| roversgonemad | Posted: 26 Feb 2022, 05:25 PM |
|---|---|
 Member Posts: 3 Joined: 26-February 22 |
Like a moron, I grabbed a power supply I had lying around that had the same barrel dimensions as the clock connector. I failed to realize that it was a 12V power supply. I now have nothing but the GPS unit functioning. Where should I start troubleshooting to find what I fried? Update: I now have eights displaying in all fields except the hyphens for the date. I am thinking I killed the display drivers... Update 2: The displays do not consistently work. I have ordered new display drivers and a new ATtiny 4313, along with the USBASP adapter. I didn't have a rework heat gun either, so I also ordered a heat gun. Overall, really disappointed in myself for not paying closer attention to the voltage of the power supply. This was a completely avoidable stupid error. Last edit by roversgonemad at 26 Feb 2022, 07:15 PM ------------- |
| [top] | |
| mit | Posted: 26 Feb 2022, 08:05 PM |
 yeah whatever Admin Posts: 538 Joined: 4-May 16 |
Hi, don't beat yourself up too much about it, these things happen. If the ATtiny is destroyed, then it's very hard to know whether the display chips are still working. When the clock powers up normally, the displays won't show anything until they first receive serial data to turn on. If there's no serial data, then they either remain off, or show that 8888-88-88 pattern sometimes. With luck, plugging in a new ATtiny should fix everything. ------------- |
| [top] | |
| roversgonemad | Posted: 5 Mar 2022, 04:27 AM |
|
Member Posts: 3 Joined: 26-February 22 |
The parts showed up and after some fiddling I managed to get the 4313 programmed. I had to move my jumper to the 5V position to get it to work otherwise it wouldn't set the SCK. Unfortunately it's still not working properly. I get a hundredths of seconds rolling and then the tenths and seconds freeze on 1. The semicolons flash, but no MST or DST illuminates and all the other digits are dashes. I have replaced the 7219 display drivers as well. I suspect I have killed the GPS unit. I have ordered another from eBay. In 10-14 days we'll see if that works... I don't suppose it matters if the chip is a 4313 PU vs a 4313 U? I found PU and not U parts. Last edit by roversgonemad at 5 Mar 2022, 05:53 AM ------------- |
| [top] | |
| mit | Posted: 5 Mar 2022, 12:39 PM |
|
yeah whatever Admin Posts: 538 Joined: 4-May 16 |
Part numbers - It shouldn't matter. The different versions of the chip should all behave the same (the ATtiny2313 and ATtiny2313A will work fine too). Programming the chip, this is probably my fault for not giving detailed instructions on it. The "how to change the timezone" section of the instructions assumes the chip is already programmed. A blank chip from the factory needs a few more steps. There are option bytes on the chip (called fuses, but they're reprogrammable). From the factory they ship with the DIV8 bit set, which limits the clock speed and also makes programming them more difficult. The first command I send to a new chip is: avrdude -c USBASP -p t4313 -B2000 -U lfuse:w:0xe4:mThe B2000 sets the programming speed to very slow. The "low fuse" is then set to 0xE4 which disables the DIV8 bit. Be careful setting fuses, one of them is the option to disable ISP, which is very hard to undo. Try again once you've got the DIV8 bit cleared on the chip. It's quite possible the GPS unit is OK, but the serial baud rate wasn't matched because of the attiny clock speed. There is another step I do for each ATtiny chip, which is calibrating the oscillator. It usually isn't needed, most chips are fine out of the box, but I calibrate all of them to be on the safe side. If the display shows almost the correct time and date but has a few junk characters, that might mean the oscillator needs calibrating too. I can help with that if needed. ------------- |
| [top] | |
| roversgonemad | Posted: 5 Mar 2022, 03:16 PM |
|
Member Posts: 3 Joined: 26-February 22 |
It's funny, I had a middle of the night, "a new chip probably needs additional programming"... That got it running! The only weird thing now is that previously it was 100% in sync with my cell phone and now it's about three to four seconds fast. There are no junk characters and everything displays properly, it's just not exactly synchronized with the phone like it was before. Just to double check, I verified against a hand held GPS unit and the clock is three seconds ahead of the GPS as well. Maybe that's the the oscillator that needs to be calibrated?? I'll leave it running for a while and see what it does. Thank you for all your help!! A few minutes and it synchronized. All is well!! Last edit by roversgonemad at 5 Mar 2022, 03:22 PM ------------- |
| [top] | |
| mit | Posted: 5 Mar 2022, 03:25 PM |
|
yeah whatever Admin Posts: 538 Joined: 4-May 16 |
That should disappear within a few minutes, it's the leap-second offset correction (which the satellites only broadcast once every 12.5 minutes). The ordeal probably wiped the battery-backed ram on the gps module, once it's found it, it should stay correct. I've been meaning to update the firmware to not flash the colons unless the leapsecond offset is known, but it only happens when doing a cold powerup so it hasn't been high priority. Edit: Ah great. Also, congrats on getting it working! Last edit by mit at 5 Mar 2022, 03:26 PM ------------- |
| [top] | |
| MistakeHaver | Posted: 19 Mar 2022, 07:55 PM |
 Member Posts: 4 Joined: 19-March 22 |
I made the mistake of putting 12V to the clock as well. I'm slightly relieved that someone else made this mistake as I felt like quite the dingus. My clock has a section of display burned out, or the thing that transmits power to the display sections has burned out. I've attached a photo of the problem--sorry about the dimness of the display. The main problem description here is that on the left side of the display, the bottom left segment is no longer lighting up. On the right side of the display, the top segment in the 8 segment display is no longer lighting up. Otherwise, the clock functions perfectly and neatly turned over when DST happened a few days back. Any ideas on how to proceed?  ------------- |
| [top] | |
| mit | Posted: 20 Mar 2022, 09:32 AM |
|
yeah whatever Admin Posts: 538 Joined: 4-May 16 |
I spoke with roversgonemad a fair bit over email so not all of the conversation is here. The image you've posted is very interesting, it's the type of thing that's normally seen if there are soldering mistakes on the MAX7219 display driver chips, but if it worked beforehand then it does seem like the display drivers are damaged. Replacing them should fix the problems. Luckily I made a tutorial on how to replace the display driver chips here: https://mitxela.com/projects/precision_clock_mk_iii/max7219_issues ------------- |
| [top] | |
| MistakeHaver | Posted: 19 Apr 2022, 02:44 PM |
|
Member Posts: 4 Joined: 19-March 22 |
Well, it's only been almost a full month, but I finally procured a heat gun and de-soldered my chips. How would you recommend that I purchase some new ones? Is there a website you recommend, or can I get some off of you? Thanks in advance. ------------- |
| [top] | |
| mit | Posted: 20 Apr 2022, 09:15 AM |
|
yeah whatever Admin Posts: 538 Joined: 4-May 16 |
I don't have any in stock that I could send out right now, the last batch I ordered from China was entirely dead on arrival, again. There may be sellers on ebay/amazon offering the part with fast shipping. Otherwise, digikey or mouser sell the part at a significant premium (plus postage). If you go for a possibly non-reputable seller and want to test the chips before soldering them in, it's possible to just bend the legs enough to fit into the footprint, without soldering it. The springiness of the legs will maintain temporary contact. You need to fit both chips in order to test the display, one on its own won't work. ------------- |
| [top] | |
| MistakeHaver | Posted: 29 May 2022, 01:25 PM |
|
Member Posts: 4 Joined: 19-March 22 |
Okay, once again delayed. I was just looking at my newly fixed and mounted clock, and realized I hadn't written a final post. As I hate it when people either write "Nvm I fixed it" or disappear completely on a forum, I thought I'd finish this off. The chips were indeed the issue. I hit up eBay for some replacements. Alltronics had a pack of 3 chips for $15.55 USD, and they all worked perfectly out of the bag. After some finicky desoldering with the hot air rework station and some even more finicky resoldering, the clock went right back to working how it should. Thanks for the assist, Mitxela. ------------- |
| [top] | |
Sign in to post a reply.