Great news man. can’t wait to throw one of these in my hrap3
did you ever try the zener diode route to run it on 3v for maximum compatibility?
I did find a 3.3v zener, and ran some simulations in LTspice to check how it would do. No matter how I configured the shunt, any setup that would allow 60mA (15mA for four LEDs) at 3.3ish volts would end up draining like 250mA+ when connected to 5v source. If you can think of a setup that would allow it to work as expected within a reasonable tolerace, and not expend more than 50mA through the shunt, I’m all ears.
Since the boards will be preassembled, I can safely put in a surface mount DC-DC and components onto the board, whenever I get to that point.
I might be able to help you solve that issue, gimmie a few days and perhaps I can come up with a quick schema.
What’s the requirements?
3.3 (psx) - 5 (usb) volts IN?
3.3 volts OUT?
Current draw = 60mA - what, 150ma?
Let me know, I don’t mind helping you on that front, circuit design is my specialty even though my only product so far has absolutely no components 
Output would be going to four LEDs in parallel, along with a current reducing resistor for each. The LEDs have a typical voltage drop of 1.2v, and should have a minimum current of 15mA each, maximum current of 20mA each. So you can setup the shunt and the LED resistors however you like. If the maximum current is anything over 100mA, I’d rather stick to a DC-DC step up converter. I just dont think its mathematically possible, since the shunt would have to drain more than 40mA at 1.7v over the zener rating.
I take it that you’re testing the new Flash PCB with the octo, square, and circular gates?
(Specifically talking about your circular gate in addition to the Sanwa JLF-standard gates…)
Engage distances are different amongst those.
If what you’re talking about with the LEDs is the case, wouldn’t the input results be dramatically different depending on the gate you install in the JLF?
Sounds like the prototype is moving a bit faster than you thought.
Good news! Not expecting to get mine before next year, though…
You’ve got time to perfect where I’m concerned… :china:
I realized that the oversesitivity may be because the receptors were pushed further away from the center; I have some older boards that dont have that, but I needed some more surface mount parts to put one together. So I’ll be trying with another board probably Thursday when those parts get here. So the next test will be sooner than it would have been with a new batch of pcb’s. I should still get those made though.
Toodles,
Just take the time you need to get this right… Most people will wait until it’s good-to-go.
And thanks for listening to the posters!
My best guess… a Dreamcast controller PCB, named after the Agetec “Green Goblin” stick.
I swear to the almighty fetus of gawd that this is one frustrating project…
Parts came in, I assembled one of the earlier boards that don’t have the receptors shoved back away from the center, and…hella flaky. Muther puss bucket crappy POS damned invisible light crap.
To pour salt in the wound, I ordered the new board last night. The one with the receptors back in the middle like this one.
My current conclusion I’m working on: the receptors pulled in DO help with light pollution. I swapped the last revision board in, and it remained perfectly solid with the exact same housing, so the housing and the recessed receptors are doing their job. Now I need to adjust where they’re activating. Which means I need to adjust each pair perpindicularly to move them further from center.
Which likely means another massive rewiring job of the board, and about 2 packs of cigarettes to wrap my head around the math involved to guesstimate how far to move them. Or make a massive panel of boards, each a slightly different amount of adjustment, assemble them all and test.
FML. Work continues.
:u: What I was thinkin too.
You’re gettin closer toodles. The end result will make all the stress tolerable.
Difference in spacing is so small as to be almost unnoticeable even right on top of each other. As near as my calipers can tell, distance between the up and down set of components is off by about 0.35mm. So each direction was 0.175mm too close to center. I was off by the thickness of an index card.
Wow, what a PITA. And I suppose trying to copy original flash-1 sensor placement would not help, seeing as how it uses a different type of sensor altogether? Perhaps that sensor type was more forgiving? Sorry for talking out of my ass, I know nothing and like to speculate, since I would like to purchase this toy very much 
Since you said you had wired one of these to a jamma I have a question that is somewhat related. I have an ASCII optical and plan on getting your Spark one day in the future. Now here is my question and hopefully you can answer it so I don’t have to visit a very dead section of neo-geo.com. I have a Rev.1 Hyper NeoGeo 64 which requires the +5 to be tapped in order to run from the MVS harness. Now if I wanted to hook any optical sticks up to this as well where could I tap it to power the stick without blowing out the Hyper which can die if you look at it funny. I’d be willing to test something with a much less expensive Happ360 but one day would like to put in a couple Sparks. Plus I want some of your expert advice since even though I can get another Hyper Board if needed the shipping from Europe is just insane.
Now you need to team up with NeoTurfMasta, smkdan, and Raz and revolutionize the custom game electronics world. (Unless you are one of those guys and just use a different name here on SRK)
I have nothing to add but to say that this thread is unbelievable. I can’t believe you are doing this. Toddles. I think I will buy one just to support your work when It become available just to support you. O.o;
v8.0 boards arrived this morning, assembled two of them and through them into the cabinet. Diagonals are a little better, but still way too tight. The cardinal directions feel like they activate at about 50-60% travel, instead of the 33% travel the original feels like. The layout of the traces and parts on the boards is just freaking beautiful though, so I’m definitely proud of that. Almost entirely 32mil traces, only 4 vias on the whole board, clean placement and routing. It may not make any difference to most everyone, but it’s like when you see a stick with uber clean wiring; that’s what looking at the board feels like to me.
So, I need to open the layout and start adjusting the placements some more. They were moved 0.1mm from the previous; I’m going to see about layout out 5-10 or so with additional adjustments of about 0.05mm each. I’m not too worried about overshooting, in fact that may be a good thing. It’d be pretty sweet to end up with a super sensitive layout, and make a small GT-C to match to low throw. No promises, but its something I’m definitely looking out for.
So, progress for sure, but not there yet.
v8.1, 8 different versions, each with with separation distance in 0.05mm increments. I grabbed the biggest separation one, threw it together, put it in the cabinet, and…damn, its perfectly playable. I put a second of the largest separation versions together for me to play on, and I just beat SFA3 arcade on it without any difficulty performing my moves. The first one is currently on loan to some locals who are throwing down today so I can get some unbiased input. Personally, I like it. The engage feels to be like about 40%, maybe less. It feels like it matches JLF microswitch board where the directions activate.
Now I’m wondering if I should make some different layouts with even further spaced LED and PT’s. The further spaced they are, the more sensitive, so I wonder how sensitive I can make it while still being usable.
Thats awesome news Toodles!!!
Got a job interview monday.
If things go my way I will buy prolly two of these things.
Nice, keep it up, Toodles