjustment.


In the past unipivots have been very popular, then with the advent of the 'battleship', ultra-stiff, gimballed brigade exemplified by the ITTOK and then the SME V they fell from favour. Now they are back to the point where they are 'flavour of the month'. The thing is that there are unipivots and unipivots as we shall see.

At first a unipivot seems a daft idea. Instead of two sets of ball races at 90 degrees to each other we have a simple point bearing, pointing either upwards or downwards, resting in a cup (a few like Hadcock use ball bearings) and placed at the pivot point of the arm. Bonkers - No torsional rigidity whatsoever, twist the headshell and the arm just twists?

But... What if we could balance the masses around the arm in such a way that when the cartridge tries to twist the arm AT MUSIC FREQUENCIES the inertia of the arm alone holds it steady. WOW! There's an idea worth considering...

So how can we do it? The simplest way is to hang weights off the bottom of the arm, either as low slung counterweights or bearing housings with a lot of weight under the pivot. This gives the arm some inherent stability, the cartridge will tend to hang downwards though cueing will still be a wobbly affair.

Spotted the snag? If the centre of gravity of the arm is well below the pivot, as the arm rides over a warp the centre of gravity will have to be lifted, this increases the downforce at a the worse possible moment, driving the stylus up towards the cartridge body as the arm resists upward movement. This is a BAD THING...

But lots of unipivots are made his way, in fact probably the majority, the Hadcock and Kuzma being two I have here at the moment. The Kuzma adds a thick bath of silicon gloop that the skirt of the bearing housing sits in.
This increases stability still further but the gloop increases the arms reluctance to ride warps to the point where the cantilever and suspension of the cartridge do much of the work. This is an extreme case of damping in a unipivot, but it is a commonly used aid to resisting the twisting forces imposed and also to damp resonances in an arm.

The other way of building a unipivot is to have the centre of gravity only just below the pivot so that when riding warps the downforce doesn't alter. In this case the same force is needed to deflect the arm up/down or sideways producing a 'perfect mass'. Such designs are much more wobbly to cue but at music frequencies it is the moment of inertia of the arm itself which resists the twisting - a tough piece of tuning all round.
Some designs like this use a 'false' unipivot where there is some physical restraint to stop wobbling. This takes the form of a bearing housing that acts like the bearing housing of the turntable by holding the bearing vertical in the twisting plane, but unlike the turntable bearing offering no resistance to 'nodding' up and down movement. As this restraint takes no load, friction is practically zero and it gives the arm the same user friendliness as a gimballed arm.
Of course there is the problem of the restraint rattling against the bearing but as it has to carry no load it can be made of some low resonance material such as Delrin and as it is a very loose fit friction doesn't enter into the equation. Audiomeca's Romeo arm is a 'false' unipivot in this mould but the delrin sleeve can be easily removed to produce a 'true' unipivot when you feel brave enough.

Just looking at a unipivot can show you which path the designer took. Low slung - Hadcock, Kuzma, VPI, Moerch, ARO - or those with a high centre of gravity, often with weights either side of the bearing housing to increase inertia against twisting like the Audiomeca and Graham.

But what unipivots all share is a bearing with close to zero friction, which (if good quality) will be self adjusting and devoid of any bearing chatter. Apart from this the problems of arm resonance and the like are much like a gimballed arm.

Parallel trackers

When an analogue master is cut it is done on a lathe where the head travels in a straight line across the disc on a sort of sled. Pivoted arms track across the record in an arc which inevitably means that for most of the time the stylus will be at a slight angle to the way the groove was originally cut. This introduces some distortion.
Also because the cartridge is mounted at an angle at the end of the arm it tends to pull the arm towards the centre of the disc, and this force has to be resisted by some kind of 'anti-skate' mechanism, otherwise the stylus will be forced against one wall of the groove harder than the other with a reduction in sound quality and inevitable uneven wear of stylus and disc. This force is applied in many ways, threads and weights, magnets or springs, but all are compromised and potential sources of colouration.

The parallel tracker tries to mimic the cutting lathe and has an arm which tracks across the record whilst the pivot is on some kind of moving sled. If the cartridge is perfectly aligned then there should be no tracking error and no need for anti-skate. Also the arm can be made shorter and so will be stiffer and have a lower moment of inertia.

The downside is of course the sled. It has to move in unison with the stylus. Eccentric records (all of them...) mean it will have to move back and forth as well as across the record.
Either the arm has a complex series of sensors and motors which spot deviation and shuffle the sled back and forth, or a very low friction slide, such as an air bearing, allows movement with little friction. Both these solutions are hideously complicated and the fact that anyone can make them work at all is a source of constant amazement to me. Needless to say they are expensive, fussy, need dust free environments and perfect set-up. But when they're right some people say nothing else comes close.

An interesting and to my mind more practical variation is the Audiomeca where what looks like a true parallel tracker is in fact a pivoted arm which has only a very small arc of operation. A small sensor detects when the arm moves beyond the arc and a motor shuffles the sled along a bit to compensate - neat as the pivot does the small variations 'passively'

Lastly - Cartridge compliance and "effective mass"

You might think that in theory it'd be nice to have an arm that only applied downforce and offered no other resistance to change in direction. Unfortunately having a mass that needs to be accelerated the arm has inertia and will resist the movement. The more massy the arm the greater the inertia or 'effective mass'.
There was a time when the lowest possible effective mass was seen as perfection but this resulted in flimsy resonant arms which needed very high compliance cartridges in order to work, with the result they tended to bounce all over the disc when presented with a warp or even cued...

This was because a cartridges compliance and an arms effective mass must be matched. The compliance of a cartridge is the 'springyness' of its suspension. Imagine it as a springboard at a swimming pool. If the springboard has a very light person on it it will bounce up and down quite quickly. If it has a real 'fatty' on it this springing will be slow. To get the fat guy to spring at the same speed you need a stiffer (low compliance) springboard.

What has this got to do with cartridges and arms? Well you want the combination of arm and cartridge to resonate, or 'bounce', at a certain frequency i.e. below that of any music signal (<20Hz) and above the frequency of record warps (>7Hz). So a very compliant cartridge such as a Shure V15 likes a low mass arm, a low compliance cartridge such as a Koetsu likes a high mass arm like a Bruer. Nowadays most cartridges are medium compliance and most arms medium mass but it still pays to make sure a given cartridge works well with your arm.


So now you're thoroughly confused? Some people will tell you unipivots sound sweeter, gimballed arms more dynamic. All I have to say is that I've heard great gimballed arms and great unipivots, all I need now is for someone to lend me a parallel tracker...

© Copyright 2002 Geoff Husband - http://www.tnt-audio.com

[ Home | Staff & Contacts | HiFi Playground | Listening tests | DIY & Tweakings | Music & Books ]