Back

Anyway, re-seating all the ZIP chips didn't improve stability, so I ended up removing all of them.

The Z3660 uses the Z-Turn can map 128MB of its own RAM to the Amiga address space.

Using modern hardware to sidestep hardware issues on the A3000 board feels a bit like cheating, but ZIP RAM modules are rare, expensive, and... a stupid idea even back in 1990!

#commodore
#amiga
#vintagecomputing
#retrocomputing
#electronics

With on-board Fast RAM disabled, the system is stable enough to run Lotus Turbo Challenge... up to a certain point!

This crash is 100% reproducible. I don't know whether it's caused by a hardware problem with the ECS chipset or this game is simply incompatible with the 68040.

Last friday I started debugging the ZIP RAM bank.

Someone on the Acill Classics repair chat suggested running ziptest, an awesome tool probes each ZIP slot to debug issues like mine:
https://aminet.net/package/util/misc/ZIPTest-1.1

I populated the first 4MB of ZIP slots populated, and booted OS 3.2.2, which didn't recognize any Fast RAM in the expected range, $07000000-$07FFFFFF.

The output of ziptest makes me think that the entire bus is busted. Perhaps I have a broken Ramsey?

Meanwhile, I also checked the reset logic, which I suspect to be quirky.

The A3000 uses a PST518, a small 3-legged IC which looks like an ordinary transistor, but monitors the +5V power line to assert a _FAIL signal when it's too low.

This capture shows what happens when the power supply is switched off: _FAIL (above) drops sharply when +5V (below) falls below 4.4V. Looks good.

On power up, +5V (still below) raises quickly to its stable level, while /FAIL (above) looks like the typical capacitor charging curve.

Am I seeing the effect of R112 (3.3kohm) C112 (47uF)?

Ugh, I need to refresh my high-school electronics on RC circuits:
https://www.electronics-tutorials.ws/rc/rc_1.html

So the time constant (τ) is 3.3 * 10^3 * 47 * 10^-6 = 152 ms

The scope captured only 350 ms of the curve, which grows to ~4.2V. Looks about right.

But is this what a good reset signal looks like? I expected it to be driven sharply between 0 and 1 logical states.

I found the PST518 datasheet here:
https://www.alldatasheet.com/datasheet-pdf/view/87726/MITSUMI/PST518B.html

My understanding is that the PST518 output is meant to be fed into a Schmitt-trigger to get the sharp power up reset signal that I'm looking for.

But on the A3000 PCB, /FAIL rruns to Gary's and Ramsey's POWERUP pins.

Perhaps they have an internal comparator for this purpose.

The Kickstart 3.2.2 roms arrived today, and I swapped them on the board.

Oddly, these EPROMs are 1 pin too long for the A3000 sockets. The seller clipped pins 1 and 42 to suggest the correct alignment.

Now I can tell the Z3660 to use the on-board ROMs and... IT BOOTS!

This confirms that the original ROMs had indeed gone bad, while Gary and other bus logic still works... which is very lucky

With the new #AmigaOS 3.2.2 ROMs in place, I can even pull out the Z3660 and use the on-board 68030 and its companion 68882 FPU, both running at 25 MHz.

Since there's hard-drive, I'm welcomed by the #Amiga bootscreen, except the familiar Amiga "rainbow checkmark" logo has been replaced by the Escom-era "boing ball"

But I forgot to cover how I fixed the stability issues.

Testing the corroded traces around the battery, I found that DRD9 was open between pin 46 of Denise and pin 5 of Agnus.

That's one of the data bus lines; it's surprising anything worked at all! Restoring DRD9 (white alligators) fixed the vertical lines as well as the frequent crashes. Woot!

The yellow alligators reconnect Paula's INT6 to Even CIA.

The trace was badly corroded around U190, the Ricoh realtime clock, and my clumsy attempts to clear those pads finally killed INT6 for good, causing AmigaOS to hang at boot

Since I couldn't find a suitable pad beyond the Varta blast area, I'll re-route INT6 all the way across the motherboard to CIA.

As a side note, the test clips I got from Amazon for $6.89 a dozen won't attach firmly to DIP pins, and the tiny copper hooks keep bending and breaking.

I was told I should have bought these fancy Tektronix IC grabbers, but they cost a whopping $29 for a bag of 10:
https://www.ebay.com/itm/274793389279

I wonder which alloy they used for the hooks?

Anyway, this is my first attempt at repairing traces with thin rigid wire.

I used polyimide tape (aka Kapton tape) to hold it down while soldering, but I should probably glue it firmly to the PCB.

I'm not sure which glue is safe to use on a vintage board, but I can ask the experts on Acill Classics.

Yesterday I also started pulling out the corroded passive components right above the battery.

These are part of the Amber scan doubler, which works just fine now, but would eventually start malfunctioning as the corrosion progressed through the solder joints.

Clearing the pads from the hardened metal oxides was extremely difficult and fatally damaged a couple of traces which were half corroded anyway.

But you gotta do what you gotta do...

And here are the new components: the blue resistors, the two diodes and some the yellow ceramic capacitors are mine.

You might notice that solder joints look bad around some of the pads. That's mostly because there's no longer a pad for solder to firmly attach to.

Connections will have to be restored with fly wires on the bottom side...

I'm not too proud of my first ever attempt to repair a trace with magnet wire, but that thing is *tiny* and even with magnifying glasses I had a hard time seeing what I was doing.

The disaster are is a corroded ground pin. It looks bad, but it will hold just fine for now.

More worryingly, several vias look badly corroded, but I have no idea how to reconstruct them. I can't simply suck the dark goo from the hole, so perhaps I should drill through?

Last week I soldered back the Ricoh RP5C01 and the D flip-flops which latch the 4-bit address bus.

I decided to power up the board halfway through, before reconnecting the RTC to its 32768 Hz quartz oscillator and the power supply circuit.

Surprisingly, the CPU was still able to write the date/time registers and read them back. Of course, the seconds wouldn't increment without a clock source!

To finish the job, I ordered this little kit which adapts a "coin battery" to fit onto the pads originally occupied by the nasty Varta:
https://www.ebay.com/itm/355497835640

The PCB uses a MOSFET to prevent current from flowing into the non-rechargeable lithium battery when the system is powered.

Older designs used a diode, but the ~0.6V drop across it would bring VCC too close to the lower limit for the Ricoh.

Perhaps it would just be simpler to disable the original charging circuitry by removing D190 and D191.

This way, the Ricoh would be powered by the battery through R192, even while the system is running. And perhaps, replacing R192 with a wire would source less current from the lithium battery, extending its life.

Since my knowledge of #electronics is limited, to play it safe I might end up using the CR2032 adapter as designed. Suggestions welcome.

While my last retrobrighting experiment proceeds quietly at home (*), I went to the asylum to finish restoring the #Amiga 3000 battery-backed clock.

My kit requires a P-channel MOSFET. Instead of the recommended IRLML6401TRPBF, I found a similar part with higher current (fine) but higher drive voltage (2.5V):
https://www.digikey.com/en/products/detail/infineon-technologies/IRLML6402TRPBF/811437

To confirm it will work, I taped the mosfet to its tiny terminals and measured +3V on the output pins.

(*) https://mstdn.io/@codewiz/113710433519948326

Since it's my first attempt at SMD soldering, I watched a couple of video tutorials and then tried to solder the same component on this practice board.

The result doesn't look great, but it will suffice for this easy job: as long all 3 pins are connected to their pads and not shorted together, the CR2032 will deliver its +3V output to the RTC clock.

And that's it... phew!

Now I just need to solder the battery holder on top of it and the 3 legs underneath...

And voilà! The RTCPWR pin is receiving 3.3V and the oscilloscope confirms that the little quartz is ticking!

Today I also did a quick 3D-printing test for a pass-through bracket to bring HDMI and Ethernet ports of the Z3660 to one of the Zorro slots behind the A3000 case.

Since I suck both at CADs and 3D printing, I downloaded an STL file for a dual-HDMI bracket as a starting point.

I didn't expect the first try to be perfect, but I was surprised to find that someone had designed and shared a bracket skewed the wrong way!

Or... do Zorro and PCI have opposite chirality?

Parametric #CADs let you specify constraints on the sizes, distances and angles between certain lines, which can be adjusted at any time.

After drawing the outlines of the HDMI and ETH connectors on the front face in the Sketch workbench, I wanted #FreeCAD to punch them as holes through the bracket's thickness.

For instance, even with the early #Amiga ray-tracing packages such as Imagine and Lightwave, I'd start by drawing a 2D shape and then make a solid by "extruding" this flat silhouette along its normal. #cad

You could extrude a profile around an axis to make radial shapes such as a vase, or along a curvy path to make a pipe. And you could even scale the section along the way to make a snake.

Over 30 years later, extrusion remains a fundamental construction technique provided by all modern #CAD software.

Real 3D, one of the first solid modelers featuring constructive geometry, boolean operations and smooth surfaces, also debuted on the #Amiga.

Real 3D was very advanced for its time, but also harder to learn. I never went beyond modeling simple shapes with it, but I had an older friend who was using Real 3D for architecture.

Lots of great 3D modelers and ray tracers originated on #AmigaOS:
https://sites.google.com/site/3dtimeline

But I digress. I went to the Artisans Asylum with this rough #FreeCAD design to print it with one of the three Prusa MK3 available in the Digifab shop.

A couple of nice people helped me setup various slicer parameters and the printer itself.

A 0.10 mm fine print with common PLA filament takes about 40 minutes, and I had to iterate a couple of times to adjust the position of the screws, add scaffolding for the mounting screw hole on top, etc.

Everything fits together, but I couldn't find suitable self-tapping screws to attach the HDMI pass-through.

Aside for the color, I'd say it's a passable surrogate for a proper metal bracket.

I noticed now for the first time that the top Zorro slot of the #Amiga 3000 is slightly taller than the other three. This is also the position of the A2000-compatible video expansion slot; perhaps the extra clearance was needed to accommodate preexisting cards?

To complete my #Amiga 3000 restoration, I still need to solve a small mechanical issue: there is barely enough clearance under the drive bay for a Z3660 with a Z-Turn board on top.

My Z3660 card is also slightly tilted upwards when plugged into the CPU slot, and the RJ45 connector of the Z-Turn contacrs the bottom of the frame, putting stress on the connector.

I could just cut an opening right over the RJ45 connector. So I took some measurements and marked the area to cut off in orange.

Luckily, there is already a hole in the frame from which I can start cutting around with a hack saw.

How hard could it possibly be?

Ugh. Cutting square holes with a hack saw isn't as easy as it seems

I guess that's the original meaning of the term "hacking":
http://www.catb.org/jargon/html/H/hacker.html

After some filing, that hole looks a little smoother

With tape covering the rough edges, the result is almost presentable

To be honest, #Commodore designed a terrible drive bay for the #Amiga 3000.

It's so bad that someone felt the need to design a new split frame with much improved ergonomics:
https://www.amibay.com/threads/amiga-3000-drive-bay-frame-split.2444489/

I should probably get in line to obtain one of these when the next batch of production happens.

The hole in the drive bay is almost a perfect fit for the RJ45 connector of the Z3660.

I can start reassembling the A3000 chassis...

Since I couldn't figure out what happened and how to repair the FFS directory structure, and re-transferred the #AmigaOS 3.2 system I normally use in UAE.

The problem never reappeared, so perhaps it was just FFS being prone to corruption when writes are interrupted by a crash. And this #Amiga was rebooted numerous times while I was chasing the remaining hardware and software stability issues.

Anyway, these are some of the joys and sorrows of #vintagecomputing

Found the perfect companion for my #Amiga 3000: this beautiful NEC MultiSync with a horizontal range of 31-110 kHz!

I pulled it from the basement of my house and tested it briefly: it still works, and the picture quality is perfect after degaussing the CRT and adjusting a few display controls

@batteman @codewiz it's too common for ffs to corrupt itself after a crash in a way that its disk-validator can no longer fix it on its own. 3.2 ships with a rewritten diskdoctor, which might be useful. I've been a pfs believer since pfs2, but it is not perfect either. I guess keeping backups is the universal solution. :-)

@jope @codewiz You're totally right. I don't use SFS on my classic Amiga, but PFS3 AIO (all in one) !
I use SFS under MorphOS ;-)
I can't remember what filesystem I use on my A1200 Pistorm with AmigaOS 3.2.2.1 (but it seems Emu68-Imager uses PFS too ^^).

I also had three SFS partitions on my hard drives: Work, Store and BBS.

However, my RDB still contained the original SmartFilesystem 1.84, which couldn't mount the partitions and was likely the root cause of mysterious crashes of the "UAEDH0" task.

Many of my favorite #Amiga system tools are still in active development today: SFS, P96, DOpus5, ViNCEd, MCP, AmiSSL, MMULib...

@batteman @jope #retrocomputing

Plus, #AmigaOS 3.2 integrates and obsoletes many other hacks I installed decades ago.

I still use TinyMeter and MagicMenu. Are there any actively developed alternatives for these?

I'm slowly catching up on 30 years of #Amiga development

@batteman @jope

@codewiz @batteman those are still the ones to have. I believe 3.3 will support floating menus natively, but let's see once it is out. :-)

@ij @jope @batteman I did not think of scanning the corrupt partition the new diskdoctor bundled with OS 3.2 to see if it would at least report a problem with the Tools directory.

@jope @ij @codewiz Never tried it (and don't use FFS anymore since a long long time ^^)