Results 1 to 7 of 7

Thread: The Dreaded LED Clock Strikes Again....

  1. #1
    Premium Member

    Join Date
    Jan 2010
    Posts
    1,164
    Thanks
    705
    Thanked 315 Times in 188 Posts
    Rep Power
    315
    Reputation
    4800

    Default The Dreaded LED Clock Strikes Again....

    Hi all,
    Hope everyone is ok.
    Wotnot had better take a heart attack pill... LOL!

    This clock is NOT the one that Wotnot helped out on ages ago. Apparently that one is working fine to this day!

    This one (and also, not another one that is working ok). This one, I did for myself

    I have not looked at what might be causing the fault - yet. As to throw it out as a curiosity thing.


    The top left segment is now lit, as others.

    But it seems to be only the top left segment on the other digits..???

    1) The Time is 10:33


    2) Temp is 22C



    So...

    Is this:
    1) a Digit(s), issue.
    2) Over voltage to a particular rail?
    3) An IC issue ?
    4) Or a Short Circuit Issue?

    I have plenty of spares and even unmade clocks, so parts aren't an issue.

    Ohhh, on looking at the photo - and going back to the clock now, as typing - there are some LED's that are on and shouldn't be.
    So the whole clock face, LED's and the Digit's are failing.... oooh arr..... LOL!!

    What on earth could be causing this...?

    All good fun, as always...
    Cheers,
    GT250.



Look Here ->
  • #2
    Premium Member
    wotnot's Avatar
    Join Date
    Nov 2019
    Location
    Scenic Rim, SE Qld
    Posts
    3,486
    Thanks
    1,532
    Thanked 3,169 Times in 1,628 Posts
    Rep Power
    1461
    Reputation
    64249

    Default

    Bad MCU ; incorrect positioning of MCU in socket ; something hinky in/around the socket itself ; solder bridges...

    Wotnot had better take a heart attack pill...
    I don't operate at that level ~ nothing 'shocks' me anymore..

  • The Following User Says Thank You to wotnot For This Useful Post:

    GT250 (03-10-24)

  • #3
    Premium Member

    Join Date
    Jan 2010
    Posts
    1,164
    Thanks
    705
    Thanked 315 Times in 188 Posts
    Rep Power
    315
    Reputation
    4800

    Default

    Hi Wotnot,

    Bad MCU?

    I'll take apart the case and have a look. The others (three of them), are all working fine. Lucky ol' me, it's just mine.

    With the MCU, is it the sort of thing to gradually fail? I'm used to IC's just going, but not slowly get worse corrupt!
    But I guess if it's a re-writable thing - and as it's top notch Chinese quality - some of the gates I guess can start to fail.

    I think I have another cannabalised - maybe two - clock bits box somewhere. The MCU's should be with them.
    I'll try and have a shot at it later today

    Cheers all,
    GT250.

    PS: We have just finished a Canton Tower (32level one). Pretty good. very pleasant to look at as it goes through its random cycles. Have it in Sound Spectrum mode at the moment, which is very nice.
    She's making a KS25 at the moment - teeny weeny R's in it, LOL!

  • #4
    Premium Member
    wotnot's Avatar
    Join Date
    Nov 2019
    Location
    Scenic Rim, SE Qld
    Posts
    3,486
    Thanks
    1,532
    Thanked 3,169 Times in 1,628 Posts
    Rep Power
    1461
    Reputation
    64249

    Default

    With the MCU, is it the sort of thing to gradually fail?
    They were likely production rejects from the get-go ...it's just like the chinese knockoff OP-AMPs... you can buy 50 of them and put them in a test circuit, and maybe 3 out of 50 will be the same, and none of the 50 will be on spec.

    You have to understand the manufacturing process...a batch is likely 10,000 units, of that there is a predicted failure rate...this is called the AQL (acceptable quality level)...most of the time it runs at around 5%...ie; 500 of those 10,000 units will be 'bad'... that's 'bad' as in unacceptable quality for usage, which cannot be sold for general use, because the chip is too far out of spec, or it's cycle/time in usage MTBF is too short. In the western world, this is called production loss ...effectively you've manufactured ewaste...but in china, it's a ...

    A Brand New Product Suitable for Experimenters and Component Welding, and for those who wish to discover what's wrong with the ICs.... (because we've got no idea..we bought them for 100RMB (AU$20) in a box of 5000 rejects...we're only in it for the $$$$ =)

  • The Following User Says Thank You to wotnot For This Useful Post:

    GT250 (03-10-24)

  • #5
    Premium Member

    Join Date
    Jan 2010
    Posts
    1,164
    Thanks
    705
    Thanked 315 Times in 188 Posts
    Rep Power
    315
    Reputation
    4800

    Default

    Well,
    It all got a lot more interesting...

    Ok: First, I replaced the MCU (I assume that's the STC, the main big legged, IC). And got the same fault...!
    Still showing the lit top left segments and the same LEDs now on permanently.

    Humm. I then swapped over the actual LCD display. Nope! Still the same segments and LED's...

    I thought about an over voltage sort of thing and could see - in my basic rudimentary way - that the Transistor next to the USB power supply has power going through it.
    So I took it out and tested it. I had another clock PCB kit and the tranny was a different code. So I tested the original one and tested the new one and they had the same 0.6xx V across whatever legs gave a reading

    I then started to look at the only other IC on the whole god damn board!! The DS1302... I measured the pins 1 and 4 and they read 5v.
    Problem for me in replacing this DS1302 - just to change every IC - is that the one on board is a SMD and the other spare clock kits I have are normal big legged things. I could in a pinch (done it before), is unsolder the SMD and then use lead wired to solder it back to the pads..!!!

    This seems to be only a timing clock chip - so I don't think it's related to the LCD segment issue...


    I checked the all the R's and they read ok.

    So, bugger me!
    From memory, it was only the left digit that first showed the lit segment. It has now grown to all four digits - and - from memory, there were no faulty LEDs. Now there are plenty - but they are not increasing, so far, LOL!!!




    Thinking about it - as I type - maybe it's a solitary LED causing some sort of electrical feedback to upset it all...?
    There is a dead LED or maybe two...

    It's been sitting near the bar fridge for yonks. hardly look at it - you know, in the background thing, then Bev goes "What's that..."?

    Oh, the joy...!!
    All good fun guys.

    Cheers,
    GT250.

  • #6
    Senior Member
    Reschs's Avatar
    Join Date
    Aug 2010
    Location
    Outside a few schooners
    Posts
    3,134
    Thanks
    963
    Thanked 2,219 Times in 1,098 Posts
    Rep Power
    971
    Reputation
    34678

    Default

    Lets deal with the problems one at a time.
    According to the diagram from your previous thread, segment segment C is permanently on.
    So follow the track from U2 pin 25 and look for shorts.

  • The Following User Says Thank You to Reschs For This Useful Post:

    GT250 (08-10-24)

  • #7
    Premium Member

    Join Date
    Jan 2010
    Posts
    1,164
    Thanks
    705
    Thanked 315 Times in 188 Posts
    Rep Power
    315
    Reputation
    4800

    Default

    Hello all,
    I'd like to report that the clock is working

    Following Reschs's reply;
    Quote Originally Posted by Reschs View Post
    According to the diagram from your previous thread, segment segment C is permanently on.
    So follow the track from U2 pin 25 and look for shorts.
    Whilst there was no short and the track was ok from pin 25.

    I decided to check ALL the leg tracks from the LED display. They all read ok!!

    NOW: Some went directly to the outer LED's. Of which became ultimately the 'weird' fix....
    All the LED's that the Display went to were working fine.
    Some tracks went to the STC Chip.

    Luckly, I had a spare PCB board and was able to use that to correctly trace the tracks (a god send the microscope set up I got!).

    Now the thing to remember, is that on various display patterns, some of the Blue/Red LED's were remaining on. A different pattern, they would differ. Some others would be off - then on with a different pattern - even though the pattern chosen would have them lit or unlit - if that makes any sense, LOL!

    Well after testing that the LED display pins were ok. I decided to tell Bev that we were going to replace the DS3201 chip. it was to be her first use of the Hot Air Gun. In my mind I had to work out how to shield the poly caps from the heat.
    She was not looking forward to it... LOL!

    As she was sat there - she moved the display pattern to a FULL on with only the Second 'movement' turning the LED off as it wound around the face. (This pattern is extremely bright and I never use it).
    The 6th minute and 32nd Blue LED's were the only ones not lit. These LED's had no connection (in the LED Segment display tracks testing anyway), to these LED's.

    Though in hindsight they might have (Wotnot, might an idea about that though). but in our testing we did check the LED's on the left and right of the testing track to make sure we had the correct track.

    Bev then says, why don't we just replace these two Blue LED's anyway?
    I go, Ok, but we have to replace that SMD after...

    We remove the faulty LED's. Test them and they are dead.

    She tests the new ones before popping them in - a CR2032 works wonders . soldered them in, plugged the USB power supply in and, and, and
    the LED Display is working fine....!!!
    NO WAY!!!!
    Can't be!!!

    But sure enough, it's perfect...
    Every Blue, Red LED is working great...
    OMG!

    Well, I don't know. Somehow those two LEDS, not seemingly connected to the LCD Display. But I assume they were connected to the STC and that 'talked' to the LED Display creating that Segment to show.

    I wanted to replace the faulty LED's back in and see if it was the 6th or 32nd one (or was it the two of them), that caused the segment to light. But Bev said NO!!!


    Again guys, a most sincere thank you for your expert advice.
    It's clear to me that I have to become better at fault tracking/analyzing - a skill that comes only after time.


    All good fun
    GT250.

    PS: Wotnot, do a Canton Tower. we did the 16 level one (ball ache bending and soldering the LED's). We got the sound module.
    It's really, really nice when finished.

    We decided to order the 1mtr version yesterday after our success with the clock (overconfident, LOL!), that has the LED rings PCB's premade. A guy on YouTube has made a large, in parts, video on putting it together. He says this 1mtr is for 'advanced' people- like Bev and me!!!

  • The Following User Says Thank You to GT250 For This Useful Post:

    Reschs (08-10-24)

  • Bookmarks

    Posting Permissions

    • You may not post new threads
    • You may not post replies
    • You may not post attachments
    • You may not edit your posts
    •