PLUG-AND-PLAY SWITCHING REGULATOR MOD FOR CLASSIC BALLY/STERN MACHINES

Judging interest on a plug-and-play mod for Classic Bally/Stern games that would upgrade the 5v in those games to use a switching regulator. This would be a plug-in board, that connected to the J3 header (power input) on the Solenoid Driver Board. This would cause the on-board linear 5v power supply and high voltage supply to go inactive. Less heat, less energy usage and makes an easy way to fix a board with a non-working 5v linear regulator circuit or upgrade it to something that's more efficient & has higher amperage capability.

I've updated the project page for this some since I first created this thread. Instead of this plug-in board only handling the displays, I'm at this point going to work on a board that takes care of eliminating the linear 5v supply entirely from the SDB circuit. This would dramatically reduce load on the 11.9vdc circuit and one of the larger culprits of heat in the backbox.

More on this project over at http://www.pinitech.com/lab/ballystern_leddisplay_powersupply.php

This is a really basic mock-up just for a picture Essentially it'd be a smaller board that has board-to-board connectors that plug into J3 on the Solenoid Driver Board. Then it has a header so you can plug the harness that normally plugs into J3 into that. Easy plug-in solution. Note: One of the 25-pin headers would have board-to-board connectors instead of male headers.

ballystern_vrboard_(resized).jpg

Price Estimate
Around $35.00 (may be slightly more/less). Estimate was updated now that I'm talking about eliminating the SDB's on-board linear supply completely. It could still be about what I originally estimated ($30 or so) but will need to see what the material costs end up at. Regardless, I think I think that'd be a great price for some peace-of-mind that this would achieve when using some of the less-efficient aftermarket led sets.

Looking For Beta Testers!
If all goes well with the prototype as far as testing & no other issues are found, I'll be looking for a handful of people to beta test these. So let me know via this thread if you'd like to be a beta tester.

Interested? Get signed up on the Waitlist!
Let me know in the comments & also sign up on the waitlist over at http://www.pinitech.com/lab.php

---
http://www.pinitech.com - "Pinball Inspired Technology"
Kits, upgrades and test equipment for pinball machines

Yep, this is where I think some people have been concerned..

4x "less-efficient" aftermarket displays in a machine, running display test. Solid 5.0v off the boards (meter in the background) and the FLUKE setup to measure "MAX" reading of amperage pegged it at near 2.5A load on the 5v regulator. And that's just what's "measurable" -- technically it's about 3A load spiking the regulator (just happening quick due to multiplexing). That's been my understanding of it anyway. So.. add an additional display..or 3x for 6 player games..and.. yeah

4displays_(resized).jpg

Not a problem at all, happy to have the feedback!

In terms of a machine that already has an led display set installed, the 11.9vdc load should actually be less since the linear regulator is inefficiently using the 11.9vdc to generate the same +5v output for the displays that the switching regulator would.

I agree, connectors are the most troublesome points. The connections that seem to be the larger concern (as far as amperage running through them) would be the +5vdc, +11.9vdc and GND. Some of those are going to be a problem too regardless, unless the recommended bulletproofing mods are done.

The displays are getting their voltage from J3 pin #17.. that's what made this type of mod appealing since you can break their normal connection to the +5v linear (at J3 pin #17) and substitute the EZSBC.

None of the aftermarket displays (to my knowledge) allow for an alternate source for the 5v. Some of them [or maybe just one of them right now] are more efficient than the others, think we all know which

Connections at J3 of any substantial current would be the +11.9vdc / +5vdc. Regular 0.1 contacts are rated at 3A, whereas the board-to-board connectors are 2.5A per pin. I'd like to think tying pin #11 & pin #12 of the board-to-board connector together (via traces on the plug-in board) would help make up the difference there. I'll do some load testing on the 11.9vdc connection when the 5vdc is at 2.5A .. it'll be interesting to see where that's at with the switching regulator & then maybe compare that to when the linear is in-place. From what I understand the linear is less than 50% efficient whereas the switcher would be 80-90% efficient. So quite a bit more load on the 11.9vdc should be taking place with the linear.

One other recent development..

Daniel @ EZSBC announced he actually has a 5A switching regulator available now (model PSU7) in this thread: https://pinside.com/pinball/forum/topic/lets-get-technical-ballystern-led-display-power-consumption

For people that have the capability to upgrade or have the SDB out for other repairs, it probably makes the most sense to just upgrade the linear 5v regulator with the PSU7 (http://www.ezsbc.com/index.php/products/psu7.html). Really great news for pinball/arcade to have a 5A regulator replacement available now.

I'd imagine this "plug-and-play" mod I'm judging interest on in this thread will still have some interest as an alternative solution (ie. if you aren't going to swap out the linear off the board.. or don't have the capability). Still.. seeing how there is now a 5A switching regulator available, I may be better off seeing what additional interest is generated on this thread & via the waitlist before getting in too deep. Maybe just order up some prototypes & if all goes well, send a few out to people that want to test. Then take it from there.

Ran some tests on the 11.9vdc load with both the EZSBC (3A PSU5 LM323K replacement) and the linear regulator (78H05 5A). Significantly less current when the EZSBC was used. More on that over in my other thread -- https://pinside.com/pinball/forum/topic/lets-get-technical-ballystern-led-display-power-consumption

EZSBC PSU5 @ 2.5A 5v output, 11.9v draw was about 1A
Linear 78H05 @ 2.5A 5v output, 11.9v draw was around the same 2.5A

Which means.. in a machine without any mods done and 5 of these types of displays (again, picking on the more power hungry of the bunch) would be near 2.7A measurable current (calculated spikes of 3.35A). With 6-player machines that have 7 displays it'd be exceeding 3A measurable (calculated spikes of 4A).

In terms of the "weakest link" -- I'd say it's the single wire bringing over the 11.9vdc from the transformer. With the 0.1 terminals rated at 3A.. having led sets adding near or over 3A load means the only way pin #12 at J3 bringing over the 11.9vdc isn't getting toasty is because it's sharing load with pin #11 (11.9vdc input to the MPU) and the board traces/header connecting those two pins.

The way I'm looking at it right now, with all this info.. is a plug-in board dedicated for just the displays would allow the linear to supply the typical 1.6A (maybe as high as 1.8A depending on the game) to the machine's boards. The extra load for the displays going through the EZSBC would drop additional current requirements on the 11.9vdc. Instead of supplying 2.5A from the 11.9vdc in the scenario above, it'd probably be more around 2A [measurable]. Probably a bit more (2.3A load on the 11.9vdc maybe?) that can't be measured due to the multiplexing.

Another thought here I've had for the plug-in board, perhaps a better one.. is just to have it replace the linear all-together.

I'm not sure if the 11,700uf filter cap normally on the 12vdc would still be needed with the EZSBC's.. but could have 2x of the EZSBC 3A to-220 regulators on the plug-in board.. one handling the machine load, the other handling the display load. Just beef up 5v, GND and 11.9vdc traces on the plug-in board.. probably add some other type of large ground connection or two on a ring terminal that can be bolted onto the solenoid board's ground bolts. Then you don't have the 11.9vdc running through the SDB at all. You don't have to do anything with the caps for the high voltage or 5v linear regulator.. or mess with any of those components. The only thing that gets used / powered on the SDB are the circuits used to activate solenoids. So 5v still needs to go to the board (and ridiculously it needs 800mA or more for the pull-up resistors on the NPN solenoid transistor bases).

Then the load on the 11.9vdc line drops to around 1A as shown in the photo below where the on-board regulator was replaced with the EZSBC and 4x led displays are going. In that sense.. it'd significantly drop load on the 11.9vdc from the transformer.

EZSBC 3A regulator used to power the machine's 5v + 4x led displays
ezsbc_displaytest_4displays_(resized).jpg

I'm going to work on a 2x plug-in regulator design that'd render the original linear on the SDB as inactive. I have a thought on how it could be mounted pretty securely on the SDB as well, if that's needed. Worst case, I have another neat test board to plug in a machine when it's doing something odd.

And as far as keeping the large filter capacitor on the 11.9vdc in the chain.. it doesn't seem it'd be needed. The EZSBC website says no external capacitors required and additional input capacitors do no harm. So.. that would make for a pretty light and simple solution.. you gain efficiency and less energy use by leaving your old ratty linear regulator on the board, leave the old filter cap there.. and pull the fuse for the high voltage as well.

Just thinking out loud myself too

But yeah after some consideration, why leave the original linear active at all? Leaving it just means it's pulling 1.6A through the 11.9vdc connection.. so then you'd add the extra amperage (maybe 400mA measurable) for the EZSBC to power the displays, and that's where the board-to-board connector pins being 2.5A starts to look a little dicier (but again, pins #11, #12 of J3 are connected together.. at the header.. on the board.. so load is split *some* between those pins). But without the linear as part of the equation (ie. powering 5v entirely from switching regulators on the plug-in board) that same 2.5A is only drawing 1A from the 11.9vdc line. Then 5v output splits out from there. The SDB gets 800mA or so. The MPU gets 500mA on its own wires. Displays get 1A in most cases (the power hungry displays might spike at 1.75A). The Lamp Driver Board has its own 5v wire, and that is light on mA (maybe a few hundred mA or less, haven't measured that in the machine). Sound boards are probably *relatively light*.. maybe a few hundred mA on something like the Stern SB-100/SB-300 but again, haven't tested.

In the end it's 1.0A draw on the 11.9vdc line, in this fairly "worst-case" scenario I've been testing, in order to supply 2.5A 5vdc (that is split out to MPU, Lamp Driver, SDB and Displays). That scenario (fully replacing the linear regulator) actually makes me feel better about the idea. Nowhere near 2.5A max of those board-to-board connector pins on any one connection point.

Here's a downside.. a plug-in mod with the 190v disconnected means you can't mix and match plasma & led displays. I'm okay with that though. Personally if I was buying a product, I'd rather know it ditches the 190v supply entirely because I'd go FULL LED in the machine. I can pull the fuses at the rectifier board (high voltage would already be cut-off to the SDB since the plug-in wouldn't carry over the connection). The 190v connection *could* be passed through.. but I'd just assume keep it off the board entirely and have it double as a dummy-proof way of shutting down someone's high voltage rectifier if they leave the fuses in-place.

What I think's cool/appealing about this is.. instead of worrying about fixing up a SDB that has aging large filter caps on the high voltage or unfiltered 11.9vdc.. or any aging components really, you're completely shutting down the 11.9vdc & high voltage regulators. It'd actually be a fairly cost effective way to upgrade a board versus even buying the caps / rebuild kits. No 190v will go to the high voltage regulator. No 11.9vdc will go to the input on the 5v regulator. I like it

Ah that's good. Really haven't looked a ton myself. The Stern SB-100 gets its supply for U1-U5 thru the MPU header. 11.9vdc unregulated also comes into the board for the amps.. and then there's another 5v regulator on the board for powering the ICs that are part of the sound section. So I guess Stern was smart enough there as well.. heh.

Yep. Those are the two solutions. The plug-in mod has the advantage of not having to rebuild/replace any of the aging components in the linear supply circuits on the SDB at all unless there's a header issue, or problem with something in the solenoid-driving circuitry.

Just hadn't planned on getting into a switching regulator design, but it would certainly reduce BOM cost.