First Decisions - The Philosophy of the Design

The first decisions were about the playfield, the mechanisms, cabinet and computers.

Decisions

Unique Playfield Design

It was never going to be easy to showcase resin-based art by using an existing playfield design. It would also easily allow unique mechanisms to be incorporated. These two points set us in the direction of designing a playfield from scratch.

This did not make anything easy¹

Pinball Hardware and Mechanisms

We needed some equipment so I scoured the Internet for opportunities. I missed lots due to caution and thrift.

People were buying anything for top dollar when I started looking.

I eventually purchased a 1980’s Gottlieb pinball machine. Both playfield and cabinet riddled with wood rot and borers.

It had to be wrapped in plastic for transport because it was crumbling apart.

The playfield had no upper edges, all the way to the insets.

Some mechanisms fell off the playfield, followed by a few more.

The adhesive that held the laminations together had been washed out of the the wood, and had set on the mechanisms.

The backbox fell off, and the cabinet sides fell in.

It was perfect.

Cabinet

I purchased a nice Gottlieb cabinet from a man down the road, he was kind enough to deliver it to the door.

Playfield

First pass is built on 12mm hardware shop marine ply. It’s good enough to do the job, and when done, will showcase what an inexperienced man can do to an innocent piece of wood.

Computer and Microcontroller

Computers because media, even simple sound and images, are easily managed with computers. We want to make a pinball machine, not invent a media system.

Microcontrollers, because pinball machines are fast, and computers doing multimedia, are not so much.

Computer

A reconditioned Wyse 3040 Thin Client is the plan. Cheap and effective with a shiny new Linux operating system on it. Since I bought this piece of equipment, the screen on my laptop failed, so that looks like a replacement for the 3040 :-)

Software

I’m in two minds over the high level control system².

I was going to write my own from scratch, but the Mission Pinball Framework (MPF) offers a lot, and promises integration with everything.

In practice it’s a royal pain because of outdated documentation, regression bugs, and over-promising.

I’m advised that for a simple interface to another I/O platform (a thing the documentation says I’ll get help with) I will have to write myself because the system is ‘mature’ and nobody is working on interfaces.

I want to make a pinball machine, not a software interface.

Adrian⁴ from Haggis told me they were still useing MPF, so I will persevere till I reach a dead end.

Microcontrollers for MPF

The computer system needs input and output electronics to talk to the electrical equipment, lights, solenoids, motors, etc.

Selecting MPF means in practice there is limited choice in the style of electronics. You need to pick something MPF already works well with, unless you love writing code.

After a lengthy evaluation I selected the Open Pinball Project (OPP) for the microcontroller, base software, and electronic design.

It’s:

• Mature and offers tight integration with MPF.
• Has potential to seriously reduce cost.

For low level control I’m using discrete MOSFET modules directly attached to coils, leading directly to microcontrollers mounted under playfield.

I’ve programmed the microcontrollers in a similar way as Cobrapin - so I have an easy upgrade path if my cost reduction plans also reduce reliability.

Changing to Fast with minor changes is also an option.

I’m keeping the connector count low, using headers as sockets for microcontrollers, and using generic circuit boards where possible.

Microcontrollers for Special Equipment

We will make a few special mechanisms, and having them operate stand-alone seems wise as:

• MPF is not going to integrate with them at anything smarter than turning wires on and off.
• I might walk away from MPF (or be driven off) for a bunch of reasons.
• It’s good for testing, as a bunch of LEDs and switches can stand in for a fancy mechanism.
• I can test the mechanisms without the pain of tricking MPF into talking to them first.

For complex systems, like the power control and watchdog, The Raspberry Pi Foundation’s 2040 microcontroller is great.

The ease of using the second core overcomes a lot of potential timing problems. It’s also very economical and it has a high clock rate for processing.

For stepper-motor devices, an Arduino UNO clone and CNC board are a good choice. Very flexible and reliable hardware, economical, and great community support.

MPF will talk to steppers, but a visit to the Internet shows just how wise that isn’t. Once I add the complexity of limit switches and calibration it’s very wise for the smarts to be in the device and well hidden from MPF.

OPP Power Electronics

This is a potential gold-mine of savings. Control systems for pinball machines are expensive.

Noname electronics from China, not so much.

If you are not fussy, and don’t want a control system that looks like a control system, then there are options.

I considered lots of FET control modules, settling on FR120Ns on a little circuit board with opto isolation and some minimal circuitry.

Oddly enough, it does not seem to matter what I order from China, I get sent counterfeit electronics.

The only way to be certain that something will work is to test, test, test. And test at extremes.

Not what I wanted, and an argument for purchasing purpose-built-for-MPF boards. As luck has it, one of my orders was filled by a product that exceeds all the ratings of the FR120N, so I’ll be moving ahead with them.

Warning! Testing electronics to failure can lead to fire and poison smoke. Use care!

Because the chance of MOSFET failure is quite high when using counterfeit components, it is important to consider how you will protect your machine (and home) from explosions, fires, nasty smells, etc.

In my case I was intending to create a supervisory and control system to manage power, watchdogs, etc. I brought the development of this system forward to deal with this issue, and some unreliability³ of the MPF and OPP combination.

¹ We knew it would make it hard, but such is life.
² I still am.
³ Unreliable in that OPP coils enabled directly by MPF won’t disable if MPF crashes, but will, if their current is sufficiently high, sit there and burn away. You can’t rely on OPP disabling coils if MPF fails. Autofire coils are shut down, but explicitly enabled coils are not. This may feel like a bug, but it is part of the design of OPP.
⁴ This was from before the unpleasantness started.