Mitsubishi DT-4700 Cassette Deck

 Mitsubishi DT-4700 Cassette Deck

 

Bought early January 2024 for just £25 including postage, I decided to gamble on restoring this good-looking, well made deck, rather heavy cassette deck.

Original Advert:

 

 

Today, the Mitsubishi DT-4700 has been fully restored.

Some of the work undertaken is summarized below.

DT-4700 Fascia & Controls: Washed and dried. All chassis screws frequently used had their threads 'bathed' in oil - making them easy to take out, and and return. The tape activation controls were cleaned with a mixture of metal polish and water 'worked in' with a cloth until they were shiny, then finally polishing the controls with a dry cloth.

Switches & Potentiometers: All cleaned with contact cleaners.

Pre-amplifier, Dolby Board, and Record Amplifier: Completely re-capped, and re-transistorized.

Power Supply Unit: Re-capped, some new diodes, new regulation transistors.

When working through the above, it is always important to periodically check system voltages - PSU, and amplifier stability before moving to the next batch of components.

Mains Switch Failure:

The original mains ON/OFF latching switch eventually failed, internally a plastic pin had broken. It became obvious that repairing this was out of the question; when working with high voltages it is advisable not to take repair risks.

 

Luckily I found some 250v rated single pole, single throw (SPST) switches on ebay that could be adapted. 

Shown below is the original mains switch and transformer circuit. Note the 10nF suppressor capacitor is across the Live and Neutral. I may add an additional suppressor across Live and Neutral later!?

Failing Auto-Stop:

 

At the rear of the cassette transport is a rotating disc that effectively interrupts an infrared beam from an excited, forward-biased LED, which is then picked up by an adjacent photo-transistor. The design is to generate light interruption and hence develop a square wave which then feeds an RC timer circuit. If the period of the wave is too long, then an auto-stop is generated by energizing a solenoid to force a Stop.

 

I had problems with the auto-stop unit, and had to run the deck with this feature disabled for a time until the problems were diagnosed.

After replacing transistors in the auto-stop circuit, it later became apparent that one major source of the problem was the photo-interrupter circuit; this being the light emitting LED and photo transistor combination.

Testing the unit 'offline' suggested that the LED was faulty - I could not forward bias the diode to conduct.

A not-so-perfect solution was then to re-house the unit with a new infrared LED/photo-transistor combination. The unit had to be part-drilled to enlarge the aperture so that both LED and transistor could be fitted. Unfortunately this did not produce a great fit, so gluing the two  semiconductors was the only viable solution - a 'glue gun' was used instead of a permanent 'super glue' bond.

The LED and photo-transistor had little effective depth, and so gluing was the only option, the 3mm bore also proved to be too wide.

Front side of IR LED and Photo transistor Circuit Board

Solder side

An infrared LED, infrared photo transistor, and mounting screws
are 'buried' inside heat-gun glue.
It's not 'pretty', but it's strong and stable.
(I may later re-work this?)

The work done doesn't look professional or 'cool', but it is functional and very stable.

The circuit diagram below illustrates the basic operation of the circuit.

Note - the IR LED is permanently on when the deck is switched on.

The original in-line resistor (1W rating) was 680Ω, but was later replaced by a higher 820Ω, 2 watt rated resistor.

Since both IR LED and photo transistor are closely coupled, I believe I could still further increase this resistance to 1KΩ, and hence lower the IR beam intensity, and still maintain a good beam to the photo transistor. I'll try this later.

The value of the IR LED current was initially about 23mA; while it's now around 18.3mA. From general experience with LEDs, I don't see why I couldn't lower the excitation current further, thus reducing the likeliness of later unexpected IR LED failure? Remember, this IR LED is permanently conducting, even if the deck's motor is not active.

 
The oscilloscope trace below shows the collector voltage of the photo-transistor during Rewind operation. Accessing and part-rotating the sensor holder behind the cassette transport will alter the duty-cycle of the waveform; therefore, an ideal 50% duty-cycle was sought.

Auto-stop now runs perfectly, with a switch off time of about 3 seconds.

Auto-Stop Warning:

 

The Mitsubishi DT-4700 cassette deck's auto-stop unit works off the supply reel, and not the take up reel. Which, in my opinion is not a good design.

Reasoning - if the 'take up' spool stops turning, the auto-stop will not detect this event, and so the right hand side of the cassette tape will eventually jam or seize with tape!

Recording Pre-emphasis: Otherwise referred to as Recording Equalization. This sets the level of high frequency recording emphasis (gain) when the audio is applied to the recording head. The objective is to drive a near constant recording current with rising frequency. The deliberate rise in voltage amplitude to the record head begins at around 300Hz to 500Hz. 

Pre-emphasis can be trimmed from a ‘resonance circuit’ in the emitter leg of Q501/Q601 – to increase or decreases this pre-emphasis, the ferrite core L501/L601 inductance can be decreased or increased accordingly.

Here, the pre-emphasis of both channels was set to match each other.

The overall pre-emphasis gain of current signal level is indicated in the service manual - from the chart, the gain over 400Hz at 10Khz is about 7.5dB.

Since some sibilance distortion was noted, pre-emphasis was reduced a little to try to 'kill off' high frequency saturation..

Playback De-emphasis: Originally R114 was a 3.9KΩ fixed resistor, I swapped both channels for a 10KΩ trimmer potentiometer so that both channels could be matched on playback exactly. Prior to removing the old resistors, their values were measured, and the potentiometers that replaced them were also set to the same values. 

The level of de-emphasis on this Mitsubishi is approximately the same as found on many late 1970s decks in my possession.

Dolby Level calibration points as suggested in the
service manual are also shown.

Tape Speed Variability

The deck's transport was surprisingly stable as the graph shows over a period of just over 10 minutes.

The target speed from the ABEX reference tape is 3150Hz, as can be seen, I accidentally neglected to check this! However, this is unimportant since (a) I can rectify this easily, and (b) the objective of the test was to check the deck for speed variations.

Wow & Flutter

Even with the original pinch roller, and a slightly temporary narrow drive belt, wow and flutter figures returned by WFGUI.EXE were very satisfactory, averaging at about 0.065% wrms.

Additional Interior Shots

February 2024.