Here are my thoughts to your questions:

1) If your EoS switch is Normally Open, you can apply the voltage across the hold winding when you see a closed/shorted EoS and shut off the voltage to the power winding via the MOSFET method you described above.

2) FL-11629 coil power winding resistance is 4 Ohms and the hold winding resistance is 132 Ohms. Looks like on the Williams WPC games they used 50V for the power winding voltage and 18V for the hold winding voltage (http://www.pinballmedic.net/coil_chart.html)

3) I would not use a timer, but rather look at the flipper button & EoS switches to know when to activate and deactivate power to each winding on the flipper coils. If you tried to use a timer, what would happen if the user pressed and released the button in rapid succession? Your timers would either have to be reset and you wouldn't know when to switch to the hold winding vs. power winding...keep it simple since complicating it doesn't really buy you anything.

4) You can use two separate voltage rails as Williams did. If you crunch the numbers, it would save you ~167mA. V=IR (40V/132Ohms = 303mA vs. let's say 18v/132Ohms = 136mA). I would be more concerned about using a Power Supply with only 30-40V. You may find this doesn't have the power that you are looking for. Personally, I like strong flippers.

5) I am no expert by any means, but it would be cool to take the microcontroller out of the loop entirely and just have the circuit control the power switching from the power winding to the hold winding via the flipper button and EoS combination. I'd have to look more into it, but I would imagine this has been done. Take a look at the manuals for the WPC games and others to see how the circuits controlled the flipper coils. You can grab these off of ipdb.org.

Good Luck!

Take a look at pinballcontrollers.com
Fast pinball
Mission pinball framework

good call, just read through this: http://pinballcontrollers.com/wiki/Controlling_Flippers and I'm swaying back to being software controlled for the benefits listed.

Good stuff!

I'm using the open pinball project's open source hardware to drive flippers, pop bumpers and kick out holes for an early 80s playfield I have using 3 computer PSUs giving 36v to the flippers and it works great.

Of course the play field doesn't have any ramps or anything that would require stronger flipper power.

You will need a power supply with a very tasty reserve of current availability. It won't be enough to have a switching supply "rated at XX amps" - those ratings are maximum continuous NOT surge as you will experience with a pinball. Switching supplies used as you propose are very likely to either shut down or explode.

Forget switching supplies - they are not suitable for your application unless you do additional things (as with the SPIKE system).

Traditionally, flipper coils use the same voltage supply for both windings - a high surge of current is drawn to pull in the flipper and then a much lower hold current is needed to hold the flipper in the actuated position. The exact voltage is less important than the current (amps) availability.

I use the Antek one from pinballcontrollers. then i added a 6v (for logic) and 12V switch power supply. i personally don't like using ATX supplies.

Just my 0.02 but definitely try controlling the flippers from software first, I had a ton of problems with voltage surges from having high voltages across contacts. If your CPU isn't fast enough to do it without lag, you're going to want a faster CPU anyway. I'd also stay away from using shift registers for controlling solenoids, or for low-lag inputs like flipper buttons or pop-bumpers, unless you can find some really fast shift registers. the Mega's got 56 IOs, save the shift registers for non-momentary inputs and lights

Ha! you're one of the other 2 users

My project:
http://homebrewpinball.blogspot.com/2015/01/fets-happen.html

I can't wait to see video of Hugh's presentation
http://www.pintasticnewengland.com/speakers.html

switching power supplies are indeed fine to use. I'm using a meanwell 48v psu for my skyrim pin. DONT_ use cheapo chinese switchers. meanwell is a good brand and work fine. I have not looked at lambda tdk switchers in a while. Lambda used to make fantastic quality big iron bench power supplies way back

you will need to check that your mosfets are logic level and that if you drive them via shift registers that the shiftregs can supply enough power per pin to the mosfet, something like IRL520 or IRL530.

Meanwell is a Taiwanese company and the majority of their products are made in China, admittedly in their own factory, but in mainland China nevertheless. Taiwan, by the way, is R.O.C. (Republic of China)

Why do so many people try and pretend they are not buying Chinese stuff when it can't be avoided?

^ Bingo! ^
Exactly what he said!

For coils - don't use a switcher, they can provide a high, relatively constant current but not intended for rapid, extreme current changes like you will find in a coil circuit. Get yourself a simple transformer and rectifier. In the long run, it will be cheaper and far more reliable than using a switching regulator and will quickly respond to high current spikes. Get a big, beefy transformer well in excess of any current you *think* you'll need - these can pack a good punch to keep things flipping quickly.

Meanwell stuff might well be made in china, but is not actually Chinese, but its a well regarded name brand using quality components vs the no name chinese switching junk that is made in China, hence my 'dont use cheapo chinese switchers' which makes meanwell niether a) cheap or b) a chinese switcher.

Thanks for schooling me on the whole Taiwain is actually the Republic of China, of course, as an Australian growing up in a Chinese neighborhood of Melbourne, I guess I just had no idea of all the different south east asian countries that existed at all. None what so ever. Who would have thought there was a whole big world out there?

There is a huge difference in buying something made in China from a company that uses legit name brand quality parts and get proper safety and testing certifications vs something from a nonname factory in China that uses counterfit caps, reclaimed + recycled parts and other low quality non qc'd parts and dont bother getting their products certified.

hell you guys have a factory over there. maybe you should pop down to Shenzen and play spot the counterfit ic's for amusement.

My factory 'is' in Shenzhen and I live here - I know all too well EXACTLY what and how things are done here. There are literally thousands of very high quality switching supply manufacturers here and we use a couple of them with great success. To make a blanket statement that all Chinese switchers are no good is simply incorrect.

Maybe if you only want to pay $3~5 each you will get what you pay for?

Meanwell is Chinese no matter which way you want to try and argue about it.

quick, better call Gary Stern and tell him his Spike system should not be using a 48VDC switcher since its a bad idea.

hey, its funny, I just PWM'd two flippers with my switcher. something must be wrong, it appeared to work.. and its been working for the last two and a half years!

In any case let's not derail the OP's request about a power supply for his homemade machine. My advice (backed by Ed) is to get a chunky power transformer, bridge rectifier and a chunky Chinese capacitor and he will find that to be super reliable.

I clarified about the Spike system in my post.

Well done to you but excuse me if I don't follow your example in my commercial equipment.

hey! do you think they might be those 'cheap chinese switchers' I was telling you about? I think so! we have a winner, folks.

You are all so clever here I really don't know why I bother..........

c'mon, I'd expect another aussie to understand sarcasm better than you do, we eat it for breakfast. its rare I get to use it, since nobody over here gets it.

do you plan on doing a reveal on the internals of your system sometime (in another thread?)

And you can put a fiat engine into a vette - that will also work and probably be just as funny.

5amps at 120VAC is giant! On my project, I'm running 7amps at 48VDC and coils can draw more (resetting the power supply), unless you limit them, which is why I have resistors and caps inline from the power supply output.

You are confusing the current that can be supplied by the wall outlet with the current that can be supplied at the OUTPUT of your stepdown device REGARDLESS if it is a power transformer or a switching power supply.

I am not saying that a switching supply won't work, what I am saying is that you would be far better using a mains transformer with a big bridge and a large cap.

Look at it this way. You are experimenting right? You are very likely going to put the wrong wires in the wrong place at some point and if you do, BANG will go your switching supply. A mains transformer will (in almost every case) suck it up and survive until the mains fuse pops.

How many transformers in all of the hundreds of thousands of commercial pinball machines out there actually give trouble? Practically NONE!

How many phone chargers, printers, laptop power supplies etc etc etc have you chucked out because they 'stopped' (all using switching supplies, granted of varying quality and cost)?

It's your call and maybe it doesn't matter much as you are only doing it for fun.

It seems pretty clear to me when two engineers who have commented in this thread (one well respected here and the other mostly ignored) both are basically telling you the very same thing and the ones that suggest switching supplies seems to be hobbyists who have obviously had a good run using switching supplies for this purpose BUT that doesn't make it the right thing to do.

Nothing really wrong with a switchmode power supply. they tend to be lighter, and more efficient. But yea, it's really hard to kill a linear transformer (unless you short the hell out of it and melt through the coating on the wiring). Of course, so long as you aren't shorting wires (or have a fuse to protect), shouldn't be a problem. I bought my power supply (used for servos) off ebay for $30.

Both types of power supplies will work, but Homepin and G-P-E are correct that a transformer setup will be way more forgiving and less technically complex to implement.

IMO, If you go through with the 48V switching power supply, the way to control any coil regardless of size/resistance/etc.... is to drive them via PWM. That gives you enough control to modify the power for most any situation you are going to run into.

pwm the digital signal. fire it for 100ms or to eos etc, then duty cycle it on for 35ms off for 200ms.. rinse and repeat etc.

You let it kick full power for half a second, then use PWM for holding power (so you don't need a separate 12v supply and a separate coil winding). You're pulsing the 48v supply to give equivalent of 12v (sort of like how you pulse LED to dim them). The only drawback is that PWM'ing makes a buzzing noise.

If the OP wants to make something just for fun I think there is a lot of "overthinking" going on here with PWM etc etc.

I would just use a standard "two winding" coil and a standard EOS switch so that the full 48V is switched to the high current coil and then (when the EOS switch opens) the low current coil is switched in series and the full 48V is then applied across both windings.

This works perfectly well for most EM and SS machines and I can't see why it won't work perfectly well for the OP.

All the coils will work fine from 48V there is no need to lower the voltage.

is this really a viable way to save money on a pinball machine? honest question.

i think it would take me at least 2 years to build and code a pin.

If you get a transformer, it'll probably have multiple taps. The last one I looked at had everything from 5V-30V, I'd assume a 50V transformer would also have some stuff in the middle, in case it turns out you don't want some coil running off 50V.

I just finished my simple one, ended up costing me ~$700 and took two years of nights and weekends, but half of that was me blowing up ICs, etc because I'm clumsy. If I were to build another one, I'd say that the time spent designing the playfield, art, and rules (the fun part!) would be as long as the time spent building it, probably less than a year total.