-
Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.
LC 575 board (in Color Classic) - funky vertical lines
- Thread starter KnobsNSwitches
- Start date
KnobsNSwitches
Well-known member
Bolle
Well-known member
Put the RAMDAC back in place, the corrosion cleaned up nicely:
I was still getting the same artefacts as before so either the chip was bad or there was something else wrong.
I took out my working 575 board to check a few things. I started figuring out the pinout of the VRAM SIMMs to have a starting point for tracing things backwards. Then traced all connections going to and coming from the RAMDAC.
The LC475 dev note says that video data is shifted to the VRAM from the main data bus (D0-15 to SIMM 1, D16-D31 to SIMM 2). Addressing is done by the memory controller which has 9 separate video address lines. /RAS, /CAS, /WE and the clock signal for the serial output ports of the VRAM ICs are coming from the memory controller as well.
Video data is fetched by the RAMDAC through the serial outputs from the VRAM.
With all that info put together in a messy spreadsheet I started going through all the connections on the broken board.
Video address line 1 seemed to not make it from the memory controller to the SIMM sockets. All address lines traced fine to the resistor packs (acting as filters) on the back of the board, however from there VA1 did not make it's way to the sockets.
It was not a broken trace though but a faulty solder joint. I don't see how that could have happened. There was no solder on there at all. This must have been a factory error, unless someone already put their hands on the backside of that board:
Soldered that little guy down and resoldered all the other resistor packs as well for good measure.
The result was a better, but still not completely good picture:
The lines started to fade away and come back while the machine was running, so there must have been intermittent contact somewhere or a trace with higher than normal capacity.
I went through all the serial data lines between VRAM and the RAMDAC twice. On the first time everything checked out just fine but on the second go there was one data line (luckily) not giving any continuity.
If it would not have decided to stop working right when I went for a second check I would have been looking for the problem forever... not sure if I would have done a third turn going through the same signals.
It looks like the trace connecting to the solder pad was causing the issue. I put in a short patch from the pad to the next via and got a steady beep from my multimeter now
And the result:
Now that was fun
I was still getting the same artefacts as before so either the chip was bad or there was something else wrong.
I took out my working 575 board to check a few things. I started figuring out the pinout of the VRAM SIMMs to have a starting point for tracing things backwards. Then traced all connections going to and coming from the RAMDAC.
The LC475 dev note says that video data is shifted to the VRAM from the main data bus (D0-15 to SIMM 1, D16-D31 to SIMM 2). Addressing is done by the memory controller which has 9 separate video address lines. /RAS, /CAS, /WE and the clock signal for the serial output ports of the VRAM ICs are coming from the memory controller as well.
Video data is fetched by the RAMDAC through the serial outputs from the VRAM.
With all that info put together in a messy spreadsheet I started going through all the connections on the broken board.
Video address line 1 seemed to not make it from the memory controller to the SIMM sockets. All address lines traced fine to the resistor packs (acting as filters) on the back of the board, however from there VA1 did not make it's way to the sockets.
It was not a broken trace though but a faulty solder joint. I don't see how that could have happened. There was no solder on there at all. This must have been a factory error, unless someone already put their hands on the backside of that board:
Soldered that little guy down and resoldered all the other resistor packs as well for good measure.
The result was a better, but still not completely good picture:
The lines started to fade away and come back while the machine was running, so there must have been intermittent contact somewhere or a trace with higher than normal capacity.
I went through all the serial data lines between VRAM and the RAMDAC twice. On the first time everything checked out just fine but on the second go there was one data line (luckily) not giving any continuity.
If it would not have decided to stop working right when I went for a second check I would have been looking for the problem forever... not sure if I would have done a third turn going through the same signals.
It looks like the trace connecting to the solder pad was causing the issue. I put in a short patch from the pad to the next via and got a steady beep from my multimeter now
And the result:
Now that was fun
Last edited by a moderator:
techknight
Well-known member
Finally. Someone out there thats as good as I am at this stuff.
So now I can finally retire on repair to focus solely on my day job as it is an ever growing time sucking black hole.
Been wanting to for awhile but found it hard.
So now I can finally retire on repair to focus solely on my day job as it is an ever growing time sucking black hole.
Been wanting to for awhile but found it hard.
Last edited by a moderator:
joethezombie
Well-known member
As always, @Bolle, most excellent work. Very well done!
@techknight, no. not allowed. [] But seriously, I for one can definitely understand the time sink. Get wrapped up so far in work and repairs, then there's no time to to actually play. Although, there is a certain degree of satisfaction when something clicks and it finally works. Thanks so much for the years of contributions on here, and your youtube videos. And I hope your retirement is sporadic at best. [}]
@techknight, no. not allowed. [
] But seriously, I for one can definitely understand the time sink. Get wrapped up so far in work and repairs, then there's no time to to actually play. Although, there is a certain degree of satisfaction when something clicks and it finally works. Thanks so much for the years of contributions on here, and your youtube videos. And I hope your retirement is sporadic at best. [}]
techknight
Well-known member
until we can build a strong enough budget and capital at work so I can hire employees to take over parts of my job, I Will be in hiatus for awhile as far as being able to sit down and do repairs. This is primarily why my youtube channel has been inactive.
KnobsNSwitches
Well-known member
WELL DONE.
thats great dedication and a great result!
thats great dedication and a great result!
superjer2000
Well-known member
Could you attach the spreadsheet to his thread? I have a centris 610 with a potentially related issue. Thanks for the great write up!!I will clean up my spreadsheet a little bit, so it will make sense for everyone else as well and send it to you.
Should make checking things easier for you
Similar threads
