Wednesday, 19 February 2025

Sony TC-520CS

The Sony TC-520CS

Thought I'd chance a purchase of this 'portable' cassette recorder made around 1976.

I was under no illusion about the deck, it's specification isn't high, but the price offered on ebay was worth the risk.


Primary Issues:
  • Worn Sony PP128-3602C Record/Play Head
  • Worn main drive belt.
  • Dusty internally.
  • DC Motor demanding too much current on PLAY, RW/FF etc.
  • Noisy controls and potentiometers.
  • Pinch roller had 'hardened' on one side.

Sony PP128-3602C Record/PB Head

The head was working in both channels but the wear was significant; so much so that extra back tension was required to hear playback clearly!
 
This reminded me of an old Sanyo M2000 cassette recorder I had many decades ago where I eventually had to change the heads.


The dilemma was either to get the head lapped professionally as there are no obvious direct replacements online, or seek another similar head.

The reader needs to be aware that this head's 'seating' on the platform is different to many similar replacements out there on the internet. The head is 'set back', where other similar heads are set further forward on the mounting platform.

Luckily I had a record head from a Sony TC-136SD 'parts machine' that was a direct replacement, although electrical characteristics appear to be a little different.

Pinch Roller

The old roller was taken out, and tapping out the 2mm pin prove easy. 

The original was a 13mm x 8mm x 2mm roller; this was replaced with a 13.5mm x 8mm x 2mm roller that works perfectly. No clearance issues.


Photographed when the Sony TC-520CS was first opened up.
Pictured are the original pinch roller, and the record/play head.

Showing Original DC Motor - top right.


Later when a new roller and record head was fitted.


The rear of this version of the TC-520CS indicates that this deck was for the UK market, there are no multi-voltage selections available here.

DC Motor

The original motor's power is sourced from the smoothed 6v full wave rectifier, but without any voltage regulation. Indeed, there are no further circuits to facilitate DC motor speed control.

So then, what mechanism is employed to control motor speed when PLAY loading momentarily changes?
 
The answer to this is - the original Sony motor works on a centrifugal force principle to regulate motor speed. The only means to adjust motor speed was to open up the said motor and make careful adjustments to the motor speed governor. However, this is not recommended since this cannot be achieved without disassembling the motor every time; some centrifugal motors have an 'access window' so that speed adjustment can be made. 
 

Tests also show that although the motor worked quite well (after cleaning the commutator), the current demand was at least 240mA during PLAY! Both RW/FF operations were much higher at 300mA and beyond. I've no doubt that this is well in excess of its design specification - perhaps the motor was momentarily stalling?

To compare this with another centrifugal motor I have in my stocks, the stocked item 'free running' current demand is only about 50mA.

Continuing on the subject of the original and somewhat high current demanding centrifugal motor system, I was initially concerned that the power supply was having to deliver excessive charge to the original motor, which would also increase rectifier mains ripple. 

This centrifugal dc motor also induced additional electromagnetic interference into the audio chain, which was noticeable on playback.

A replacement was sought, and in my stocks I had a 2mm spindle DC motor which was previously employed on a Technics RS-630US deck.

Using a 'home made' 2mm to 2.5mm shaft diameter conversion, this worked well, but is rated at 9V and not 6V. Speed adjustment is very sensitive and as a backup I had bought some (2mm spindle) clockwise ('CW') Mabuchi EG-530AD-6F 6V rated motors and so later carefully fitted these hoping that the result would be equally good?


With the 'new' Mabuchi (Chinese copy?) motor fitted, PLAY current is around 90mA, and RW/FF approximately 150mA or more. These a little higher than expected, but nevertheless acceptable. 


An extra circular strip of high relative permeability
is spring-wrapped around the motor to
reduce motor noise interference.

With the new Mabuchi motor fitted, speed control is facilitated by the internal potentiometer which is easily accessible. However, there was a drawback to this mechanical configuration - speed control wasn't broad enough. The minimal speed allowable is fixed at around 1.5% to 2% above the correct tape speed of 4.75 cm/s!

Why is happening? 

The answer is simple - the working pulley diameter to flywheel diameter ratio (D1:D2) is too large; this flywheel's diameter is just too small. A larger diameter belt wrap would have given a lower running tape speed.



The recently fitted square belt has 1.2mm x 1.2mm cross section, and so a smaller 1mm x 1mm belt has been ordered. A smaller cross section square belt should lower the effective gearing ratio by an estimate of 1% to 3%. We shall soon see!

So far then, on order: 1.0mm x 1.0mm x 76mm square belt.
 
Autostop:

The auto stop system that Sony designed for this and other similar decks is not electronic, but is mechanical. 

Within the system a worm drive is used, which is very low-geared. However, despite the low gearing, the collective torque load on the DC motor does increase periodically (approximately once per 3 seconds), so much so that a higher current-demand designed dc motor is needed for good speed regulation. Here the relatively small torque-current surge due to the auto-stop system would not affect the steady running of the motor, since the rise is largely 'buried' in an already high free running motor current. In addition, the brush to commutator contact area is generally larger thus keeping 'contact resistance' low as possible.

In audio terms, the current Mabuchi EG-530AD-6F dc motor and servo struggle to maintain a high degree of speed regulation, and so it is quite possible that the auto-stop mechanism will have to be disengaged later?

This every-three-second speed drift does not show up in wow and flutter figures since the 'wobble' is of very low frequency, possible 1/3 Hz? Nevertheless, it is perceivable on certain types of classical music. 

As an example of the speed 'wobble', the ABEX 3.15Khz test tape can deviate as much as 5Hz to 15Hz every 3 seconds. In contrast to a average deviation of 2Hz, or 3Hz at worst! 
 
Wow & Flutter:

The original specification for this Sony TC-520CS portable cassette recorder quotes a 0.26% (RMS weighted) wow and flutter figure.

The smooth running of the new Mabuchi motor, together with a suitable fitting drive belt yielded dramatically lower wow and flutter figures.


Power Supply:

The following only applies to the UK version.




New Square Belts Fitted:

Just fitted a 1.0mm x 1.0mm x 77.5mm square belt, then removed the autostop belt, and we have a 'transformed' Sony TC-520CS. Correct speed within tiny margins is now attainable.

Removing the autostop belt lessened the sporadic speed variations due to the every-three-second autostop generated speed 'wobble'. 

I strongly suspect that belt thickness does indeed impact on the 'wobble' when the autostop loading periodically increases belt tension, then releases it; is a 'stiffer' belt is less likely to flex during this time? However, a thicker square belt does tend increase wow and flutter. Which, in my opinion, is the reason why most manufacturers eventually opted for flat drive belts.

With a ABEX 3.15Khz reference tape inserted, WFGUI playback data suggests the very low frequency speed wobble is as much as 12Hz, and sometimes more. I have no choice at this moment other than to use a narrower square belt to get the PLAY speed down to acceptable levels.

The danger of not having autostop means no autostop for PLAY/RW/FF etc, but an additional risk that - if the take up spool fails to turn (for whatever reason), then there will be an unpleasant tape jam into the capstan to pinch roller line. However, the take-up spool is driven by the main drive belt and looks very unlikely to fail.

Speed Drift & Wow/Flutter:

Speed Drift with auto stop coupling ...


Speed Drift without auto stop coupling ...


Note: ±10Hz drift is approximately ±0.3% drift.

Wow & Flutter 
without auto stop coupling ...


Mean wow and flutter now under 0.08% wrms.

Supply Spindle

The supply spindle is 'capped' via a push-on 'end cap', but this lost its ability to fit securely when returning it to the spindle. A 1.5mm e-clip was then used as a substitute.








(24/02/2025: Corrections and additions may follow where and when necessary)

cassettedeckman@gmail.com




Sunday, 1 December 2024

Technics RS-671

The Technics RS-671

A risky purchase?, possibly?, possibly not? So, as a precaution I studied the aluminium fascia for scratches and corrosion, and then estimated the potential corrosion or rust internally. After some time I decide to take a chance and conveyed my buying concerns to the seller, and offered to buy for 35% lower than his asking price.

There's little doubt that this machine had been left in 'storage' for 10, 20, or more years. There was even a small twig to be found inside the machine next to a smoothing capacitor - my only conclusion was that it had been left in a garage for a very long time.
 
Some Original Advert Images:
 

The paintwork on the top of the deck was so poor, I could easily scrap it off, but concerns were more directed towards the inside of the deck.


More original advert pictures ...





I have to confess, I did not take any photographs of the insides of the RS-671 - I just went to work on diagnosing the main problems and fixing many issues.

Internally it wasn't as bad as the external condition of the deck; just as I had calculated!, but much work needed to be done. Smoothing capacitors were probably leaking since during initial testing, only 10v was available to the motors, when it shoud have been around 14v to 14.5v.

The audio section PSU was cleared of debris, and the moldy areas were cleaned and allowed to dry. Later the main power supply capactors were replaced. Even the zener diode on the audio regulator circuit was replaced with a slightly higher voltage value; 20v instead of 18v. I don't have any 18v zener diodes in stock at this moment.



The usual cleaning and lubricating practices were applied to the mechanisms, including all rotating components of the cassette transport.

Since reporting on this Technics RS-671 repair is very time consuming, what follows below, is not generally in the order that these repair jobs were done. Indeed much of the tedious repair work is not written about.

Loose Take-up Spool Idler Tyre

This had worked loose, although the outer rubber was fine, the inner diameter had apparently stretched!

There was no obvious replacement available, so I bought a few idler tyres to try. The actual diameter of the aluminium rim on which the idler will sit, is 14mm. 
 
A 16.4mm x 11.8mm x 3mm was the best option, but I had to 'sand it down' to an outer working diameter of about 15.5mm. The reason being - the idler would interfere with the take-up spool/reel when the Technics was set to RW mode.

The effective idler diameter was reduced by employing a tight fitting drill bit, and putting the drill into a drill stand while I carefully applied a sanding block to the rotating tyre. Later the tyre was cleaned and Rubber Renue applied.





 
Rewind/Fast Forward DC Motor
 
Is a DC motor which appears to work on a centrifugal force principle to regulate it own speed. At a certain angular velocity enough centrifugal force is generated to throw out 'weighted contacts' and temporarily disengage the DC supply from the motor. An access port is provided if angular velocity settings are to altered by the user - in this case I didn't change anything.
 
The motor was thoroughly cleaned with the aid of a glass fibre pencil and switch cleaner.
 
Before Cleaning.

Showing it Access Port.


 The motor has to be isolated before returning its original casing.
 
The Capstan Motor

Based on the uPC1003C2 botor controller; a 'hard to find' IC that (if found) will be costly, so beware!

Note: the photographs that follow, are not in the order in which I worked on the motor.

Initially I just replaced the two electrlytic caps -
both indicated with a white mark.





Both motor brush (left) and commutator (right)
were cleaned with switch cleaner and a glass fibre pen.

The rotation of this motor is clockwise.

Motor current demands are dependent on load torque, and if we ignore the basic electronics of the motor controller circuit,
Imotor ∝ Tmotor.

At a set speed (usually around 2400rpm) the total load torque is derived from the take-up slip friction action, the flywheel mechanism, and other
mechanical frictional losses,

Initial DC current demands from PLAY action were around 125mA, which from my experience is considered too high. This will put a mild current strain on the brush-commutator interface. At that time I used a temporary flat belt from the motor to drive the flywheel. The belt was really too small - 122mm (diameter) x 5mm (width) x 0.6mm (thickness).

After fitting a similar belt with a slacker diameter of 135mm, the load current dropped to a mean of 100mA; much better! This figure was further improved, when the 'new' modified take-up spool/reel idler was fitted to replace a temporary 'O-ring'. Now the mean PLAY DC load current was hovering around 89mA to 90mA.

This figure could be improved still further if and when I decide to reduce the amount of take-up spool slip friction. This can be achieved by effectively reducing the force offered by the compression spring within the take-up hub.

Of interest the motor plus controller load currents with the motor running in PAUSE mode ~ 52mA, and in PLAY mode (without a cassette 'load') ~ 60mA.

With the take-up hub removed and disassembled, I attempted to reduce the amount of compression that the default spring offered. This was more difficult than I first anticipated; how do you reduce the size of a compression spring effectively?

However, with a little effort I did manage to get the PLAY load current into the motor controller down to around 79mA ... 85mA, but a mean value of about 75mA was achievable with another third-party compresson spring that I opted not to use, as it was slightly ill-fitting.

Since this is effectively a 'top loading' cassette deck, it is worth noting that total take-up spool friction during PLAY will statistically be a little higher than for a front-loading cassette machine. So having a good amount of take-up torque is essential, but of course being mindful that it must not be excessive.
 
Original Compression Spring for the Take-up Hub

DC PLAY loading current ': 79mA ... 85mA.


Although the motor now works well, I am not 101% confident that all is good; I've had issues with it, and yet have not been able to pinpoint minor problems. So far it's stable with little speed drift, typical figures are usually within ±3Hz (±0.1%) from a 3150Hz ABEX reference tape over 10-15 minutes.

Wow/Flutter
 
Better than expected considering the state of the deck when I bought it.

After careful 'sanding' of the pinch roller, and a slightly reduced load on the motor, wow & flutter now hovers between 0.04% to 0.07% (JIS WRMS), with an occasional excursion beyond those figures.



All these fixes are provisional, since the RS-671 has so many pending issues to address. Two such issues that are on the list are: (1) Pinch roller replacement, (2) Record-Play multiswitch, which does need to be removed, and internally cleaned due to occasional 'pegging' of the right VU meter.

03/12/2024
 
Record-Play Multi-switch Fix 05/12/2024
 
The said switch was carefully de-soldered, and very carefully dissassembled. Then the internal contacts cleaned with a glass-fibre pen and switch cleaner. This procedure is indeed risky, but with patience and perserverence it can be achieved.

Solder side view with switches removed.

Retrieving the switches - note the different sliding switches

After cleaning with contact cleaner and a glass fibre pencil.

Preparing to return the switches.

All good, no sporadic oscillations in either channel.

System Control

This version of the RS-671 has a set of 4 x 20mm fuses post-transformer, but pre-rectifier.

Attention was brought when the 2000mA 'slow-blow' fuse open-circuited after many testing cycles of the RW/FF circuit. Of interest to the reader - RW/FF DC motor loading current was typically anywhere between 125mA and 225mA, and would rise to approximately 300mA when the RW/FF motor reached the end of the RW/FF operation. This is effectively a stalling condition, where we can expect armature current to be in that 0.3A region.

The circuit below illustrates some circuit loading ac current conditions under either RW or FF, and indeed the large rise of secondary current when the autostop (the final of two phases) is initiated.

Original Schematic from the RS-671 Service Manual
Note: No fuses are suggested?   

 
18/12/2024
 
Rewind/Fast Forward Mechanism (Revisited) (22/12/2024)
 
I wanted to improve the torque for both RW and FF, especially RW as sometimes on difficult 'inertial' cassettes, the RS-671 would struggle. However, most cassette tape are fine - no problem.
 
There is a sliding mechanism which on the activation of the RW/FF DC motor will swing leftwards for RW, and to the right for FF. 
 
The area underneath the mechanism was cleaned (again) with metal polish to make it as 'friction free' as possible. The guiding mechanism originally had some small rectangular pieces of woven nylon bonded to the mechanism to act to raise the arm, and also to offer a very coeficcient of friction.

The main nylon-like mesh was removed, and underneath the mech a cut piece of polished PTFE (0.5mm thick) was bonded using double-side 'stick tape'.

During RW/FF it is quite possible that any roughness may inhibit good idler to idler contact, hence the motive to change the previous set up.



So far, there is a noticeable improvement in RW/FF torque since Rewind is almost 100% likely to finish the Rewind task for one difficult cassette tape I have in my possession. The torque may be further improved if the knurling on the motor drive was deeper? However, for now, it is unlikely that the machine is going to be opened up again for a while.


Wago Connectors
 
Supply to both DC motors is now made via four Wago connectors. This is more convenient so that measuring loading current during PLAY, RW and FF operations can be done without resorting to de-soldering and re-soldering.
 
 



 
(Article is subject to corrections, and additions without notice.)