As I've mentioned in some other posts, I'm helping a high school engineering class with starting a group dedicated to designing and building pinball games. We're starting with a couple of rebuilds to show the kids the tech side of it.
With a small group over the summer, I'm rebuilding a Fireball home edition. Picked it up cut down to be a coffee table. No back box. At least the playfield wasn't butchered.
Anyway, the controller is being replaced by an Arduino mega feeding events to a Raspberry Pi running MPF.
First order of business: replacing a burnt flipper solenoid and resoldering some loose switches. We thought about ditching those big boards and mounting switches directly to the underside of the playfield, but it seems like overkill for this purpose.
Pro-tip: When soldering on the kitchen table, always cover the table with something to avoid spousal displeasure in the morning

20170712_234417 (resized).jpg
20170712_234508 (resized).jpg
20170718_000623 (resized).jpg
20170718_000615 (resized).jpg

Yeah, I've seen his series of videos and read the blog. I'm going for something similar, with the caveat that my goal is an easily modified Arduino control system that the engineering students can implement on other games.
The second machine is a Zizzle PotC.
This is the Arduino based controller I'm using for this project:
ebay.com link: New Small Meduino Mega2560 R3 Pro Mini ATMEGA16U2 Female header for Arduino

It's more compact and, IMO, easier to work with(it doesn't have the weird pin spacing that makes a regular Mega such a pain).

The playfield has some damage from the screws. Since it's not actually plywood(just what looks like MDF with some thin veneers on the top and bottom), the screws created some compression damage. Also, the inserts aren't perfectly flush. I suspect that they never were. I'm going to grind out the damage with a dremel and fill it with epoxy putty. Still debating about a vinyl cover for the playfield, and then using a playfield protector rather than trying to clearcoat it.

In addition, I discovered that the lights are tied up to the switches directly. Since I'm switching to Neopixels, I'm just going to cut out most of the traces on the main boards and wire the switches up directly to a matrix.

A small update.
Got the final part to fix the flipper(replaced the plunger and crank assembly). It's now good to go.
Dropped the idea of a switch matrix. They always bugged me as being prone to error. It occurred to me that I had already solved this problem on an earlier project(a Z80 shield for an Arduino Mega). I used shift registers to read and write to the Z80 data bus. But there is no reason they can't be reading switches. So, last night, I acquired some SRs and started wire wrapping a board. I should be at a testable state for the registers tonight.
Didn't think about taking a pic, or I would show off my oddball way of prototyping boards. I'll take a couple tonight.

It's my preferred way to prototype boards. This is an early picture of my old Z80 board.

FB_IMG_1501008290194 (resized).jpg

Switched gears a bit. Moved from wire wrapping to a regular protoboard after my new wire wrapper kept breaking wires. This will be cleaner anyway.
One step closer to a switch reading board...

20170727_012535 (resized).jpg

Well, CAX slowed me down, but I can't bitch too much, since I also picked up a couple of populated playfields for (near) future projects(Williams "Flash" and Bally "Space Invaders").
I have also settled on a shift register approach to the solenoid driver board. Makes life simpler.

Built the first of four interface boards for switches. Worked on the shift register board some more. I'll be testing the shift register code tomorrow evening.

20170802_010222 (resized).jpg

And I'm rethinking the idea of shift registers for the solenoid board. While it makes the main software loop easier to grok, it would preclude easily adding PWM to allow for user settable power levels on the flippers. For a simple machine like this(or the Zizzle), it likely doesn't matter. But for something more complex, it could be an impediment. Also, I would have to have exceptions(like servos and steppers for playfield and backbox toys) to the driver config, so it doesn't make sense to lock it in.
Any thoughts on my rambling notions are welcome

The shift register board worked the first time.
"I love it when a plan comes together!"
Not the prettiest wiring, but it works for a quick sanity check.
So close to reading the actual pinball switches that I can taste it

20170803_005454 (resized).jpg

Finished the switch boards and started mounting them on the old PCBs. Cut the traces to isolate the switches from the lights.

20170806_232957 (resized).jpg

20170806_232900 (resized).jpg

20170806_232912 (resized).jpg

So, some explanation of the last pics.
The Fireball Home machine ties lights directly to switch activation. So, there is no way to independently activate lights on this machine. I used a dremel and cut the traces, isolating the switches. I have noticed that the switches are not always responsive. I may end up drilling out the rivets to disassemble and clean them, since finding this particular type of contact switch is a bit hard(unless someone is parting one of these out anyway....).
I mounted the interface boards a bit high, but it was the place that was least likely to damage other things on the PCB. I ran out of standoffs, so I opted to stop measuring and soldering until I had more.
The little red board is a MOSFET driver. There will be one per solenoid, mounted close to each one.
I've opted for two separate power supplies, a 24V for the solenoids(this was originally a 22V system), and a standard ATX PC supply for the 5V that will power the Arduino, the switch boards and the Raspberry Pi. I'm looking for one of those power strips that have a switch separate from the body, so that I can just plug them in and mount the switch to be accessible from the back of the machine.
If nothing goes horrible, I should be able to flip on this beast tomorrow night.

Well, nothing horrible happened except that I opted to sleep last night instead of work on this.
Anyway, got the big switch board assembled, with the interface boards wired on. These will handle half of the switches. I need mount and connect the second board now.Also got the little power distribution board done.

20170809_002422 (resized).jpg

Took some time from circuit design.
I got my Neopixel clones in the mail today, and used OpenSCAD to whip up a widget to attach them to the playfield in place of lamps. The first pass was a little too tight to fit right. I'll fix it and reprint tomorrow. Gotta love having a 3D printer for parts like this

20170815_232101 (resized).jpg

20170815_232031 (resized).jpg

20170815_233144 (resized).jpg

In case anyone was interested, this is the rendering from OpenSCAD. I like designing objects programmatically instead of visually, so it works well for me. Once I have it refined a bit, I'll post the SCAD code and a link to an .stl file, in case someone else wants to play with it.

led_holderv0.5 (resized).png

That would be awesome! I'll ping you after work! Thanks!

I'm just a parent who went nuts when the teacher asked for help.
But, I hear ya. I wish I had had something like this when I was in school. Or college, for that matter.

Thanks!

Makes total sense. I'm thinking of making the Neopixel hole a cone shape, so that I'm not blocking light on the sides. I'll play with it some more tonight. Thanks for the suggestion!

Very cool! Those LEDs are pretty much the same as I am using. I'd love to check out the STL.

Cleaning up the playfield, removing extraneous wiring and getting the switches mounted on the cabinet connected to the switch boards. I was angry I couldn't find spacers anywhere, so I just used some 3mm bolts as spacers.
So, TODO to wrap this up enough for the kids:
1) Mount MOSFET boards for flippers
2) Mount Arduino and the shift register boards to the playfield. Run wiring between them and the switch boards.
3) Finish Arduino code, and get it talking to the Raspberry PI(basic ruleset)

After that, I'll pass it back to the class, and help get that project kicked off again.

playfield_cleanup (resized).jpg

While waiting to experiment with DDDwingmaster's holder, I went searching for non printed devices, after I was inspired by his design to rethink my design. I ended up with a bunch of possible solutions, with the best being....plastic covered jumbo paperclips. The LED clips right in, and the covering prevents shorts.
It seems to work well, and hold the LED just at the edge of the hole, making for a nice even light spread with no spotlight effect:

20170820_234713 (resized).jpg

Finally had a chance to start wiring the LEDs. Wired three, then hooked them onto the top of the playfield to run the Adafruit strandtest example. Works great.

20170823_203350 (resized).jpg

No pictures today, just text.
I've decided not to finish wiring the full LED set. The kids can finish that.
I reattached the old PCBs, and all of the new boards. I tested a new 24V PSU. Worked like a charm, and it was nice to watch a pop bumper firing. I didn't want to cut up the larger Antec PSU I'm using for 12V and 5V. Instead, I cut up some of 4 pin Molex extension cords and made power adapters for the LED strand and the Arudino Mega. Since the new Raspberry Pi only uses a micro-USB for power, I need to dig up a cord to butcher to splice into the 5V.
To get power to all of this(and to a hypothetical VGA monitor to act as an LCD), I'm using a 4 outlet power strip with an external button, which makes it easy to power the whole thing on at once, with a single button.
I dug up an old Arduino Pro micro. I'm considering it for running the lights. That leaves the Mega with 3 functions:
1) Shift Register reading
2) Solenoid firing
3) Serial communication

The Pro Micro will deal with:
1) NeoPixel lighting
2) Serial Communication

Really, if I don't see any serious lag with the shift registers once the game is looping in real time, I could downsize the Mega to a smaller device.

Both will be present as a new platform under MPF.

Now, it's mostly down to finishing up the Arduino code. And documentation. Lots of documentation. *sigh*

Worked more on the wiring.
All the solenoids are connected. All the switches are wired in. Discovered that the tilt bob doesn't have great connectivity. I might need to polish it or solder a wire directly to the arm.
Flipped the playfield over and ran my LED test. Looks better.

Almost ready for the kids.

20170831_002103 (resized).jpg

That is the truth. It gets messy fast. I'm still working on a couple of routing issues.

Real life intruded on my project. Mostly buying a new car because my son got into a fender bender and took out my little Toyota. Alas...
So, my earlier choices have caused an issue. Driving the solenoids from individual MOSFET boards is great from a conceptual point of view(I think it's easier to understand from the kid's point of view), but sucks from a wiring point of view, because now I have 21 wires to run instead of 14. I built some secondary power supply boards to minimize the length of the wire runs as much as possible, but it is still messy. Lesson learned. If I go this route again, I'll build a single driver board and just run a single line to the solenoids.
So, I'm down to the solenoid power and signals. I also sketched out a design for a parallel shift register controller. Not sure if it is necessary yet, but I've got it in my back pocket as a possible method of speeding up the main loop.

I attached a pic of the driver boards I used. In addition to the main power, which I did daisy chain, each one has a +5, signal and ground for the low power side. So, really, each solenoid requires 5 inputs(+24v, grd x 2, and +5v and signal). If I had instead built a single board, all of the switching would have happened on the board, only requiring +24v and grd for each. Bad design on my part, trying to keep the component as discrete as possible for easy instruction.
In retrospect, I should have ditched the original PCBs, direct mounted switches and done a centralized driver board. Live and learn.

150924085912258608167 (resized).jpg

Well, a short somewhere in the 24V section fried a MOSFET board. I discovered this after trying to test a single solenoid, getting one bump and not realizing until the magic smoke was coming out that the coil was still firing.
Not a big deal in the grand scheme of things, but it is disheartening. I'm contemplating ripping the thing apart, throwing away the PCBs, mounting switches directly to the playfield and wiring it the way I'm already using for my next project. I think I'll sleep on it before doing anything rash.