# Power Supply for two 50V/12.5A Solenoids

I'm in the planning stages of building a custom pinball table and one of the sticking points is powering the flipper coils. They're 50V DC coils with a resistance of only 4 ohms during the stroke. That's 12.5 amps, and there are two of them so potentially up to 25 amps.

However, they are only high power for the length of the stroke, when an "end of stroke" switch uses the full coil resulting in 133 ohms resistance bringing it down to a much smaller 0.376 amps which is all it needs to hold the flippers up.

"Real" pinball machines use a massive transformer with a full bridge rectifier. They're unregulated (I believe), which is fine, and have several taps for other voltages I'm not concerned with generating (much easier ways). But I'm not having much luck finding something similar for the 48-50 volts that isn't several hundred bucks.

Trying to design something though I'm not sure what to aim for. The 12.5 amps (each) is for such a short duration it seems like overkill to design something that can handle that much power continuously. However, I don't know how a modern switching power supply would handle such an impulse. I feel like it would think there is a short circuit and shut down to protect itself.

I've heard of people using ATX power supplies from PC's, just daisy chaining 4 of them together to get the +12V rails up to 48V (which is close enough). These can be had for fairly cheap (~\$15/ea) and have a lot of power (25+ amps).

How would these handle an impulse load like that? Can they be connected serially without other issues? Is there a simpler solution?

A transformer is probably the simplest option. It doesn't need to be exactly "110 in 50 out". Transformers are ratiometric, so you just need about 2:1. I would get a 110 to 230v transformer and run it in reverse. 110 * (110/230) = 52.6v. I saw some cheap, high power transformers that fit the bill on ebay.

You almost definetly can't hook up 4 atx power supplies in series, because they're probably grounded.

You may be able to use a smaller power supply, plus a cap bank. You're correct the supply will probably trip almost instantly if you overdraw it, even with a bunch of caps. The reason is that for the caps to provide power, the voltage has to drop. If the supply is any good, it's going to fight that drop and basically take the whole load itself.

The way to counteract this is to place a beefy current limiting resistor between the power supply and the caps.

simulate this circuit – Schematic created using CircuitLab

Calculating the exact resistor is actually kind of tricky. But you can just buy a handful of 1 ohm power resistors and figure it out by trial and error. You could also play around with a circuit simulation, but you'll need to know the coils inductance. I think if you click on this schematic it'll bring up circuit labs, which has a simulation feature.

First, you might not need that full 12.5 ampere, because a solenoid has inductance, and your computation is for the maximum current after some time. The solenoid will act quick and short, so the actual current will be lower.

Then, you might not need 50 V. What the solenoid needs to get moving is current, not voltage. So why not charge a nice inductor with, say, 5 A and force that through the solenoid? Like the answer of Drew, but then different.

An ATX power supply gives lots of amperes at 5 volt. Make sure the DC resistance of L1 limits the current to something like 5 A. When you open the switch, the current through L1 will continue through the diode and the solenoid.

The inductance of the solenoid might cause the voltage to rise to, say, 50 V, but who cares. The amperes will make the solenoid to move, which is what you want.

You might want to implement the switch with a FET transistor, which might be cheaper than a mechanical switch. And to control your pin-bal machine you probably use more stuff, like an arduino, which might interface nicely with a FET. But that are implementation details. Main point is just try an inductor L1 and a modest power supply instead of overkill 50 VDC and 12.5 A.

Good Luck!

ATX supplies are no good for this.

First of all, ATX supplies cannot be connected in series, unless their internal circuitry is modified so that the output ground is floating, or they are left ungrounded at the mains side, which again is something that should not be done.

Another thing is they barely work when only one of the output voltages is loaded, and they work extremely poorly with pulsed loads, especially if the load is suddenly just disconnected. So they need a lot of extra work to make them able to handle arbitary pulsed loads from single output voltage.

Almost any other solution would be simpler, like having a capacitor to provide the large surge current and smoothing the current taken from supply with an inductor.