Skeeball

[galgon]

So after a few years of passively searching I recently picked up 2 used 13 foot skeeball machines (model S) for the basement.  Beyond needed some deep cleaning and light bulbs they work quite well.  They were in active arcade use prior to us buying them.  The scoreboard uses 1 watt #74 wedge based bulbs typically used in car dash board lights.  I had the bright idea of upgrading the bulbs to LED so that I would never have to change them again.

I searched online and found a place called super bright leds and picked up a couple of different types of the #74 led bulbs to see which worked better.  Quick testing and this one won out: You are not allowed to view links. Register or Login   During testing the lights blinked at the correct times, shutoff and generally worked good but I only had 5 LEDs in a scoreboard with 28 blubs.   

Fast forward to this week when I got the full set of bulbs in the mail.  I put them all in the scoreboard and now all of the lights are always on.  They briefly dim but you cannot read the score at all.  It occurs to me that the difference in wattage .2 vs. 1 is likely the problem.  Thus my question: any thoughts on how to solve this little issue other than switch back to the old bulbs? 

[vairmoose]

Suggestion...   and this is only a test.     Replace one of the LED bulbs with the original incandescent.    and see what happens then.     I had a similar situation in my kitchen.  I replaced all of the incandescent bulbs that were on a dimmer with new dimmable LED bulbs.  Prior to the change I could ramp the bulbs all the way down to zero light.   After replacing all of the bulbs, at the lowest setting on the slider, the bulbs were still on  (probably about 1/4 of original lumans).  If I turned power on with the switch at lowest setting, no light, but after moving the slider up, the lights would be on when moving the slider back down.   In the kitchen this was ok but I replaced one bulb with incandescent and I seemed to have full control again (but different appearance with the one bulb so changed back to all LED).       

Larry

[dmaccole]

Suggestion...   and this is only a test.     Replace one of the LED bulbs with the original incandescent.    and see what happens then.

+1.

\dmc

[galgon]

I tried the replacing one bulb test.  I replaced the center light on the right most set of lights.  In doing so this set of lights started working normally.  I believe this worked for these lights because the center one is never actually fully lit up.  During the awaiting player stage the numbers have a swirling 0 pattern where one bulb out of the 6 outside bulbs is out and the other 5 are on.  The center light is dimmed very low.  This worked fine for the right most number.  When playing the game you only score by 10s.  So the right number is always 0 and this looked fine. 

The other numbers were not affected by changing the one one bulb.  I tried changing the center bulb in the middle set of lights but this did not work as well.  The middle set works fine in the awaiting player stage.  But during the game if the center bulb is litup the other lights are also lit.  So a 0 and a 1 look fine but a 2 ends up looking like an 8.  The same behavior happens with the left most set of bulbs. 

Any thoughts on the next test?

[tbone321]

The problem is that this device was not designed to work with LED's and the dimmer circuits have way too much leakage for them.  Adding the conventional bulb to the circuit added enough of a load to pull the leakage voltage back down below what was keeping the LED's lit when they were supposed to be off.  You are going to need to find each one of these circuits and add a load to then to suck up the leakage current and of you are using bulbs to do it, when that bulb burns out, the rest of the LED's on that circuit are going to light back up again.

[galgon]

The problem is that this device was not designed to work with LED's and the dimmer circuits have way too much leakage for them.  Adding the conventional bulb to the circuit added enough of a load to pull the leakage voltage back down below what was keeping the LED's lit when they were supposed to be off.  You are going to need to find each one of these circuits and add a load to then to suck up the leakage current and of you are using bulbs to do it, when that bulb burns out, the rest of the LED's on that circuit are going to light back up again.

So if I understand you correctly I need to find a way to consume about 1 watt of power per circuit. This is assuming each one of the set of 7 lights making up a number is a circuit which the testing seemed to indicate.  This power consumption has be be after the dimmer switch controller which based on the diagram of the board wiring it looks like it is.   

So my two questions then are where in the diagram do I need to add load to?   And more importantly what should I do with those few watts of power that would be more fun than hiding a few bulbs behind the scoreboard?

[n1ist]

I'd try removing the resistor between pin 9 and ground on the ULN2003s.  Normally, the diodes on the 2003 outputs to pin 9 are used as flyback diodes for inductive loads, and pin 9 is tied to the load's supply voltage.  In this case, they are using them, along with the 330-ohm resistor to leak some current through the bulbs to keep the filaments warm, so they turn on faster.  LEDs draw much less current, so they are illuminating with this current.

That's my guess from looking at the schematic.
/mike

[myoung008]

I would try Mike's idea first, but if it doesn't work here are a few more solutions:

#1.  Try adding a 1K resistor to ground on each of the inputs to the darlington arrays (ULN2003A pins 1-7).  That is an attempt reduce the leakage current by forcing the transistors completely off rather than mostly off due to leakage in the 74C164.  You can try just a few to start out with, but it will only affect the circuits that have the resistor.

#2.  If that doesn't work add a resistor on each of the outputs (ULN2003 pins 10-16) to +12, 1/4 watt 2K should be fine, but experiment if need be.  These resistors will get warm when the corresponding bulb is on.  A lower value will waste more power but be more likely to fix the issue.  If you go below 0.7K you should look to 1/2 watt instead.

#3.  Of course, you could always replace the darlington arrays with discrete mosfets and be done with it, but that's quite a bit more work, and you may still need the pull down resistor from #1.  This is more of a nuclear option rather than just a simple circuit modification.  If you decide on this route it will probably be easiest to place the resistors and mosfets on a proto board and jumper to the main circuit after removing the ULN2003A chips.

Hope this helps.

Marc

[galgon]

Mike,

   Just to clarify should remove the resistor and add a jumper where the resistor was or just remove the resistor and do nothing else?

Thanks,
Joe

[n1ist]

Just remove the resistor.  Since the loads are not inductive, you can just leave pin 9  unconnected.  If the ULN2003 are socketed, just bend pin 9 out so it doesn't go into the socket.

/mike

[galgon]

Thanks for the help.  I desoldered the resistors (the uln2003s were not socketed although the 74c164s were).  This solved the majority of the issues.  The numbers are now clearly readable with no lights staying on like they had previously. 

However there are still a few behaviors that are different than the original lights. It seems that any time the scoreboard changes all of the lights very quickly light up.  You can see it when you play (when each of the 10 point switches are hit with a ball the non-lit LEDS briefly flash.  I can also see this when the unit is waiting for players.  In this mode the scoreboard has is showing 0s with one of the segments from the 0 not lit up.  The section moves around the 0 in a fairly quick circular pattern.  Typically the center light is not lit during this process but now the center LEDs quickly flash with each change.  Here is a short video showing the problem: You are not allowed to view links. Register or Login

What is odd is that previously when using all normal bulbs the center bulb would be barely lit during this mode.  The user could not see it but if you took the cover off the lights you could see the bulb faintly lit up.  Any thoughts on this one?  I included a pic of the board with the resistors removed.

[n1ist]

The center bulb being barely lit was caused by that resistor; by keeping the filaments warm, it reduced turnon time, and made the bulbs last longer by reducing thermal shock when turning them on.  As for the brief flash, it sounds like there's some sort of glitch in whatever is driving that board that was narrow enough that a lightbulb didn't light (they need time to get hot enough to glow) but the LEDs will turn on instantly.  This may not be a faulty part; it may be a software issue that didn't used to cause a problem but now is visible.

/mike

[tbone321]

I would say the these are not problems, they are part of the design.  This system simply was not designed to work with LED bulbs.  They use much lower current which means that they don't put the expected load on the system.  They don't need to be pre-warmed, have a very different dimming curve, and turn on and off almost instantly which is very different than the icans that the system was designed to use.  What you are calling glitches and software issues are probably creative design choices that take advantage of the properties of the Icans that the board and sotware was designed to work with. 

[galgon]

So after having these for a few weeks now I am starting to think:
1. wouldnt it be nice to have a high score board somehow - thinking an external TV display
2. what about other types of high scores - most 50s in a game, hit all holes in a game, etc. 
3. well if we are going to do that what about new games entirely - skeeball cricket (the dart game) sounds interesting, lowest number of balls to hit all holes, only score points when hitting a certain hole, etc. 
4. "Fix" the display to properly use the LED bulbs in it.  Hopefully this is just modifying the data going to the dispaly board and not the board itself.
5. Lights (after all that is what we are here for) - would be cool to place some smart string strips up the lanes and aether-ish spotlights on the board itself.  Would be really cool to have a spot light highlight the hole you have to hit in order to get points (as a new game) but would be difficult to get a spot to hit the lower holes because of the angles involved.
6. Updated/different music

Given all of that it is pretty clear that I will need to do a major overhaul on the machines.  I was hoping to get some thoughts on how to do this/what hardware I should use.  My first thought is to use a raspberry pi and then wire all of the inputs and outputs into that.  Would likely need a few additional I/O boards as there are a lot of inputs.   But the rasberry pi has audio and video outputs which will solve #1 and #5.  I am not really sure what the best way would be to control the lights.  Perhaps a Zues 8 somehow controlled by the Pi. 

Any thoughts, advice or ideas would be great.  Hopefully I can finish by the end of the decade.