One of the LEDs that shows time on my alarm clock is behaving badly -- 7 Segment LED Display

[whiteoutage]

These first two pictures OF THE CLOCK were taken right after radio is plugged in an outlet, and clock is flashing on(Pic #1) and off(Pic #2).

First photo shows the 2nd digit has an led that never lights up during either the 12 AM or 12 PM hours.
Second photo shows faulty LED working, but only when it is supposed to be in the off state, so the led is not burnt out, but seams to be off when it should be on and vice/versa.

Wondering how to reform that diode?

PS - ignore the two pins with solder blobs in that long row of 30 some pins - they are not shorted - I just added too much solder when I thought they were bad joints.

 

[Bluejets]

What does it do at 2,3,4 5,6, 8 and 9...??

 

[whiteoutage]

Just put it back together and checked. As you maybe suspected, those digits/times are messed up too since they also need that middle LED. The only digits that don't use it are of course 0,1 and 7 - so it's messed up 67% of the time,

Do you know what would cause that LED to be in the wrong state most of the time ? Maybe bad diode, resistor, processor, logic programming - no clue here?

 

[whiteoutage]

By using the diode checker on my meter I was able to turn on every led including the bad one by probing each pin to the second to last pin on that long row. The bad pin is pin number 6, or pin 4 if you start after the empty string of five pins. I could not locate a schematic for this cheapo alarm clock, but did trace pin four to the 42 pin IC (Sanyo LM8364 - no exact match on datasheet located on discontinued chip). I read that each LED usually has a separate resistor for protection. How to trace faulty pin 6 to its resistor to test it after trace goes to IC, or am I barking up the wrong tree as usual?

 

[bertus]

Hello,

There is some info on the LM8364 sanyo chip on the radio museum site:
https://www.radiomuseum.org/tubes/tube_lm8364.html
You will have to register to download the larger images.

Bertus

 

[whiteoutage]

Hello,

There is some info on the LM8364 sanyo chip on the radio museum site:
https://www.radiomuseum.org/tubes/tube_lm8364.html
You will have to register to download the larger images.

Bertus

Unfortunately, Radio Museum charges $25 fee for membership. Even though it is just a onetime fee, that price is still too steep for me. Sanyo does have similar 42 pin dual alarm clock chips with available data sheets, but wondering without either an exact datasheet or a schematic if repairing this faulty LED is a lost cause?

 

[Harald Kapp]

Maybe bad diode, resistor, processor, logic programming - no clue here?

• It can't be the LED, as it works - only not as expected.
• Resistors: highly improbable. There is usually a single resistor to limit current. Here I can't see a bank of resistors (one for each segment would be expeced). It seems the current is controlled by the clock chip.This seems to work judging from the brightness when the segment is lit.
• The clock chip could be defect. I would judge this to be an unusual failure. Unusual but not impossible.

Carefully check all solder connections. Make sure there is no short circuit anywhere. Clean up the solder blobs you created. There may be an issue even if not visible to the naked eye. Not much else you can do without a schematic and some defect tracking skills.

 

[whiteoutage]

Just blew up radiomuseum's tiny.png preview and could not make out much of anything other than it controls an "LED display (common cathode)", along with controlling other functions of the radio. I did not blow up and post here due to copyright laws - the rest of the picture is not readable but still may be of use to someone smarter than me.

Looking at the circuit board, appears that most of the 21 pins on one side of the IC are dedicated to the LED display. Only test I could think of doing was to lamely check continuity between bad LED pin and its connection to IC, it's of course good, but don't know what else to check?

PS. I did clean up the blobs last night, checked for shorts between LED pins and checked to make sure each LED bridged to big IC chip.

 

[whiteoutage]

Just blew up radiomuseum's tiny.png preview and could not make out much of anything other than it controls an "LED display (common cathode)", along with controlling other functions of the radio. I did not blow up and post here due to copyright laws - the rest of the picture is not readable but still may be very useable to someone smarter than me.
Looking at the circuit board, appears that most of the 21 pins on one side of the IC are dedicated to the LED display. Only test I could think of doing was to lamely check continuity between bad LED pin and its connection to IC, it's of course good, but don't know what else to check?

PS. I did clean up the blobs last night, checked for shorts between LED pins and checked to make sure each LED connects, has continuity, to the big IC chip.

My gut tells me there is maybe an issue with the internals grounding? or something of the four character clock LED - not scientific at all I know. When you first plug it in, the clock flashes 12:01 then time LED flash off - they all work except the one that flashes off when it should be on and vice/versa. Off chance of programming logic error which would not be fixable unless I could get same chip from same radio -not going that route for sure. If above LED internal issue is correct then finding replacement also probably not economically feasible either. Ran out of ideas, but as you all may know I'm an electronics knob so hoping my wild guesses are wrong, and hoping someone has a gestalt.

 

[whiteoutage]

My gut tells me there is maybe an issue with the internals grounding? or something of the four character clock LED - not scientific at all I know. When you first plug it in, the clock flashes 12:01 then time LED flash off - they all work except the one that flashes off when it should be on and vice/versa. Off chance of programming logic error which would not be fixable unless I could get same chip from same radio -not going that route for sure. If above LED internal issue is correct then finding replacement also probably not economically feasible either. Ran out of ideas, but as you all may know I'm an electronics knob so hoping my wild guesses are wrong, and hoping someone has a gestalt.

Thinking there may be a way to rule out defective parts out by somehow taking voltage measurements but I have no clue as which pins should be probed and what expected voltages should be? Maybe the problem pin six is getting a negative voltage instead of a positive voltage? Any advice on taking voltage measurements, etc. would be greatly appreciated.

 

[Martaine2005]

I would guess the IC is damaged.
Try spraying it with a freezer spray, it may well function correctly for 20 seconds or so.
If you have a can of compressed air, you can invert the can to get a freeze from it.

Martin

 

[whiteoutage]

I would guess the IC is damaged.
Try spraying it with a freezer spray, it may well function correctly for 20 seconds or so.
If you have a can of compressed air, you can invert the can to get a freeze from it.

Martin

no freezer or air here would it hurt to stick in the freezer for half an hour?

 

[Bluejets]

All "g" segment down I'd say the ic.
Multiplexing would be internally controlled.

 

[whiteoutage]

All "g" segment down I'd say the ic.
Multiplexing would be internally controlled.

Consensus is IC is knackered. Anyhow, took some voltage measurements for practice. Voltage drops from resistor 103 (the big honking resistor next to the IC) to 9 LED pins of the 1st two digits (first digit only has two pins and 2nd digit has 7 pins) . Measurements were taken while the clock was flashing, 1 on the first digit, 2 on the second digit. The voltages were all consistent except for faulty pin 6 (when LED flashed on normal voltage was 1.8v, when flashing off V was zero; however, pin 6 flashed between 1.8 and 1.2 volts in on/off states - don't know if this helps confirm our theory or not, or if maybe I should continue voltage testing, etc?

 

[Bluejets]

Voltage testing is a waste of time on multiplexed outputs.
A scope would be ok but again waste of time on a 2 cent clock.

 

[whiteoutage]

Voltage testing is a waste of time on multiplexed outputs.

OK this is a waste of time - so maybe the reason pin 6 has faulty voltage is cause like you said IC is bad and sending wrong voltage? Please remember I am a complete knob.

A scope would be ok but again waste of time on a 2 cent clock.

Yes, it is a 2 cent clock, but maybe an OK way to cut your teeth on electronics repair, at least for hillbillies like myself?

 

[Harald Kapp]

Yes, it is a 2 cent clock, but maybe an OK way to cut your teeth on electronics repair, at least for hillbillies like myself?

The damage is already there, so trying to fix it won't hurt. The least is: you learn something. When you get the clock back to working, the better for you and our environment.

 

[Bluejets]

Thing is he has zero reference to go by....maybe if there was one working to compare with for scope readings.
Trying to find out what is going on inside a multi-pin ic chip would be next to impossible.

 

[Harald Kapp]

Trying to find out what is going on inside a multi-pin ic chip would be next to impossible.

I totally agree - and even if one could find out, there's no way to repair a defect within a chip. But on the admittedly slim chance that the defect is on the pcb, trying to track it down can nevertheless be an exercise in debugging hardware. It's up to teh op to decide whether he wants to invert more time here.

 

[whiteoutage]

Even a hillbilly can understand when it is time to throw in the towel - I hope. You guys are the best!

"God grant me the serenity to accept the things I cannot change, courage to change the things I can, and wisdom to know the difference"
Karl Paul Reinhold Niebuhr (1892–1971