Amiga 500+ (Rev 8A) - long black screen display duration before kickstart boot screen

[solarmon]

Hi,

I had recently got an Amiga 500+ (rev 8A) with battery damage repaired - as per this thread:

http://eab.abime.net/showthread.php?t=98371

However, the boot up sequence is a bit strange and intermittent.

It will show an initial black (or is it dark grey?) screen with a strange red vertical streak on it:



The duration of this is a bit random and it can take a while - up to and beyond 10 seconds? (sorry, I haven't timed it) before it proceeds with the rest of the boot sequence and then showing the kickstart boot screen.

I assume this is not normal/common? My A500 (Rev 6A) does not behave like this - it goes to the boot screen in a reasonable/normal time (less than 5 seconds?).

Can somebody explain this behaviour (funny black screen, and long boot time) and to where I should look to try to diagnose and repair the issue.

[kipper2k]

you could have a few things going on there..

with no floppy attached or failure of floppy there will be a delay before showing the splash screen. The red stripe could be agnus/denise related or even something else, tough to find.

[solarmon]

Originally Posted by kipper2k

you could have a few things going on there..

with no floppy attached or failure of floppy there will be a delay before showing the splash screen. The red stripe could be agnus/denise related or even something else, tough to find.

Thanks for the pointers.

This behaviour is when a working floppy drive (physical or Gotek) is connected to the internal floppy interface.

I'll look around Agnus and Denise areas. But since I had a corroded pin 3 (Red) on video connector could that possibly be related? Too much of a coincidence - the red colour?

I did also find that the resistor for the OVL signal from RP101 was reading about 1.2K ohm instead of about 4.7K ohms - not sure if this is related. But I got a replacement 4.7K ohm resistor pack that I will replace to rule it out.

[solarmon]

I found this post on another forum regarding the Amiga boot sequence:

https://www.lemonamiga.com/forum/viewtopic.php?t=13978

It states one of the steps is:

5. Change the screen to Black, screen stays Black (or with red Stripes) if CPU fault.

Which sounds like my issue - although mine only has one or two stripes. I'll check around the CPU then to rule that out. I'll also test with the CPU from my A500, but this CPU from this A500+ was previously tested in the A500 and I didn't notice this issue.

Or it could just be a socket or trace issue associated with the CPU, so I'll have to check those too.

[nogginthenog]

Dumb idea: You have a non-A500 Kickstart and it's looking for an IDE hard drive.

[demolition]

Originally Posted by solarmon

This behaviour is when a working floppy drive (physical or Gotek) is connected to the internal floppy interface.

So is the floppy drive actually ticking when the KS screen appears? And will it boot from a disk? If the floppy cable was oriented incorrectly, it would cause this large delay, but obviously it would not be able to load anything.

And I really doubt that the CPU is at fault if it is able to show the KS screen eventually.
It could be that the RESET signal remains active for a long time.

[solarmon]

Originally Posted by demolition

And I really doubt that the CPU is at fault if it is able to show the KS screen eventually.

Yes, I believe the CPU itself is OK - it tested OK on another A500 (not plus) and tested the known working CPU from that machine in this suspect machine and it had the same issue.

Originally Posted by demolition

So is the floppy drive actually ticking when the KS screen appears? And will it boot from a disk? If the floppy cable was oriented incorrectly, it would cause this large delay, but obviously it would not be able to load anything.

The floppy drive is fine. When it does get to the boot screen it is able to load, for example, Amiga Test Kit OK.


I had done some prodding about, and some more cleaning of sockets and pins, and reseating of the socketed chips.

And right at this moment, it seems to be OK and getting to the boot screen quickly as expected.

However, whilst I was prodding about I came across the following issues, which may or may not be related.

* Twisting the Denise chip produced the following graphics corruption:



Further investigation showed that pin 4 of the Denise socket is a bit loose and moving pin 4 of the Denise chip slight made it loose contact. This is not really an issue as pin contact OK normally

* Pushing and pulling at the Kickstart chip produced the following scrampled graphics corruption (example at boot screen):



The kickstart chip socket pins had some corrosion which I had cleaned up during the initial repair. I did another cleanup and I cannot reproduce the error again - so either the cleanup or the re-seat may have resolve the issue.

Originally Posted by demolition

It could be that the RESET signal remains active for a long time.

Yes, this is what I suspect too. I've got a logic probe that I will need to learn to use, but I need to understand the reset logic. I believe that the RESET pin on the CPU (and the custom chips) will be held low for a brief period and then held high?


In addition to this, there is another problem - Amiga Test Kit says it cannot detect the battery backup clock:



It had been able detect it previously:



But I have noticed that this intermittent too.

This might be because of the U11 and U13 logic chips which I had not replaced yet, or the MSM624B clock chip itself.

[solarmon]

Just an update after some further troubleshooting.

The issue with pin 4 of the Denise socket was 'fixed' by just bending that pin out just a bit more than the rest - so now just lightly prodding the pin won't cause a loss of contact.

I haven't had any occurrences of the boot up delay - so I think I'll put it down to dodgy/corroded contacts on the pins/sockets of various chips, probably on the reset signal ones.

The issue with the undetected real time clock seems to be resolved itself. However, I had the issue of it not 'ticking' along. I could set/reset the date, but it would not tick on - i.e. the seconds did not increment:



In Workbench, I could not get the setclock command to work:



I tested the real-time clock on an A501 expansion board connected to the A500+. I had to bridge jumper JP9. (I actually manage to lift both pads in doing so, so I had to improvise).

The Amiga Test Kit could then see the external real-time clock OK and the seconds were ticking along - this example shows it 48 seconds in:



From this, I guessed that it was probably the U9 (MSM6242B) real-time clock chip on the main board that was faulty - so I proceeded to order some of these chips with the intention of swapping it out, and putting in a socket too.

After that, I thought maybe some of the solder on the pins on U9 and surrounded components just needed a reflow, since the issues seem a bit intermittent.

So I undid JP9, and reflowed the pins of U9 with addition of some new solder, and any associated components near it.

To my surprise (and annoyance for ordering before I did the reflow) it started working!

The real-time clock was detected and it was ticking along. Workbench can now see the time and the setclock command is working as expected:



Maybe a reflow of the other chips and components might be a good idea overall.

And I'll now have some spare MSM6242B chips should I need it some time in the future...

[solarmon]

Spoke too soon. This morning I tried again and I experienced both a long boot delay as well as the clock not ticking along.

There must be some other cold joint and/or broken/dodgy trace?

[solarmon]

Originally Posted by solarmon

Spoke too soon. This morning I tried again and I experienced both a long boot delay as well as the clock not ticking along.

There must be some other cold joint and/or broken/dodgy trace?

So, since then I have set up my logic probe ready to check the reset signals. I had added pin headers to CN14 to get access to 5V and GND for the logic probe and even put white electrical tape on Gary, Paula, Denise and CPU to mark out the reset pins.

However, since then, I have not had any occurrences of the boot up delay. Typical.

What was still a problem was the clock. It was still intermittent. The battery voltage when the A500+ was turn off to the RTC chip was an acceptable 3V (I think about 0.3V was dropped via the 1N4148 diode I had put in as replacement for R913). So something else was wrong with the RTC setup.

I then suspected the Y9 oscillator and TC9 varicap - so I reflowed the pins for those (which I hadn't done the last time round).

I had previously used the 'setclock reset' command, and the clock reset option in Amiga Test Kit, but that did not have any affected. I then found this article from another forum:

https://forum.amiga.org/index.php?topic=51562.0

that suggested to reset the clock using RestartClock by Thomas Rapp.

So I did that last night:



I set the clock the current date and time, and this morning both Workbench and Amiga Test Kit showed that it had kept time (whilst the A500+ was switched off) correctly:





We'll see tonight whether it is still keeping time...

[solarmon]

I tried it again this evening after work, and the clock was still keeping time.

So it was the reflow of the oscillator and varicap and/or the RestartClock command.

Either way, it seems to have fixed it so I'm happy!

And still no signs of the boot up delay.

[solarmon]

Tried it again yesterday evening and the clock had stopped earlier that morning.

Back to square one again. I'll leave this for a bit whilst I tinker with my A1200, and revisit this issue when I get some motivation for it.

[solarmon]

Got some motivation/determination to revisit this issue.

Since the time was not ticking, I thought that the Y9 oscillator might not be good, so I replaced it.

I put some kapton tape under the oscillator (see image below) as there seems to be a bit of the solder mask exposing the ground plane underneath it and possibly the original oscillator casing may have been touching it.

However, this did not resolve the clocking keeping time issue.

So, I took the plunge and replaced the U9 MSM6242B RTC chip itself. I added an IC socket to it and had to repair some of the damaged (during extraction) pin holes:



And I used RestartClock tool to reset it, and Amiga Tool Kit to set the time.

So far, it has been keeping time, but I will hold my breath until I see it working for much longer.

[solarmon]

Before the RTC clock replacement, I did come across a black screen issue again. This time it was seemed to be a permanent black screen rather than delay. Or maybe it was just a very very long delay.

That time I had my logic analyser ready and I was seeing that the /RST signal on the CPU, Paula, Gary etc was flipping between low and high.

I unplugged the keyboard to make sure it was not the /KBRESET signal coming from it. No, it wasn't and /RST was still flipping between high and low.

I probed the U42 555 timer chip and even the trigger (TRIG) and threshold (THRESH) input pins were flipping between high and low. So maybe an issue with the R71x/C71x components?

I was probing U37, U15 and U36 pins to see if anything else was flipping. When I probed pin 7 of U36 (74LS05), which is just a ground pin, the black screen disappeared and the boot screen was displayed.

I don't know if this was a just coincidence or how touching of the probe on this pin resolved something. I could not get it to black screen again, so something changed.

I decided to reflow the pins of the U42 (555), U37 (74LS32), U15 (74LS157) and U36 (74LS05) chips and the R71x/C71x components.

I haven't had the black screen issue again, but it has been intermittent and can re-appear after a few days.

If it reappears again, then I will take more time looking at the R71x/C71x input components to the U42 (555) chip. And possibly replacing the U42 (555) chip, if that is deemed to be causing the issue.

[solarmon]

Since the last update, I had left the A500+ turned off for over 10 days. I checked it again this evening, it booted up without any issues and it was still keeping time, although it had gained about 1 minute.

Since I had put a different oscillator in, it probably needs calibrating. But since I don't have an oscilloscope to do that, I might just put back the original oscillator, which probably didn't have any faults with it.