Well, that explains a lot of things.

The only barriers to west coast shows are cost of travel and how-the-heck-do-I-drag-a-pin-that-far. I'm not a rich guy, so the closer to one show per year I get, the better off I am.

I didn't even think of that, but yes - that would probably work!

Well, nobody will ever be complaining about the ramp switch again, and it's thanks to a new part that is now available here!

https://pinside.com/pinball/forum/topic/mrs-for-ramps-a-universal-solution-to-plastic-ramp-microswitches

The center ramp switch was just BAD. No matter how I adjusted it, it would either trap a slow ball or miss a fast one.

IMG_20200216_154045343 (resized).jpg

After hearing about M&M Creations' new switch, I showed some interest in trying it out, so I bought one. It showed up yesterday, and I got my chance to install it this afternoon. The switch includes a diode built in, and pre-tinned wires in green and white to match the traditional WMS/DE colors. Of course, this didn't actually match the pre-redone ramp wiring, but apparently I at least had the foresight to match green to green.

One thing I found about the new magnetic reed switch is that it's very, well, magnetic. As a direct replacement for the microswitch, it had enough power to grab and hold a slow-moving ball. This is probably a non-issue for the games it is designed for, and it's flexible enough to address this issue too. Since it is self-adhesive, it can be mounted anywhere - so I just placed it in the bend, at which the ball is already going to be fast enough to get by:

IMG_20200216_154325828 (resized).jpg

The result? I love it. It doesn't trap the ball at any speed (once it has been positioned appropriately), it always detects it passing by, and the detection range isn't too wide that I get false positives from anything nearby. The ramp walls are stainless steel, and thus non-magnetic and don't affect it. I wouldn't fill a whole game with these $30 switches, but for individual applications, it's perfect.

Disclosure: I bought the switch myself and was not paid to write this review. I am just really happy with it.

Necro'd.
Back from the dead.
Reports of this thread's demise were greatly exaggerated.
"But it refused."
OMORI did not succumb.

Haha, it's cute to look back at the end of the third page, all hopeful for bringing this game to MGC for 2020 with new code.... ha... haha... yeah...

BUT... after two years of convention withdrawal symptoms, I'm finally preparing to go back to one of my favorite places, and my famous build needed some refurbishment before the big weekend. With the game sitting at the Pinball Collective surrounded by useful tools, Jesse and I pulled the playfield. I don't own a rotisserie, but I do own sawhorses and an old lockdown bar receiver - and that was good enough for us.
IMG_20211023_132522984 (resized).jpg

My rewire of the game as detailed on page 2 included a whole bunch of Molex connectors, and this was a lifesaver as I disassembled the top quarter of the playfield. Subassemblies came off with just a few screws and a pull of a plug.
IMG_20211023_132537517 (resized).jpg

The exception was the center ramp, which shared a wire harness with the Librarby award light. Not a huge setback, just had to remove that socket to get the ramp off.

With the upper playfield finally exposed again, I could focus on the things that needed fixing. "If you can read this, the game isn't fully assembled yet "
IMG_20211023_132533117 (resized).jpg

What needed fixing, you might ask? Awful shot rejection! The Resort VUK and the Skeleton House saucer were particularly bad offenders of this. I considered various methods of addressing the design flaws, and managed to come up with some solutions. The VUK took some experimenting, but I made it work. I curved the lower guide wall up to reduce the area the ball can bounce, then added a one-way gate (tumble-polished to perfection) just off to the side. Between the bounce off the wireform, the bounce back from the moved lower wall, and finally the gate, Mettaton's resort rejects a lot less shots. I tried a bunch manually and got almost every shot.
IMG_20211023_135308660 (resized).jpg

On the other side, the saucer needed work one way or another. There was a little wireform guide that was literally just a bent paper clip, which I immediately discarded as it kept falling out and blocking the ball. I noticed that the shot often bounced back and diagonally to the left, so I added a mini-post. To help deflect the ball down, we added a rubber post sleeve at the mount screw - invisible from normal play - and it seemed to work extremely well. It still rejects occasionally, as does the resort, but both shots are quite a bit better now.
IMG_20211023_140951494 (resized).jpg

Once the tougher maintenance was done, we dropped the playfield back into the game, hooked all the wires back up, and gave it a quick test. To my great relief, nothing broke or failed or fried as a result of pulling the playfield.
IMG_20211023_155348424 (resized).jpg

The bottom two flippers are getting rebuild kits, after EOS troubles on the left and a broken link bolt on the right, and all four flippers are being replaced with black bats with green rubber. It follows the battle grid color scheme and I think it will come out looking nicely. A once-over with some Novus 2 and a fresh batch of color-matched LEDs for the inserts should really bring the game up to the visual standards I hoped for.

Past visitors who have played the game will probably be happy to hear it's gotten a code update as well. Besides the graphical improvements promised for the 2020 show, the new code includes new music cues, new random encounter multiballs, and some bug fixes.

The last two improvements are a new speaker panel - now that I can make them look nice - and a new lamp driver board. The game's lamp matrix has been annoyingly dim since the electrical redesign, and I only recently realized the reason why. The row drivers are switched with some optoisolator chips I had laying around. I didn't think much of this until I looked at the specs for those chips, and discovered that they are far, FAR underspecced for driving a row of LEDs. Mouser order is incoming for eight proper drivers and the perfboard to mount them to. The backbox circuitry is gradually becoming a mess, but it's a mess that works.

Day two of refurb. Playfield surface was pretty scratched and dull. Fortunately, what better thing is there for polishing plastic than Novus Plastic Polish?
IMG_20211024_105722055 (resized).jpg

As I hoped, it came out amazingly shiny once again, and played massively fast once everything was put together. And I'm loving how the new color scheme is looking.
IMG_20211024_114639752 (resized).jpg

I took a couple of hours playtesting it to see what could be fixed. There were several things I had to throw onto the list of needed improvements - not the least of which is add some stucky-protection under the ramp.
IMG_20211024_133920499 (resized).jpg

Also, the new code includes a new top-secret music replacement easter egg. Massive internet points for you if you recognize this...
IMG_20211024_131458972 (resized).jpg

I fitted the slingshots with slightly tighter rubber (2-1/4" rather than 2-1/2") and will actually need to crank up the power level on them now, as well as the skeleton house saucer. The waterfall scoop, on the other hand, could probably come down in power, as it has a pretty forceful kick that makes it hard to take control from the flipper it feeds.

The parts are ordered for my updated lamp driver, and a quick update to the firmware code will hopefully fix the significant ghosting issues.

Nope, sorry to inform - money's a little tighter this year and MGC wins on proximity...-ish.
...I don't suppose you're going to MGC? If you are, I'd love to give Sonic a flip!

Oh, definitely. I could imagine this remix further remixed with the synthesized electric guitar track that Williams used until DCS.

The real mind-blower: Despite how famous it is, and how much the song is considered synonymous with Undertale... the song Megalovania is not present in this pinball game. So far, no one has noticed.

Today's "lunch" break.
IMG_20211028_132956044 (resized).jpg

A mostly drop-in replacement for the lamp matrix row drivers, now actually capable of driving the necessary current. Hoping it works, as it was arranged by studying other circuits and not actually by testing on the solderless breadboard...

Front header comes off the lamp board via a new connector I made up; rear header should be plug-and-play with the existing connector that runs to the lamps. 18V is applied via the center traces, and soldering a power wire to the mount on the left should do the trick to get it there. The controller board that actually runs the matrix has been updated with a ghosting fix so obvious I should have seen it long ago - previous columns were not being turned off as the next row is turned on. Whoops.

Nope, no Expo - although at this point I think we all realize that. I did have a secondary excuse for not going - company-wide inventory day is today at work, and I'm the helpdesk for all six branches statewide.

But let's see what I can get done in the week until MGC. New speaker panel build is in progress; now that I know how to build a decently nice one, I expect UT's backbox to look slightly less homemade. Slightly.
IMG_20211028_222038447 (resized).jpg

Up next will be filling, sanding, cutting out a panel of perforated sheet metal for the speaker grille, then painting everything.

Filling and grille are now complete, with the grill cut to fit the flat face and ten screw holes drilled into it. I filled the corner joint with fiberglass filler - same as I did for the Volcano Blast cabinet - and tomorrow I'll be sanding it smooth and applying the primer.

In other news, I added up all the stats files I had, dating up to the 24th of this month from sometime in February or March of 2019. This machine has survived 1,761 documented plays now!

New speaker panel got the first coats of paint. I really liked how the speaker panel for VB turned out, so I went with the same process on this one.
First of two coats of gloss black on the wood panel.
IMG_20211030_200257084 (resized).jpg

Single coat of flat black on the metal grille.
IMG_20211030_200336068 (resized).jpg

More software and firmware to deal with. Found a couple bugs while play testing, some coil power levels need adjusting, etc.
And now I have a rush order coming in for a new flipper coil, because guess who MELTED ONE today. Of course that couldn't have happened last week; it has to be six days before the show.

We do not speak of the lamp driver.

That's not a bad idea at all, and I'm definitely starting to understand why he does that. Might order a batch of those for the future.

Was fairly busy today but did run some new code down to the shop to fix the software/firmware issues I ran into during last test run. Speaker panel got its second coat, probably will get the whole thing assembled tomorrow and then I just need to scrounge up a decent speaker amp board.

Also, there's a new music replacement easter egg. Hold right flipper as you start the game.
IMG_20211030_141636714 (resized).jpg

If you get this reference, I love you.

Not quite installed yet - struggling to make time around work - but new speaker panel is almost ready to install. Just need the wires and a bit of pre-drilling on the mounting blocks.
The old one:
IMG_20211023_163200536 (resized).jpg

The new one:
IMG_20211102_073419694 (resized).jpg

I think the new one is just a tad bit nicer than the old - and with the speakers nearly twice as big, I'm fairly certain this game will finally be audible at the show. This panel even has the backglass lock properly centered!

Thanks so much to all who voted!
I'm exhausted tonight but I'll be back to this thread later to share stats and more.

IMG_20211107_102757762.jpg

Stats and more, I had said. So here you go! Finally got the game unfolded today and took the opportunity to grab the stats file - in addition to finally locating the wireless mouse that had gone missing. (The mouse left the hole and got the cheese; I am filled with determination.)

The game took on 240 plays over MGC weekend, which is down a bit from previous shows - but to be expected when it was down for several hours. Oh yeah, more on that later for those who missed the fun... The top score of 1,508,810 went to the skillful AAAAAAAAABCEH, who played for nearly 14 minutes and filled up as much of the boss-fight heart as anyone could this time around, in addition to a multiball.

Toriel's mode continues to dominate in points awards, averaging just under 77,000 points each time with the best target basher milking it for nearly 200K. The Snowdin saucer continues to disappoint me, taking a total of 64 successful hits and trapping balls an unfortunate number of times. Ball traps were still a major issue here, with one unlucky match taking 11 ball searches. The problems here are 100% top-right; balls can get stuck in the area around the Snowdin saucer in the event of a bad eject bounce, however several balls also got stuck underneath the center ramp, wedged between that and the Lesser Dog plastic near the bumper. I had to free up several stuck balls from this area, and I think the only solution is going to be a very large airball shield to close off the space.

Nobody managed to face Flowey or even Asgore at the show. Additional game rules like Random Encounter Multiball may have played a part in this; in general I think this stat is okay as it's called a wizard mode for a reason. Once again, the game shows a slight bias towards the right-hand side, and whether this is due to its design or if the game's leveling was skewed right isn't entirely clear. I did adjust the levelers some way into the show, which I suppose could have made up the difference. 167 left drains vs 172 right, and 15,167 left flips vs 16,049 right.

While my "improved" saucer's 64 hits were disappointing, some shots definitely took a pounding - with the most commonly hit target actually being the Grillby scoop this time around at 869 hits - and the center ramp a distant second place at 495 hits. I believe this is due to the flipper configuration having changed since the first two shows; following installation of a rebuild kit, they aim a bit lower now, with center ramp shots being slightly less likely. Half as many shots again (222) went to the Lab saucer - which, again, was a poor design that I couldn't do much to fix. One thing that did make me happy, though, is seeing exactly the same number of successful shots to both the Hotland passageway and the Resort kicker. While the Resort probably did reject a fair number of shots, it was hit more times than it was at the game's debut show, despite the game being played less times in total.

Scores averaged just over 190K, making me think the replay score ought to come down a bit or I could look into auto-percentaging. In total, the game took on slightly more than 13 hours of active play time.

So, you might be wondering what was with the downtime I mentioned earlier. What happened there? Well, obviously not flipper problems, as everything was working great as I packed up the game, and I'd been given such good advice about adding fuses.

I mean, other than how I got busy with other things at the show and never actually took the opportunity to install those fuses.

So I get a text from Josh, a relatively local Pinsider, telling me my left flipper was weak. (Thanks, Josh!) Of course, I was on my way out of the Wisconsin Center for Saturday dinner at that time, so I did what any sensible person would do and said screw dinner and went right back in to do service work. I mean, how hard can it be to fix a weak flipper, right?

Turns out, it was a massive struggle. After multiple attempts at EOS adjustment came up fruitless, I chopped out the switch and picked up a new one from Rob Anthony's table. There, surely that would solve the problem. I buttoned up the game, left for dinner, and came back to find the flipper limp again.

At this point, Matt Benzik - builder of "Motor City Taxi" and generally great guy - paid me a visit, ending up helping me troubleshoot flipper problems - for at least an hour. We tested the drive circuitry, voltage, grounding, and the coil. When the coil itself left us suspicious, we replaced it with one of Rob's, during which I dropped my soldering iron and burned my finger. Even then, troubles - although jumpering across the coil to bypass EOS proved the switch was, in some way, the culprit. It finally took Rob doing some solder-based bulletproofing on the EOS switch itself before the game was finally fixed, and with the flipper back to full power, I closed up the game for the last time that Saturday, confident my troubles were over.

Sunday I came down to the arcade to find the game shut off and unplugged.

And when I found the troublesome left flipper now firmly bound in one place, I had only myself to blame. Not enough time spent on EOS adjustment led to, at some point, the switch failing to open, burning up my second flipper coil in a week. Old-style flipper EOS setups just aren't worth it, guys. Do it right. Let a computer take control of those coils.

But setbacks or no, it was generally another great run for Undertale, with several attendees actually reporting they'd come to this show SPECIFICALLY to play this game - which made me feel warm and fuzzy. Marco TV put the homebrewers in the spotlight with on-stage interviews, with the last one going to Undertale - fortunately moved to Sunday after I'd fixed all the flipper problems. It showed up in one of their recorded streams at the 2:04:00-or-so mark.

Well, it's just over three months until Midwest Gaming Classic, and if you've been reading up on the history of this game on this thread, you know what that means - it's time to hurriedly take it apart and give it some sort of desperately-needed overhaul.

Today's topic is the upper shots. Again. I made a decent attempt at this in October, but the fixes I came up with weren't quite as good as I'd hoped - and, in the case of the skeleton saucer, apparently made it worse. Eleven ball searches... I'm sorry, whoever you are.

What I ended up doing for the resort and lab shots was a bit of a re-engineering to work around the fixtures I can't remove. I'll post the design later, but I came up with an Inkscape layout that manages the ball's speed instead of letting it bounce straight back or around corners. From there it was a simple process of sending the file to Pololu and ordering a bag of spacers from Pinball Life.

Time to disassemble the playfield once again - and this time, without a two-week deadline looming over my head, I hope to fix a bunch of oversights from the original design.
IMG_20220121_204051189 (resized).jpg

- Topside wires are all thin and solid. This will be replaced with stranded wire to avoid breakage - and in the correct colors, as well.
- Left ramp components are crooked. I will need to make some minor changes to the wireforms to correct this. Once done, the left return wireform's mounting loop shouldn't dig into the side of the cabinet anymore.
- All three major shots at the back will be redesigned to absorb ball momentum instead of bouncing it back.
- I may move the Grillby and Librarby lights to look more professional, instead of having prominently visible lamp sockets.
- The wiring for the ramp will get moved and rerouted so the ramp can be removed more easily.
- Further addressing the pop bumper - which is almost perfect but can still get a ball stuck on top once in a while.
- I am going to try re-orienting the skeleton saucer so it kicks down instead of into the bumper; I'll also experiment with removing the one-way gate to add some randomness to the ball's return to the open field.
- The plunger lane will get an airball shield, finally, after years of the ball getting stuck under the ramp.

Removing the playfield finally gives a free look at all the little details that people usually miss. Like Lesser Dog.
IMG_20220121_205017434 (resized).jpg

And it also reveals things that make me question my own sanity. Like adding a free-hanging GI socket....directly over a normal one. WHY??
IMG_20220121_205402251 (resized).jpg

Nexy inspected the topside GI wiring and agreed we need to redo it.
IMG_20220121_213707371 (resized).jpg

Pulled a saucer back-stop from the parts bin - tumbled to a shine a few months ago and fitted with a brand-new rubber stopper. It took some bending to make this work where I needed it, but it will get the job done.
IMG_20220121_220950751 (resized).jpg

As usual, sitting down and working is an open invitation for Clark to make a bed out of my lap. He also took the time to inspect my work.
IMG_20220121_221015473 (resized).jpg

Thanks, I'm glad you like these updates!

I do actually have one of these - Mettaton, whose single wheel is a black standup target - which itself is due for a refresh because the sign shop originally printed it on white instead of clear. If I end up doing anything like this for the scoops, most likely I'll make text-based signs directing the player to those shots.

Today's free time fixed a few more of the issues. One of them required modification to some wireforms.
IMG_20220122_134813883 (resized).jpg

Normally I would question removing supports from a ramp, but the left return wireform seems pretty well built and is mounted to very solid posts. It had three supports. The thing is... the middle support was too tall. This adversely affected the position of the ramp, both dropping the ball from unnecessarily high into the inlane, and causing the upper portion to hang too far to the left. I'll post photos of the finished left ramp assembly later, but just hacksawing off this single support really improved it a lot.

Speaking of improving things a lot... I have won at impromptu saucer repair.
IMG_20220122_203059691 (resized).jpg

Obviously this will require further testing with the game physically running, but the updated skeleton saucer feels like a HUGE improvement over everything that came before it. Between a moved rubber post, new backstop set into the playfield backboard for extra space, tweaked position of metal wall, and the playfield overlay cut away behind the saucer, pretty much everything here has been modified and now genuinely feels like it won't reject any more shots.

As far as the kickout itself, I went ahead and re-oriented it to kick straight down, and based on testing the mechanism with my finger, the trajectory is PERFECT. It slightly bounces off the rubber next to Lesser Dog, passes through the gap where the one-way Snowdin gate used to be (already removed), and heads for the bottom flippers, passing just close enough to the upper right flipper that you can send it flying towards Hotland if you time it right.

Thanks, glad you've enjoyed checking it out!

The overlay is 1/16" polycarbonate with a vinyl backing sticker containing the actual art. Below (technically "above" in the order it was printed) the artwork is a black masking layer that keeps the "inserts" only lit within their boundaries. Performance is....so-so, to be totally honest. There's definitely some noticeable warping when you look at any reflection in the plastic, though not enough to affect the movement of the ball. The playfield's a tiny bit thicker than standard, which played absolute havoc trying to tweak the jet bumper to respond properly (and it's still not perfect).

Given a chance to do it again, I would absolutely spend the extra few hundred to contract the job out to a company that knows how to produce a real playfield. Which, in fact, is what I am already planning on doing for Volcano Blast.

A couple orders arrived in the mail today, so I'm hoping to get most of the playfield improvements done fairly soon.

Midwest Gaming Classic has been the tradition since the first year this game was playable. It's also been to Expo once (and returning to that show isn't entirely out of the question). Not sure about going any further out than that as I'm kinda tucked away in the northern woods and living under Mt. Ebott.

Today's update notes will follow each photo.

IMG_20220128_224230775 (resized).jpg

I think the shot rethinking is complete. Besides the right saucer's fixes posted above, the other two problematic shots have also been made about as good as they're going to get. Pictured is a momentum-killing loop right before the VUK, which slows the ball down considerably on entry, and a fairly standard damping setup for the Lab saucer, which we as an industry have already understood for decades... But anyway!

IMG_20220129_221737404 (resized).jpg

My babies came to inspect my work, and decided I needed to be done for the night, enforcing this by using my lap as their cuddle platform.

IMG_20220130_095738860 (resized).jpg

So... wiring. I did a whole bunch of rewiring prior to the debut show, but it actually wasn't everything. The stuff on the topside of the playfield was left as it was - solid core wire in an old set of colors that I wasn't using anymore. I didn't even realize, until I'd taken it apart and really took a good look at things, that it was such a mess. Lots of kinks in the wires and a huge lack of zip ties. And yes, the solid core came back to bite me, as the waterfall flasher (blue) sporadically quit working prior to MGC 2021, the fault of an internally broken wire.

IMG_20220130_095953049 (resized).jpg

I always love getting new tools that turn out to be amazing, and a Molex Mini-Fit Jr pin removal tool did not disappoint. I freed up several used connectors the day I got it - which included the liberation of a 20-pin motherboard connector from a long since discarded power supply. In the case of the UT refresh, it was a big help saving me on plugs as I redid the topside wires. It was also nice to have insurance against accidentally pinning a connector incorrectly, as it's an instant undo where once I would've had to discard the connector.

IMG_20220130_095127539 (resized).jpg

What's "Rillby's," anyway? What's that got to do with Undertale? This was one area of the playfield that technically always worked, but never really made me happy. The Grillby's indicator light is out in the open, not only threatening to cause a short somewhere but just looking ugly. Printing up a sign is apparently out of the question, as I ordered "clear" vinyl sticker sheets that turned out to be white. It's not enough of a priority to mess with right now, but... what else could I do? How can I indicate to the player that they should shoot for the Grillby scoop, without a bulb sticking out?

IMG_20220130_212704144 (resized).jpg

Fortunately, I had the right parts on hand to play around with this concept, and I hit upon a solution: Flood the scoop with light. I dug around for the brightest bulbs I could find, settling on a 4SMD red wedge and a very bright white Flex SMD. I love Comet. The red shines into the inside of the scoop through the hole for the switch, and the flex bulb is mounted topside in a way that just peeks around the scoop, out of harm's way from the ball. Speaking of harm's way, both the top bulb and the white flasher got a new wiring harness that runs underneath the ramp and out the backboard. This is a very significant improvement over the old arrangement - wires running behind the red target and half peeking into the scoop, and already showing significant damage to the insulation.

IMG_20220130_212651756 (resized).jpg

In general, wiring harnesses were given a lot of love for the refresh, everything kept together and routed whatever way is most hidden away. There are a few wires visible from the top, but the unavoidable ones are still much better than tucking them behind targets and into spaces that their connectors can't even fit through.

And you might wonder how the new Grillby indicator light looks.

IMG_20220130_214000121 (resized).jpg

I think it'll be quite clear when you need to shoot for it.
The Librarby indicator lamp (orange, pictured at far right over the "T" target) was also moved from the post behind the scoop to mounted underneath the ramp. As it shares a connector with the ramp flasher, this just made sense. I intentionally reversed the lamp socket's polarity so that I could swap in the pictured orange retro-style bulb, which are one of the few Comet offerings that are positive/negative polarized. (Apply voltage to bottom of bulb, ground path through bayonet base.)

My start button was broken. I have no earthly idea how it broke, but it's been like this for a long time.
IMG_20220131_212322809 (resized).jpg

I wish you could buy individual pieces for a pushbutton, but on first glance I didn't see anything available. So I decided to pick up a new one - this one in red, like the classic Undertale heart.
IMG_20220201_205148320 (resized).jpg

Yeah, that's just become a running gag over this whole game's development by this point. Anyway, the cabinet wiring is done.
IMG_20220201_204651842 (resized).jpg

Quick explanation on that. I upgraded to a new audio amplifier because the old one just didn't have the power for proper-sized speakers. The new amp works fine on 12v wall wart power, but when I tried to use the built-in 12v rail I already had in the game, it heated up the amp to unsafe levels. To solve this, I reworked the SSR switching layout a bit, so that the main power relay switches a full duplex outlet; one plug goes to the playfield power supplies and now the other is the amp's adapter. I think I did a decent job making this all look clean, and the addition of a 6-outlet splitter on the main line ensures I still have an open utility outlet when needed.

IT'S ALIVE!!
IMG_20220201_212208014 (resized).jpg

Granted, it's not perfect. Still have some lighting issues, and two flashers are backwards/swapped. I'm not changing the wires, I'll just change the code, as the way it's wired right now makes a lot more sense. Lab and snowdin saucers did give me a couple of bounce-outs but in general they are a lot better behaved than they once were. The resort, on the other hand... it's a little hit and miss. I will continue to leave it out of the vital rules.

Nexy disapproves of my design, and I don't blame her.
IMG_20220131_210519156 (resized).jpg

The power supply idea isn't bad, but I'm not sure I can take advantage of it for the absolute kludge that is this game. There's quite a few unrelated power rails to be able to consolidate much.

48v for coils
18v for lamp matrix, rectified from 12vac on that little transformer
12v from computer power supply, driving monitor and flashers but apparently not appropriate for the sound system
12v sound system adapter
6.3vac for GI, the other voltage on the small transformer.
5v from PC power supply for..... something? Trough optos, I guess.

Anyhow, yesterday I fixed a softlock, today I recorded a satisfactory gameplay video. I'll get it edited eventually - for now it's time to fold up the game for transport to the local lumber store for location play!

That's... sound advice, no pun intended. I didn't actually realize the ATX power supply was a noisy power source - could that be the reason why plugging the amp into it resulted in the amp heating up enough to cause first degree burns?

I might end up looking into the power situation further, then. I suspect the 5v may be enough for GI (especially if I were to fit it with brighter bulbs) rather than trying to go with 12, then I'd have a good reason to use 5+12.

The lamp matrix is directly lifted from the System 11 spec, which uses 18v for the main supply, but the duty cycle of the lamp strobing results in the individual bulbs only getting an equivalent of about 6v. The lights are a little dim right now, however the cause of this problem is that my row drivers (pre-bulb on this design) are VASTLY underspecced. I attempted to address this before last year's show, but (as perhaps implied by my post about "we don't talk about the lamp driver") designed the replacement driver board incorrectly. I've studied lamp matrix drivers in more detail since then and will soon make another attempt to fix it.

The button's pretty cool, although would it fit within the recessed opening meant for a standard button? I hesitate to leave it sticking out the front as it makes it easy to press it by accident when leaning over the game.

In other news, it's out in the wild again.
IMG_20220207_082232097_HDR (resized).jpg

And now we do.

The pic I posted of an updated lamp row driver board was in high hopes, but it just straight-up didn't work. The reason would be because I made an assumption about how to use a transistor to control voltage supply (as one does in a matrix) rather than grounding path. You can't use N-channel for that...

Fortunately, I was able to find an example of how to do a lamp matrix properly with currently available components, and after testing on a solderless breadboard this time, I assembled this, pictured here before it was safely mounted in place.
IMG_20220326_080526046 (resized).jpg

I'm not sure I actually posted the reason I needed this auxiliary driver board for the lamps. So basically, for the last 3 years, the lamp matrix rows - I think I got it backwards, but in this case the rows are off the 18v supply and the columns connect to ground - have been driven by two very underpowered chips. Optoisolators, specifically, rated for about the amperage of one bulb MAX. I seriously cannot believe those chips survived all this time. As soon as more than one light in a row is supposed to be lit, the power takes a nosedive and all the lights end up dim.

The results?
IMG_20220326_080533663 (resized).jpg

Although not all the lights are lit in this photo, it's pretty obvious which ones are, and this is in a very brightly lit room. Although I have one issue to work out (looks like a wiring issue, row 5 seems to be out), the new lighting is a HUGE improvement over the old.

Now to fix the upper left flipper EOS switch, the one that broke in half. That flipper mech been flipped over 110,000 times, so I suppose it was bound to happen soon.

See you at MGC!

This is a gameplay video made from clips taken a few months ago, thus not including recent code and lighting fixes... but look no further if you want the latest playing tips.

I put in a final (hopefully) code update this morning, fixing a handful of bugs and generally making things just a tiny bit more pleasant. Since last show we have:
- Pretty much all new EOS switches, caps, and flipper protection fuses.
- Bug fixes
- A new lock animation
- Additional entries in the secret soundtrack
- Fully capable sound system (no more desktop stereo sitting on the floor under the game this time!)
- Playfield lights bright enough to actually see
- Shots you can actually hit

This year's homebrew section is shaping up to be pretty exciting, I hear, so I hope the latest improvements keep things fun!

Games are loaded up. Counting the hours until we depart, and I hope you all enjoy the latest improvements.

IMG_20220428_162957879_HDR.jpg

You're welcome, guys! I'm so glad you enjoyed playing. Ryan, you were awesome and I hope we get more opportunities to play each other's games. Oh, remind me to send you that soundtrack to Pinball Construction Kit.

So, after a day of rest and unpacking, here's the summary for anybody who missed it. This year, the Tale of the Under did not fail me. Or, well, it failed a little tiny bit, but otherwise did not fail.

I only had three known issues with UT at this show. One was a blown GI fuse, which was completely my fault. Ball got stuck in the top corner during transit (mistake # 1, didn't remove balls for transport). Booted up game, noticed missing ball, reached finger into corner area to dislodge it (mistake # 2, didn't turn off machine before digging inside). Got the ball out, but shorted it against a GI socket in the process. No real damage done, just a slap on the wrist in both pride and a couple dollars of expense at Rob's table.

The second issue was a broken solder joint, a very easy fix resolved within 15 minutes of my receiving the text about the issue. Definitely suffered some PTSD from that text though, which told me "lower left flipper is broken." After multiple hours spent solving flipper problems at the 2021 show, I felt some pretty significant dread as I raced up to the arcade hall to diagnose the issue. I was incredibly relieved to open the coin door and find the flipper's power wire hanging free, if a little disappointed in my own soldering during my last attempt at repair.

The last problem wasn't a game breaker, necessarily. A software bug - annoying that it happened, but I was glad it happened at the moment that I was watching someone play. The "collect extra ball" event decrements lit extra balls as part of its routine - however, if the function is tripped multiple times while another event is in progress, multiple extra balls could be collected even if the shot is only supposed to be lit for one. Saw a guy get about half a dozen extra balls in one shot during Dog Residue multiball, and wrote that one down in the book of bug fixes.

All three issues were, in the grand scheme of things, pretty minor, and quite a few hours of uptime made it onto the stats. Can't check them yet - need to unfold the game back at 41 Lumber, where everyone loves it, before I can get a flash drive plugged in - but I'll be sure to post whatever interesting bits I see.

Marco interviewed the homebrewers again, with Undertale once again getting the stream spotlight! (1:02:00 mark)

And, having set everything back up again and retrieved my flash drive, it's time for Stats for Nerds.

WE HAVE BEATEN OUR PLAY COUNT RECORD!
MGC 2019: 310
Expo 2019: 229
MGC 2020: Nobody played anything
MGC 2021: 240
MGC 2022: 325!!

Data field summary... this will be for all plays between April 29 and May 1. The lowest-scoring game was 33,030, the highest 2,039,230. To all who scored well, I apologize for forgetting to reset the high score table to manageable levels - it would've been nice to see people lay down some names. On average, players were scoring around 288,000, so I think my 500K replay score seems reasonable...?

Really wish I knew who managed the highest score, because that person was filled with determination. Their game lasted 19 minutes, during which they faced every boss including Flowey (though didn't beat him). They managed two extra balls, 160,000 on Toriel, a Miniboss Multiball, and two random encounters, as well as stacking up two bonus items.

In total, the game took on nearly 20 hours of active play time, a pretty big fraction of the total time it would have been running. The shortest game lasted a minute and 10 seconds. This would've offered a consolation extra ball at the Ruins scoop, but sadly, the player missed that chance and only managed a total score of 52,050, though they did score higher than a few other players.

Only that one grand champ reached Flowey, but the person who got the number three highest score of 1,345,600 managed to fill the main six minigames of the heart, the only other person to get that far. Second place went to someone with 1,614,040, beating 5 bosses - including a record-setting 300K from Toriel.

The ramps took their usual beating, 321 on Waterfall and 793 on the center ramp. A dozen or so games had one or two ball searches, with a small handful having 3 (replacement). I wish I had some way of knowing where the ball got stuck, although since I never had to pull the glass for one, I would guess behind the jet bumper, as this is the only known place where it can still happen. Even after a lot of reworking, that bumper never did get totally perfect about detecting the ball resting behind it, and unfortunately, sometimes the ball lands right back there after a ball search. I'm actually wondering if it would be worth adding an opto across the top of the bumper (from under the rubbers nearby) to detect when a ball gets stuck back there for a second or two and pulse the bumper automatically.

The maximum number of extra balls awarded was 5. The maximum number of extra balls you can receive, according to the software rules, is 4. My previous post detailed how this happened, and I'm really glad I was lucky enough to be around to see that software bug happen.

I ended up putting back the one-way Snowdin gate prior to the show, and it was probably the right call to make as it greatly improves a player's chances of hitting Hotland and Resort. These shots got 420 and 395 total hits, respectively, implying a far lower rate of shot rejection. That was my goal! Left/right shot flow generally seemed pretty good, with a very close balance of 22,282 left flips and 23,128 right. However, for at least the first day, the game was a bit off level, leaning to the right, resulting in 306 left outlane drains versus 269 right. I fixed this at some point - Saturday, I think and hope - when I tested out the game myself at opening time.

Most people got Temmie's awards at least once, with the second highest scoring player also taking the record of 4 Tem Shop visits. A fair number of players - 59 - managed the skill shot (although I have no way of knowing how many used a soft plunge and a flipper shot), with the record of two skill shots going to a player that scored 371,170, facing Toriel and Papyrus.

Only 24 games had players collecting items at the left ramp, and I think I could take that as a sign that I need to make that award more intuitive to collect. Adjusting the inlane switches to reduce bounce-overs off the ramps would probably help as well.

The most commonly played minigame was with Sans, as probably expected since the Grillby scoop focuses his mode. I'm actually kind of thinking that maybe the Grillby scoop isn't the best shot to focus that mode, as it's the shot that collects modes itself, leading to somewhat of a bias towards him over the other bosses. Library scoop might be a better candidate.

Major shot hit rate breakdown:
Least-hit: Ruins scoop, 286 hits.
Librarby scoop, 304.
Left ramp, 321.
Waterfall scoop, 346.
MTT Resort, 395. Really happy that people are able to hit this shot with relative ease now - it took a lot of reworking to make that happen.
Skeleton saucer, 402. This one had only gotten 113 hits at the debut show, pretty bad at rejecting. Much better now!
Hotland passage, 420.
Alphys's Lab, 692.
Center ramp, 793.
Most-hit: Grillby, 1671. A fairly central shot that you're often directed towards, and pretty much the easiest thing to hit from the left flipper.

Six players were able to get Miniboss Multiball. Considering how standard this multiball mode is, I think I should tweak some rules to make it easier to get, or at least more intuitive to figure out. McSquid asked me in person how to get Muffet during one of our conversations, so that seems like a pretty good indication that the game needs to make it more obvious. But for those looking to get a multiball mode, a lot more had luck with the random encounters, with 98 games experiencing at least one. Some players hit enough shots to qualify a second one, with 121 total random multiballs awarded in total.

Despite the number of upvotes I got when I posted about the secret soundtrack mode on Reddit, only one single person (besides myself) played the game with the alternate music. Maybe I should make a new pricing-type card that also includes a note about this Easter egg, because while it's not the most well-known game, there HAD to be more than one person at the show who was a fan of the incredible psychological horror RPG, "OMORI".

All told, I think it was a great run, and it seems like a lot of people enjoyed what I brought. If you were one of those people... thanks for playing.

It is indeed - although I've considered ways of digging up even more useful data, such as a complete switch matrix change map to finely track every shot. I could use such a thing to identify, even after the fact, possible places where the ball is getting stuck, as well as expected switch hits that are missing.

But anyway... here's the way the stats work. The game keeps track of a whole bunch of progress variables, with a set of each per player. Skill shot level, number of times multiball was achieved, extra balls awarded, flips on each side, etc. After every game finishes, a comma-separated list of all these statistics gets dumped into an ongoing stats.csv file located within the game program's directory, which I can retrieve with a flash drive if desired.

The resulting file never needs to be loaded into memory and rewritten, reducing the likelihood of data corruption when the power is spontaneously cut to the game at 8:00 PM on Saturday. It's just appended. An old sample file looks like this when viewed in Notepad:

Date,Score,Name,SecondsPlayed,MonstersCompleted,Run,Ramps,BallSearches,TotalExtraBalls,LeftFlipper,RightFlipper,LeftDrain,RightDrain,TotalTemAwards,SkillShotLevel,AttackItem,DefenseItem,TorielTotal,SansTotal,PapyrusTotal,UndyneTotal,AlphysTotal,MettaTotal,AsgoreTotal,FloweyTotalShotsWaterfall,ShotsLeftRamp,ShotsHotland,ShotsCore,ShotsLab,ShotsCenterRamp,ShotsGrillby,ShotsSnowdin,ShotsLibrarby,ShotsRuins
2021/10/20 21:35,1244000,OMORI,797,4,0,30,23,2,0,0,0,0,2,5,0,0,90000,48000,0,40000,0,120000,0,0,7,12,3,3,8,18,9,1,0,5,1
2021/10/20 22:25,35000,SNUUY,22,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0
2022/02/15 20:16,151000,XOUZO,189,1,0,1,8,0,0,0,0,0,1,1,0,0,40000,0,0,0,0,0,0,0,2,0,0,4,0,1,1,0,0,2,0
2022/04/14 22:55,79000,AAAA,46,0,0,2,1,0,0,0,0,0,0,3,0,0,0,0,0,0,0,0,0,0,0,2,0,0,2,0,0,0,0,0,0,1,0

It's missing a few data fields, but you get the idea. With this information all dumped into the file, it's simply a matter of loading that file into Excel or a compatible spreadsheet editor and working some formula magic on it. I usually throw a few summary rows onto the end, giving me Min/Max/Avg/Sum of all the data fields, with this last time including >0 and =0 at the end so I could count the total plays that (for example) made the skill shot to any level, or did not get multiball.

Since the data is stored per-game (per-player, rather, so a 4 player game counts as 4 unique data rows for each player's performance), I can trace back to a particular play to find out how far that player actually progressed. So if my summary MAX for the score field comes to 1,244,000, I can search for that score in the data table and then look at the other fields to see what that player accomplished on the way to that score.