Category Archives: Tech Tips

DIY LED Work Light

LED work light attached to a Star Wars
My sort-of home-made work light.

My friend Chris says the best tool for working on pinball is “more light”, and he’s right. Chris uses a classic clamp light. These need wall power and take a lot of space in the toolbox, too much for when I’m not in my garage.

As LEDs have gotten cheaper and more portable, I looked for a way to have a lot of light in a small space. Then I hit on this post on Adam Savage’s Now, I built one of these to Adam’s directions, but it didn’t work for me. I don’t have his workbench to attach to.

I did try attaching to a pinball machine’s siderails with Magswitch doodads, as shown in this subsequent video. Magswitches are amazing, but not in this application. Their steel bases will scratch the siderails, and depending on the game, attachment is either mediocre or nonexistent. Not all siderails are magnetic, and WPC siderails just don’t have enough iron to hold the weight.

I prefer something simpler. I got this this mounting kit and a camera-flash-type mount. (Update: Try this, a kit of just the necessary bits!)

The neck is Loc-Line, which is a modular hose. It supports a fair amount of weight on its own. I use a clamp to attach the hose to a game, a gender changer to get the right end on it, and then the camera mount to give me somewhere to attach my light.

Add a light like this. I believe this is the one Adam Savage uses, but you can experiment. As long as it has a 1/4″-20 thread, you’re all set. These are pretty cool, with adjustable brightness and color temperature.

These lights take a F550 camcorder battery (included). Newer batteries now seem to have USB-C plugs for charging, which is fantastic. I have a few of these and I move them around my work lights.

But for longer sessions, a power adapter is helpful. The lights have a barrel plug. I needed a power adapter with a 2.1mm x 5.5mm barrel plug, 12V, center-pole positive, like this. With this plug, the light can run without a battery. (Adam Savage never uses the batteries, and he runs the power cable through the hose. Since I don’t permanently attach my lamps, I don’t.)

I use 7-9 segments of hose (depending on my mood) plus the gender changer. This is just enough to get whatever angle I want on the lamp, but not enough to stress the clamp. Once you learn to disassemble and reassemble the hose, it’s easy to make a change here if you want to change length. I found I liked the lamp just as much with fewer hose segments.

Unlike the original, since I move my lights around, I found a shorter neck worked well, and I didn’t need to reinforce it with armature wire.

To save on the camera mount, I have also used 1/4-20″ hardware and the end plug with a hole in it, but the camera mount is easier to live with and thus recommended. If you decide on just the end plug, you will need a short screw, a nut, washers, and some luck.

I find the clamps here don’t mark the cabinet, and can be positioned with enough freedom to get light where I want it. If the squeeze clamp isn’t to your liking, sells other clamps.

One problem: Cost. The light panel by itself is $45, the mounting hardware is another $31, and extra batteries are a thing, too. But they’re awfully nice to be able to have in my only moderately large toolbox. And if one is good, two is better.

Tagged , ,

Troubleshooting Counterforce drop targets

Chris refers to the Gottlieb drops as the Cadillac of drop targets. I’m not so sure, but they do play well. Drop targets are an annoying mechanism to work on, especially those with added features like trip coils.

Williams drop targets have a failure mode I call slo-mo drop targets. You smack one with the ball, and instead of clicking down, it sinks slowly into the playfield. This comes from dirt. If they’re really dirty, you can hit one, the ball can move far away, and then it sinks. It’d be cool if it happened on purpose.

I got this Gottlieb Counterforce. And it had slo-mo drops. I’ve never seen this in a Gottlieb game. There’s nowhere for the dirt to hide. So, while shopping out a bunch of stuff wrong with this game, I did the drops.

I replaced all the targets (which didn’t match). I followed PAPinball’s guide, which is very good. I didn’t expect this to fix anything, but I hoped it would. It didn’t.

Toasty coil sleeve in my drop target bank.

I cleaned the coil sleeves for the two reset coils, which didn’t really help. So I removed the bank and did a deeper (but still half-assed) cleaning. When I was done, I still had slo-mo targets. (This sleeve is cooked, but it still works OK. Perhaps I’ll replace the coil, but for now, it’s working fine.)

It became pretty clear the problem was with the reset bar not sinking. If one target was already fully down, any other target would drop quickly. But if no targets were all the way down, you’d get slow action because the drop target had to push the reset bar down. So once the reset bar was down, targets would drop quickly.

On this mechanism, there is some freedom to adjust the coil in the bracket. Coils can be twisted slightly relative to their plunger and the reset arm. The coil needs to be square to the bracket, and both coils need to be at about the same depth. The coils can be adjusted slightly so the reset bar doesn’t over-extend the targets out of the top of the playfield, but they need to be square to the mechanism to work correctly.

Now, as drop target banks go, this one has all the bells and whistles. There are seven drops, but there are seven trip coils on the top, too. These are triggered by the computer when it wants to drop a target. In Counterforce, these are triggered by rollover switches and standup targets scattered all over the playfield. Each drop target has a sense switch, but it also has a lamp voltage switch used for “when lit” indicators for the trip switches, and also for the big falling-bomb gimmick.

On my game, one of the trip switches didn’t trip. This appears to have been a coil that was open. The coil was A-18642, but research indicates there are a couple other usable substitutes, one of which was in Chris’ parts bin. So that’s what got used. (Pinball Resource can identify the substitutes, if necessary, but they also have the correct part number.)

I was able to identify that this coil was open by using a multimeter. In the diode setting, I was able to measure the diode of the bad coil — but not of the good coils. For the good coils, when I try to measure the diode, I measure the coil instead. So that was my indication that the coil was at fault.

So far, so good. Resoldering these coils is kind of a pain because there’s so little clearance in this mechanism.

My Counterforce is still torn apart waiting for other parts, but the drops all now work correctly.


  • Make sure the reset arm drops completely, and freely. If not, look for a twist in the position of the reset coils.
  • The reset coils should be positioned so that the coil stop stops the plunger before the drop target goes up too far and shears off the foot.
  • Find the criteria for remote-trip coils and verify they’re all working.

This all applies mostly for Gottlieb drop targets. In any case, I haven’t generally had problems with them once they were rebuilt in home use.

On Williams games with the horseshoe mechanisms, the problem is the horseshoe mechanism. Those things are terrible.

On Williams games with optos, and I suspect those with leaf switches at the bottom of the column, they get dirty and sluggish. I cleaned my BK2K’s targets once and they’ve been great, even after the game was operated for six months at the PPM.

On old Stern games, there’s only one spring that drops the target, so it can be twisted by the spring. If you can get that adjusted correctly, they work great, and can be swept by a good shot. I love these.

Old Bally games are similar, but I don’t have as much experience with them.

I have seen two different drop mechanisms on new Stern games. On my WPT, there’s a service bulletin to enlarge the catch shelf for targets that has cured most of my woes. On my Simpsons Pinball Party, I had a loose connection — otherwise it’s been quite reliable.

Tagged , , ,

SPP drop targets are too generous

I have more or less finished a shop job on my Simpsons Pinball Party over six years in the waiting. One problem that I have been frustrated with was the drop target bank got confused sometimes and would award “drops completed” several times. Careful observation over an extended period of procrastination and play revealed that the middle target seemed to be at fault.

I actually went as far as buying some new microswitches, just in case one was flaky. Then I remembered to always look for a bad connection first.

The diodes on these switches are mounted on a terminal strip a few inches from the drops. I thought for sure one would be cracked, but they looked fine.

So I pulled the drop target bank out of the game, and one of the QD connectors fell off of its switch lug. I used pliers to squish the switch a little bit together and slid it back on. So far so good!

Tagged , ,

Installing a knocker in Stern’s Star Trek

I have had a Star Trek (Pro) and the parts for a knocker install, for well over two years. I decided it was time. I think this general plan would go for any metal-head game.

Get a real manual from an old SAM game. This is helpful in finding places to pick up the needed voltages. I used WPT, because I have one. Stern manuals are terrible after about 2008, but they were superb before that.

To tap voltages off existing connectors, get an IDC punch-down tool for .156″ pins. The trick to these is to run the wire into the connector just far enough. If it’s too far, it won’t punch down on the far side, no matter how hard you push.

Star Trek never had a high voltage board for its DMD, so there are some mounting posts in the head not being used. Maybe it’s different in the LE. I installed the same step-up board that I used in WPT and Simpsons, except it now requires +20V. I just put it on one post and held it down with a #8 nut.

The real pain point in installing the knocker is figuring out where to pick up the needed voltages. The 520-5254-00 step up board requires +50V, +20V, solenoid ground, solenoid input (from Q24) and solenoid output. (Older versions of this board omit the +20V but sometimes pull in at power on. Like a System 80 game!)

The cable dressing here is pretty terrible. I may have to re-do this due to poor crimping anyway, and if I do, I will trim all of the wires to the same length and dress the harness neatly.

+20V is available on J7-1, right next to another source that the game actually uses. In my Star Trek, it’s a fat orange wire. I suspect most SAM games have available +20V pins.

Solenoid ground is available from J10-2,4,5. On Star Trek, pins 2 and 4 are not populated, so I stuffed a wire into pin 2. It appears the game generally uses black for this. I suspect most SAM games have available grounds.

+50V is not as easy to get on Star Trek, as there are no spare supplies. On WPT, the dedicated magnet supply wasn’t used, so there was an unused pin in the J10 connector. Star Trek, however, has a magnet. I tapped the magnet with a suitcase connector (Scotchlok). I had some trouble getting this to make contact. I either didn’t squeeze hard enough, or the second wire slipped out of position before I clamped down. In either case, next time, I’ll check continuity immediately before moving onto the next step.

I also crimped two wires into the +50V pin for the step-up board so that I’d have a +50V wire to feed the coil with. I probably won’t do that again. Another suitcase connector would have worked just as well. (In my WPT, I fed +50V to the coil, then daisy chained it to the step-up board, which is a better idea when putting the knocker in the head.)

I ran a long cable with three wires out of the head and into the cabinet. One wire picks up the Q24 output and gives it to the step-up board. One is +50V and one is the output from the step-up board to fire the solenoid.

Knocker mounting in cabinet

Knocker mounting in cabinet

I screwed the knocker in the bottom and firing at the cross member near the coin box. I used screws that are a little too long and pierced the bottom of the cabinet. The cabinet bottom is not great plywood. I may change this over to T-nuts in the future. I believe I based this particular mounting on APB Enterprises’ directions on how to mount a knocker for their kits. For my WPT mount, I eyeballed it and it’s relatively close to the coil–and wonderfully loud.

If I make any further changes, or I do this again, I want more connectors to make the harness a little easier to work with and debug. Having the connector near the solenoid makes cable management easier. Having connectors for the patches into the existing connectors would make removal easier.

Another possible improvement: Does the shaker motor offer a place to hook in for +50V?

Tagged ,

Installing a knocker in Stern’s World Poker Tour pinball

I just installed a knocker in my World Poker Tour. I used Stern’s step-up board, the one that is used for the kickout from the bumpers, available from the usual suspects. I got most of my information from this page.

A few random notes:

I used a 520-5254-00, available from Pinball Life, and documented in the WPT manual.  Because the knocker drive transistor can only sink 20V, a step-up board is required.

Three connections to the game are required: +50V, solenoid ground, and signal. Okay, you could probably use +20V if you wanted a wimpy knock from a wimpy coil. You can source the signal from the connector under the playfield that is used to drive the coin counter that nobody installs. +50V and solenoid ground can be found on J10. On WPT, it appears that pin 8 is not used, but it is fused by F7. Great deal! Three pins on J10 offer solenoid ground.  I used pin 2, but 4 and 5 would have worked. (You can pick up ground anywhere, but it is probably better to pick up the officially blessed solenoid ground.)  I happen to have an IDC punch-down tool.  I’ve had it for fourteen years, and now I’ve used it.  I didn’t have to solder anything (other than the coil) because I picked up all of the connections from connectors that already existed, which makes for a nice clean install.

I installed a Williams knocker, in the head. Mine was apparently from a High Speed, but it’s the same mechanism Williams used until they gave up on knockers.  The geometry of the Stern head leaves a lot to be desired from a knocker perspective. I will probably install the next knocker in the cabinet, and even that can be trying as I have discovered previously.

I put the head in the upper-right mounted to the side of the cabinet.  This was a tight squeeze for the assembly, and it was hard to install and it will be hard to remove for service.  I mounted the knocker on a block of scrap wood, then screwed that to the cabinet.  I don’t like the adapter block, but because of the diagonal brace in the corners of the head, a the knocker wouldn’t have clearance if it screwed to the head directly, and there’s no room for it on the back of the head.

I set the Q24 option to “knocker” and the “knocker volume” to off.  I discovered the “knocker” test drives both the real knocker and the fake knocker and it’s loud as hell. There is no other test for Q24.

Next time I install one of these, should I ever get another Stern game, I’ll probably try the Gottlieb knocker.  It doesn’t require a separate strike plate.  Running on +50V, it’s probably not as anemic as in a Gottlieb game. Or, perhaps I’ll mount a Data East-style unit in the head, down below the computers, where there’s space.

A modern Williams knocker fires up and has no spring.  Bally, Data East/Sega, and old Stern Electronics knockers fire at the side of the cabinet.  The bracket used for Data East is the same as Williams, but the plunger is different.  Old Williams knockers, like on Firepower, and Gottlieb knockers fire at their own bracket, and are mounted in the cabinet. Stern Pinball has never used a knocker.

The sound of a Stern knocker is different than a Williams knocker.  It’s more of a thud than a crack.  I suspect the coil pulse is longer. I prefer the knocker on my Twilight Zone, but any mechanical knocker is better than the pop on a Stern game.

Adding a knocker to my Simpsons Pinball Party was similar, but I used a “Data East” knocker assembly (really, it appears to be a kickback assembly) and mounted it in the body. This gets even more of a “thud”, and it’s not nearly as loud, but not as satisfying.  Finding solenoid ground and +50V was more difficult and I did a bad job.  I should clean it up.

Update: there may be two versions of the 520-5254-00 board, one that takes a 20V input and one that doesn’t.  I haven’t used the 20V on yet, but that’s the one I want for future games. I asked about the difference to Stern’s technical support, and they responded:

The newer step up driver boards use 20V and 50V. The coil will fire using the 50V power. The 20V input is used to sense weather or not the voltage is present. This is important because if you have 50V without 20V, the coil will turn on until the fuse blows. I would recommend the new board!

I’m trying to confirm that the two boards have the same part number.  There are pictures of a five-pin version online with the same version number, but the parts sellers only picture the four-pin version.

With the four-pin version, sometimes my games knock on power-on, I assume due to the problem outlined above. (A little like the System 80 thunk, I’m sorry to say.)

Tagged , , , ,

a note about rebuilding Gottlieb “fat boy” flippers

In general, I don’t rebuild Gottlieb “fat boy” flippers as part of shopping out a game. This is in contrast to Williams solid-state flippers, which do wear out. The exception is the EOS switch. The lever arm will chew a hole in it over a couple decades of play, and it’s worth checking first. It may be fine, or it may need replacement.  I have also seen the plunger/link crack apart. But otherwise, these flippers just work.

On my Monte Carlo and couldn’t make the Firepower-esque lock shot, up the left side and into a saucer. Without this shot, there’s no multiball, and no way to spin the roulette wheel. The flippers seemed fine otherwise, but were clearly missing some oompf.

I replaced the sleeve because it’s cheap and checked out the link assembly, which seemed to be fine. I decided to get more drastic. I have a couple MA-989 upgrade kits with NOS link assemblies and new EOS switches that should make the EOS switch also nearly indestructible. I decided this was a fine occasion to use one. Ultimately, I replaced the EOS and lane change switches, the coil, the coil stop, the plunger/tip, and the lever that clamps around the flipper shaft. Still, it was weak, and in fact it was a little worse.

But in the process, I had installed the flipper with an angle that was too horizontal at rest. I think I was working off the angle of the left flipper, which wasn’t necessarily right to begin with.  I went back and took a look the flyer and went for a droopier angle, and happily, I can make the shot!

I suspect that flipper mechanism develops a lot of power at the end of its stroke.  Plus, the angle made the flipper play funny, and I may have had a hard time adjusting.

I will keep all the parts I took out as spares—I don’t think they’re bad. And I will pay more attention to the angle in the future.

I am not enamored with the MA-988/MA-989 upgrade kits. They require some care to ensure that the return spring isn’t touching the switch. The geometry of the flipper just doesn’t allow this fix to be very good. Plus, my NOS parts came with NOS factory rust in non-critical locations. The original flippers work pretty well without the fix.anyway. I will probably use these kits up and switch back to the original-issue parts once I exhaust my small supply.

Tagged , , , ,

Flash lamp weirdness

I chased down the flash lamp problems on my Monte Carlo.

The flash lamps next to the roulette wheel were locked on. I found a out-of-spec transistor on the driver board.  The under-playfield transistor (MJ2955) tested fine once the driver board was disconnected, so I just left it.

The other one was weirder. The “left dome” flashers did not work. Lamps themselves were fine. The transistors tested okay, and there was voltage at the lamp socket, but grounding the lamp socket did not cause the lamps to light.

I found that the two lugs of the lamp socket were shorted together. Apparently the resistor in circuit that steps down the voltage is good enough to protect the transistors from failing, or I just got lucky.

I might have caught this sooner if I’d noticed that the last person to re-rubber the game had looped rubber around 89 sockets rather than reading the manual for the correct rubber ring sizes.

Tagged , , ,

Williams EMs Sometimes Have Interlock Switches

I decided to spend a few minutes trying to fix my Spanish Eyes.  It just wasn’t getting power at all.

I stared at the schematic for a while and chatted with Chris a bit, and he said, “Is there an interlock switch?”

Yep, sure enough, there is.

Gottlieb and Williams games, particularly those with three-prong power cords, have interlock switches and the coin door must be closed for operation. It is even rarer to find one that has not been bypassed.

And it’s not in my schematic!  I don’t know why.  I might guess that it was a factory option, on some games, but I know of other Spanish Eyes that don’t have the interlock switch. It looks factory, and I can’t imagine anyone ADDING an interlock switch.

(An interlock switch cuts power when the coin door.  Most video games have them on the back panel, and some games, particularly early ’80s Atari games, have them on the coin door.)


Spring Break restoration notes

Old notes are bulleted, with inner bullets updating the work I did:

  • Battery corrosion.  This had one of the “DataSentry” style black batteries of doom.  … I have so far removed the battery but not cleaned the damage.
    • Ultimately, I just stopped.  I have parts to rebuild the reset section, but it’s working; why fight with it?
  • Computer has been trashed.  Like every 80B computer, this has a piggyback board arrangement where the solder joints fatigue and crack. … [T]his involves desoldering the board so I can work on the underside.  Unfortunately someone decided to go to war with the thing on the top side.  I did finally get the board off, but I am afraid I have damaged the plated-through holes.  I obliterated at least one pad on the bottom and damaged another.
    • I bought one of the GPE daughterboards.  Highly recommended.  The redundancy that this board adds helps make up for my mistakes, and Gottlieb’s mistakes.
  • Ramp cracked at entrance badly.
    • I have a new ramp, but it’s not installed.
  • Sounds lacking.  So I can hear some correct sounds, but mostly it’s static.  Can’t hear the music. I want to take a look at this, but there’s no reason to until grounds are verified.
    • I did put this off until after the ground mods, but the ground mods weren’t required.  All the grounds in this game are one big common (eventually) and that was easy to verify.
    • The actual problem was one or more bad ROMs on the sound board. I verified this by swapping parts with my Monte Carlo, which had a working sound board. This was much easier than trying to go at it with an oscilloscope… which I also did.
  • Ground mods not done.  I suspect this is causing sound problems.  At least one set of pins (this game has five) is burned badly.
    • I did the major ground mods.  Didn’t fix the sound (not surprising).  I’m not happy with them, though, because it means permanently rerouting a cable in a way I just don’t love.  I may try and come up with another method the next time I do this.
  • Lots of lamps missing.
    • Bad sockets and bad bulbs. I need to solder some sockets “shut”, but that’s about it.
  • Driver board has transistors replaced with the twisted-leg hack.
    • Replaced with CEN-U45.
  • Black rubber on playfield.  Ew.
    • Yeah, still haven’t had time to shop the game.
Tagged , , ,

Maybe they’re all blown.

In the vein of the Maybe they’re all burnt out rule, I recently got a Gottlieb Monte Carlo  and none of the pop bumpers worked.

I got really lucky with this one. Each pop bumper is fused separately, but each and every one of them was blown individually. I could have gone checking the power train back to the transformer, but I was sort of smart and checked the fuses first. I say sort of smart because if I was legitimately smart, I would already have pulled every fuse on the game and made sure they were as originally specified, and I haven’t.

All of that said, why did three fuses die simultaneously? Well, the game was moved several hundred miles on the back of a truck. The drop target reset fuse had also blown. Maybe the fuses were just old, or perhaps I have another problem. In any case, though, they certainly needed to be replaced.

Tagged , ,