Novag Robot Adversary

  • Year: 1982

  • CPU: Z80B - 6MHz (8-bit)

  • RAM: 5KB

  • ROM: Program 32KB (Up to 40KB); Control 8KB; Integration 2KB

  • Programmer: David Kittinger

This machine arrived to my hands in such a bad condition, that I was not sure I could make it work. That's why I didn't take exhaustive photographs, and I can't detail the whole repair process. Even so, I thought it would be interesting to publish the material I have: there will always be someone who can take advantage of some idea, photo, etc. or simply discover what this curious chess computer looks like inside.

It arrived to my hands in this condition: with the robot arm broken in two places, without an external power adapter, without one of the lower covers, etc.

Disassembling...

It is immediately obvious that this machine has already been serviced at least once. Someone tried to repair it but left the job unfinished and messy.

The NiCd battery has leaked and damaged the surrounding circuitry.

ROM bank.

Cables cut at the power input connector!

All the cables have been cut and reconnected, but they are still in bad condition.

With the help of a Lab Power Supply, I feed the motherboard and start testing: There is no response from the electronics and there is some kind of short circuit that causes my Lab Power Supply protection to trigger.

DIAGNOSIS:

I think what happened to this Chess Robot was the following:

The "limit switch" (see below) installed on the central axis of the robot arm broke down. This failure caused the axis to rotate non-stop, always in the same direction. The bundle of wires began to twist on itself, and some were stretched so far that they broke completely (even some wires that looked fine had broken copper inside them).

Eventually the shaft was stuck, the motor started to draw a lot of current and to heat up, until some component of the power supply, such as the 2N3792 power transistor, burned out.

Subsequently "someone" tried to make a repair, but leaving the job unfinished and messy.

The rest (the robot arm breakage, how the external adapter got lost, etc.) is a mystery to me.

I start with the repair, disassembling and removing the motherboard...

Motherboard in detail:

CPU Z80B

6MHz Quartz Crystal (= System Clock)
Therefore this board corresponds to the original design.
In 1984 Novag introduced a 7.5 MHz version.

8 x NMOS STATIC RAM of 1Kbit x 4 (TMM314APL-1)
2 x CMOS STATIC RAM of 1Kbit x 4 (TC5514AP-8)

I proceed to disassemble and recondition the PCB of the ROMs...

The contacts of some EPROMs are completely blackened!

I continue with the repair of the power supply...

I use a slightly more powerful equivalent transistor (MJ15016G instead of 2N3792), applying thermal paste.

I also replace a bias resistor, and the filter aluminium capacitors.

I check the power stage that controls the DC motors, based on TIP30 transistors (PNP type), replacing all the aluminium capacitors, too.

The first phase of the repair is complete. The motherboard is ready to be tested...

Before testing the motherboard, I decide to restore the power supply wiring, as there are cut cables...

Special attention must be paid to this area. The cables must be kept out of the way as shown in the following photo, so that they do not interfere with the rotation of the central axis of the robot arm...

After mounting the motherboard, the electronics works: the keypad works, LEDs light up, sounds come out of the speaker, etc. :-)

It's time to move on to the next phase: REPAIRING THE ROBOT ARM...

The bunch of cables is unsoldered.

Using a suitable tool, the main shaft nut is removed.

Here you can see the main shaft assembly as well as the damage.

In addition, the robot arm has another broken part.

Here are some photos of the disassembly of the robot arm...

Repairing the main shaft...

The crack is "clean": The pieces fit perfectly.

After testing, I found that this plastic welds perfectly with SuperCeys Unick (cyanoacrylate with other components, such as rubber).

Surface finishing.

Result after installation of the upper friction bushing.

And now, repairing the other piece...

This piece has a very fragile design.

Rebuilt part, but it is obvious that it will break again, unless we reinforce it. I decide to attach the metal part to the plastic body, but with a screw perpendicular to the existing one, so that it is attached to the "centre" of the plastic part, and not to the ledges.

Machining of the metal part: 2.5mm drill hole.

Machining of the metal part: 2.5mm drill hole.

Making M3 thread.

Making M3 thread.

Checking that the thread has been properly made.

Detail of the new threaded hole.

Detail of the 3.2mm drill hole in the plastic part. The screw will go on the other side, taking advantage of the fact that the piece is hollow.

Now, the metal part has two fixings.

Once the structural damage has been repaired, I start with the revision of the sensors and motors, and then move on to the rewiring of the entire robot arm...

Different coloured cables to make work easier.

All parts involved.

The metal part has to be placed after inserting the plastic arm into the angle, as shown in the following picture.

Double screwed part. So it won't break again. :-)
The new screw is inserted where indicated by the arrow.

Small arm completed and tested: it has a motor and a micro-switch in the head.

This DC motor does not work properly: it has little power and turns irregularly.

I proceed to disassemble it, after removing the cam wheel and the micro switch.

Disassembly of the DC motor.

The commutator is a bit in bad shape, and some tin had come out of the solder joints!
The tin was melted and stuck to the inner walls of the engine casing.
I took the photo after I had restored the coils, so I don't have a photo from before.

Conditioning of the commutator: sanding and removing carbon from it.

Wiring all components.

The cables leading to the head must be routed as shown in the picture.

The support of this spring was damaged: I fixed it and reinforced the area.

Detail of a photodetector, which controls the rotation of one of the DC motors.

Ready to close.

Assembly completed.

Now it's time to mount the robot arm on its support, after preparing it...

This microswitch, which is responsible for detecting the "parking position" of the robot arm, is damaged.

On the right the new microswitch: Cherry DB1

I hold the support upside down on the workbench, so that the robot arm can be inserted from underneath, screwed to the central axis, in order to be able to solder all the cables.

Cables routed through and central shaft screwed without tightening.

Before tightening the nut, it is advisable to loosen the main motor holder to move it sideways. Once the large nut has been tightened, the pinion gear can be brought closer to the crown wheel by adjusting the clearance.

Work completed: gears adjusted and all cables soldered.

At this point, I connected the robot arm to the motherboard and everything started to work, but poorly: The robot arm was not quite aligned, the sound was distorted and very loud, the pieces were not detected on the board ("Piece Discrepancy" LED always on), etc.

The first thing I did was to adjust the new microswitch, and then some other adjustments that you can see below...

By adjusting VR1 we can vary the Gain of the audio amplifier.

VR3 is used to adjust the sensitivity of the sensory board. After several adjustments, I discovered that the potentiometer was faulty.

Installation of a new 10Kohm trimming potentiometer.
I got the best sensitivity on the board by setting it to 3.8V.

Not being able to find a NiCd battery like the old one (left), I decide to buy 2 x 1.2V cells (right) and build it myself.

New battery installed on the motherboard. It is in charge of maintaining the memory.

Testing ways to better cool the power transistor of the power supply, as I feel the original heatsink is insufficient.

Everything is now working fine. All that remains now is to install the sensory board...

Some squares on the sensory board sometimes fail, in particular those in columns A and B. That is, sometimes the computer does not detect that a piece has been positioned on top of the square. Inside this board there is a membrane similar to those of the Novag pressure computers, with tracks made of conductive paint. When a piece is placed on top of the square, the magnet at its base attracts another magnet underneath the board. Both magnets join the contacts of the membrane, which is sandwiched between the magnets.

It is possible that this membrane has suffered the same degradation (oxidation?) found in many Novag computers (with pressure sensors) of the time.

And finally I complete the assembly...

PHOTOS AND VIDEO: