Sony CRF-320 Repair

 Not much wrong with this very heavy portable radio sold between 1976 and 1980 and which is now much sought after even though its price is boardering on the ridiculous I even spotted a secondhand back cover being offered for £200.

The customer reported the main tuning dial was out of kilter and new LED lamps were no longer working.

 

 Above front view and below with the front cover detached.

 

 I removed the front and panels to access the circuitry.

 

 

I powered the radio from the mains supply and checked the various observatios made by the owner. Noticeable was missing illumination for the clock, meter and FM dial. The LW/MW dial and the SW dials could both be lit when these modes were selected and were very bright. Radio 4 tuned in at something like 190KHz instead of 198KHz but MW and, in particular short wave, dials seemed OK. I did notice however that the SW dial has a cursor adjustment and if this is offset it would give out of sync readings between the digital readout and the dial calibration. Sensitivity on LW/MW wasn't particularly good no doubt because of component aging.

 

 The first step was to align the LW/MW dial. Using my Tiny SA as a signal generator revealed the lowest frequency I could tune was about 160KHz against a dial reading below 150KHz. The receiver has a standard analogue design in its long and medium reception modes and tweaking the oscillator brought in 150KHz and further adjustments corrected tracking. Alignment is straightforward once one is familiar with the layout of the coils and trimmers. The manual is strangely pretty unhelpful because it shows two drawings, one each for long and medium wave adjustments with neither showing all coils and trimmers, and it has the positions of the medium wave oscillator trimmer and coil reversed which was unhelpful to say the least! To aid alignment I've shown above the locations of the complete set of coils and trimmers.

The standard method of alignment is used ie. trimmers at HF and coil slugs at LF. The dial markings are very vague and Sony didn't bother to indicate alignment points.

The FM range seemed to be OK as was the shortwave range which employs a digital synthesiser and a red digital readout. Tuning upwards by a single step from say 3.999 gives a display of 3.--- rather than 4.000 which is a trifle annoying but which I assume to be normal. Tuning to a frequency standard proved the dial read perfectly.

Next I need to figure out the problem with the LEDs which I understand replaced small filament lamps.

 
 

 This is the connection panel for the various lamps used in the receiver.. being Clock, Meter, FM dial, LW/MW dial and SW dial with common black and five red each feeding a ballast resistor for dropping the LED voltage chosen to suit its forward voltage and supply.

The customer reported that the LEDs had failed one by one which suggests poor reliability rather tan a broken wire or bad solder joint.

 LEDs are pretty well anonymous, as they are rarely marked with any identification, so you need to check things like the package style, lead mechanical shape and internal structure to identify the manufacturer and type. The blue types fitted in the Sony were installed in 2018 and appear to be similar to a 5mm Kingbright range in a package called "T one and three-quarters" and, looking in a supplier catalogue for this period, suggests the following options.

The supply voltage measured as 14.75 volts. VF = forward voltage drop, IF=forward current, P= LED dissipation, Lv-mcd= a measure of LED brightness.

 Make

 Type

 VF volts

 Lv-mcd

Supplier Code

 IF mA max

 IF mA nominal

 P mW max

 Ballast resistor

 Kingbright

 L7113PBC-A

 3.2-4.0

 280-1000

 SC08060

 30

 20

 120

 560

 Multicomp

 OVL-5523

 3.2-4.0

 1950-4300

 SC08062

 30

 20

 120

 560

Kingbright

 L7113QBC-D

 3.3-4.0

 1700-2700

 466-3548

 30

 20

 120

 560

 I removed one of the open circuit LEDs and found it was fitted with a 180 ohm ballast resistor. This would drive the LED current to something like 60mA with a dissipation of close to double the maximum. This being so explains the poor reliability of the existing LEDs if indeed they have the spec given above. Clearly the remedy is to replace the LEDs with a set having the best rating for the job. I selected a high brightness white LED made by Nichia, type NSPW510DS which has a forward drop of 3.2V at 20mA having a luminous intensity of 6.1cd and a viewing angle of 45 degrees.

The table below gives the original LED operating conditions, with basically any type having the same actual over-dissipation, together with the Nichia used to replace the originals.

 Make

 Type

 VF volts

 Lv-mcd

Supplier Code

 IF mA max

 IF mA nominal

 P mW max

 Ballast resistor

 IF actual

 P mW actual

 Kingbright

 L7113PBC-A

 3.2-4.0

 280-1000

 SC08060

 30

 20

 120

 180

 64

 211

 Multicomp

 OVL-5523

 3.2-4.0

 1950-4300

 SC08062

 30

 20

 120

 180

 64

 211

Kingbright

 L7113QBC-D

 3.3-4.0

 1700-2700

 466-3548

 30

 20

 120

 180

 63

 214

 Nichia

 NSPW510DS

 3.2-3.5

 6100

 713-3996

 30

 20

 105

 510

 22

 70

 I chose a ballast resistor of 510 ohms which will result in a current of about 22mA (dissipation etc about 70mW) and fitted these. Apart from the clock LED, the others are glued into rubber rings which fit snugly into apertures in the different assemblies. When I checked I found the LEDs produced a pleasant greenish-hued illumination due to characteristics of the plastics used in the receiver.

With the panels refitted and the battery leads plugged back (only the clock D-cell is used) I noticed the clock LED was pulsing at around a one second interval.
 

 Return to Reception