520STFM C070523-001 REV.D upgraded to 1MB - black screen, not working

[cleoo]

 

 

I acquired this 520STFM a while back and started getting parts to get it to work since it would just show a black screen with white border.

 

The PO said he upgraded the board to 1 MB as you can see the row above the original ram has been populated.

 

 

So, I thought it was the PSU since it was old and those were problematic. I followed Jan Beta’s video and installed a Meanwell RD60 PSU.

 

When I booted it, the only change I noticed and it was very brief, it was a very distorted white and green display-glimpse of a desktop?

 

Anyway, i quickly flicked it on and off and it went back to the black screen, white border.

 

So, I went ahead and started to replace the caps.

 

Console5s caps were a little different than what was on the board

 

On my board C35 is occupied with a radial capacitor 470 uF 16v. Console5 supplied an axial capacitor 2200 uF 16v. Is that still acceptable? Not sure if it will even fit.

 

I also can’t find C124 on the board to put this 100uF 16v axial capacitor. Highest one on the PCB is C119. If you can direct me to it because I can’t find it.

 

Also, looking through this board I see these resistors. Doesn’t look right.

 

 

These look burnt and there’s a clear coating over part of the board.

 

Any ideas with what’s going on with these mods?

 

Thank you.

 

[cleoo]

So I turned it without putting the remaining two capacitors and got to this...

Then when I went to attach the mouse. It went back to the black screen with white border.

[TGB1718]

Have you tried re-seating all the chips that are socketed, I would also give the MB a goo clean-up and remove that "stuff",

maybe re-flow those solder joints on those resistors

[tjlazer]

Remove all socketted chips and clean sockets and chips well and put them back in to see if that helps.

[cleoo]

Thank you both for the suggestions. I will give those a try.

I was emailing Luke at Console5 and he believes this board was one of the early Rev D boards in the transition from Rev C because of the 470uF cap.

He sent me a pic of how that 2200uF was installed.

[Robson]

Burn marks on resistors look like they're from the soldering iron, lazy option...soldered from component side, hence ugly work. The green is a capacitor (not resistor) though. The large cap is just a filter, the board should be operational without. You can fit at the end between +5V and gnd.

Black screen with white border is ram or ram related problem (MMU/latches or lines). I had one of these, previously repaired, butchered boards in the past...it was a nightmare with multiple problems (sorry to say) to fix.

However yours can be as simple as a failing ram IC. A diag cart + a pc with serial (running hyperterminal)  could help you a lot even when the screen is black. Unfortunately they did not socket the upper half meg ram when expanded the board so it's a soldering job if you find the location(s). After cleaning, reseating socketed, checking for suspicious pins and reflow them, I'd disable the top Bank (lifting up ras/cas lines and tie to +5V). On this board I think it's the RAS lines that are not through resistors to ram chips just wired. 

Edited September 25, 2021 by Robson

[cleoo]

Thank you. It had powered up to the green desktop after making most of the recap but something caused it to go back to the black screen. I’ll try those suggestions of yours as well.

I’m nominally proficient at soldering so I was having difficulty removing the rock hard solder, even after applying fresh solder to the small vias.

I soldered the large cap this way.


The C124 cap looks jerry-rigged in the picture Console5 Luke sent. Can someone take a picture of their unmodified C124 cap for proper placement?

It looks like the negative of the axial cap is on the small via near the and the positive is connected near the C124 via. Is that the positive?

[cleoo]

Robson, how would I even disable the top row of “new” ram? I don’t have a schematic to figure that out right now.

[Robson]

The large cap seems ok, the negative points to GND (large plane) and can see that the other end is connected to the plane that has the + silk mask mark.

To disable the banks what you normally do is to lift the Ras and Cas resistors by the pins that are closer to the MMU, then join those lifted legs and connect them to +5V. This way the MMU ras/ cas lines are not connected to anything and the RAM pins are pulled to  Vcc thorough the resistors.

Schematics:

https://www.gossuin.be/index.php/520-et-1040-stx

 

You'd need a multimeter and a schematic to track the signals. Even though there is no specific schematic to this rev board (as far as I know) you know the MMU pins (18,21,22) so you just check for continuity. I can't remember but they are within the R90, 93, etc resistors (probably the first four from top but needs checking).  You can see there are 2 "missing"...those points are connected, so you might need to cut a trace there temporarily. Before all of this I'd try getting a diag cartridge.  Where are you based? I'd offer to borrow mine but it's not only the cart you need: Another ST and a serial cable or a serial to usb adapter and a serial cable to connect it to a PC that runs hyperterminal, to get all the messages outputs through serial. But this only happens if the boards boots to the point where the cart takes control. Hard to tell what is wrong with yours but the fact that you've seen the desktop before is not a bad sign. Clean on and off is probably 1 or more ram IC fail (I mean no garbage on screen, etc) not something complex that I had to face with. 
Definitely general cleaning and diag before cutting traces. 

For blobs,  soldamop and a good solder sucker are your friends. I hope it helps somewhat. 

Edited September 25, 2021 by Robson

[Robson]

On this photo there is my board. You can see that I marked at the bottom which IC means which data bit.  So it's not straight forward numbering from left to right. D8-D15 ICs first then D0-D7.

[cleoo]

Thanks Robson for the suggestions.

I replaced the main capacitor with a 2200uF and did change the C41 capacitor with a 100uF at the suggestion of Luke at Console5.

But then I started to get distorted screens and got this 20 bomb screen once it settled down.




I replaced the main cap back to a NEW 470uF cap I had for the Lynx and removed the C41 cap back to the original one.

The screen was again distorted. Then I started to seat the large chips to see if they made a difference and I noticed the bottom chip (blitter?) was changing the screen display when I applied pressure at certain points.

But then it finally cleared up and showed these screens.




I reinstalled the 2200uF cap and also
removed and reseated those two chips and now it boots onto the GEM desktop!





So, the only issue I have is obviously the floppy drive needs to be fixed. It beeps and spins but nothing shows up on the desktop. I had to unplug it so it could get onto the GEM desktop.

Each of the keys makes and audible sound on each press that I can hear through the monitor.

The other issue I’ve experienced is the left mouse click doesn’t always work.

I can now look into other options and getting it to work fully.

Thank you.

[Robson]

Glad it was just as simple as contact errors.

Edited October 2, 2021 by Robson

[cleoo]

Duh, I actually waited longer before giving up. I need a manual.




Left mouse button still is non-responsive at times. Wonder what is causing the issue...

[snarkdluG]

Resolder the mouse and joy port. It probably got intermittent fault due to crack in the solder joints over time. Or open up the mouse and put some deoxit (contact cleaner) in the switches or IPA. Or just buy new switches and solder those in.

[cleoo]

I’ll try that. I noticed when I shut it off and turn it back on, the mouse button became active again.

[cleoo]

That was a good suggestion about the fault with the keyboard. Sure enough there were breaks in the solder. I applied new solder and checked continuity.



OMG. There were 30 little screws to remove. What a pain.

[cleoo]

Unfortunately, I ended up breaking the reset switch.



I tried to superglue the pieces but it ended up breaking when I pressed it. Which left the pins shorted and it booted into a white screen without reaching the desktop.

Too bad I don’t have a 3D printer to make a replacement.