Rialto (1964, Wulff)

I only now realize I never properly wrote about Rialto, so it's time to rectify that.

This is a slot machine made in 1964 by Wulff (Germany).  Slot machines became legal in Germany in the early 1950s, but that came with a set of limitations that made for some fascinating design choices.  Games could only offer a maximum 10x payout, and most importantly games needed to take a minimum of 10 seconds to play.  

So no putting in a coin and getting a result within 1-5 seconds, they had to design machines that offered a full 10 seconds of gameplay.  It is for this reason that I absolutely adore the 50s and 60s German machines and all of the small / quirky ways they innovated within this rule.

This machine was difficult to arrange for, and I knew before it arrived it would be a project.  The first round of repairs on this machine were done by Sylvain, who graciously documented the repairs. (archive)

There was a number of broken switches and busted parts.

This round, my friend Robi was able to tackle the issues.  But what were the issues?  The main issue was that it seemed like the motor itself was gunked up so bad that it only advanced at a snail's pace, such it could take a few minutes to make a single revolution.  He took out the motor and cleaned/re-greased it, but that was only part of the issue.

disassembly

There were also switches that were dirty, reducing the power that got to the motor.  Even when clean, the motor would sometimes still start sluggish, and then proceed to speed up.  What was happening there?  Well, it was winter and my game room is the back room in the basement.  Even a few degrees in temperature seemed to make all of the difference with the motor's initial acceleration. 

The motor's force needed to be maintained across a series of switches for each stage of operation, and any of those being dirty adversely effected it's progression.  But the motor itself did not appreciate being below ~ 20 degrees Celsius.

2 photos of the right side of the machine.  As they say in the collecting world, "check out that beautiful patina!"

coin slot at the top, coin reject button, and coin return

licensing information card

Canadian nickels are readily accepted in lieu of 10 pfennig coins

10 pfennig
21.5 mm diameter
1.7 mm thick
4 g weight
Brass clad steel (90% Iron, 10% Brass)

Canadian nickel (5c)
21.2 mm diameter
1.76 mm thick
3.95 g weight
Nickel-plated steel (94.5% steel, 3.5% Cu, 2% Ni plating)

Rialto vs Bingolet

The Rialto model is similar to Bingolet, except it has an automatically firing plunger.  The player would have a short window to fire the ball, but the motor is advancing and will trigger a solenoid that fires the ball for you.

Bingolet

So while it is nice we have the Bingolet schematics, they can't be trusted to be entirely accurate to the Rialto.  The addition of the autofire mechanism and timing circuit change things ever so slightly.

from an old pennymachines thread (archive) here are the internals of a Bingolet:

inside a Bingolet

The main noticeable difference is that the Rialto has an extra coil below the payout mech.  This is the solenoid that fires the ball automatically.

Bingolet payout area

Rialto payout area with extra solenoid to automatically fire the ball.

Here is a video (someone else posted) that shows it firing automatically:

Replacing rubbers

The original rubbers in this machine are all quite old and hard.  The peg rubbers measure 4.22mm OD and 2mm ID.  Well, the 4.22mm is the rubber, the 2mm is the width of the pin itself.

The top bumpers are 11.34-11.40 OD.  I have not yet replaced these.

For the pegs I first ordered the red/orange silicone tubing.  2mm ID, 4mm OD.  It was easy to work with, but the softness of it meant the ball would hit it on the top pegs and barely bounce.  This is a game of chance so you want a bit more chaos at that top row.  I decided to use silicone for the other pegs, but something harder for that top row.

I found polyurethane tubing to use for the top row.  2mm ID, but only 3.175mm OD but that is fine.  A much bouncier top peg, and the gaps between any two pegs increases by 1.65mm as a compromise.

replacement material!

As you might notice in the photo above, I did manage to pull one of the old rubbers off intact!   It is 33.81mm in length.

it looks pretty good, but the actual pegs themselves are blocked by the paint on the glass.

Updating the lights

The Rialto glass is right up against the lights.  I am not going to spray the glass to preserve it, but I am interested in turning down the heat so it doesn't bake so much.

Rialto uses 12842 3V 12V bulbs, and I found an equivalent contemporary replacement is DE3021 bulbs, which are available in LED.

DE3175 / DE3021 / DE3022

The original bulbs are 30mm long.  The above spec sheet lists the DE3021 as 31mm.

When I measured the new bulbs, they measured as 30.45 mm which is just a .45mm difference, but in the enclosed light areas it can be a tighter fit when you put the plastic shields on around them.  I haven't tried replacing all 4 bulbs in a single column with LEDs, and I have a funny feeling the plastic shell might get stuck.  Before I do that I should probably get the shells 3d scanned so a fresh one can be printed if it breaks.

I think if the bulbs were actually 31mm long they should not properly fit within the light shields at all.  The sockets themselves are springy and can easily take 31mm, but the rigid light boxes do not.

The light barriers were super dirty and soap and water didn't seem to to the trick, but Simple Green and paper towels got off the final layers of dirt/soot.

the LEDs I got were .45mm longer than the original bulbs

I swapped in 6 LEDs as tests

There are still incandescent bulbs available.

For the time being I am going to replace the 6 GI bulbs with LEDs, and just ensure all of the game bulbs are working.  I'll also go at them with a clothe and remove their decades of dirt to help minimize their heat a bit.

Math!

Let's look at how the numbers are distributed:

1, 2, and 3 each appear 3 times each.  4 appears 5 times.  Guess which number is least desirable?  But it's actually not a crushingly unfair placement of numbers, as the left-most and right-most 4 are statistically the rarest to enter.

If we ignore the leftmost and rightmost 4 slots, it is the 1 and 2 that are the next hardest to hit, or at least have the lowest odds that the ball will bounce in to them.

If you want to aim, shooting for the left-half of the playfield is your best bet ( 1 3 2 4 3 1), and the right-side is more 4-heavy (4 2 4 3 1 2), offering lesser returns.  The makers were banking on players shooting far more to the left than to the right.,

If the odds were even to hit each of the 14 holes, we'd write it like this:

• hitting a 1: 3/14 = 21.4%
• hitting a 2: 3/14 = 21.4%
• hitting a 3: 3/14 = 21.4%
• hitting a 4: 5/14 = 35.7%

That would give us a wins percentage like this:

A 10 pfennig play would average a 7.54 return

But with the edge numbers, it's not such an even distribution,  Let's change the weightings to better reflect the odds of hitting each hole.  Please note, this is not scientific weightings, more just an arbitrary feel.

This slightly alters the odds of hitting each number:

1: 20%

2: 23.64%

3: 24.55%

4: 31.8% 

1 is reduced in odds, which you'll note makes the highest payout scarcer.  2 and 3 are increased, and having the 4s at the outer edges reduces the odds of hitting a 4 by 3.9% with these estimates.

So while note huge, than gives an extra .2 pfennig edge to the house, or to say it another way, reduces the expected payout percentage for the player by nearly 2%.

The 27.29% is the odds of making ANY win in a round.  73.4% is the expected payout rate.  This is actually abysmally low compared to modern machines.  Think of Japanese pachislo machines, where the lowest default rates you can set are still within the 90% range!

Video

Here's some video of the machine in action now

One final photo