User avatar
Mr.Toast
Roseville, CA

Posted Sat Jun 15, 2019 7:21 pm

A few months back, an AmigaLove.com user contacted me about help getting his Amiga 2000 up and running after a decade plus of hibernation. I haven't worked on an A2000 yet as most of my experience has been the A4000, A500s and my A1200. I offered to take a look at it, but he would have to send the board from FL and have to be extremely patient!

A few weeks later, the bare board arrived. I knew it was going to have battery damage and cautioned him that it might take me several months to get around to it. Too many projects with my personal clock cycles limited due to work, baseball season, domestics, and the occasional bout of apathy. But, I also wanted a bit of a challenge, as well as the desire to see another Amiga live again.

The board revealed was terrifying. Corrosion was intense. This was going to be a heroic salvage.
IMG_20190302_111928.jpg
ERmagahD

What surprised me was the extent of the damage. The accumulation of years of storage and the evils of Varta (Varta, Destroyer of Worlds ©) had done their best to render this board unusable. I also had the fortune to have access to the Sacramento Amiga Computer Club's (SACC) stash of Amiga parts-including several Amiga 2000s in various working states. One of those boards was a badly abused and stripped down rev 6 (312722) PCB. I originally was advised to chuck the board but I am glad I didn't. It became a valuable reference later on.
IMG_20190302_173929.jpg
MC68000 haz sadz

The corrosion process had crept all the way to the CPU and kickstart ROM sockets. Despite the damage, it was important to verify the current working state. My basic test setup consists of a known working A2000 PSU and my trusty OSSC (self-built, mind you :D). Unfortunately, black screen with the video signal sync at the normal 263P signal @ 15.73khz output. This was both good and bad. It let me know the clock was working and that Denise was doing her job. But, it also meant that without even getting to a CPU startup, we wouldn't even get some color codes to help with troubleshooting.
IMG_20190302_183213.jpg
True story: Varta's secret weapons are actually microscopic dwarves who mine the copper in your PCB and pay you back by giving you cobalt.

Removing the CPU and kickstart revealed damage on many of the interconnecting traces and my troubles would likely be there.
IMG_20190302_210928.jpg
Carpal Tunneling for Amiga Enthusiasts!

Several washes with white vinegar, followed by denatured alcohol not only signals the start of restoration, it also promotes damage to one's sense of smell! The fiberglass pencil is incredibly useful to stripping off the damaged PCB mask and removing the corrosion. But after a short while, your hands will start feeling like you have knots for muscles and that is no fun. Once all the components were cleared out of the area, it was time to break out the big guns.
IMG_20190511_192842.jpg
If only bare copper stayed this pretty!

Enter MR. DREMEL.
Yes, that's right. The hobbyist's best friend. The Dremel. And today's ammo of choice is the 511E Finishing Abrasive Buff, starting with the 180 grain, and finishing with the 280 grain buff. Very careful and gentle touches on the bare board quickly stripped the remaining corrosion and solder mask down to the bare metal. Additional cleanup with of the damage and many many washes of denatured alcohol were the needed to remove the last of the crud.

We're still a long way from a working system, but at least most of the corrosion was done with. I say most because, later on, minute amounts of corrosion would re-emerge from the vias. The bottom of the board was pretty healthy. I noticed a slight amount of corrosion opposite of the battery and keyboard connector. Letting it sit for a few days and studying where new corrosion would emerge was valuable in chasing down the last of it. This was going to take a while.

Meanwhile, my mostly-trusty desoldering station (a ZD-915, pretty much the same a lot of the budget ones floating around) suffered from an unfortunate accident. The glass tube that contains the molten solder in the gun had rolled off the workbench one day while cleaning. Needless to say, I knew I'd be down for a bit while waiting for a replacement. This gave me a chance to keep an eye out for any new corrosion signs-which did happen and were mitigated immediately-until we finally saw no more new corrosion.

Did I mention that Varta is an evil Djinn?

User avatar
Mr.Toast
Roseville, CA

Posted Sat Jun 15, 2019 7:49 pm

(cont)

With the corrosion finally under control and stripped down, we were ready to do the build up. I had originally intended to clean and strip the keyboard connector, along with the axial caps and ferrite beads around the keyboard connector. Disassembly, cleaning and reassembly was successful-but in the back of my mind, I was still worried about VartaJuice® and wondered if she would be hiding her evil in places that the fiberglass pencil would not reach.
IMG_20190511_193453.jpg
Ahh, so that is how those connectors are made!

IMG_20190511_203634.jpg
That looks better... but, did I get it all?

I also had a 64 pin socket on order from China that took about a month to get here. I have a spare pull one, but it looks a little shoddy. I decided to take a chance with what I had for a while just to get on with the troubleshooting.
IMG_20190526_205354.jpg
With UV curable solder mask, it looks messy but is great for sealing up our work.

It was also time to start laying out the solder mask. A tube of UV curable mask is a bit like painting with a thick ink. It will never be pretty with this approach, but at least is should be a step to a working config.
You'll also note that I finally got some replacement glass for the desoldering station-which made for some replacement of all our capacitors while I "have a think" about what will need some troubleshooting.
IMG_20190526_205428.jpg
There were about three or four damaged spots on the bottom that were potentials for problems in the future. Kapton to the rescue!

This was a good pause because as the board was explored for capacitor soldering, it gave me a good chance to spend some up close and personal time with both the top and bottom of the board. Several imperfections on the traces were found and buzzed out. A few needed bodge wires and a few just needed a tiny strip of Kapton tape to provide some protection for exposed traces.
IMG_20190529_235903.jpg
YSOD?

I luckily have a working spare A2000 to test the custom chips on. Post cleanup, the 68k CPU and all socketed ICs worked great on the test system. These ICs were marked with green dots from a paint pen. However, when we put them in our unit under test, we were seeing a yellow screen-which generally indicates a CPU fault. This was to be expected as much of the battery damage was located around the CPU, it was likely down to some damaged traces that were not obvious on first inspection.

Lesson learned: I should have put the solder mask on AFTER completing all the trace verification. In hindsight, I don't know why that didn't occur to me.
Last edited by Mr.Toast on Sat Jun 15, 2019 7:52 pm, edited 1 time in total.

User avatar
Mr.Toast
Roseville, CA

Posted Sat Jun 15, 2019 7:50 pm

IMG_20190529_234455.jpg
On the hunt for broken traces!

The first couple damaged traces were pretty easy to identify and bodge up. These traces were damaged around the network resistor adjacent to the 68k. But, we were still getting that dreaded yellow screen on startup.
IMG_20190607_073418.jpg
GOTCHA!

For several evenings I poured over schematics and probed and traced and probed and traced... Something had to be damaged and …. dammit, where are you????

As mentioned earlier in this thread, there was one particularly challenging trace problem related to the data lines. In this case, it was the pin 1 (D4). This line traces to the kickstart pin 22. From there, it should go on to Gary. This line showed good continuity several times during test. However, when I went back and did a full check of all 68 pins, I noticed that this particular line was a bit more intermittent on the second pass. Or maybe it was the third pass. Or fourth. Or... you get the idea.
IMG_20190607_073249.jpg
ChuckyGang was here? Thank goodness for DiagROM!

This was the eureka moment for me! Additional probing confirmed, followed by some close up inspection under the microscope, the trace led to a via ring. The via ring then lead to another trace that had just a tiny fragment of copper connecting it to the ring! YOU BUGGER.
IMG_20190608_100348.jpg
/me does victory dance. Then checks to verify webcam is off.

Out come the jumpers and next thing you know, we have LIFE!!!!
DiagROM is a really awesome tool that is a must-have tool for working on these old machines. Its a modern ROM that goes in your kickstart socket. While its not quite a "dead test" like cart, it does provide some functionality assuming a minimum working system. It will allow you to some hardware functional health checks, and has the ability to be connected to a PC via the serial port and run a remote terminal to do testing via TTY. But, CPU/ROM functionality is critical for it to work at all. Fortunately, with that final trace jumper in place, some of these advanced features of DiagROM became a moot point.
IMG_20190608_135223.jpg
Making the changes permanent.

With our intermittent trace confirmed, Kynar made for a simple jumper to bring D5 to KS pin 22. A few dots of hot glue and its on to testing!
IMG_20190608_112849.jpg
And now for some gaming! (I mean "testing")

I was thrilled with a WB boot that went by the books! And, not to be content with that, we had to spend a few hours throwing down some floppy games. Enough time to be utterly content that we had a working and reliable A2000!
Last edited by Mr.Toast on Sat Jun 15, 2019 11:25 pm, edited 2 times in total.

User avatar
Mr.Toast
Roseville, CA

Posted Sat Jun 15, 2019 8:17 pm

[EPILOGUE]
One thing was bugging me. That keyboard connector and the surrounding passives (ceramic capacitors and ferrite beads) were salvaged from the original board. I just had this sinking feeling there would be some manifestation of corrosion in the metallurgy that would continue on despite my repair efforts. As the repair dragged out for several months due to cycling through other projects, it kept bugging me despite having a working board now. Every now and then I would notice some light green powder residue. It might have been the solder mask curing-or something more sinister. Either way, it had to go.
IMG_20190615_150839.jpg
Oo shiny!

Luckily, aside from the custom ICs, there are often many modern manufactured replacements (we call them "Form Fit Function" or F3 in my line of work) for logic chips, passives (resistors, caps) and even most connector types. Along with some brand new passives, I think we have finally defeated Varta at her sneaky game.
IMG_20190615_150859.jpg
HAIL TOAST. VANQUISHER OF VARTA!

Installed the sexy new keyboard connector and new passives, it was time to freshen up some of the solder mask. I've found that cutting it with a bit of denatured alcohol reduces the globbiness and reduces the texture. But, to be honest, it still looks like green chutney. Which is not exactly a pleasant aesthetic.

Mmm, pass that samosa, please :D
IMG_20190615_160711.jpg
Keyboard test is good! Its just a shame my A2000 keyboard SUCKs at some keys.

While the keyboard functionality checked out good, to be honest, I haven't got a refurbed A2K keyboard yet. The ones in my possession all have some quirky keys and smell of years of storage shed neglect. I really need to fix one up soon.

All in all, it has been an exciting learning experience! And, I think the owner will be thrilled with having a working rig again and be back in the game.
Last edited by Mr.Toast on Sat Jun 15, 2019 11:40 pm, edited 1 time in total.

User avatar
intric8
Seattle, WA, USA

Posted Sat Jun 15, 2019 9:30 pm

EPIC post, Toast! Incredible journey of a master at work.

User avatar
Bulletdust

Posted Sat Jun 15, 2019 9:45 pm

Good job, scope and everything!

I normally just sit the board on top of a moderately powerful light and inspect the traces under a magnifier!

User avatar
oggie

Posted Sun Jun 16, 2019 3:54 am

I have a 2000 with the same problem I need to look at. I sure wish there was a guide on what traces to test based on the color (green for me).

User avatar
obitus1990
USA

Posted Sun Jun 16, 2019 5:34 am

Excellent job! I love reading your posts, with all their nerdtastic sense of humor.

I know your pain of the fiberglass pen (not to mention all the times I forgot to put gloves on and wound up with nearly invisible fiberglass splinters in my fingers)... so I started using sandpaper, manually. Then I went with the dremel after watching John Hertell's videos of stripping the soldermask off various PCBs with a belt sander (or was it an orbital, I can't remember). My fingers have thanked me since going with rotary :)

User avatar
3D-vice
Germany

Posted Sun Jun 16, 2019 7:25 am

That is an awesome documentation of the restoration! I like that fact that the electronic components have "the good old size", a size where you can solder stuff yourself and work with without the neccessity of using magnification glasses. ;)

I've once tried soldering SMD components, but this is beyond my capabilities, lol.

User avatar
G8rduc
North Florida

Posted Sun Jun 16, 2019 4:22 pm

Little history on this poor A2000! It was given to me by a remote office that used it for data analysis. I'm not sure on when it was used, but I'm going to guess it was in the early 90's. I was about to pull the trigger on a used Amiga when one of the IT guys sent out a post asking if anyone wanted an A2000 with monitor that had been sitting for over a decade. Otherwise, they were going to junk it. I immediately replied and waited about 2 months before it finally made its way up to me.

What I got was an A2000 with a 1084s. The A2000 has a GVP 8+HD in it. 8mb of RAM and a 40MB hard drive. No keyboard or mouse, so I'm looking for replacements. Immediately, I ordered some other things, like a CF card replacement, newer KS rom, Amiga mouse to USB convertor and some other knick-knacks....Chris was more than kind enough to offer me his service and I was definitely not in a hurry and here we are. To say I'm excited is an understatement. This Amiga, along with a Coleco ADAM, are waiting to go next to some arcade machines that I've got, in my soon to be near-completed (but not really ever finished) man cave. It's hitting the hottest time of the year plus I just tore a muscle in my calf, so I'm a little slow right now, but hopefully, within the next month, I can finish it off and get these amazing toys set up. In the mean time, the Amiga will reside in my office at work and do a little playing alongside modern computers.





Return to “Hardware”