Sony TC-136SD Restoration
Date Began: 12/02/2021: an on-going progress report, now finished.
Bought off ebay about a week ago at a slightly risky price of £47 including postage, I looked forward to owning this deck with its Ferrite & Ferrite Heads - the one that was supposed to last up to 200 longer than conventional heads at the time.
This deck was manufactured between 1975 and 1977?
The
seller was unable to assess if or not, it was working properly.
However, I knew from the advert that the seller didn't realise that the line out connections were intended for an amplifier, and not directly to loudspeakers.
Immediate Issues
- Dirty internally
- Noisy switches
- Slippy idler tyres
- Plastic Idler 'rough up' with a soldering iron!?
- Servo Motor soon began to short circuit - causing a fuse to blow.
 |
| I
decided to check if the old motor was faulty and not the controller - I
wired in contacts to my variable power supply. Varied motor voltage Vm from about 2.5 - 6v and examined the current Im. Most of the time motor current was low (Im <100mA free-running), but then randomly Im would surge to over 500mA and then the motor stopped. Somewhere, internal short-circuiting was taking place. |
- Previously repaired Eject, Pause, and Record buttons were again failing to engage.
Positives
- Heads in very good condition - no notable wear on either erase or record heads.
- Pinch roller - also in good condition, just a little dirty.
Repair Tasks
(1) DC Servo Motor Replacement
The old motor ran at ~2400rpm, but in a clockwise motion using a 2.5mm shaft/spindle. The task now for me was to obtain a motor replacement - but my stock of motors were all 2mm shafts!
Luckily, I found a suitable, but perhaps not an ideal solution? - a 'sleeve' that would fit snugly and allow the old 2.5mm pulley to work correctly on the new 2mm shaft.
The decided 'new' motor was a '12v' Matsushita CCW motor, but with all the internal electronics desoldered and removed, except for a potentiometer.
I had to modify the motor and employ an alternative controller to supply the deck with 2400rpm clockwise motion.
 |
| Matsushita 12v DC motor: now with just connections to its brush contacts. |
In an earlier blog I employed the AN6651 DC motor controller chip to another project, I was going to do the same for this cassette deck. https://cassettedeckman.blogspot.com/2020/05/external-motor-controller-experiment.html And why not, it worked exceptionally well for the Sansui SC-1300, so why not this machine?
After soldering together a new motor controller circuit, I then tested the circuit's correct functionality before I temporarily wired it into my TC-136SD deck.
Of course, the Matsushita motor must turn clockwise, if not, reversing the connections to the motor solves the problem.
The circuit works - audio is very stable, so now the next task was to make some permanent wiring solution.
 |
| I had no 100Ω potentiometers for 'coarse' speed setting, so a 200Ω was used. |
With the controller circuit fitted, the speed was adjusted using both a 400Hz, and 3150Hz test tone reference tape, which was then compared to the equivalent digital reference.
The speed of the TC-136SD is very stable, no audiable wow and flutter, apart from the occasional drop-out transient.
Fixes to date
- Removed all dust and dirt.
- Cleaned all mechanisms and springs with isopropyl alcohol.
- Lubrication of mechanisms with PTFE based spray, and fine oil for the capstan shaft.
- Switch-cleaned (Servisol Super 10) all switches, and contacts.
- Rubber idlers all cleaned, carefully 'roughed up' and Rubber Renue applied.
- New belts fitted.
- Power supply unit: all electrolytic capapcitors replaced.
- Old servo motor removed, and new modified clockwise dc motor and driver circuit fitted.
- Potentiometers switch-cleaned
- Level meters read '0VU' at the old DIN standard of field strength measurement quoted at 160 nWb/m.
- Internal playback levels balanced
- Internal input levels adjusted for approximately TDK D sensitivity, and balanced.
- Record, Pause, and Eject keys all carefully repaired using Gorilla Epoxy resin. Gave about 48 hours to fully cure. All other keys are in good condition.
Replacing Electrolyic Capacitors
The power supply had all its capacitors replaced early in this restoration project.
Currently
I am in the process of replacing all electrolytic capacitors in the
pre-amplifier, equilisation, and playback sections. I usually change
about 6 capacitors (symmetrical replacement) before testing the deck in
the upgrade cycle process. (20/02/2021)
VU Meters
I intend to make the VU display scale appear white for a more vivid appearance.
 |
| Sony's TC-136SD - working, but still a work in progress. |
LED lighting for VU Meters
 |
The original circuit to illuminate the VU meters utilised the 50Hz alternating 6V tap off the transformer, protected with a 500mA fuse.
The
VU Meters now have white LEDs to illuminate the display. I could have
wired a circuit that would work similar to half-wave rectification, but
decided to experiment with a rectifier circuit that also delivers double
input voltage. A half wave circuit also produced flicker which I noted
immediately.
The
voltage doubler is not strictly necessary, but since I have never wired
such I circuit together, I though it would be interesting to see how
this circuits responds to the LED load in detail.
*************
New LED Circuit
This new circuit was to excite the LEDs with near constant current, in contrast to a flickering 50Hz display.
 |
| Actual Circuit Used This circuit simulation shows the respective current surges for each component as indicated by colour. Once the capacitors are charged, the power supply merely 'tops up' the charge in the capacitors to meet the demand of the LEDs. An AC ripple current component can be seen running through the LEDs - larger valued capacitors will lessen the ripple. The average LED current is approximately 6mA. |
LED Circuit Calculations
Most LEDs are rated somewhere between 1.8v to 3.5v at around 20mA for a rated brightness level. However, white LEDs are typically 3v or higher. At 20mA current, this will be too bright for the needs of this display, so a trial and error approach was adopted.
Since the peak (non-loaded) AC voltage was about 11volts, I could expect the voltage doubler to reach approximately 22v, minus diode voltage drops, and further small voltage drops due to internal transformer resistance under loading conditions. Under a loading of about 10mA-20mA, the voltage doubler delivered approximately 18 volts.
To estimate the series resistance is relatively simple: we have 18 volts across the Vr+2ᐧVd circuit. Since we can expect the voltage across each LED to be approximately 3v, that leaves Vr ~ 12v. Then if the LEDs are sinking ~20mA then by Ohm's law {V=IᐧR}, we have R= 12/0.020= 600Ω.
As mention earlier, the brightness at Id ~20mA will be far too high, so I then approximately quartered the current (and brightness?) to 5mA by increasing R to 2400Ω, but later finally settled on using a 3300Ω resistor. The display isn't too bright, nor too dim.
Other Pending Tasks ...
- Complete electrolytic capacitor replacement (as yet unfinished .. 28/02/2021, 13/03/2021)
- Complete signal transistor replacement, probably replace all 2SC1361/2SC1363/2SC1364 with KSC1815-GR series?
It
turns out that they were not 2SC1362 as stated in the
service manual, but 2SC632A NPN transistors!
Looking at the
specifications of the 2SC1362, 2SC632A, and comparing to a modern approximate equivalent - the KSC1815-GR, I decided to replace the first 4 transistors for both Left and Right in the playback pre-amp circuit.
A quick hFE test
on my pack of KSC1815-GR revealed they had high hFE (DC current gain at a
quoted Ic current) at about 300 - this is higher than the 2SC632A - at least on paper? However, I have just measured the 4 removed 2SC632A transistors, their (DC current gains) were: hFE= 274, 362, 424, and 432! Without mathematical analysis it may be difficult to appreciate that these variations in DC bias hFE figures will not affect the quiescent collector Ic or emitter Ie currents significantly, nor Vce that much, but changes will be detected at the input, in particular Ib and less so Vbe. (Proof needed?)
Maximum Collector Rating: The maximum rated collector current of the KSC1815-GR is quoted at 150mA, but there is
little evidence that these 1815s are going to be driven hard. The
original 2SC632A collector Ic quiescent currents in each circuit (according to the service manual diagrams) will be
about 120uA (12v/100,000Ω), and 1.4mA (8v/5600Ω) respectively.
Tested the deck after 4 transistor replacements for record and playback - all is fine! No issues to report. 👍
02/03/2021: Further transistor replacements (four: 2⨉ Left Ch then 2⨉ Right Ch) as marked above. Both 2SC632A and 2SC634A replaced with the Fairchild KSC1815-GR. Again hFE measured at ~ 300, although hFE variations will not make much difference to biasing, nor signal gain, since the negative feedback network should be dictating gain. Also Q103, and Q203 were replaced with KSC1815-GR.
04/03/2021: And now Q107/Q207,Q108/Q208 have been replaced with the KSC1815-GR. The Q107/Q207 are muting transistors.
It is very important to clean both sides of the circuit board where desoldering and soldering have taken place - it may be better to dry-clean the board with Q-Tips. Any residual flux may act as a conduction path, and in circuits like these which react to small currents, additional currents will flow if there is insufficient isolation between circuit board tracks. The results are somewhat unpredictable, but the audio will suffer!
 |
| The Sony TC-136SD has an additional sheilding board which is earthed via the screws on the right. |
The Sony TC-136SD is working very well, and after replacing 14 transistors in the playback stage, the machine is sounding much better than when I first switched it on. The playback amplifier's background noise is significantly less and the treble is cleaner. Very pleased!
More Transistor Replacements
I ran out of stocks of KSC1815-GR and so decided to continue using the KSC1815-Y. These 'Y' variants have a lower quoted DC hFE rating, but are sufficiently high enough so as not to disturb bias and signal. As stated earlier, reasonable variations (~±33%?) in hFE and hfe very probably won't alter biasing, nor engineered signal gain effectively (when negative feedback is employed) - not proven here.
I've also replaced more electrolytic capacitors too.
All transistor replacements so far are shown, note the late inclusion of C1845FCK ...
- Later: replace power supply regulator transistor and zener diodes?
===================================
Bias Adjustment
The TC-136SD employs a kind-of switchable discrete capacitance approach to bias adjustment. There are four discrete capacitors: 15pF, 18pf, 22pF, and 27pF which can be 'wired in' in parallel, and so in addition to the variable 8pF trimmer. The latter gives fine control over bias. Initially this Sony TC-136SD was wired so that 22pF and the 8pF trimmer controlled the level of bias.
However, this setting was over biasing Type I ('Normal') tapes, as I couldn't get the TC-136SD to reproduce a flat frequency response between 1Khz and 10Khz at -10dB, or -20dB ref: 0VU.
Circuit Board with Modification
I eventually settled for the minimal capacitance setting of 15pF + Δ8pF trimmer. The 'Δ8pF' term denotes the trimmer is variable from 0pF to 8pF.
14/03/2022: Bias Modification Circuit
Recently I've settled for this option - all the capacitor banks have been disengaged, and one 10pF~60pF trimmer capacitor (100v rating) inserted for each channel.
This setting now allows the TC-13SD to reproduce a differential of only 0.5dB in frequency response at 1Khz and 10Khz. And at both modulation levels of -10dB, and -20dB, ref: 0VU, 160nWb/m DIN standard. (I am not suggesting the FR is flat between 1Khz and 10Khz - I may investigate this later?)
However, at -20dB, the frequency response at 12.5Khz was approximately -2dB down, as somewhat expected for this deck.
=====================
Additional Images
New Pinch Roller
Finally, I nearly forgot to put in a new pinch roller!
 |
| Chinese made pinch roller working superbly well. |
This circa 1975 Sony TC-136SD is working superbly well. The sound playback is clean and balanced. Background playback amplifier noise is low - probably slightly better than experienced back in 1975 thanks to lower noise-figure KSC1815-GR and KSC1815-Y NPN transistors.
The record/replay head azimuth is correctly aligned, playback levels are equal, and have been calibrated using a dolby-level reference tape: ~+2.7dB above 0VU. The internal record levels have been calibrated to match that of TDK FE90 cassette tape.
*************************************
Pause Key
The Pause key broke again, clearly a design weakness as previous owners would probably agree.
This time I cleaned away the epoxy resin from the key's underside and drilled a pilot hole to fit a nail that would be used to drive the Pause control into action. Clearly this unit requires quite a lot of force to engage Pause, so this time I had to make sure a revised Pause key wasn't going to fail again.
 |
| The nail extends through the drilled hole, almost to the end. |
 |
| Pause now works very well - action feels 'solid'. |
 |
| 29/03/2021: Main Circuit Board - fully recapped (marked in red), still some transistors to replace later (marked in green) 30/03/2021: All transistors replaced, including the Dolby circuit. |
Calibration
I have finally calibrated this deck based on an full track ABEX (10Khz azimuth) , and an ANT-AUDIO.CO.UK (Dolby Level) reference tape.
The frequency response of the TC-136SD is around 30/40Hz ... 13,000Hz ±3dB at -20dB, ref 0VU for Type I tapes.
The frequency response is still fairly flat at -10dB modulation (ref 0VU) from 100Hz... 10,000Hz, but falls sharply after that.
The deck now can produce quality recordings, which are very stable. My only criticism is the obvious high frequency saturation and aliasing-like distortion that occurs recording from a high frequency spectrally aggressive CD source. Recordings from stereo FM are usually fine.
END
These blogger pages are subject to potential corrections, and possible minor additions. 30/03/2021.
Small technical correction: 02/04/2021.
05/04/2021: Q101/Q201 = C1845FCK.
