Binson Echorec Memory System
by Phil Taylor
When Binson Echorec echo units were first introduced in the late 1950′s, effects technology was limited to studio effects, tape delays, plate and room reverbs and tube compression. The advantages of the Echorec over other proprietary echo machines of the day were sound quality (wide frequency response), quietness, and stability (very low ‘wow’ and ‘flutter’). The Echorec was considered to be the top of the range echo. The founder of Binson, Dr Bonfiglio Bini and principal engineer, Mr Scarano Gaetano undertook a research effort to develop a signal delay storage medium that was superior to plastic magnetic tape. The result was a unique device, utilising a specially designed steel/alloy disc or drum, which carried a durable flat metal magnetic band. The drum was driven by an A.C. motor via a rubber idler wheel (similar to a Garrard 401 turntable). Record and playback heads were arranged around the edge of the drum. The stability of this transport was a significant improvement over tape with reduced wow and flutter. However, this hi-fidelity did not come cheaply, when one considers the price of a ‘Baby’ Binson in 1962 was around £140-00. Compare that with a Vox AC30 at £119-00, or a Strat for £160-00.
The Binson magnetic drum was constructed from three separate parts, the steel axle, balanced wheel and aluminium thread ring. Approximately one hundred turns of very fine 0.1mm (0.004″) diameter recording-grade stainless steel wire were wound around the circumference of an aluminium thread ring – a very tricky operation. The wire was wound on to the drum using a modified coil winding machine which Binson built specially to do the job. The wire had to be kept tight against the drum face with no overlaps and no kinks and the beginning and the end the wire was secured by a small pin (the external pin is visible on the edge of drum). Even more challenging was the following operation of using a lathe and very fine (280 grade) tungsten carbide grinding wheel to accurately mill the rounded edge off the wire so that it was almost, but not quite, semi-circular, creating a very smooth, flat wire surface that could be properly magnetised by the heads. This was a delicate and crucial operation and the milling had to be performed very slowly as the wire would all too easily break. A new worker on the job would break a lot of wire before finally getting the knack – this was among one the most difficult operations at Binson factory.
Magnetic drums are no longer manufactured and it’s rare to see used, let alone new old stock ones for sale. Consequently, even an old or damaged drum with worn surfaces is still a valuable commodity – it’s irreplaceable – and it’s perfectly feasible to rework it. The drum can be rethreaded with new wire and resurfaced to restore it back to factory-new condition. The milling and finishing work is vital if the hi-fidelity performance of a brand new machine straight out of the Binson factory is to be achieved. If the wire is left round the head will only be able to make direct contact at one small point on the wires’ circumference as shown in the cross-sectional drawing (a) below. It can be seen that just about all the head to wire contact is lost to air gaps (coloured in blue), which results in a weak signal being ‘printed’ on to the drum. In practice this means there’s no possibility of attaining those lingering, crystal clear repeats as heard on Pink Floyd’s 1973 track ‘Time’ from ‘The Dark Side of the Moon’ and the signal to noise ratio of the Echorec will be lousy too. Drawing (b) shows how milling off the wire rounds maximises surface area of the wire (coloured in orange) creating a wider recording track for the head to magnetise.
On an historical note, wire recording machines used to zip along at a rapid 2 feet per second (24ips). Even though the wire was passing the heads at speeds higher than a modern professional studio tape machine, the audio performance was still poor. By increasing the effective surface area of the wire Dr Bini improved the fidelity of the Echorec in comparison to the old wire recorders.
The drum is beautiful example of precision engineering, being accurately milled to fine tolerances. Although the drums were manufactured in the Binson factory, the high level of precision machining required between the drum axle shaft and axle housing meant that the lapping and calibration work had to be undertaken by an external company that specialised in building injectors for diesel motors. The drum and housing were manufactured as a pair, matched one-to-one with the same serial number so it’s important to note that if purchasing a spare drum that the axle housing is included as well. Never buy a spare drum without the axle housing: it will never work properly.
Early Binson machines were fitted with drums manufactured from solid cast iron. These were very heavy, weighing in at almost 3lbs. This bit of over engineering was almost certainly intended to reduce ‘wow’ and ‘flutter’, that is, slow and fast pitch variations due to changes in the rotational speed of the drum, to an absolute minimum. All that mass meant that once that metal drum had attained speed it would be incredibly stable – the angular momentum of that spinning chunk of iron acted like a mechanical low-pass filter effectively filtering out any transient fluctuations in mains voltage or unpredictable dynamic frictional changes, reducing wow and flutter to vanishing point.
The same approach was taken to minimise wow and flutter in idler wheel driven turntables, such as the Garrard 401, where the mass is an alloy platter rather than an iron drum. The wow and flutter of a 401 is measured at 0.05% – right at the limits of what test gear could measure at the time. Technically there’s no reason why a properly serviced Echorec should not approach this figure or, to put it another way, wow and flutter should be inaudible in a properly serviced Echorec. Later on Binson began fabricating drums from hollow mild steel to significantly reduce their weight to just under 1lb. My guess is that Bini decided that all that mass was overkill and he eased off on the specs.
The picture on the right shows an Echorec head that has been machined to give it the correct face profile. The head profile needs to be different from a standard 1/4″ tape machine head, to aid ‘zenith’ alignment so that it makes proper contact against the hard surface of the magnetic drum. The drum must be kept lightly oiled to reduce the wear where the heads contact the drum to ensure a long service life. Any light oil such as sewing machine oil can used. The lubricating oil must not be allowed to get on the rubber idler wheel otherwise it will slip against the drum causing excessive ‘wow’. The record and playback heads are positioned around the circumference of the drum, the first of these is the record head and the others are the playback heads giving four possible different delay taps. There is a delay of approximately 75ms before the first playback head to reads the recorded signal from the drum, 150ms for the second head, 225ms for the third head and the longest delay is 300ms for the fourth head. Heads are selected with the knob on the front panel which gives 12 separate echo selections from one head alone to complex multitap effects. The table below shows head resistance measurements for three different models of tube Echorec machine serviced and repaired by Binson technician, Mr. Scarano. From this data it can be seen that the impedance of the playback and record heads installed in an Echorec such as the Baby or B2 is between 600Ω—700Ω.
|ECHOREC MODEL||RECORD||PLAY 1||PLAY 2||PLAY 3||PLAY 4|
|BABY TF3-A TUBE||666Ω||584Ω||598Ω||602Ω||618Ω|
If the Echorec’s drum has been resurfaced then the heads will almost certainly require re-lapping (recontouring) as the drum will have caused uneven or adverse wear on the heads. This results in a loss of high frequency response of the delay repeats and a higher incidence of dropouts because of poor head to drum contact. Head relapping is a highly skilled job – a thorough knowledge of the materials used in head manufacture (mumetal, ferrite, recovac, vitrovac, sendust, etc) is required before attempting restoration. The materials used in the head must first be identified in order that the correct method of relapping can be determined – different head materials require different specialist lapping compounds, however as long as the heads aren’t too worn they can be reworked to “as new” condition. If the heads are beyond repair then replacements can be obtained from Photovox company in Italy – the original supplier of heads to Binson back in their heyday in the 1950s.
The original Echorec memory system transports were all fitted with a powerful A.C. induction motor that rotates at fixed speed. The synchronous speed of the motor is determined by the A.C. mains frequency and number of poles in the motor winding (not voltage!). This is 50Hz in UK, Europe, Austalia and most of the world, except America where it is 60Hz.
The speed of the motor can be calculated:
ω = 2 × 60 × f / n
ω = pump shaft rotational speed (rev/min, RPM)
f = frequency (Hz, cycles/sec) = 50Hz
n = number of poles = 2
ω = 2 × 60 × 50 / 2 = 3000RPM
Echorecs manufactured for the American market were fitted with motors that had smaller diameter spindles to correct the gearing ratio so the idler wheel was driven at the same speed as European models. Consequently the motor speed in a European Binson Echorec will be about 20% faster when operated in America resulting in the overall delay time being 20% shorter. Altering the diameter of the rubber drive wheel or changing the voltage WILL NOT alter the speed of the disc. Binson, UK can undertake a modification to add a D.C. motor and vari-speed control circuitry to alter or fine-tune the delay repeats (tempo matching). The motors are high quality units manufactured by Trident Engineering Ltd. Part number: 2-40/1212.
A few vital statistics from my own Binson Echorec model T7E:
Measured drum diameter (d) = 4.7″
Drum circumference = π × d = π × 4.7 = 15″
Measured rotational speed of drum = 71RPM, so every minute, 71 × 15″ = 1065″ of the magnetic band on the drum surface passes across the heads. Converting this to inches per second:
1065 / 60 = 18ips, which is very respectable tape speed for hifi reproduction
The motor spindle circumference can be calculated from:
Drum ω = 71RPM
Drum circumference = 15″
Motor spindle ω = 3000RPM
Spindle circumference = 71 × 15″ / 3000 = 0.36″
The measured the spindle diameter was just over one tenth of an inch, which gives a circumference of about 0.31″
The delay time can be calculated from the measured distance between heads which is 1.3″ (34mm). If the drum speed is 18ips = 457mm/s then delay time for the first playback head is:
First tap = 34 / 457 = 0.074s = 74ms
and the Second tap = 2 × 74 = 148ms
and the Third tap = 3 × 74 = 222ms
and the Fourth tap = 4 × 74 = 296ms
Now here comes the really clever part – these playback head distances are not arbitary. A considerable amount of thought was put into the positioning of the heads around the drum by Dr Bini, the designer of the Echorec. They were chosen so that the delay times are musically related to one another. If the fourth tap is considered as representing a quarter note then the third tap is a dotted eighth note. This is a very useful and inpiring repeat pattern and has been utilised by guitarists such as David Gilmour and ‘The Edge’ to build some fantastic riffs. Continuing on, the second tap represents an eighth note and the first tap a sixteenth note. These shorter delays can be heard in slapback and rockabilly music from the 1950s and, again by Gilmour to good effect on tracks such as ‘Time’ (The Dark Side of the Moon).
The Idler Wheel
A rubber idler wheel transfers the rotational velocity of motor shaft to the drum. The idler wheel isn’t directly coupled to the motor shaft, it is mounted on a metal baseplate that is free to slide towards the motor shaft and the circumference of the magnetic drum. Pressure of the idler wheel against the motor shaft and drum is maintained by a spring-loaded mechanism. Binson’s use of a powerful, high torque A.C. motor and idler wheel in the design implies they were aiming to ensure accurate and consistent drum speed in the Echorec.
A useful tip to keep the idler wheel in tip-top condition: If the machine is going to be stored or is unused for any period of time then it’s a good idea to make sure that that idler wheel is not left in contact with the motor spindle. If the metal spindle is left pressing against the rubber it will deform it, leaving an impression or dint in it. To prevent this from happening, make sure that the rubber wheel is pushed clear of the spindle by removing the tensioning spring. Deformed or hardened idler wheels can be sometimes reworked by soaking them in hot water as this softens the rubber, but it’s better to prevent this happening in the first place. The idler wheel is another one of those irreplaceable Binson manufactured parts, N.O.S. ones rarely come up for sale and if even when they do, they’re expensive.
It’s also worth mentioning another new idler wheel design made by Italian engineer Marcello Patruno for his custom ‘Special Binson‘ Echorecs built from N.O.S. Binson parts. His Echorec idler wheel is fabricated from aluminium and has a thin rubber rim, rather then being solid rubber, a little like a low profile car wheel. The idea is to prevent its shape from distorting, therefore yielding lower ‘wow’ than the original Binson solid rubber idler wheel. His new-build Echorecs also feature a wide frequency repsonse of 45Hz to greater than 18KHz, but the best thing about his machines is that they’re capable of much longer delay times than a factory Binson – delay range is 150ms to 850ms (a standard Echorec is approximately 350ms). His machines also feature manual bias oscillator control, variable motor speed and are genuine improvements over the original Echorec 2° machines.
Overhauling the Memory System
The picture above shows a completely refurbished memory system. This work was undertaken by Eric Snowball, Binson UK in 2009. At the time Binson UK had the capability to reface the drum, relap the heads and clean and replate any corroded metal parts so that the memory system could be restored to “factory new” condition – the original audio performance of the Echorec is restored! All you had to do was remove the memory system from your machine and mail it to Eric to work on. However, the work was specialised and labour intensive costing around £700. Make sure you look after your machine! Binson, UK could also undertake a modification to add a D.C. motor and vari-speed control circuitry to alter or fine-tune the delay repeats (tempo matching). The motors are high quality units manufactured by Trident Engineering Ltd. Part number: 2-40/1212.
Special thanks to Luigi Amaglio for supplying excellent close-up photographs of Echorec parts and the table of measurements for playback and record heads.