Jump to content
Due to a large amount of spamers, accounts will now have to be approved by the Admins so please be patient. ×
IGNORED

Robotron 2048 Repair Log


Recommended Posts

Had @Godzilla's Robotron 2084 PCB set on the bench recently.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_3558.jpg

 

Being an early Williams PCB it follows the pinball modular style where the system is spread over multiple PCBs, (CPU Board, ROM Board, IO Board, Sound Board) all connected with ribbon cables and hooked to power individually. Not only is this a pain to power up on the bench but it also means the board is littered with physical connectors that really start to cause problems with nearly 40 years on the clock.

 

These boards also use an old type of tri-rail DRAM chips, that need three power supplies to work (+5v, -5v and 12v) and if any of those are missing the chips have a high chance of being killed. There is a specific order the power rails need to be applied, or disconnected, but almost all equipment just uses the "instant on" approach. But factor in flaky old sockets and old power supplies and RAM faults are fairly common.

 

When Williams boards boot they go through a self test process and draw the (in)famous "rug" pattern on the screen a couple of times before either reporting the fault out via the LCD digit display, or going on to run the game.

 

This board was not getting very far, drawing half a rug on the screen...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_3583.jpg

 

...before giving the error code 1-3-1 on the ROM board 8 segment display.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4225.jpg

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4226.jpg

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4227.jpg

 

Which according to the manual means it has found a RAM fault (denoted by the 1), in RAM bank 3, at chip 1.

 

The RAM for this board consists of a big slab of 24 DRAM ICs on the main board, arranged in 3 banks of eight, giving a total of 48KB onboard.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4244.jpg

 

While the first 1 is helpful, the 3-1 that follows is slightly misleading.

 

In this case, with a working board the PCB starts testing at bank 3, then moves on to bank 2 and finally bank 1. Each of the 8 DRAM chips in a bank contributes 1 bit to the bytes being read, so bank 3 chip 1 holds the very first bit of the very first byte being read. All the testing is doing is trying to store a value at the first address and then trying to read it back, if the values match then it passes the test and moves on, but as soon as the system finds a fault it halts and gives its diagnosis, in this cases 1-3-1, after which it reboots and tries again.

 

Although you could have a problem with that chip (or socket) in reality these "fail at the first bit" faults can also mean the PCB can't access any of the RAM if it reports the 1st chip it tries. In which case a fault in the logic that handles the data flow in and out, or the logic that controls the chip enable, and write enable lines is more likely.

 

To get anywhere with this sort of issue you need either a logic probe, or an oscilloscope, and in this case the fault was on the DRAM address bus, with line A1 showing no signs of life.

 

Address line A1, pin 7..

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4238.jpg

 

..silent as the grave.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4242.jpg

 

Compared to address line A0, on pin 6...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/afa66e7a-fa9c-4df0-b7fc-76a7270d37cd.jpg

 

...alive and kicking.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4243.jpg

 

All the other lines on the DRAM bank were active and had healthy looking signals, so the fault on A1 was likely the one that finally took the cabinet down.

 

Pulling out the schematic this line tracks back to a 74153 multiplexer at 4E...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/Schematic%20Fault.png

 

...which is this bugger.

 

66ed5dea-94bb-4e4d-9ef5-752ab82b7bfe.jpg

 

...and poking the scope at the relevant output pin showed the output was just as dead as the input pin on the DRAM, so this wasn't an issue with a broken track.

 

Piggy backing a 74LS153 from a scrap board lit line A1 back up.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4229.jpg

 

This proved the chip had lost the ability to drive that PCB line, rather than something else tying that line low.

 

The board would then "rug"...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4224.jpg

 

... pass the self test...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4231.jpg

 

... and start attract mode.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4236.jpg

 

That should be it, but I've met Williams PCBs with Signetics branded TTL faults before, and when one fails the others are right behind it, ready to fail a few days later. I've been burnt before so the other three 74153s on the address bus needed to retire too.

 

As I was placing a fairly large order for logic ICs I just added four 74153s to the order, I could have fitted the more common LS type 153s but fitting one LS in amongst other non-LS ICs can in theory cause timing issues, I also didn't have 4 74LS153s in stock. Normally this stuff takes a week or so, but 6 weeks later, lots of chasing, one split order and a partial refund later they arrived.

 

a81f5982-9c7d-4c82-b2fa-2a130330e2d7.jpg

 

The culprit comes out...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4247.jpg

 

...and new one goes in

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4250.jpg

 

Well not "new", all this stuff is decades old, I imagine they stopped making base 74 series in the 1980s, most 74LS production stopped mid 90s with only a few still being made into the 2000s. I've bought "new" TTL ICs in Jaycar recently that have date codes from the 90s

 

The last of the four refused to cooperate with the desoldering station so it was cut off and the legs desoldered individually, and the remaining three replaced.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4267.jpg

 

With the address bus fully restored the game would now boot happily, and would run fine for ages but if I knocked the desk or tapped the board it would instantly crash, re-rug and start up again.

 

This is very common in old multi-PCB games, and even inside the cabinet I would expect this game would be pretty prone to random issues and crashes from any knocks and bumps.

 

These old Williams multi PCB games have a load of weaknesses that later single PCB games don't, namely all the PCB interconnects and the old style "single wipe" IC sockets which rely on metal pressing up against metal. This works fine when everything is new and clean, but over time a layer of oxide and dirt builds up on the surfaces increasing the resistance until the contact lost. With the 40-way ribbon connection between the CPU board and the ROM board there are over 100 metal-on-metal contacts, from the vampire taps that bite into the ribbon, to the socket and connector pins at the ends. When things are never removed or re-seated this oxide is never scraped away so the contact point can slowly degrade to the point that it is barely hanging on. The old type single wipe sockets have a second issue, they rely on the springyness of the IC legs to hold themselves against the socket pins, usually the inner face. As the chip ages the legs can relax a bit and the pressure is lost. It seems that the larger the single wipe socket the more problematic they get with age, as I think the plastic also loses its strength. The more modern dual wipe type has a contact on the inner and the outer face of the contact so the leg is held much better, but the sockets on this board seem to be in a pretty good state.

 

For this board the issues seemed to be mostly down to the ribbon cable between the ROM board and the CPU board, as any movement there instantly caused the board to crash.

 

The pins on the board were pretty tarnished and it's almost certain that the same degree of crap is up inside the connector too, plus the ribbon cable itself was old and stiff, possibly with oxidisation at the vampire tap points too. All up there are 120 contact points in this link carrying the data bus, address bus and ROM controls lines, any glitch on those will be fatal.

 

Also the connectors on Williams boards always seem to be missing their locking levers, if they ever had them, and as they also function as eject levers their loss means the only way to disconnect the cable is to pull on the ribbon cable, which actually pulls the vampire taps out of the ribbon over time. The fact you can see the metal pins...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4252.jpg

 

... are not fully pressed into the ribbon, and the fact this much daylight gets through..

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4253.jpg

 

... means the contacts are probably only just hanging on anyway. Its fairly likely that the constant pressure and movement from the vampire taps have chewed the cores away so there isn't the diameter of metal there once was.

 

Even a night clamped in my mini-vice didn't fix the issue...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4254.jpg

 

...within an hour the daylight was back as I suspect the plastic housing is slightly bowed.

 

I've had limited success with refurbishing connectors like this, even with a polish and contact cleaner they they usually only work for a while but then fall over again, even without the obvious physical issues. It's far better to replace the lot, the only part thats pure un-obtanium is the PCB mounted connector on the ROM board side.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4265.jpg

 

It is possible to buy a whole new ribbon cable, with both connectors fitted, but only from the US and that would add another month to this repair, it's simpler head to Jaycar and take the DIY option.

 

The old 40 way PCB connector came off...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4269.jpg

 

... and a new 40 pin DIL connector goes in.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4281.jpg

 

Not the same style as the old one, but for some reason 40 way with latches is one size Jaycar no longer stock at all, plenty of options other than 40 pin. I'm guessing they sold out in the IDE era and the current stuff they have had for decades. Same with the ribbon cable itself, no 40 way available, so I got 50 way and peeled the outer 10 cores off.

 

The old ROM board connector clip comes off, revealing the vampire taps that bite into the cable cores. This ribbon cabled was peeled off and added to the "retired" parts pile.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4274.jpg

 

With the vampire teeth exposed I could get in with a glass fibre pen and some cleaning solution to get them back to a full shine.

 

Clamping the new 40 way IDC connector onto the ribbon was easy enough for the free end of the cable, I just used the minivise to press the cable into the pins, but for the PCB end that wasn't an option. There's no way I could apply enough force by hand to get all 40 of the teeth through, the only option was to bring one of the big gun clamps I used to clamp up water damaged woodwork...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4275.jpg

 

and use some soft-ish wood underneath the PCB that the pins would sink into...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4277.jpg

 

...so I could put force up through the flat PCB but not onto component legs. With another piece of wood on top to spread the pressure it all went together without causing any damage.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4276.jpg

 

With that done the board is now as solid as a rock, I can kick the desk all day long and it keeps running. Even flexing the ribbon cable and moving the ROM board around no longer has any impact.

 

The rest of the boardset got a polish on all the pin headers, and all the socketed ICs were removed, were put through my leg former, and had the inside edges of the pins polished up to a shine the glass fibre pen.

 

One decoupling capacitor was slightly domed...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4283.jpg

 

...and was starting to let go underneath...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4284.jpg

 

..so it was replaced. I took an identical one off the IO board, and fitted that to maintain a matched set of 3, mainly for aesthetics. The IO board got a Rubycon one from my parts box.

 

With all that done the board should be solid as a rock for many years to come!

 

All that remained was to name the guilty...

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/IMG_4288.jpg

 

... and get this back home to it's cabinet.

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/Cab4.jpg

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/Cab1.jpg

 

Glorious sight!!

 

http://i40.photobucket.com/albums/e230/Womble76/Robotron/63cee8a7-0058-4237-ac8c-921125d21412.jpg

Edited by Womble
Link to comment
Share on other sites

Oh cool, thanks for the heads up. Half the battle is knowing what the hell they are called! :)

 

RS is out of stock at the moment too, but I'll stick a couple in my next order for sure!!

Edited by Womble
Link to comment
Share on other sites

  • 2 weeks later...

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...