Audiomods Rega conversion

The RB250 arm has been around so long now

that it’s easy to forget how radical the design was
when it first appeared. Firstly, the geometry was,
within the practicality of rounding the
measurements, correct. This might appear an
obvious starting point but, amongst the top
twenty 1970s tonearms, only the SME III and
Michell had a correct offset angle, all the others
being up to five degrees out. The armtube design
with fixed headshell was radical and, in an S-
shaped market, influenced just about every
design that came after it.

The great strength of the Rega design is the

armtube. It’s a very well executed thin-wall
magnesium/aluminium alloy die casting with a
one-piece headshell and double taper. Die
casting is a far better way to make an arm than
using a tube or even machining it from solid but
it’s prohibitively expensive unless, as in the Rega
case, it will be a mass-market product with a
long life.

The rest of the arm is exceptionally well

designed for production at its price point but
that’s also its great weakness when we are
looking for real performance.

The bearings and the way that they are fixed is

very compromised by price. The design of the
RB250 depends upon careful setting up of a
loose-tolerance bearing. There is no way to
improve upon them unless we completely rethink
the bearings and the way they are held

The search for an improved arm and the resulting product was never, and isn’t, a
commercial project but it has resulted, after some years of experiment, in a very
limited quantity of bespoke, hand-built arms as well as DIY kits for fellow
enthusiasts.

At commercial rates the design is simply too complex and time-consuming and uses
some rather expensive parts to be translated into a series production product so it will
only ever be hand made in tiny quantities.

www.audiomods.co.uk 1
[email protected]
 Audiomods Rega conversion

Modifying the arm

The starting point for the design is that it is cheaper to use the armtube from a rough
or damaged RB250 than to create one of equivalent quality. Casting is impossible for
small quantities, tubes are compromised and machining from solid is both very time
consuming and gives us the wrong grain structure.

The design targets were to be able to use the arm tube casting as a starting point but to
use very high quality bearings of lower inertial mass than the originals and to mount
them in a way that gave better control and much higher precision.

Bearings and arm yoke

Can we fit high quality, close tolerance bearings to a standard Rega arm?
Not really. The RB250 is designed around bearings that have a degree of “play” so
that they can be set up within the limits imposed by the way they are held in place. A
very close-tolerance bearing wouldn’t be an
improvement and is more likely to increase
resistance.

After lots of experiments we’ve settled on small

ceramic hybrid bearings (ceramic balls in steel
races) for our rebuilds. These are very high
precision, wonderfully smooth right from the start
and the whole installation has been designed
around using these properly. They just wouldn’t
work as substitutes in the standard location.
Unfortunately, these bearings are maybe forty
times the cost of the ones going into the original
arms, but worth it!
This type of bearing has big advantages:

• Very low “sticktion”. It’s almost always said that an arm should have very low
friction, but that’s not really the case. We are, after all, using ball bearings in a way
they really aren’t designed for. The 10x3mm ones shown here are designed to run at
up to 50,000rpm with a 50kg load. Then the running friction would be important. Our
bearings make tiny, continuous to-and-fro movements so what we are interested in is
the starting friction. As part of the load “seen” by the cantilever, bearing friction is
small compared to arm mass inertia. What we really want is low “sticktion” – the
tendency for a bearing to stick-and-go – which results in jerky, rough movement. The
smoother the bearings, the better they can be adjusted.

• Lower internal inertia. The ceramic balls have lower mass. This might seem
insignificant, but an arm makes constant, tiny movements so the direction of the
bearings is always reversing. The lower the ball mass, the less likelihood that the balls
will skid rather than roll. (This is a big weakness of the originals.)

• Very high precision bearings can achieve a play-free setup with less pre-load.

www.audiomods.co.uk 2
[email protected]
 Audiomods Rega conversion
The disadvantages are:
• Very high precision bearings need very high precision mountings.
• Hybrid bearings are far more expensive than the best quality steel ones.

The final design relies upon making the bearing housings easily adjustable to achieve
the precision we need. It isn’t necessary to measure the final bearing adjustment, but
it’s a factor of probably 10x finer than the original setup.

From serial number 21, the horizontal and vertical bearings have different lubrication
applied because of their different loadings - all the horizontal bearing balls are evenly
loaded whilst the upper balls of the vertical ones bear no direct load.

The arm yoke is machined from solid HE30 aluminium with removable bearing caps,
rather like car main or camshaft bearing journals. The journals are made to very fine
limits so that the bearings will lock into place with very little pressure on the screws.
The caps “float” and are located by the bearings themselves.

The bearings are held in the arm tube by press-fitting a solid aluminium carrier with a
one-piece, ground silver steel shaft for a very high-precision fit.

The bearings are now much farther apart for stability(41mm centres rather than the
original 19mm) and the arm yoke is more massive than the original. This has the
advantage of adding inertial mass in the horizontal plane. We tailor the yoke mass to
match the cartridge compliance range. See the section on antiskate and bias control to
understand why this is so important.
The vertical bearings are also
completely new.
The Rega vertical bearing has two
weaknesses:
• The bearing position, which is
dictated by two factors: a large distance
between yoke and bearing because of
the position of the antiskate magnet and
a narrow distance between bearings
because of the need to accommodate the
wiring plug. These are just the opposite
of what we want for stability: minimum
overhang and wide spacing.
• The base is also the bearing housing so
the bearings are coupled directly to the
armboard/plinth, perhaps creating a
feedback loop.

The redesign of the vertical bearings is intended to address these issues. The result is:
• Better shaft design with positive bearing spacing
• Much wider bearing spacing (30mm centres, originally 15mm)
• More accurate, adjustable location from a split housing
• Bearing housing decoupled from the base
• Built-in VTA
• Fine adjustment of mounting distance

www.audiomods.co.uk 3
[email protected]
 Audiomods Rega conversion

Counterweights
There are a lot of counterweight modifications available for the Rega arms, each of
which might help the performance to a degree.

In the standard arm, the plastic end stub

tends to decouple the arm from the weight.
This is not particularly good because the
mass of the weight doesn’t control
resonance in the armtube and the difference
in materials properties tends to stop transfer
of energy from tube to stub. If we simply
fix a metal stub rigidly to the armtube, then
the tube will tend to dissipate energy by
trying to vibrate the weight as well but the
weight can releases that energy and feed it
straight back to the armtube, slightly out of
phase with the original. We hear this as a
smearing of the sound. So the aftermarket
weights that use solid stubs screwed into
the armtube may create more problems than
they solve.
The Origin Live counterweight achieves a
real improvement by coupling the stub to
the armtube in a way where the stub’s
natural frequencies are very different from
the arm’s and it now uses low-area contact
for the weight (early versions didn’t).

Our design also alters the path from armtube to weight. The stub retaining screw is
threaded into the bearing carrier, not the armtube. The result is a rigid structure but
one that uses the different characteristics of the interfaced parts to control but not
reflect the energy.

Materials choice
The way in which sound is transmitted across the interface between two parts depends
upon the “acoustic impedance” value of the materials they are made from.

So, if we have an aluminium-to-aluminium interface such as the one between the

armtube and our bearing carrier, the AI value is the same so the two parts will tend to
react rather like one.

www.audiomods.co.uk 4
[email protected]
 Audiomods Rega conversion

In the stub and counterweight the aluminium/brass/lead/brass interfaces have very

different AI values, so energy will reflect rather than pass through. Like this:

The values can be expressed as:

Where RE is the reflected energy and Z1 & Z2 are the acoustic impedances.

Material Acoustic
Impedance
Aluminum 17.3
Brass 36.5
Glass 17.4
Lead 27.2
Magnesium 9.90
Perspex 3.2
Stainless Steel 44.8
Tungsten 100.0

The weight is a complex layered structure: brass, 3 layers of lead, then brass or
stainless steel, held under considerable pressure by 3 high tensile allen bolts. It’s
machined in such a way that only the lead and the nylon locking screw is in contact
with the stub. We do this by accurately boring and reaming the weight, then pulling
the bolts down to the final torque. This “spreads” the soft lead a tiny bit, leaving it
slightly proud of the reamed hole. Having two dissimilar metals bonded strongly
together stops “ringing” and tends to convert mechanical energy into heat. The effect
is to absorb but not reflect vibration from the arm. We can see from the table that
tungsten would be a very good substitute, but this is not a commercial project and the
complex machining and threading is simply too difficult.

www.audiomods.co.uk 5
[email protected]
 Audiomods Rega conversion

Low-cg counterweights
If we rigidly fix loads on the end of a pivoted beam, the more they are below the pivot
point, the more “stable” the beam is, ie the more it will tend to return strongly to
horizontal when deflected. If the mass were exactly on the centreline the beam would
be happy to stop in any position. So if we make a tonearm whose counterweight has a
very low cg it will tend to return strongly to level when it is displaced by a warped
record. The downside of this is as the cartridge goes up, vtf increases; as it goes down,
vtf decreases. This effect is most serious with highly compliant cartridges where the
cantilever will be most displaced. At the other extreme, if the arm is balanced exactly
along the centreline it will only have the vtf of the cartridge to return it to the correct
position and exactly the opposite will happen.
A very low cg might also help prevent chatter in poor bearings by evening out
dynamic bias (torsion) forces, and this may be the reason that such weights seem to
offer a quick bolt-on improvement for Rega arms. The real answer is to get better
bearings!

If you look carefully at the Michell counterweight, you will notice that, as well as
lowering the cg, it moves the mass outward in plan. This will have an effect on the
yaw inertia, the same as we do with the heavier arm yoke, and for the same reasons it
will help the performance. It’s a clever design.

For the best compromise we need to create a counterweight whose mass is biased a
little low to exert some control without over-stabilising. A precise calculation would
depend upon the exact cartridge compliance, cartridge mass and the amount of warp
we find acceptable. So we just take an average.
Our counterweights are lightened above the centreline and the fine adjustment weight
sits well below it to put centre of mass about 15% below the centreline, useful for
controlling warp tracking without overstabilising. By changing the mass of the fine
adjustment weight, we can tailor the arm for different cartridge compliance ranges.

www.audiomods.co.uk 6
[email protected]
 Audiomods Rega conversion
Antiskate
The standard Rega antiskate works OK. You can’t really trust the numbers to set it up
by and it doesn’t apply a really consistent force across the whole record, but this is a
budget arm.

Antiskate is much more important to good reproduction than is often thought, because
the forces we need to counteract are created by the music, not just the track of the
groove. It’s a very complicated subject.
Why do we need it? Passing the stylus through the record groove causes drag. If the
cartridge were aligned straight down the arm and at a perfect tangent to the groove
this force would act squarely on the bearings with no sideways component at all. This,
rather than the lack of tracking error, is what makes linear trackers so good.
But it doesn’t work like that. The cartridge is at an angle to the armtube to make the
geometry work, so the drag force becomes a torque on the arm that’s translated into a
movement towards the centre of the record. This torque is transmitted to the arm
bearings, the cause of chatter in poor bearings.
If we don’t have any antiskating then the inner (left signal) groove of the record gets
pushed against more strongly than the outer groove and we hear distortion in the right
channel. So we must add a little counteracting force outward to stabilise it.
The problem is how much.

www.audiomods.co.uk 7
[email protected]
 Audiomods Rega conversion

The amount of antiskating force we need to apply is affected by many things:

• Groove friction. This is affected by the friction coefficient of the record and very
much by the stylus profile. This effect is stronger at the outer grooves where velocity
is highest.
• Tracking error. Going across the record this starts high, reduces to zero, climbs, falls
back to zero then finally climbs again. Just how much this affects the drag depends on
the stylus profile.
• Groove modulation. The more energy put into the stylus by the signal, the more
drag. This is a reaction to the energy used to excite the cantilever and the energy used
to create electrical current in the generator. This effect is more marked at the end of
the side than on the outer grooves and low frequencies create the most drag.

So we have a force constantly varying as the record

is played that we must counteract without knowing
its exact value. Setting up on a blank disc is not
accurate because it doesn’t reflect the real drag
value of the cartridge in the groove or an average
value of the dynamic drag. Setting up with a test
record is much better but here it’s important to set up
at a number of points across the record. Careful
listening with a known record is the final test. Listen
carefully at outer and inner grooves, around the null points
and halfway between them. Distortion from bias setup can be
identified because it appears on one channel: right channel, underbiased,
left channel, overbiased. A slight mismatch might be heard by stereo
image moving to left (under) or right (over).

Our quadrant antiskate does help you

to optimise the force across the
record, weighting it at the outer and/or
inner grooves. The setting arrived at
will be influenced by the kind of
music you play.
Depending upon the stylus profile,
simple acoustic music will probably
be more neutrally biased, whilst
orchestral or opera that tends towards
crescendo on the inner grooves might
need a bias weighted towards the
record’s centre. A high-compliance
cartridge might need a bias slightly
weighted towards the outer grooves.
Only listening will tell.

www.audiomods.co.uk 8
[email protected]
 Audiomods Rega conversion

Arm tube modifications

The most important modification is the effect of the solid pressed-in bearing carrier,
which changes the characteristics of the RB250 armtube to rather like the RB300, but
the other changes also contribute significantly to the sound.

Polishing an arm tube

Unless you intend to carry out other mods to the arm tube, removing the paint will
make it sound far worse than standard.
There is no sonic advantage in polishing the arm tube. It looks nice.

There are advantages in bead-blasting the arm tube Michell style, because that affects
the surface hardness a little. It also makes good commercial sense for them, because
it’s quick and hides surface imperfections.

Drilling an arm tube

This does have sonic advantages. The pattern and number of holes isn’t critical, but
there are mechanical and sonic advantages in the spiral drilling with graduated hole
sizes that we’ve developed. The exact layout and size depends on the target mass and
the range of cartridge compliance aimed for. Michell and Origin Live now drill/slot
their armtubes. We started in 2003. Enough said.

You could choose either spirals or a row of holes underneath (Michell) style.
Spiral holes have the advantage of reducing mass if that’s needed for a high-
compliance cartridge.
Slots: I have reservations about slotting the arm tube, though I have no test or
experimental data to support it. The strong point of the Rega arm tube design is that it
is deliberately very rigid. It would be expected that reducing the rigidity by cutting a
long slot would tend to introduce bending vibration modes. Maybe this is OK, like
second harmonics in a unipivot.

Earthing the arm tube

The arm is originally earthed by a strap fixed into the counterweight thread. This can
cause problems even on standard arms.
If you are using our upgrade kit then the armtube must be earthed – ceramic bearings
won’t conduct!
Our method is to drill the armtube to add a very small copper stud onto which the
earth wire can be soldered. This is kept completely separate from the signal wiring
and taken direct to ground.

Arm cross bracing

This is a very effective way of controlling resonances in the arm tube.
Aluminium discs are fitted into the arm tube approximately 1/3 and 2/3 along. They
need to be of a very precise diameter to fit properly. The width is varied according to
the target arm mass. Calculation suggests that the normal mode of the tube changes
from around 850Hz to 130-190Hz in the sections.

www.audiomods.co.uk 9
[email protected]
 Audiomods Rega conversion

References
Much of the really technical analysis of the behaviour of arms and cartridges was
done many years ago when the vinyl record market was huge. Some very
knowlegable engineers devoted serious study to the problems and their work still
remains the best basis for understanding the subject and designing solutions.

A number of the papers below are available for download from www.audiomods.co.uk

“Analytic treatment of tracking error” HG Baerwald

“Pickup Design” JK Stevenson (Wireless World, 1968)
“Tonearm Geometry and Setup” M Kessler & D Pisha, (Audio, Jan 1980)
“Determination of sliding friction between stylus and record groove” R Pardee,
Bell Corporation
“The skating force phenomenon”, J Kogen (Audio, Oct 1967)
“Pickup arm design techniques” TS Randhawa (Wireless World, March 1978)
“Sensitivity of Phonograph turntables to normal loads” TS Cole,
(AES Journal May 1968)
“A stereo groove problem” G Alexandrovich (AES Journal, Jan 1961)
“The cartridge alignment problem – a new approach” RJ Gilson,
(Wireless World Oct 1981)
“Resonance in a tube” Victor Reijs
“An introduction to acoustics” RH Randall
“A textbook of sound” AB Wood
“Properties of sound waves” JC Drury

Reviwed/revised June 2009

Jeff Spall
6 Nutfield Rd
Redhill
Surrey
RH1 4AU

www.audiomods.co.uk 10
[email protected]