Amiga 3000 Resurrection

[RDP]

Hello all,

After about 15 years I have decided to turn on my Amiga again and get it going. Let's just say that I do this out of my passion for the machine. Just opening it up and messing around brings back amazing memories. Nothing beats this! Anyway, after 15 years I have some..... "problems". Some, I have already fixed, one is a showstopper.....

When I used my Amiga I was proud of my hardware selection. After getting caught up on the forums it seems I made the right choices back then:

A3000 desktop (made in U.S.A.)
Warpengine 3040 040@40 w/64megs of ram
Seagate Hawk 2gb on the Warpengine SCSI
Cybervision 64
Emplant (don't really need this anymore....)
Mindlight 7 (not used yet, still trying for a working system....)
Original keyboard
Amiga Technologies black mouse
Some keyboard adapter, cannot remember what it's for, think it might be for using PC keyboards on the Amiga, or vise-versa.....
SD floppy drive (A3000 style) - broken, threw it away.
HD floppy drive (A3000 style)
OS 3.1.4
Kickstart 40.68
Gotek with FlashFloppy

So begins my resurrection:

1. Gotek with FlashFloppy. Needed this as I cannot find my WB disks, or my WarpEngine, or Cybervision floppies. This went reasonably well.

2. Spent DAYS trying to get the HD to working with the WarpEngine SCSI Tools software. After going crazy I finally stumbled onto just changing the device to warpengine.device in the standard SCSI HD tools on the WB Install disk. Wasted days.....

3. Installed WB onto the HD. Went well.

4. Cybervision software, I just couldn't get it to work. Gave up, and am now using the Picasso96 shareware software. Seems, slow, maybe,? I have larger problems now....

5. Installed BetterWB.

6. Cut out the old battery after hearing about horror stories. No idea how I can get this taken care of, I need to find a tech.....


I now have what should be a clean system, but I have massive problems with my mouse:

For the first few days after running perfectly for some minutes, the mouse would start to only work on the far left of the screen, you could scroll to the right, but then the pointer would just back to the far left. If you scroll to the left, it would jump to the same position about 100 pixels to the right, then scroll left. While up and down worked find, once this started you couldn't do anything productive. However, turning the machine off for some time would usually fix this. Now, recently the problem starts almost immediately, and turning the machine off for a very long time doesn't fix the problem. Removing the expansion cards, swapping the CIA chips, none of this helps. I live in Tokyo, so buying a new mouse is not cheap or easy just to do a simple test, but if I want a USB mouse with Deneb, then I need a regular mouse to get it installed. Maybe I should just try a regular PS/2 to Amiga adapter first?

If anyone has any ideas I would love to hear them!

thanks,
RDP

[talybont]

Here's some info from someone knowledgeable in regards to Amiga 500, maybe this is relevant:

Only fire button/left mouse button is connected to CIA.
Right and middle button is connected to Paula.
Movement: Denise via U15.

Source: http://eab.abime.net/showpost.php?p=867144&postcount=6

Here's the schematic for the A3000: https://www.amigawiki.org/dnl/schematics/A3000.pdf

The joystick/mouse ports are on page 9. The component referred to as U15 above is probably U401 in the A3000.

Looking at pictures of the A3000 motherboard the Denise chip is right next to the battery, so maybe something has been damaged by the battery leaking. If you can take some pictures of the area maybe someone here can suggest what to do next.

[Kin Hell]

Take some pics of the Motherboard area when you chopped the battery out & post them on this thread please. - Take your time & make sure they are in focus & as higher Resolution as possible.

[Toni Wilen]

It sounds like mouse counter (Denise internal register) problem. I don't see how external components can cause this kind of mouse problem, especially if distance is always same (exactly 128 pixels?).

Does keyboard mouse control (Left Amiga + cursor keys) work normally without random jumps? (If it works normally: it is guaranteed hardware problem)

If you run some game (probably need whdload) that supports mouse in joystick port (for example Lemmings), does it also jump around or is the problem restricted to mouse port only?

[buggs]

Sadly, the A3000 battery is located in a very fragile spot between Denise and Paula. In most cases I've seen so far the "soup" of the leaky battery did its main destructive work at and under U480, U477, U195, Denise and further (Flickerfixer RAMs, Paula, even Agnus etc.). Especially the vias in that area need serious attention in addition to obviously broken traces.

Example of an A3000 mainboard after fixing and clear coating (OCS Denise installed):


I honestly don't like to bring such news to somebody who would like to return to his hobby. The bright side is: the machine and especially its unobtanium chips like Agnus 8372B are alive. It is fixable and just needs attention.

[TjLaZer]

Get some good cutters and carefully snip off the battery ASAP!

[alexh]

Originally Posted by TjLaZer

Get some good cutters and carefully snip off the battery ASAP!

It was in the OP that he already did that

[RDP]

Wow, you people are awesome, what an amazing response so quickly!

I will get a good picture soon as requested of the area with the battery gone, the machine is closed up now due to space restraints in my small room. For now find attached a useless picture I took before the battery was removed. When I did a quick visual around the area I did not notice any corrosion, I think I might - might have been lucky.

I will also do the keyboard test as soon as possible and respond back.

I cannot do the joystick test, I don't know where my joystick went..... and I only have this single mouse.

thanks!
RDP

[mattsoft]

You definitely have some on pins 25 & 26 of Denise and pin 7 on U480. But overall, doesn't look too bad!

[buggs]

Originally Posted by RDP

I will get a good picture soon as requested of the area with the battery gone, the machine is closed up now due to space restraints in my small room. For now find attached a useless picture I took before the battery was removed. When I did a quick visual around the area I did not notice any corrosion, I think I might - might have been lucky.

You have (in a way). Still, your A3000 needs attention. Compared to other A3000s, the PCB is in good shape.

Please have a look at the spots I've marked in the photo you've provided. The two vias near the battery are almost always broken (internally within the hole, sometimes also on top). The two logic chips also need to be (temporarily) removed/replaced in order to fix the vias underneath and nearby.

Denise exhibits battery acid stains, which points to the expectation that there is damage below it's socket (see also Toni's comment on page1). Maybe you can temporarily fix your apparent issues by re-seating Denise, maybe not. In either case, battery acid in Amigas has the tendency to continue damaging the PCB when not treated properly.

I would remove the Denise socket, strengthen/fix traces and vias underneath, then handle the two aforementioned chips (and their nearby vias). For good maeasure, the resistors/caps with acid stains should be replaced as well. In total, the material cost for parts is <$5 - unless there is considerably more damage than visible from your photo.

[RDP]

Wow, such knowledge. Let me know what I have not addressed.....

Important point (I think): The problem starts about 5~6 minutes after the machine has been off for a number of hours. It works perfectly, then the horizontal starts going wonky, then finally it sort of restricts itself to a narrow horizontal range.

To address all the great points (thank you ALL!!!):

Toni Wilen: Thanks for the hidden tip, not just the test: Yes, keyboard mouse control works perfectly.

mattsoft: Great eyes, I missed that (not sure what I was really looking for). I have uploaded more pictures, but from what I can tell you have already identified the main areas.

buggs: Thank you for the wisdom, and also adding further places of concern that should be checked, thank you!


If I can address the corrosion issues I might not need a replacement Denise?

Most important, I need a tech, I have zero experience in this area. Thankfully there is another thread on this board about people who can do work, and someone in Tokyo has put their name up there, in 2016... . I will try to contact them.

Question: It could take a while to find a tech, is there anything I should be doing right now, on my own?

UPDATE: Uh-oh, I notice green junk on R469 and R470, and also on one of the legs of the battery holder.....

thanks all,
RDP

[Toni Wilen]

Originally Posted by RDP

I cannot do the joystick test, I don't know where my joystick went..... and I only have this single mouse.

If this is reply to my joystick port test: I meant connecting mouse in joystick port and running 2 player mouse controlled game. If both ports have identical problem: problem can be only caused by U401, connections between U401 and Denise or Denise itself. (Which matches battery location..)

Because Denise multiplexes both ports (Denise's 4 input lines are used for both ports with separate output signal telling which port is "sampled" that connects to pin 5 of U401), bad/marginal connections probably can cause all kinds of really weird mouse behavior..

Does the mouse pointer move randomly if mouse is in joystick port?

[RDP]

Originally Posted by Toni Wilen

If this is reply to my joystick port test: I meant connecting mouse in joystick port and running 2 player mouse controlled game. If both ports have identical problem: problem can be only caused by U401, connections between U401 and Denise or Denise itself. (Which matches battery location..)

Because Denise multiplexes both ports (Denise's 4 input lines are used for both ports with separate output signal telling which port is "sampled" that connects to pin 5 of U401), bad/marginal connections probably can cause all kinds of really weird mouse behavior..

Does the mouse pointer move randomly if mouse is in joystick port?

Toni: Just tested, and yes, the behaviour is the same (jumping back on the horizontal) when using the joystick port. Interesting: Until I actually swapped the mouse into the joystick port the mouse cursor was moving on its own to the right, but after plugging it in, the normal "bad behaviour" (same as when plugged into the mouse port) was exhibited.

Note: Only X (horizontal) has problems on the mouse, Y (vertical) remains fine.

thanks,
RDP

[DofD]

I had similar issue with my A3000, and went through a bunch of diagnostic stuff, and at least for me it turned out that it was the mouse that wasn't working properly. I got a new mouse and all was okay.

Your motherboard actually looks pretty good, but of course a good clean to remedy the corrosion should be done.

[RDP]

Originally Posted by DofD

I had similar issue with my A3000, and went through a bunch of diagnostic stuff, and at least for me it turned out that it was the mouse that wasn't working properly. I got a new mouse and all was okay.

Your motherboard actually looks pretty good, but of course a good clean to remedy the corrosion should be done.

By any chance did you check to see if the behaviour of the mouse was identical when plugged into the joystick port (when using a compatible game, such as Lemmings as was previously recommended up the thread)?

thanks,
RDP

[Kin Hell]

Pull the board asap & get some Lemon Juice on the areas pointed out to neutralize the acid. Don't be scared either, an old toothbrush is an excellent tool for scrubbing it all down.

[DofD]

Not quite sure, it was a few years ago now. But I think I did that and got the same results when moving the mouse around on the workbench.

Originally Posted by RDP

By any chance did you check to see if the behaviour of the mouse was identical when plugged into the joystick port (when using a compatible game, such as Lemmings as was previously recommended up the thread)?

thanks,
RDP

[JimDrew]

Wow... with the exception of having a Picasso IV instead of the Cybervision 64, I have the exact same A3000/040 setup that you do, right down to memory and the SCSI setup. Fortunately mine did not have a battery leak. This is the development system that I used for years developing things like EMPLANT and FUSION.

[gdonner]

@ Jim Drew: Hat's off to you for developing those!

Speaking of Amiga 3000's and hardware damage, my A3000's motherboard was professionally repaired recently (by Anthony Hoffman--does excellent work), the old rechargeable battery was removed and replaced with a proper CR2032 mini-board, traces fixed, clock chip(s) replaced, etc. but for some reason the clock keeps resetting on power off to a date only one year ago--which is odd, considering if it was a full reset, it would show up as the Commodore default date of 1978; not 2018.

Has anyone else run into this?

My 3000 has a brand-new CR2023 battery, I've done the "SetBatt A" routine from Shell using SetBatt-1.2 from Aminet, etc. but for some reason this issue persists?

What seems odd is that every time I load SCSIPrefsMUI, it pops up a message saying that "The Amnesia bit has been cleared or reset" (or something to that effect--writing this from work ATM).

Another oddity is that, although my system recognizes the clock, and I can set it via the Time Prefs editor, it doesn't appear to actually be saving it--i.e. I don't see the hard drive LED flash briefly when doing so, like I recall being the norm when saving time in my other Amigas (A2500, A4000 decades ago).

The only thing I can think of is that perhaps(?) there's some residual damage/battery leakage that escaped the professional repair?

[jbenam]

Originally Posted by gdonner

but for some reason the clock keeps resetting on power off to a date only one year ago--which is odd, considering if it was a full reset, it would show up as the Commodore default date of 1978; not 2018.

Has anyone else run into this?

I had a similar issue on my A4000 - it turned out the date it was defaulting to was when I had first installed the system. It was probably stored somewhere and it kept using it as a starting point. The issue went away after I finally had the time to fix the RTC circuit

I'm not sure if the A3000 is any different than an A4000 in that regard, though.

I remember using Scout to check if the RTC was properly working and it would accurately report it as non-working until I fixed it completely. Not sure if it would work in the same way on an A3000 but that's worth a shot, I guess.