NOTE: Discussion continued at http://pinside.com/pinball/forum/topic/led-ocd-advanced-lamp-controller-shipping-with-wait-list

LED OCD is an add-on lamp controller that solves the following problems.

- Eliminate the flicker or strobe effect when using LEDs in a pinball machine. This is accomplished by increasing the lamp matrix scan rate.
- Allow brightness to be controlled. "Profiles" are configured to define the on-time of lamps, effectively controlling the brightness.
- Simulate the fade on/off of incandescent lamps. By gradually changing the brightness of the bulbs, LEDs are made to fade on/off instead of their typical abrupt changes.
- Allow ghosting to be completely eliminated through a combination of additional hardware and improved transistor control.

Here's video of it in action:

The second one is a video showing me configuring it. It's quite outdated at this point, and the software interface has changed, but it gives you an idea what it can do. This one you'll have to watch on a big screen in 1080p to get much out of it.

Sorry for the formatting. I can't figure out how to get the videos to embed nicely.

It currently only works with WPC machines. I'm looking into older Williams, Data East, etc., but I'm not there yet.

_{Post edited by herg : Add summary of project}

Standard bulbs. There's a mix of CT frosted, CT premium, BCs Frosted, Ablaze, and incandescent in the videos. The cheaper LEDs actually work better since there's not extra stuff in the bulb like the premium LEDs.

Yep, customize on the PC, upload and save to the PCB, then disconnect the PC.

Maybe after I've thoroughly tested it myself. There's only one board built up at the moment. More parts are on the way.

Yeah, making LEDs look more like incandescent bulbs is the idea.

Due the the instant on/off nature of LEDs, many people see flicker or a strobe effect when the lamp matrix is scanned at 62 Hz. This PCB monitors the outputs from the lamp driver, then re-drives at 250 Hz. That's the flicker part.

Dimming is controlled by varying the lamp row active time based on the configuration from the PC. That's the dimming part.

The PCB then ramps up/down through the configured brightness values in order to make the LEDs fade on/off like incandescent bulbs.

At the moment, it's WPC only. The work to adapt to other manufacturers/boardsets would be minimal, but I only have WPC machines to test in. Also, new PCB layouts would be necessary to fit the mechanicals of the other machines.

I don't have LEDs in my AFM yet, but I'm hoping it will help with the way the lamps pulse. With all the blinking lamps being synchronized the way they are in AFM, it really drags down the lamp matrix voltage. LEDs should help.

Arduino is nice if you're just looking for a simple controller board. In this case, I'd still have to add the driver transistors, molex connectors, etc. There's also not enough I/O, nor would it screw nicely into the backbox. IMG_2995.JPG IMG_2998.JPG IMG_3000.JPG

Thanks, but it will be a while before I will be moving on to Stern. Getting WPC working 100% and deciding if it's feasible to sell is my top priority.

C, though there's a lot of direct hardware access. It's only somewhat like C in a PC environment. I used an EasyPIC7 dev board early on, and I'm still using the PIC18F45K22. Nice and cheap, but it does the job well. The most expensive part is the USB chip.

Nope, just the matrix.

You don't even know how much it would cost yet (nor do I). The biggest hurdle as far as cost is assembly. The quotes I've gotten are quite high. Our lab tech at work has agreed to help if needed, but even the two of us would not be able to crank out very many.

It's quite different actually. Sure pulsing the LEDs to make them glow smoothly is the same, but monitoring the GI triacs, rectification, string vs matrix is all completely different. Unless you require that the GI is all LED, you'd have to switch a lot of current. The software would be a lot easier, but the hardware would not. I have no plans to work on it.

We're jumping the gun a bit talking about System 11, assembly, etc. The next step is to wait for parts to arrive and build up a few more. That way, I can quit swapping this one between my machines and be able to get more repetitions in. It's not going in anyone else's machines until I'm convinced it's solid.

Assembly isn't bad, other than the USB chip/connector. I was impatient and did the first one at home without a microscope, but I won't be doing it again. Short of those parts and a few SMT capacitors, the rest of it is though-hole. I'm not sure yet how I feel about kits, though. If someone were to damage their machine, I'd feel like crap.

As far as putting it in a System 11 game, I haven't looked at it yet, but I think it would limited to mechanical changes. If you mounted the board somewhere and build cables that adapted everything, that might be enough to get it working. You'd be on your own doing that, though. Until I get a System 11 machine (anyone have a Pinbot) I won't be supporting it.

Any mix of bulbs is fine, and you can configure different bulbs to react differently. The TOTAN in the video is LED in the smoke and jewel inserts, incandescent in the rest. The blue LEDs are blindingly bright, so they're set to 24% max brightness. Actually, the smoke with sun, moon, etc are set brighter than that now that I think of it. The jewels are at like 60% max. The image at the end of the video shows the configuration app, but there's a lot of info to absorb. I might make a video showing a typical configuration process if I end up with time to do it.

Here's a video of me adjusting the settings after replacing a few incandescent bulbs with LEDs. If possible, you'll want to watch it on a larger screen in 1080p. Otherwise, you won't be able to see much.

I'm ready to have someone else test it at this point. A local friend has volunteered to test in his machines, but having another person somewhat more detached from me would be helpful.

This person must be technically proficient. Understanding how the lamp matrix currently works will be helpful. LED OCD is not a product yet. It's possible for things to go wrong. Obviously, this person will need to report any findings to me. Finally, owning a WPC machine (preferably more than one) is required.

It also will not fit with a Flipper Fidelity kit (probably Pinball Pro as well) installed. This was an oversight on my part, and I am working to come up with a different mounting solution. Please keep this in mind if you volunteer for beta testing.

NOTE: I'm covered for testing at this point. Thanks for the response.

_{Post edited by herg : Already have enough testers}

It has been a while, so here's a quick update:

- I have four of them in my machines, and they're staying there.
- PT has one. He has installed it in his No Fear.
- One is ready to be tested in a System 11 (with interconnect) once the owner of the machine is ready.
- I've nearly completed the board modifications, and I intend to have a bigger batch of PCBs made in October.
- I intend for my TOTAN (including the board) to be available for play at MrWizzo's NoVA pin party this Saturday, Sept 29.
- I've created a website using my mid 90s web skills that mostly just repeats this thread.
- I don't want to do pre-orders, but if you're interested in getting one, contact me through the website or PM. Having an idea of how many to make would be really helpful.

http://herg.homeip.net/ledocd/

The ROM patches attempt to fix the ghosting issue, but do nothing for brightness control, flicker, or fade. Obviously, I'm biased, but there's no comparison, IMO.

I haven't even looked at Capcom, so I don't know the difficulty level. It's certainly a good way down the road.

Settings are based more on the bulbs you choose to use, so presets are impossible. The default settings will improve over a game without the board, but you'll have to configure for best results.

That's great to hear, and thanks for the help.

They don't vary the voltage. The duty cycle is varied by turning off each of the row transistors at the appropriate time during the column active time. Think PWM, only more complicated.

Thanks for the info. I'm concentrating on getting WPC finished before moving onto other machines, though. Well, WPC and older Williams since those should just be different cables with no board changes necessary.

Close, but it's not limited to on/off during a column active time. At 250 Hz, with 8 columns, that is 500 usec per column strobe. To light an LED at 50% brightness, it turns off the row transistor after 250 usec. For 25%, after 125 usec, etc. That way, the LED continues to be pulsed every 4 msec, and there is no weird cadence created by trying to light the LED... say 3/10 times to get 30%.

GI is not rectified, it's just straight AC. The smoothing technique Crash mentions is common to use on rectified AC to smooth it into DC, and there's two big caps on the WPC power driver that does this for the lamp matrix.

Crash's top pic is full-wave rectified, but that's not what the GI power looks like. You could build a board with bridges and capacitors to get smooth DC for GI, but LED OCD does not attack this issue. It does not touch GI at all.

The premium LEDs have small scale versions of the full-wave rectifier/capacitor circuit, and are actually a good solution to GI flicker due to AC. GI flicker due to attempted dimming by the power driver is another issue altogether. I haven't looked into it, but I think you'd need bigger caps than would fit in a bulb.

I think the best solution to the GI problem (that would still work with an original power driver) would be an all hardware solution with rectifiers, smoothing caps, and an adjustable switching regulator. You'd end up with a lot of hardware, though.

Again, this is slightly off-topic since LED OCD does NOT address GI.

There's a section on the usage page of the website that goes into detail on this, but in short, certain combinations of bulb types and settings can still result in ghosting, but you can work around it with corrected settings. There are no false triggers.

Not quite that simple. While using a mix of incandescent and LEDs, setting the incandescents to 85% has eliminated all ghosting during my testing.

I can't say for sure without doing a specific test, but I have not noticed any issues with setting LEDs to 100%.

Edit: One issue with setting the LEDs to 100% is that I can't see to play the game due to the intense brightness.

OK, I did some tests with CT Frosted LEDs set to 100%. The only case I saw ghosting was a 2 bulb situation. 2 LEDs in lamp #35 set to 100% caused ghosting on lamp #45. I decreased lamp #35 to 90%, and the ghosting at #45 was very close to gone. At 86%, it was completely gone. This is likely dependent on the type of LED also.

Nope, CT Premiums work fine. The extra circuitry inside them causes them to not light at all until about 40% duty cycle, but if you set the minimum brightness for them accordingly, they will fade on/off just fine. Since they jump from off to 40%, it's not as smooth as other bulbs, but it's still an improvement over using them without the board.

If you already have a machine full of Premiums, rest assured you won't be forced to choose between not being able to use the board and swapping out the bulbs. If you're buying bulbs to use with the board, however, they would not be my first choice. My personal favorite so far are CT Frosted. They're relatively inexpensive, plenty bright, and work fantastically with the board.

I'm working on it. Production PCBs are on order and should be here in a few weeks. A little time to build some up and test, and I'd estimate a small number will be ready for sale by mid November if all goes well. Once they're ready, I'll be going down the list of people who have PM'd me to offer them for sale.

Yes, it has been thoroughly tested with WPC-89, and I have been using it in my HS2 for the past couple months.

I can't say what it's worth to you. That's for you to decide, though saying it "just fades LEDs" is selling it quite short.

I can say that based on my best estimates, I'm going to have to sell at around that price point to make it worthwhile for me to do. I'm not going to get rich doing it. I hope that I might be able to eventually break even on what I've spent developing it. If not for the love of pinball, this never would have happened.

On a side note, this is now bordering on "Marketplace" discussion. If it needs to be moved, feel free to do so. Can I move a thread I started?

Dan,

Thanks for the encouraging words. I'll put your name on the list, but F-14 is a System 11 game, and I have yet to be able to test with a System 11. Soon, I hope.

Or 1.55 Euro or whatever they cost for the NoFlix PLUS bulbs.

Thanks for the info. I'll look into it. Though the general idea has been discussed quite a bit since LEDs started being used in pins, this particular implementation is different than I have seen discussed elsewhere.

The software on the microcontroller is what I believe is truly unique, and that is both copy protected and copyrighted. It sucks that in a closed community like pinball that I even have to worry about it, but I guess that's life. We've seen it happen before.

I also have a design notebook and configuration management tool to track all software and hardware changes.

The number of threads discussing glowing cats eyes is quickly approaching the number of WOZ threads.

- Production boards have been received. I've built and tested one. Building more in parallel to speed things up.
- Cable sets for the first batch have been built, but I'm still waiting on keying plugs for the connectors due to me ordering the wrong quantity.
- A shipment of USB cables and another of packing materials is being held up by Sandy complications.
- I should be able to start shipping small quantities within the next week or so.

No worries. You're at the top of the list, and the minor stuff I listed is all I'm waiting before I can ship yours.

I could actually ship you one now, but it would be missing the keys in the connectors to prevent you plugging it in wrong, you'd have to use your own USB cable, and it would arrive in a Franken-berry box.

Uh, you do know it doesn't support Stern machines yet, right?

It uses a mini-B USB cable (I'm including a 15ft one), and I'm waiting on both .156 and .100 plugs.

EDIT:
I see you mentioned Rollergames a while back and that your collection doesn't include any WPC games. It has only been tested with WPC so far. I'm sorry.

_{Post edited by herg : Checked Frax messages}

A small number of kits are shipping today, and I think I'm just as excited as the guys who will be receiving them.

Here's a couple pictures of the finished boards.

IMG_4448.JPG IMG_4100_2.JPG

No, those caps are to smooth the input voltage. Since the output is a matrix, it cannot be smoothed using caps at the board.

Discussion continued at http://pinside.com/pinball/forum/topic/led-ocd-advanced-lamp-controller-shipping-with-wait-list