Archives for January 2016

Candy Speaker Replacements

Posted January 30, 2016

Occasionally I start researching speakers for the Astro City and Egret II, don’t find much, then give up for a few months and try again. The only drop-in solution I’d come across always seemed to be the Cambridge Soundworks SBS 52. Knowing this would work for the AC, I started focusing on what would fit in the Egret. There were a couple Egret tutorials for full on powered speaker solutions, usually involving modifying the original wiring and shoving in an ugly PC subwoofer behind the coin box. Yech. I didn’t want the cabs to sound unnaturally pumped, I simply wanted them to sound better. The stock speakers in both are actually pretty decent, but the Egret in particular could use a boost.

Cambridge SBS 52 are cheap on eBay so I thought I’d give them a shot in the AC. Removing the housings required a long Phillips screwdriver, with a fifth screw hidden behind the speaker grill. I removed the old solder, brought them upstairs and swapped them with the AC’s originals, using its quick disconnects. Sampling games I was familiar with I could hear maybe a very slight improvement, but the midtones had a kind of boxiness to them that wasn’t so pleasing. Or maybe they were fine and I just needed to give them more time.

I went ahead and took apart the Egret’s factory speakers which took a little more work to disassemble. They have their own brackets and speaker boxes, and their wires were soldered directly on to the driver’s tabs rather than using connectors. It turned out the Egret’s speakers are also 3″, same as the AC, and both are 4 Ohm ~10 watt. So sorta rushing things, I ordered another cheap pair of SBS 52. When they arrived I went through the same process, removing the housings, desoldering, then desoldering the Egret’s speakers, soldering in the new ones, back in their boxes with the bracket and back in the cab.

Since I’ve been playing a lot of DonPachi I threw the board in and played for a few minutes. It didn’t take long to realize it actually sounded a bit worse. I’m not sure how I thought swapping in unpowered PC speakers would be an improvement. I also realized that the Cambridge speakers were actually 3.5 Ohm — probably not a big deal but long term maybe not a great idea. In the end I put the factory speakers back in both cabs, annoyed with the whole project. Anyway, these were never meant to sit in someone’s office, but in cacophonous arcades pushed, by the dozens, side-to-side. This was another lesson in being happy with what you already have.

Galaga Restore Part 3

Posted January 24, 2016

A few items had remained to finish this project, largely the K7000 chassis. I was losing steam after spending weeks sorting out endless monitor problems. Before I could dive back in with another round of prodding, I suddenly found a box in front of our door containing a Sharp Image monitor chassis. Mike! This guy’s too nice — I’m not even sure where he found this thing. Except for a fairly menacing looking very alive spider, the chassis looked to be in excellent shape. I swapped it in and it immediately worked. Using the test pattern generator I calibrated it with a mirror then mounted it back in the cab. While I value the gained experience of working on an old monitor, in the end simply replacing it turned out to be the sanest option.

There were a couple more issues to work out. One was quick, replacing the 35-year-old 6×9 speaker. It actually sounded fine, but if I could eke out a small improvement, why not. Surprisingly I could find very few 6×9 speakers between $10 and $75, so I chose a Lanzar OPTI2698 — 8 Ohm and capable of 1190 more watts than necessary.

Now for the joystick. Even after previously spending several numerous hours rebuilding the original, it continued to feel sloppy. Worse, the two leaf switches would occasionally need bent again to help the stick auto center, inevitably causing the ship to move on its own in one direction (more on that in a minute) when the strength of both leafs weren’t exactly the same. Rather than try finding another Galaga stick, I started to consider Mike’s suggestion of using a Pac-Pro joystick. While I usually loathe the idea of swapping in modern replacements, the originals just didn’t wear well, and this was still a leaf stick. It mounted on the control panel with a Twisted Quarter Galaga adapter plate, with the oval hole in both acting as a 2-way guide. Unfortunately once the control panel was back on the cabinet, I found that it wouldn’t shut completely — the Pac-Pro base and the top leaf tabs were protruding about 1/8th of an inch too far. The tabs were able to bend 90 degrees, but I had to take the Dremel to one side of the new base. Nothing that’s ever seen, and it fit. I wish the red balltop was the same size and material as the original, but its matte finish matches the overlay well. It’s considerably stiffer but hopefully that will become less noticeable as it breaks in.

Overlapping with the previous issue of the old leaf switches sometimes nudging the ship by themselves, at some point this started to happen even when the joystick wasn’t plugged in. After crediting up, the ship would slowly gravitate to the left in random blips. It turned out to be one of the Namco custom chips, 51xx. I swapped it out with another 51xx on a spare Galaga PCB I happened to have. Previously I’d hoped to fix this second board and sell it, but having a donor board on hand seems a much better idea.

It’s great having the game finished, slid in next to Gyruss, and actually playable. I owe a big thanks to Mike for not only locating and delivering Galaga, but also tirelessly answering my tiresome questions, and problem solving from beginning to end. He elevates “a friend in the hobby” into something we should all hope to emulate.

Galaga Restore Part 2

Posted January 2, 2016

It’s been a few months since Galaga was mostly wrapped up, with one remaining snag that’s lingering on into the new year. What could it be? How exiting, but first let’s see where we left off.

I applied several coats of paint to the front and back of the cabinet, the coin door, other metal parts and screws. For a rattlecan, as it’s affectionately called, Rustoleum Satin Black actually does a decent job on smooth surfaces. At some point I suppose I really should look into an air gun. There was a stubborn bondo hump on both the coin door and the front of the cab from patched holes which took three or four rounds of sanding and painting but eventually blended in. After a few days of drying, I applied the kick plate and front art, bolted the coin door back on and added new coin door inserts and bulbs. The original Midway coin door plate, despite my best efforts to save it and the rivets, had to be replaced by a repro. Next came the control panel overlay and rebuilding the original joystick, which included replacing the bushings and tweaks to the leaf switches to try and get the stick to center with less slack.

At one point I had intended on adding a Galaga ’88 board to the mix, requiring a double jamma adapter, a Galaga-to-JAMMA adapter, and of course Galaga ’88. Eventually I found the PCB, but it wasn’t cheap and looked so good in the Egret that I decided to skip the dual setup and leave Galaga as original.

The last issue to work out was the finicky K4600 monitor. It had a magenta cast to it, was faded to almost black at the top and overly bright at the bottom, and had occasional flickers of light. I started by replacing the caps and reflowing solder on the chassis — no change. Hoping that it may just need the pots fine tuned, I spent hours calibrating it but the issues remained. Still, it was good experience for someone with no real monitor knowledge, including watching my first horizontal width coil disintegrate.

As I started looking on KLOV for local replacements, Mike offered up a K7000 project monitor. Since the closest alternative was in Sacramento, I agreed to Mike’s generous donation and began researching known issues and buying a few parts: flyback, HOT, voltage regulator, C36 safety cap, and R103. Little by little I replaced the parts that had failed, but I still couldn’t get the thing to power up until Mike pointed out that the middle leg on the HOT wasn’t connected. Fool!

While the K7000 was now firing up, the image was rather magenta again, strangely similar to the K4600 — while I could see red, blue and green individually, it somehow wasn’t mixing for pure whites. To rule out the PCB I tried Gaplus which gave exactly the same results. Then came a series of never ending tweaks and tests, checking B+ (123.6), to more reflowing and calibrations. Mike brought over a test pattern generator which helped since Galaga only displays a test grid. We adjusted convergence, then tried swapping Q201, Q202, and Q203 transistors around on the neckboard. Nothing seemed to help, then at some point we lost blue entirely. Since then some good suggestions have come through KLOV, requiring more time on this before I start hunting for another K7000 chassis.

So Galaga — it’s close! The cab was looking good, though like Gyruss I kept the sides original so it certainly looks better slid in next to other games. Thanks to the enhancement pack it’s now saving high scores like Gyruss. Hopefully with a bit more work on the monitor this project will be completed.