In 2009 I bought the OEM radio pictured above from a wrecking yard. It is from a 1990 Civic with 124k miles on it. The OEM radios in Civics from 1988-1995 are all interchangeable, so this second-hand OEM radio fit and connected perfectly in my 93 Civic DX. It was returnable for a full refund for 30 days. The data label on the radio showed that the OEM manufacturer for Honda radios at this time was Matsushita (now Panasonic I believe). I could not get the cassette player to work; it may need a new belt and cleaning. I only wanted the radio anyway, so I removed the cassette player (remove 4 screws and unfasten two connectors). The radio worked perfectly except for one thing: At night, the LCD display showing the station number was dark. What follows are the steps I took to diagnose the cause and repair the radio.
I disassembled the radio and found the lamp burned out. The factory lamp is soldered in place, so I had to remove the tiny radio station lighting circuit board to de-solder the old lamp. See photos below. To take out the lighting circuit board, remove the one screw and disconnect the 2P connector.
I looked for a matching replacement bulb but could not find one. It is probably something one could only order through Panasonic, at best. Instead I bought bulb #2721 at Autozone (12 volts, 1.2 watts), unfolded the two thin wire terminals from the bulb's base, and soldered the terminals to the lighting circuit board. I put a blue cover from an old burnt out OEM bulb on the #2721 bulb, so the color of the light would be like OEM. Here is the bulb I bought:
I tested the radio. The station backlight was still dark. But I applied 12 volts directly to the lighting circuit board's terminals, and the lamp lit. I traced the circuit on the main circuit board (good eyes or a magnifying glass required) and did numerous continuity tests. (A multi-tester with probes that have sewing needles alligator clipped to them are helpful for such tests. See the photo at Wiring Tips.)
The lamp's ground was fine, but the power supply to the lamp seemed odd. I found that the last component in the circuit board electrical track providing power to the lamp was a resistor. The left arrow below points roughly to the terminal joints of this resistor. (The right arrow below points to the terminal joints of the radio station lighting circuit board.)
This resistor is numbered "R902" on the circuit board itself. The resistor showed no continuity, but this is usual for "high" resistance. I had no idea what was acceptable resistance for this low voltage lighting circuit, though.
With power connected to the radio, I checked the voltage on both ends of the resisor. The photo below shows a better view of the resistor.
The voltage at one end of the resistor was 8 volts. At the other end (the one electrically closest to the lamp), the voltage was zero. So this is why the lamp was not lighting. I checked the resistance of the resistor. Soldered in place it read about 8000 ohms. Since indicator lamps are rated at about 12 volts and 1.2 watts, from Ohm's law about 0.1 amp of current should run through them. If this current were getting to the lamp, then the voltage drop across the resistor would be 800 volts. Conclusion: The resistor was bad.
While researching this I found a site that tended to reinforce my diagnosis and gave me a solution: http://www.askmehelpdesk.com/advice/t-8584.html. See the posts there by "mystified88." Accords circa 2003 had a radio backlighting problem, and the cause is a bad resistor. Mystified88 and others at the askmehelpdesk site said they just jumpered the resistor, and the jumper corrected the lighting problem. The R902 resistor in my Civic's radio has what looks like two orange bands and two gold bands, so I think it is rated at 3.3 ohms. See the resistor color coding chart at http://en.wikipedia.org/wiki/Resistor_code. This passes the common sense test as well: With everything working correctly and at normal operating temperatures, a current of about 0.1 amp should get to the lamp, and so the voltage drop across the resistor would be about 0.33 volts. Thus the voltage to the lamp would still be very close to 8 volts.
I am hoping the 12 volt-rated lamp will be fine with no resistor and seeing 8 volts across it. The one caveat is that the R902 resistor's purpose probably is to minimize current to the lamp at startup. At startup the lamp filament is cold, and so its resistance is much lower. I found one site that suggested its cold resistance would be as much as a factor of 20 less than its hot resistance. Its hot resistance should be about 120 ohms (= 12 volts/0.1 amp). If its cold resistance is 1/20th of this, then its cold resistance is about 6 ohms. I estimate the design startup current should be about 0.9 amp ( = 8/9.3 ). Without the resistor, the startup current is about 1.3 amp ( = 8/6 ) or about 50% higher. Worst case, I expect the bulb, with the jumper instead of the resistor, to burn out sooner.
I desoldered the resistor using the desoldering iron pictured at Wiring Tips. I used an ordinary staple (the same as what you find in the stapler in your home office) as a jumper and soldered the staple in place. I checked for continuity of the circuit, then plugged it in and tried the backlight. The backlight now works great.