advertisement


dŵr & Friends (how not to build a turntable).

InSides

dŵr
Introduction

I've dabbled with analog reproduction for more than half my life. That seems like much, but then again, not as much as many other enjoyers of this hobby. And then, dabbled seems like the most appropriate term for a non-English speaker, as what I am doing can be described most aptly as 80% enthusiasm, 10% science (perhaps) and 10% sheer stubbornness.

I was unlucky (or lucky, depending on point of view) to have limited access to turntables and the general analog reproduction techniques, as interest in vinyl had waned dramatically at the beginning of the 2000's. I did own, over the years, a number of hi-fi turntables (exclusively belt driven) and quickly understood that I would not be happy until I either (1) purchase a turntable that I could not afford or (2) build one.

Option one was out of the question. Option two required a few years of learning, research, drawing, inspecting materials, prototyping, chasing machine shops and finishing houses, and listening. Little did I understand at that time it would not be the cheaper option. :)

This thread is started in response to feedback on the System Pics 2020 thread, and is meant to describe some of my efforts related to turntable building, as well as shed light on motivation and approaches. I will be reusing material I have posted on my native forum, and translating where needed. As such, will need to do several posts before I get up to date on the current happenings.

I tend to write longer posts and typically follow up with a lot of photos. You have been warned. :)

Turntable

I decided to call the turntable dŵr.

To a lifelong fan of Orbital, the reference should be obvious. To a teenager in the 1990s, almost inevitable. As I understood it, it is the Welsh word for water, which also seemed natural a name to describe the freeflowing shape, and ease of reproduction of the turntable it came to represent. Unrelated to this, in my native language, the pronounciation of the word also refers to a key musical term.

So it stuck.

GDkZPXd.jpg


A mass-loading affair, there is no suspension to speak of. The turntable has been designed to work together with the bespoke rack, rounding of a concept to gather and sink unwanted vibrational energy instead of damping it.

Plinth. A bespoke laminate of beech and maple, arranged against the grain for structural rigidity, the shape of the plinth is a result of its function. A rounded triangular form with pillars at the apex points, and a central thicker section to house the bearing block. Carved on a CNC, sanded by hand and finished with beeswax and linseed oil to avoid lacquers.

Pillars. The plinth rests on three damped pillars which also serve as tonearm mounting points. Machined out of 6082 aluminium, each pillar consists of 8 individual components which help (1) clamp the plinth, (2) mate ball bearings with the shelves of the rack, (3) dissipate tonearm vibrations and (4) facilitate the mounting of tonearms with a wide specter of effective lengths through cantilevered mounts. The ball bearings are Silicon Nitride ceramic balls (Si3N4) with the highest sphericity rate available to me commercially. Arrangement of said balls follows the kinematic bearing principle allowing for self-positioning.

Bearing. Non-inverted single piece bronze sleeve to dispense with mating bushings. Machined of C932 bronze, the bearing sleeve consists of a single piece (the housing is the sleeve) and a machined cap housing the thrust plate. The bearing spindle is machined out of steel which is then subjected to a hardening process, but is not polished to the finest grade we could achieve in order for microabrasions to be able to retain oil along the length of the spindle. The tip of the spindle rests on a 8mm Si3N4 ball which is supported by a replacable thrust plate, typically made of Vespel (SP-21). The bearing is mounted in a single piece 6082 aluminium block which mounts to the wooden plinth.

Platter. Two platters, actually, in an effort to reap the benefits of both high and low mass designs. The lower platter is machined out of aluminium (6082), with a diameter of 320mm and a height of 70mm, for an approximate mass of ~15kg. The upper platter is machined from Polyoxymethylene (POM, or more often known by one of its brand names, Delrin). POM has similar vibrational characteristics to vinyl (PVC), and has the added advantages of (1) being easily machinable as opposed to PVC and (2) significantly less toxic when heated up (which is always a plus). Both platters are joined through three conical bronze (C932) couplers, which are tipped with 3mm Si3N4 balls, thus allowing the largest degree of platter/bearing separation I could do apart from an air bearing. The record spindle is also made of C932 bronze and can be replaced to cater for varying record hole diameters.

Drive. Based on a previously published project on DIYAudio, the drive centers around a readily available BLDC motor, driven by three individual phases courtesy of a DDS generator and 3 channels of Class D amplification. Speed monitoring and correction is done via a closed feedback loop, implemented through an Arduino Nano. In the vain of sharing, and thankful for the BLDC drive project being published, I have open sourced the actual source code (and the associated PCB) of the speed monitoring section over on the turntables GitHub repository. Power is transferred via a POM pulley and a compliant square section belt (2mm square). All electronics reside in a separate box, the form of which immitates the turntables plinth. Connections to the turntable are via Neutrik connectors (Ethercon, XLR).

dXVhtxf.jpg


The speed correction algorithm is based on parameters devised by research and experimentation. Full running speed is available approximately 4 seconds from the stop position, and rotation correction is sensitive to RPM changes of ~0.001 due to sensor limitations. Control is exerted through a wired remote control, machined from aluminium (6061) and POM, housing a small OLED display to show current RPM and additional information, as well as two pushbuttons to handle multiple operations programmed with the Arduino code.

DADqmSq.jpg


Further details about how the remote works, and what different button combinations yield, can be found here.

Tonearms

The plinth was designed to support up to three tonearms. When the turntable was first built, two were mounted.

em7JtOj.jpg


Tonearm #1

Called tân, Welsh for fire. Took the lazy way out with naming once the turntable got its name. It was just easy. :)

The tonearm implements a stabilized constrained unipivot principle, which has been around for a while. I believe as early as the SpJ La Luce tonearm, if not earlier. This has, in recent times, been successfully copied (in one form or another) to a number of tonearms including the Continuum Cobra, the Durand Telos, the VPI 3D (dual pivot modification) etc. Although there is still a single principle pivot point, the tonearm takes advantage of a secondary pivot, which is typically different in type, in order to facilitate stabilization of said tonearm during cueing and tracking.

The unipivot bearing is atop a bronze (C932) spindle, in the form of a kinematic bearing arrangement built up of three Si3N4 3mm ceramic balls, and a pointed POM pivot. The tonearm tube is aluminium (6061), damped with POM sections, and untreated sheep wool. To further stabilize the tonearm, three separate counterweights are implemented - (1) a fixed one concentrating mass around the pivot point, made of aluminium, bronze and PTFE, (2) a main one made of bronze and PTFE, and (3) an auxilliary fine weight (bronze) with a range of adjustment of 0.2 grams.

rP3rRMl.jpg


VTA adjustments are available during tonearm playback, and have been designed to avoid disturbing the rigidity of the tonearm. The spindle moves up and down in a PTFE sleeve, facilitated by a fine threaded bronze screw on the outside of the tonearm base. Once proper VTA is achived, an internal locking screw prevents misalignment.

Azimuth adjustments are available during tonearm playback, and are facilitated by the secondary pivot. This, in turn, is composed of an elliptical sliding surface, which is 3D printed, using a specific filament from Igus (Igus Iglide) designed for printing bearing surfaces, and a bronze spindle which touches on the very thin ridge of the sliding surface, thus providing for (1) tonearm stabilization, (2) progressive antiskating, and (3) the basics of the azimuth adjustment mechanism. Similar to the VTA adjustment mechanism, settings is through a fine threaded bronze screw which rotates the tonearm tube and can be locked via an internal set screw to prevent misalignment.

J4tvEJg.jpg


Tonearm parameters:
  • Effective length: 310mm;
  • Effective mass: 13.5 grams;
  • Range of VTA adjusment (at the pivot point): ± 10mm;
  • Cartridge range: 5-25 grams.
The tonearm is typically used with a DL-103, which yields further data:
  • Horizontal resonant frequency: 9Hz;
  • Vertical resonant frequency: 10Hz.
Tonearm #2

Conversion to SME 3012. A tonearm which started life as a SME 3009 S2 Improved (removable headshell). I changed the tonearm tube and end stub with OEM 3012 parts, and then machined out of bronze (C932) a full set of replacement weights. Additionally, a 10mm bronze spacer to enhance rigidity and make the height adjusment more usable, and a bronze knife edge bearing.

Took the advantage and replaced the ball races in the pillar with SKF ceramic ball bearings.

Damped with sheep wool and rewired using silk isolated copper litz wire. Additionally, machined a custom elliptical piece to house a standard DIN connector instead of the typical RCA and SME connectors.

b2amSdO.jpg


The tonearm is typically used with a Ortofon SPU #1 E, which yields further data:
  • Horizontal resonant frequency: 9Hz;
  • Vertical resonant frequency: 12Hz.
Rack

An integral part of the turntable, a mass loaded affair built of beech and maple, with an internal metal skeleton built of damped aluminium tubes (6082). With general dimensions of 700mm (width), 500mm (depth) and 700mm (height), it rests on hardened steel cones tipped with 12mm Si3N4 ceramic balls. The rack is built as a single piece and cannot be modified nor dismantled unless destroyed.

Shelfing is built up of natural white marble, a type indigenous to my home country, polished on 5 sides and abrasive on the bottom side where it mates with a combination of cork, rubber and Sorbothane to control vibrational spread.

hUZqJro.jpg


With the above, the first phase of development of dŵr concluded. It has brought many hours of pure musical enjoyment, but as it was built on a strict project management schedule (to reign me in from a plethora of ideas) it has since undergone a series of under-the-hood modifications.

To those of you who reached the end of this post, I salute you.

I shall wait to see if more details are wanted before I proceed with posting about what is current with the turntable.
 
Superb; I salute you too sir! Extremely well written up; do you have any more details about the experiments you carried out with rim/idler drive?

Rim drive measurements are coming in a few weeks - I have yet to submit the drive components for finishing (aluminium is to be anodized and the bronze is to be TiN plated) - and I will complete the details then.

In the meantime, there are plenty more details that I can share (starting later today) for what I came to call phase #2 of development. (*)

(*) Yes, I am deluding myself that at some point in time I will be "done".
 
Rim drive measurements are coming in a few weeks...

Thanks! I find the subject of different drive methodologies fascinating; I have several different types here at home:

  • Garrard 401 - Idler
  • Lenco 88 - Idler
  • Linn LP12 - Belt
  • Michell Hydraulic Reference - Belt
  • Stanton ST.150 - Direct drive.
 
So, at some point in the second half of February, having listened to the then current iteration of the turntable for more than a year, I realized that I may not be able to contain myself for much longer without changing the turntable. Apart from ideas flowing, I gained access to improved machining techniques, improved finishing techniques (PVD in the form of TiN plating) and a bit more knowledge, so decided to give it a go.

Knowing myself, a new project management plan was due, as the only way of ensuring that the upgrade cycle will have a finishing line.

To that effect, I set up a five step plan:
  1. Bearing. A completely new bearing, along with a revised bearing block, as a direct result of the work on @igor_xxxx 's turntable. Expected improvements include further operational stability, better handling of mass and a lower noise floor;
  2. Drive. A new monoblock motor pod, and a new pulley. Expected improvements include better heat handling through improved heat transfer, lowered belt slip and a lower noise floor;
  3. Electronics. Revised metalwork on the electronics box, conversion to a double decker, to include the external SMPS, add connectors for an easier access to the microcontroller and implement a master switch to conserve OLED life;
  4. Microcontroller. Integrate a new version of the Arduino Nano (Arduino Nano Every), for a larger memory capacity to allow for advanced functions as well as an automated calculation of wow; and,
  5. Tonearm. Another 3009 to 3012 conversion, to allow for another SPU, and fill up the lonely, empty, third tonearm pillar.
The steps are now complete - but as each step includes a larger writeup (I point you to the opening post of this thread for reference) as well as a number of photographs, I will be posting one step per day over the next week.

Hopefully, by then, we will have reached the current state of the turntable and can proceed with rim drive experiments.

Until tomorrow, here is a technical drawing of the new (and hopefully improved) bearing:

nHkZihE.jpg
 
Beautiful work!

Have you considered putting this on the market to see what it would fetch and then potentially starting fresh on v2? I enjoy building things and look for every excuse to take on a new project. Nothing like this though!
 

I can't decide which impresses most, the single minded pursuit of technical excellence or the exquisite finish of all the components.

Do / have you built furniture for your home Insides?

Jim
 
Step #1. Bearing.

As promised, the first step of upgrades, dŵr's revised main (*) bearing.

UEvp3KE.jpg


(*) The "main" reference will come in handy later on when we touch on the rim drive.


Back to this bearing, however, it is still a single piece affair, machined out of a single piece of bronze (C932), in line with the technical drawings I published yesterday.

I again went for a single piece sleeve, as opposed to the more typically found bushings pressed into bearing sleeve bodies. I have found, in my albeit limited time available, that a unibody approach almost inevitably yields superior results. It is, however, incrementaly more difficult to produce, and for more complex bearing sleeves, a slip-up with the reaming procedure would probably mean scraping the entire bearing, so no wonder manufacturers rarely take this path.

As mentioned, machined of a single piece of C932 bronze (along with the threaded bearing cap). Spindle is hardened steel, thrust plate is Vespel (SP-21), retaining threaded ring is machined out of Polyoxymethylene (POM / Delrin), and bearing contact is achieved through an 8mm diameter ceramic ball (Si3N4). The ball is there on the photo for reference, as well as being an integral part of the bearing assembly.

The revised bearing is a bit larger than the original one. The spindle is now 55% longer (+50mm), its diametar is greater by 7% (+1mm), the total mating surface is now greater by 77%, and spindle tolerance is now H6 as opposed to H7 for the original bearing.

Finishing is also different. The bearing spindle has been sandblasted using glass dust at a 50µm granulation, followed up by TiN plating using PVD.

To visualize the size diferences between the new and old bearings, here is a photo of the old bearing and its components (the 8mm ball is the same size, in the snapshot for reference).

STbtek9.jpg


For the earlier bearing, finishing boiled down to polishing. But as it turns out, C932 takes kindly to fine polishing, and apart from some light tarnishing, it holds up quite nicely.

Given the change in form, the new bearing cannot be mounted within the same bearing block, so I took to machining a new bearing block. A single piece of 6082 aluminium, strategically damped internaly, through six sealed chambers with a combination of leadshot and epoxy.

EnW6LWT.jpg


And, again for reference, here is the earlier bearing block.

4KmVsTO.jpg


Expectations

I've made it clear to myself that I would be as realistic as humanly possible to my expectations here. I have been very pleased with the existing bearing performance, but as advanced manufacturing techniques became rather more available, I wanted to pursue the next (to me) possible limit. I was hoping for improved performance, but fully expected to have difficulties establishing any measurable differences.

Worst case, it should not work worse than the existing bearing.

Results

Regardless of whether one is a proponent of break-in or not, due to the nature of the bearing structure, this bearing needs to be worked in. I left the table spinning for ~12 hours before any considerations. A few measurement cycles took place the next day, letting the bearing warm up for approximately ~15 minutes.

So what did we find?

Subjectively:

  1. Reduced settling time (speed stability achieved sooner);
  2. Reduced noise (stethoscope does not yield any noise caused by rotational movement);
  3. Smoother start/stop sequences - less belt stress.
Objectively:

Currently the only available vibrational measurements to me utilize smartphone gyroscope sensors - although they have turned out to yield quite consistent results. What follows is a display of vibrational measurements for a ~600 second (10 minute) rotation cycle (at 33 1/3 rpm), for the system with the old (left) and the new (right) bearing (including their individual bearing blocks).

Z axis aligns with the central rotational axis for the turntable.

8OSwl8d.jpg


Moving on to the next step.
 


advertisement


Back
Top