The Modern Firepower Pinball Project - Power Supplies

WARNING:  THE CIRCUITRY DESCRIBED IN THIS POST USES ELECTRICITY AT LEVELS DEADLY TO HUMANS.  SEEK PROFESSIONAL EXPERTISE IF YOU ARE NOT SURE HOW TO SAFELY USE THE PRESENTED INFORMATION.

One of my very first achievements on my long pinball quest was a home-built solenoid power supply.  Though it was many years ago, I still remember clearly the first time Troy and I activated an old, used and abused solenoid with my home-built power supply.  At that moment I felt like a magician harnessing an invisible, mystical power.

The bottom half of this Williams 'High Power Solenoid Circuit' actually describes most of the power supply design.  J102 is the signal coming from the Transformer, which runs through a Bridge Rectifier (BR3), and a small Capacitor (100uf) and Resistor (10KOhm) provide a small amount of line conditioning.  Not shown here is the circuitry (Switch, Fuse, Line Conditioner and Varistor) that are in the circuit between the AC power source and the Transformer.

But as with most things pinball, for the DIY enthusiast there is a great void of information for what should be a relatively simple topic.  I had spent months researching and building the power supply from scratch, and had no idea if it would really work with pinball solenoids.  It was at that moment, hearing the BAM BAM BAM of the happily activating solenoid, that I first felt capable of overcoming the challenges of building my own pinball machine.

I wish I had all the answers, but to be honest there's still several aspects of pinball power I haven't had the opportunity, or the need, to solve.  But many of the answers I will be providing here I have never seen published anywhere else.  And trust me, I looked.

Power supplies in and of themselves are not complicated beasts.  Any number of books and resources will teach you how to build any of the many types of various power supplies.  If you are interested in building a power supply, I highly recommend you read up before tackling your own project.  In case you're curious, I read 'Building Power Supplies' 2nd Edition by David Lines, which I had picked up years ago at RadioShack for other projects. 

I do not provide enough info here to guide you through the process; rather my goal here is to help you choose the correctly sized power supplies that are needed for a modern, DIY pinball build like my Modern Firepower Pinball Project.

Power supplies are needed for several items in a modern pinball build, primary among those are the solenoids and the bulbs.  In my modernized pinball build, I'm also using a power supply for an industrial USB hub (which in turn powers four connected USB peripherals), and for the ball trough opto-switches.  If you're using any special anima-tronics with motors, you might also need a different voltage power supply just for them.  Every build can have unique requirements, so your choice of components and features will ultimately drive your power supply requirements.

Since I'm using an 32" LCD TV for the backbox (which also powers the speakers), I didn't have to worry about backbox lighting and audio amplifiers.  Obviously the TV came with it's own power supply (built in).  The only other device in the pinball machine is the small PC running Windows, which came with its own power supply.

Since the only real challenging power supply is the solenoid power supply, I'll discuss the other power supplies first, and save the best (worst?) for last.

Read on for more info...

Keep in mind that not all commercially manufactured pinball machines have identical requirements.  Since we in the DIY community are often forced to choose from publicly available components (rather than fabricating our own playfield machinery and electrical components), you should pay close attention to what brands and voltages the parts you purchase are compatible with.  I'm certainly no expert on every pinball manufacturer, so if the specifications I lay out don't match your machine's components then you should do some more research from some other sources and product manuals.

Starting with the lights, it seems that most LED pinball lights run on 6.3 Volts.  I don't really know where that value originated, or if it is really shared by all manufacturers, but since it is the most commonly available pinball LED bulb I decided to base my DIY pinball build around these 6.3V bulbs.

Long before I was ready to assemble my playfield, and before I had any power supplies, I had a need to test the LED bulbs with my LED-Wiz output control module.  Being that it is USB based, the LED-Wiz already has on-board 5V power, and to my surprise the 6.3V bulbs I was testing worked just fine on 5V.  I imagine that the bulbs are about 20% dimmer on this lower voltage, but under the Firepower's playfield inserts I found that the brightness of these 4-LED bulbs are perfect even at 5V.

I also did a bit of math to figure out what size power supply I would need for the bulbs.  If I was using incandescent bulbs, then the per bulb amperage is 0.25A for a #44, or 0.15A for a #47.  I would have used the cooler #47's in a non-commercial, home environment. With approximately 80 bulbs on the playfield, multiplying 0.15 x 80 shows I would have needed a huge 12A power supply to allow all of the incandescent bulbs to be on at the same time.

But I knew from day one that I wanted to use energy efficient LED bulbs.  Not all LED bulbs consume the same power (unlike standardized incandescent bulbs), but in general they consume about 1/10th of the power that an incandescent bulb uses.  For my calculations I used 0.025A x 80 LED bulbs for a much more reasonable 2A power supply.

There are some exceptions to the 6.3V LED bulb standard - for example I have some LED flasher bulbs that are 12V rated.  I tested these with 5V and 6V power sources, but these flasher bulbs didn't light up at all until I tried 12V.  So for bulbs alone I now needed to provide two different power supplies.

Before I spent any money purchasing a couple power supplies for the bulbs, I sorted through my collection of old, obsolete electronics and computer peripherals.  When you stop and think about it, most electronics come with wall-warts and power bricks attached, and they are available in all sorts of voltages and amperages, and here is where I got really lucky.

Modern desktop PC's use two primary voltages to power internal drives and peripherals: 12V and 5V.  In my searching I came across an old external DVD/CD drive enclosure.  I bought it years ago for about $30 from NewEgg, but no longer had a need for it.  The wonderful thing about DVD/CD drives is that they use both 12V and 5V.  I looked at the power brick and knew I had won the power supply lottery:  12V @ 2A, and 5V @ 2A!  I now had a single power supply that could power all of my playfield bulbs.  Further, since I would be running the 6.3V bulbs at 5V, the actual current draw would be even lower, probably about 1.6A for 80 bulbs, leaving me a little spare capacity.

This salvaged power supply from an external DVD drive provides 5V and 12V, both at 2A, enough for the playfield lights, USB hub and connected peripherals, and the ball trough opto-switch.

But it only got better from there.  In my earlier testing I had purchased a cheap USB hub that had an external power connection, but it came without an external power supply.  One day in testing I hooked it up incorrectly to my Radio Shack adjustable power supply.  Instead of feeding it 5V, I gave it 12V, plus I accidentally reversed the polarity.  Within seconds of turning on the power I smelled burning plastic, and picking up the USB hub to investigate I got a nasty burn on my fingers.  Needless to say, my cheap USB hub was toast.

In my search for a replacement, I decided to go for an industrial USB hub.  These hubs are expensive at around $70 for the hub I picked up, but I thought it had a better chance of surviving its new home attached directly to the underside of the pinball playfield.  The odd thing about this particular industrial USB hub is that it wouldn't run off of 5V, but rather needed a power supply between 7V and 24V.  Lucky me, the 12V portion of my new-found DVD drive power supply was underutilized by my six 12V flasher bulbs, so it had plenty of 12V power to share with the USB hub.

I got lucky once again when it came time to hook up the ball trough opto-switches.  I didn't know it when I purchased the ball trough, but the circuit board for the opto-switches also required 12V power (but mysteriously produces a 5V switch output signal).

So my single power supply, producing both 5V and 12V, was more than sufficient to run my regular LED bulbs, flasher LED bulbs, opto-switch circuits, USB hub, and connected to the hub, the three LED-Wiz output controller boards and the single U-HID input controller board.

Had I wanted to, I could easily have picked up a 6.3V 2A power supply to make the playfield bulbs a bit brighter, but I see no need for more brightness with the bulbs I have, and I love the simplicity of a single power supply connection for all of the above devices.

All that remained to be provided was a solenoid power supply, and that's where the challenges popped up.  The biggest problem: there's no published standard (that I could find), nor solenoid power usage statistics that are useful for power supply design.

Are there any shortcuts for the solenoid power supply?  Sure.  You could probably find an old, used power supply from a pinball machine and, assuming it has the correct specs, use that in your build.  But most of these pinball power supplies are purpose built with multiple taps: i.e. a 50V tap for solenoids, a 6.3V for bulbs, a 12V tap for electronics, a 18V for sound, etc.  Originally I looked for some of these pinball power supplies, but they seemed old, hard to find, and expensive.  For my DIY build, I didn't need such a complicated power supply, so I had decided to go with a dedicated 50V power supply.

Are there any off-the-shelf generic power supplies you can buy?  Perhaps, depending upon your specific application. In my searching, I found the selection of 50V power supplies to be rather limiting.  You would probably do better looking for a 48V power supply, or even better using solenoids that operate at lower voltages like 24V.  I spoke with someone recently who was building a pinball machine that will only need 24V, and (assuming our calculations were correct), there were many off the shelf power supplies available, starting in the $10 range.

Me?  I had decided to build my own power supply.  There are many types of power supplies that can be built, but for solenoid use I chose to go with a linear unregulated design. Unregulated power supplies vary the output voltage with load and line conditions, so a 50V unregulated power supply may not always produce 50V, but as long as the voltage is in the ballpark it should be fine for powering solenoids.

My home-built power supply.  The 50V 5A transformer is at top, and all circuitry is protected in a Radio Shack project box.

To build a power supply, you'll need a few ingredients:

• Transformer (to drop the house voltage to the solenoid voltage)
• Bridge Rectifier (to convert AC to DC)  
• Line Filter (a standard EMI filter for AC power)
• Varistor (surge protection device)
• Fuse (to protect the Transformer)
• Capacitor (to smooth the final DC voltage signal)
• Resistor (works with the Capacitor to smooth the final DC voltage signal) 

Inside my home-built 50V pinball solenoid power supply.

The sizing of several parts, like the Resistor and the Capacitor, I borrowed from the Williams 'High Power Solenoid Circuit' schematic (at the top of this post).  The Bridge Rectifier, Line Filter, and Varistor I picked up at a Pinball supply shop, so I knew they were right for the job (these parts are rated many times higher than the house AC and pinball DC voltages, so they are somewhat generic parts, one size fits all pinball machines).  The Fuse Holder, Power Switch, and even the project box were generic items I picked up at my local Radio Shack.

Of all the parts, the transformer is the key component, and the part for which I had no technical specs, so I had to start from scratch.  But to pick the right transformer, I needed to know the requirements for powering pinball solenoids.

The first specification in question is voltage.  Is there a standard pinball solenoid voltage?  No, but there are several common voltages.  It seems most pin manufacturers had a favorite voltage or two that they liked to use.  I had already decided in my build to use primarily Williams style components, and looking through several Williams pinball manuals, I noticed that 50V was very common for the solenoids.

NOTE:  Having said I planned on using Williams components, I have actually purchased many playfield assemblies that are not Williams components, simply because I liked their designs better than the older Williams designs.  My kickers/slingshots are Data East, my ball savers/knockers are Stern, and I think the ball locks/eject holes are also Stern.  To be honest, I don't know what voltage these non-Williams parts were designed to work with, but I hooked them up to my 50V power supply and so far haven't had a lick of trouble.  In fact, during testing I lowered my power supply to 25V and I felt that all of the solenoids were too weak, so my guess is that all of my components just happen to be designed for 48V or 50V .  Just because I got lucky doesn't mean you will, so do your homework.  Pinball machine manuals are your best friend here.  When looking at available parts, read the manual for the tables that they are compatible with to determine what voltage the manufacturer powered them with.  You can certainly uses a mix of components in your DIY pinball build if you want to, but wiring gets complex fast when you need multiple power supplies for different rated solenoid coils.

So I knew I needed a 50V transformer, but how big?  In the beginning, I really didn't know, so I just purchased the largest 50V transformer I could find under $100 bucks, and it was a big 5A beast.  But this was just for testing, and for my final build I knew I would need to make sure I had the right transformer for the job.

So I worked the math.  I figured worst case I would see a maximum of 6 simultaneous solenoid firings (two flippers + 4 ball multi-ball). I tested the solenoids I was working with and measured about 4 Ohms (this measurement lined up with solenoid information I've seen published online).

Voltage / Resistance = Current, so 50V running through 4 Ohm solenoids would draw 12.5A (uh-oh, not looking good), and 12.5 Amps x 6 solenoids meant I needed a power supply capable of 75 Amps of draw.

Yikes!

The math just didn't sound reasonable, and I knew I was making a mistake somewhere (several mistakes, as it turns out).  Looking online, I wasn't able to find any published information, so I've had to do all my own hands on research using my test 50V 5A transformer.

My first clue that I had messed up the math was that the solenoid didn't pop the 5A fuse that I had on the input side of the 5V transformer.  If the solenoid was really pulling 12.5 Amps, that fuse would have popped instantly.

I didn't have an ammeter to measure the current, so I devised another test.  I put a fuse holder between the power supply and a flipper solenoid, and I tested various fast blow fuses of different ratings, noting which values blew the fuse, and which values didn't.  A 2A fuse blew immediately, a 3A fuse would fire a few times before blowing, and a 4A fuse worked all day long.  Interesting!  So 3 Amps was about the power draw of a flipper solenoid, but to be on the safe side I decided to do all my math again using a 4A value for the solenoids instead of 12.5A.

Working the math in reverse, that meant the 'energized' resistance of the solenoid was 12.5 Ohms  (50V / 4 Amps = 12.5 Ohms).  Since my particular test solenoid was drawing closer to 3A, the real-world 'energized' resistance was closer to 16 Ohms, about 4 times higher than the directly measured resistance.

But why the discrepancy?  Why did a solenoid coil that measured 4 Ohms with my multimeter behave as if it was 16 Ohms when electricity flowed through it?

Here's what I found out: The directly measured resistance of a coil is only useful for identifying bad coils, and somewhat useful for comparing one coil to another. The directly measured resistance cannot be used in the power formulas, as the resistance of a coil increases dramatically when energized. This same challenge holds true when measuring resistance for LED's (and possibly other components as well).

The resistance changes because the electricity flowing through the component heats up the coil wires instantly. Your average multimeter cannot measure the resistance of an energized coil/LED without blowing a fuse; I'm not sure that any multimeter can measure this, especially since to energize the coil, you have to put it in a circuit, and now you're measuring the resistance of the whole circuit, and on top of that the measurement signal would be lost in the power supply signal.

Since you can't measure energized resistance directly, you have to calculate energized resistance by measuring amperage and voltage.

With my new knowledge, I now understood why the 5A fuse on the AC side of the transformer didn't blow when I fired the solenoid.  But even at a real-world 3A power draw, I didn't think I could fire more than one solenoid at a time.  Even worse, using my conservative 4A value for design purposes, the math said I needed a 24A power supply to simultaneously fire 6 solenoids.

But here's the rub:  I've already fired multiple solenoids simultaneously on my playfield, at least 3 or 4, maybe as many as 6 during a 4-ball multiball test, and I've never once blown the fuse due to too many solenoids firing simultaneously.  Not only that, I did my multi-ball testing with a 3.15A fuse, much lower than the transformer's 5A capacity!

How is that possible?  Well, I've never found any technical literature or published formulas to support my theories (and I've certainly looked), so the following statements are simply my own deductions.

From what I've deduced, it seems that the current draw is not the same on both sides of the power transformer. It appears that both Voltage and Current are transformed by the transformer, and that only Power (watts) remains the same.

In very basic terms, on the output side of the transformer,  6 simultaneously firing 4 Ohm solenoids, drawing a total of 24 Amps at 50 Volts, would consume 1200 Watts (24A * 50V = 1200W).

On the input side of the transformer, 1200 Watts at 115 Volts draws 10.43 Amps (1200W / 115V = 10.43A).  This formula ignores the inefficiency of transforming voltage, so to be on the safe side I would add about 25% additional current capacity, concluding that I needed a 13A 50V transformer to support 6 simultaneous solenoids.  That also means that each solenoid is drawing about 2A of power at the wall.

A 13A transformer sure sounds a heck of a lot better than a 75A transformer, but still.... I wasn't popping my measly 3.15A fuse, even during multi-ball tests.  I felt I was still missing a crucial element to the formula.

Keep in mind that this 13A design spec includes my margin of safety.  Since my actually measured resistance of my test solenoid was barely above 3A, the real-world math shows that 6 simultaneous solenoid firings would only draw about 7.8A on the AC side, plus any overhead from transforming the voltage, for a real-world total of about 10A.

So, if 10A was a 'real-world' value, perhaps I wasn't actually firing 6 solenoids simultaneously in all my testing.  Perhaps I was only firing up to 2 simultaneously, drawing just over 3A (remember I'm testing with a 3.15A fuse).  While there is nothing in my software that would prevent every solenoid firing at the exact same time, maybe it just wasn't happening.

When I originally wrote my pinball software, I knew I couldn't leave solenoids on for long periods of time as that would burn them out (flippers being the exception to the rule, and the flippers I'm using have old style dual-coil solenoids that automatically drop to a lower current when the flipper hits the End-Of-Stroke switch).  You have to turn solenoids on and back off in a brief period of time, just long enough to allow them to extend fully before retracting.  I had no idea how long to energize a solenoid coil, but my first guess was 200ms, or 1/5th of a second.  In my early tests, that value worked just fine, so at that time I never bothered to test any other values.

In the current version of my pinball software (the Chameleon Pinball Engine), I wrote a utility that allows you to map the solenoids to the software, and to fine tune the firing time of each solenoid individually (I'll cover this in detail another day).  Raising and lowering the firing time in 5ms increments, I would test fire each solenoid while holding a ball in position and evaluate how far the ball traveled.

Note: I am not using PWM for any solenoids (or for the lights for that matter).  The 'fire time' simply refers to the amount of time the solenoid is turned on, allowing it to fully extend, before turning it back off.

To my surprise, I found that a total energized time of about 20ms was enough for almost all solenoids to exert full power onto the ball.  The exception was the pop bumpers, which seemed to work best when energized for 40ms.  Regardless, both of these values were much shorter than my original firing time of 200ms.

Taking the worst case scenario of a 40ms firing time, in theory you could fire 25 solenoids, back to back, all within a single second with no overlap.  While to the human eye (and ear) this would appear as if I had fired 25 solenoids simultaneously, but technically only one is being fired at a time.  40ms is extremely quick.

These short fire times have two implications:

1. It is much much harder to have multiple solenoids firing at the exact same time, not impossible, but extremely rare.  So even though I'd like to think I'm firing a whole bunch of solenoids simultaneously, in reality this rarely happens.  It's certainly plausible that I've never fired more than two solenoids simultaneously. 
2.  Even if the planets aligned and 6 solenoids fired simultaneously, the power draw is not a constant drain but rather a momentary spike. There seems to be some capacity for the power supply to handle momentary surges. The published power supply specs are for constant draw, not peak surge.  The small capacitor that smoothes out the DC power supply is probably buffering most of the spike.

There is one last factor to consider in my test results - the original solenoid that I did all my Ohm and Amp measurements on was a flipper solenoid.  Flippers are some of the most powerful solenoids in a pinball machine, so really my math was oversized, calculating 6 simultaneous flipper firings.  In reality, most of the solenoids are much weaker and draw less Current. 

So you made it through the confusing math, and you may have noticed that I simply shared a lot of theory, and didn't have a nice formula to button it all up to give you a simple way to calculate the power requirements for solenoids.  Sorry, this is as far as I've gotten in my testing.  I'll do some more testing in the future to refine my theories and formulas, but here are some takeaways to remember.

• Do not used the published or directly measured coil resistance values to calculate power draw!  Doing so will indicate you need an artificially large transformer or power supply.  The resistance of an energized solenoid coil is about 4x higher (at least in my testing), which means the Current draw is about 4x lower.  You need to measure Current and derive Resistance.

• The current draw on the output (solenoid) side of the transformer will be much higher than the current draw on the input (house) side of the transformer.  Multiply the calculated DC current by (DCV/ACV).  If you are using a 50V transformer, that means 10 DC Amps * (50 DC Volts / 115 AC Volts) = 4.35 AC Amps.  If you are using a 24V transformer, that means that 10 DC Amps * (24 DC Volts / 115 AC Volts) = 2.09 AC Amps.

• Transformer and power supply ratings are for continuous current draw, which is not what they will see in a pinball machine.  Solenoid coils are energized for extremely brief periods of time, causing very short spikes in power draw, and it seems that you can exceed the ratings by a small amount without causing harm.  Any solenoids that have a constant-on mode, like flippers or some diverters/posts, will operate in a low current hold mode which shouldn't be a primary concern for the overall power supply capacity rating.

• Always use fuses, both on the DC/Solenoid side of the power supply, and on the AC side of the power supply.  I have 18 fuses in my system:  One 3.15A fuse on the 115V mains, one 5A fuse on the 50V DC supply, and 16 4A fuses, one per each solenoid output.  All of these fuses are below the technical capabilities of my power supply and circuitry, and none has ever blown due to normal solenoid usage.

In my future testing, I may write a routine in my software to fire multiple solenoids simultaneously (i.e. 3 at a time, 7 at a time, all the way up to 14 at a time, which is how many I have installed in my system).  This would be the only way for me to ensure they really are firing at the exact same time (and not 20ms apart), and I could see what number of solenoids it takes to pop a fuse.

I also plan to test some lower amperage fuses on each of the solenoids, not just the flippers, to get some real-world metrics of the power draw of different solenoids.  Currently I am running a 4A fuse on each solenoid, and none of them have ever popped.

Hopefully you found this power supply primer helpful in your project.  If you have any questions or test ideas, please let me know.