John's Arcade Forum - Classic Arcade and Pinball Collecting and Restoring Discussion Forum - RETRO MAME - Nintendo Vs Forum

Hey John,

Just listening to your latest Podcast... :-)
#28

About your flashing Leds...
Ok, so remember these are effectively diodes, therefore are polarised (like an electrolytic capacitor).
i.e. *must* go in the correct way around.  Google it up.  Often then have a kink (oooerr) or a longer leg to denote the anode (+) connection.
The cathode (short led) shound connect to 0v/ground.  The ground for switches, etc. should be ok, unless there is some mad wiring somewhere?
 
As for Pinball...of course the original lights were bulbs, which can connect any way around as they are not polarised (so basically a resistor).

I am lucky, as I have done electronics since an early age (11), now I am an old feccker at 43! ;-)


Good luck dude,
Cheers,
Dave.

He has them in the right way. The issue is ether the LED needs too high of a voltage for the blink signal or it is over loading the circuitry that drives it. I believe he is going to try a different LED bulb that has a lower voltage requirement but if that doesn't work I think it is going to be time for a FET solution.

I am listening to the podcast and frustrated listening to it about the blinking LEDs. John, the way those cone buttons work is probably a BJT transistor in there. The reason why you need both +5V and Blink is because +5V is the power and Blink is a signal that turns on and off the BJT. You really shouldn't try to power a LED directly off that blink wire, you are going to overload the PCB circuit that drives that.

Oh, really?!?! So I shouldn't try it?!?

I am just nervous to do it long term. The buffer chip that is driving the cone button is probably only rated for a few 10's of mA's. When you overload the buffer it is ether protecting it self from too much current or thermally shutting down. While you may be able to get a bright LED to blink directly off the line you could speed up the death of that chip. I actually just got an atari game that has the cone button today so I can figure out how they work this weekend.

Does this make sense?!?!? Some guy on Youtube made it.

http://picpaste.com/tempest_havoc-G7ixzoXO.png


This guy also says this:

"As suggested on the previous video: you want to use only the RED and WHITE wires... not the BLACK!  The logic board cannot drive enough current, but it might be able to sink it.  Also, the RED and WHITE is what the cone button uses, so when the cone button is lit so is the roller.  If you use the BLACK and WHITE then the roller and cone flash the opposite way to each other.  Key message: DON'T USE THE BLACK WIRE AT ALL!"

And, this...

"If the roller LED won't light with just the red and white, then try temporarily removing the white from START 1.  That way, you won't be limited by the START 1 LED's forward voltage.  You can also try flipping the red and white to make sure the polarity is right for the roller LED.  If all this still doesn't work then those 150 ohm resistors are too limiting.  In that case, locate the conversion board, find the common point of the two 150 ohm resistors R30 and R31 (docs are online), and bring a new wire from there to your roller bulb holder, the other side of the bulb holder going to +5v.  Then you'll get full flasher current to your roller bulb (Q18 can sink several amps), and could use either a bulb or a LED of your choice.  BTW, that's what the original Major Havoc does... connects the roller bulb directly to the signal called LIGHT."

well kind of not what you want to hear but why not move the flashing circuit off the 1p / 2p start buttons and dedicate it back to the roller as the start buttons weren't flashing originally on this game if I interpreted everything correctly...

from there you can work on a mod to get the 1p / 2p flashing...

John what he saying makes some sense. Most often things can sink more current than they can supply. You will still need to be careful about  how much current you use though.

I really think it is best to just use a transistor driven relay. Then you don't run the risk of damage.

A circuit like this is simple and will ensure you don't damage your expensive board.

(http://i.stack.imgur.com/Oarz3.gif)

i'm not sure why you would add in a relay to this circuit. the Rds of a nice 10A logic level fet would be fairly comparable to that relay.

The relay isolates the driving voltage, but I agree that it isn't using a lot of current.

I don't know what to do now. I can't make it work. Someone give me some direction! :)

Isolates the driving voltage? Driving voltage for the LED? Driving voltage for the LED is going to be +5V or +12V and both of those are well below standard FET voltages. Also I typically don't drive relays with a FET unless the coil is quite large.

I agree a relay for this seems overkill, I would imagine you should be able to find a diode(s) that could be inserted inline the 5v or 12v path to do a controlled flash, in my experience only time you need to relay is when you need something to carry the load of the operation ( mind you, most all my electrical knowledge I am referring to is 12v electrics in Volkswagen / Porsche restorations  ;D )

Relays are cheap and easy to use. FETs take a bit more knowledge. Also Resistance of a relay is usually pretty low (only beat by some fantastic power FETs).

John - I will get something worked up for you this week. It may be crude because I don't feel like waiting 2 weeks for OSHpark.

My 2c is that John needs to make the light flash in unison with the 1st and 2nd player lights... "anything less would be uncivilized."

Someone on Youtube said this:

Try using a mosfet. Attach the gate to the pulsing wire, source to the always on wire, and drain to the bulb. An IRF510 would work great!

That sounds easy.

So, just get this and I am good to go?

http://www.ebay.com/itm/5-x-IRF510N-IRF510-Power-MOSFET-N-Channel-5-6A-100V-/220893759353?pt=LH_DefaultDomain_0&hash=item336e4b4779

Yeah that is what i was saying to do lol. The only thing is, I wouldn't use the IRF510. Let me find one real quick

And keep the ground wire on the bulb?

Okay nevermind on the IRF510. I was going to try to pick out a better FET but i realized that was a huge waste of time for how little current you are going to use. Ill scan something real quick

(http://i.imgur.com/noQSG21.jpg) (http://imgur.com/noQSG21)

Is that up arrow thing the led? Lol.  In English?

well the thing in the middle is the IRF510. I didnt draw it schematic like that one you should understand. The arrow thing is the LED.

So....

5v goes direct to the led and then the other led lead to the MOSFET? Then one leg of MOSFET to ground? The other to pulsing 5v?

yep.

Just curious... That doesn't match what the other guy said. Does it?

No it doesn't match. You wouldn't be able to turn on the mosfet if you attach the source to the LED. Its a bit complicated so I'll spare you that. But the gist is, N-type fet's (IRF510) have to be on the low side of the load, P-type's have to be on the high side (as described by the other poster)

Cool! Thanks. Looks like radioshack has the MOSFET. I'll grab one.

Ok, I stopped at Radio Shack on the way home from work and grabbed a MOSFET.

And....

IT WORKS!!

Couple of things:

#1 - I cannot use the non-ghosting LEDs. If I do, the roller light goes OFF when I start the game. The voltage must be dropping below 4V.
#2 - The roller flashes the exact OPPOSITE way the 1p+2p cone buttons do. Not horrible, I can live with it.
#3 - During the attract mode the roller pulses from DIM to BRIGHT, DIM to BRIGHT. It's never 100% off. It still looks cool. However, when I start the game it goes to 100% lit but DIM. Still looks good.

Ian, what do you think of all this?

Awesome John! This deserves the bear:

(http://me.tighelory.com/images/animated/bear.gif)

Check the voltages if you can. It must not be going all the way to ground when the led is suppose to be off.  As for it being opposite that is really lame and means they are using negative logic. There isn't too much you can do for that other than a more complicated circuit or using a different fet which I'm not sure radio shack will have. I'm going to think about this some more and might be able to come up with something useful.

Couldn't he just use an NPN transistor and two resistors to invert the signal? Would that do it?

Like this?
(http://i.stack.imgur.com/KntFP.jpg)

You could. Kindof. First I would have to find a value for r that would work and with the fet not turning off all the way is going to cause issues

just a thought, if you installed the mosfet infront of the 1p 2p buttons and roller, they all should blink at the same time, right now in your current configuration the mosfet is after the signals the 1p 2p buttons are receiving...

I made a comment on the video about this. I told John to use a pchannel mosfet on the highside to drive the bulb. It may have been possible to do this without the mosfet but not with a light bulb only with an led.

(http://i.imgur.com/5Evif68.jpg?1) (http://imgur.com/5Evif68)

use a P-MOS transistor

I tried this tonight. It doesn't work. :(

It goes from BRIGHT to kind of BRIGHT. It doesn't turn off at all. Even with the non-ghosting. It's like it's going from 5V to 4.5V. It must not be going below 4V as the non-ghosting LED doesn't even turn off. Ugh! :)

I say try the npn transistor like I suggested, it will invert the signal.

(http://i.stack.imgur.com/KntFP.jpg)

You can get everything at radio shack.

Okay so I've figured this thing out. I really should have looked into the schematic.


on  the PCB there is a latch that is attached to a 120Ohm resistor on the anode of the Cone LED
The Blink signal is attached to the anode of the Cone LED. What this means is that when the Latch is LOW the voltage at Blink will be equal to 5V-LED Vf which works out to be around 3.3V
So the Blink signal is actually going from roughly 4.7V to 3.3V, It never goes all the way to ground because of the voltage drop across the LED.

To deal with this you need to put a trigger in that only occurs when the Blink voltage is above 4V.

so Yeah in a way I am doing the inverter with the NPN but the important bit is the voltage divider driving the  NPN.

(http://i.imgur.com/r6elBX3.png) (http://imgur.com/r6elBX3)

Is if I am understanding what you have here you are using the npn (Q1) to invert the signal and the MOSFET to regulate the voltage?

I only choose to invert the signal to use the better n channel mosfet. I could have done this circuit with the p channel non inverting. What matters is the voltage divider that makes 4v be .65v at the base of the npn. The npn won't turn on until the voltage is above .65v at the base. So effectively what I am doing is making sure the mosfet gets turned off properly when the voltage drops.

John needs to get himself a breadboard!

Was this ever resolved and if so, was a follow-up video made?

No follow up video but this circuit has long since been completed if you are interested for yourself.

I am very interested!  Was it made to blink in time with the P1/P2 start buttons (vs. alternating like the last video showed)?  What parts would I need to buy (other than the LED)?  It looks like the MOSFET has been changed vs. the one mentioned in the video?

I never got this to work.