Robotron restoration: the ultimate conflict between man and machine?

CharlieG

Well-known member

Donor 6 years: 2019-2020, 2022-2025
Joined
Jul 12, 2019
Messages
4,905
Reaction score
6,009
Location
Northern, Virginia
I recently picked up a Robotron cabinet for my next restoration project. The seller claimed to have once been a small-time operator, and that this machine was the only one they kept for family use after closing down the business. It had apparently been sitting unused in a storage shed since about 2000 though, and was presumed to be non-working. But it was complete and in relatively decent condition, so I though it would make for a fun project. This is the first Williams machine I've owned, and I'm looking forward to learning how they're different from Atari and Nintendo.

BB80292D-34BA-4E82-B364-ADF8A61BF919.jpeg

When I first saw the cabinet, I immediately noticed the odd CPO that was missing the large "2084" logo on the front, as well as a few other differences from the typical Robotron CPO. We've had a good discussion about that CPO over here: https://forums.arcade-museum.com/threads/bizarre-robotron-cpo.476538/

F84D394F-B8A6-417A-A5B0-CE447B806D6A.jpeg

The bottom of the cabinet looks like it has seen a little bit of water, with a small amount of splitting along the plywood sides, but I think I'll be able to use some wood glue and clamps to get it back into shape. One of the leg levelers and its wooden base is a bit mangled, so I'll need to repair that too.

9157680F-6689-438B-B1CF-6EFACA699ED4.jpeg

The side panels have plenty of dings & scratches, and a handful of deeper gouges, but nothing that would lead me to strip, sand, and repaint the stenciled artwork. I typically go for a "players restoration" anyway, so I'll probably just clean it up a bit and maybe do a minor touch-up here or there.

161251C4-09ED-44E1-BA80-C97B1399771C.jpeg

The marquee is in great shape, but the bezel has some pretty serious flaking. When I first opened up the cabinet, it looked like there was a bunch of confetti inside. I'll spray on some triple-thick to keep it from getting any worse, and if it ends up really bothering me, I'll grab a repro or something.

A1771D94-003C-43A2-8789-FEA3FBF53F62.jpeg
 
The joysticks and buttons look and feel correct to me, but I admit I'm not sure I'd know the difference at this point. The upper back door is missing, so I'll have to make a new one. The K4900 looks all original and in good shape — under a thick layer of dust, that is.

B5911238-1953-45EC-B042-9AE231F2286E.jpeg

I removed the game boards and monitor for the ride back to my house. There was some debris in the bottom of the cabinet and I didn't want to risk any damage as the machine rode on its back in my pickup. The boards look good with the yellow ROM set and no visible damage.

6B2C3CAC-DC01-46FE-B191-94228F4EC431.jpeg

56A0BABD-B0AF-498A-BA18-108073E5F0E4.jpeg

At first I thought everything was original and untouched, but upon closer examination I noticed that the power supply board had a serial number sticker (604492) that didn't match the rest of the cabinet (572248). Connector 4J1 coming from the transformer assembly to the power supply board was burnt and broken. Part of it had been cut off and replaced with three wires soldered directly to the base of the header pins on the back of the PCB. Operator field repair? Obviously, I'll be fixing that. One of the pins on connector 4J2 also looked a bit toasty. Additionally, I noticed that transistors Q1 & Q2 on the power supply board seemed very loose, as if their legs were about to snap off. Another thing for me to take a closer look at.

749FA11D-2E45-48B8-B688-08BED2012305.jpeg

CD0A659C-C344-4A22-A2D8-455A6D2AED4C.jpeg

All of the fuses in the cabinet tested ok, so I'm guessing this machine might have been working when it was put into storage 20 years ago. However, the power cord looks like a cheap replacement — it's also frayed to the point where bare wire is exposed in one spot and the ground plug has been removed. Yikes. I was feeling a bit lucky, so I disconnected all of the outputs from the power supply board, plugged in the machine, and fired it up. The marquee light came right on, and I didn't notice any bad signs like sparks, smoke, or a burning smell.

D93D30F0-3880-4C92-A996-FC87A18180F0.jpeg

Once I replace the power cord and the burnt connectors on the power supply board, I'll start by testing the voltages before slowing connecting everything back together. The K4900 will get washed and likely a cap kit, etc. I imagine I'll need to rebuild the power supply board, and we'll see what other fun stuff pops up as I continue on with this project.

What could possibly go wrong? 🤪
 
That power cord is original. Mine looks just like that. Williams made good machines overall but they cut corners in a couple areas and the power cords was one of them.

Also, your power supply needs serious love. I just rebuilt both power supplies on my Joust and Robotron. There are a few things you should do to these to get them working well again. Replacing all the headers with trifurcon is paramount.
 
Also, your power supply needs serious love. I just rebuilt both power supplies on my Joust and Robotron. There are a few things you should do to these to get them working well again. Replacing all the headers with trifurcon is paramount.

I was planning to start with a D-8784 rebuild kit from @security0001 plus everything else recommended in the product description (i.e., cap kit, +5vdc pot, BR1 heatsink). Is there anything else you'd recommend? Maybe the SCRs too? The Arcade Parts and Repair website suggests an 8A 50V replacement SCR, but the original Robotron drawing set lists Q1 & Q2 as 8A 100V...?
 
Maybe the SCRs too? The Arcade Parts and Repair website suggests an 8A 50V replacement SCR, but the original Robotron drawing set lists Q1 & Q2 as 8A 100V...?

I did some more digging. It turns out, the bill of material in the drawing set calls for an 8A 100V SCR for Q1 and Q2 on the D-8784 power supply board... but the schematic in the drawing set labels Q1 and Q2 each as a C122F... which is an 8A 50V SCR. Go figure! So the replacement part recommended by the Arcade Parts and Repair website should work fine.

This was pointed out in forum threads a couple of times before:
 
Replacement AC power cord reference ->

 
Replacement AC power cord reference ->


Yup, I saw that thread too! The last one in stock at my local Home Depot is purchased and currently waiting for me at curbside pickup! 😂
 
I took advantage of the beautiful weather today to wash the monitor while waiting on parts to be delivered.

Here's a "before" photo of the filthy K4900 sitting on the workbench in my garage:

C3924114-E9FB-497E-A7B9-B2510B992D6B.jpeg

And here's an "after" photo of the washed monitor drying in the backyard sun:

7AEB3DC3-74BE-4A31-97F0-7218A182F7EE.jpeg

I was very careful with all of the stickers, etc. and I think they all survived intact — at least the ones that were still there before I washed it. 👍
 
Continuing to make good progress on this project. Here's what I've done since my last update...

Replaced Power Cord

The original Williams power cord was in rough shape (i.e., badly frayed & missing the ground plug), so I replaced it with the Home Depot extension cord used by other KLOV members in the past (SKU #737774):

https://www.homedepot.com/p/HDX-15-...Tap-Extension-Cord-Brown-HD-737-774/100672778

Once you cut off the banana tap, it's a near-perfect match for the original. Here's a set of before & after photos:

D6178FD7-09C1-47C1-B2E3-934C28FC1DA2.jpeg

4DA24023-0205-47BD-BDB4-01449F125911.jpeg

The entire transformer assembly cleaned up nicely, too:

9F3046B5-2F64-4AB7-BCF7-74ABB389F17C.jpeg

Rebuilt Power Supply Board

Next, I started working on the power supply board. I installed the D-8784 rebuild kit and cap kit from @security0001 at arcadepartsandrepair.com plus a few other new parts: the +5V trim pot mod, a heat sink for BR1, and new SCRs (the legs on the old ones felt flimsy and about to break). Everything went in nicely, but here are a few things I noted in the process:
  • R10 had been previously removed, which is recommended by the schematics if your +5VDC voltage is low.
  • F2 & F3 were wrong, with 3A 125V & 15A 32V fuses installed instead of 5A & 4A.
  • The following old caps were visibly bulging: C2, C5, C11, C13, C15, and C16 (which also had a cold solder joint).
  • C14 had a really bad cold solder joint on the negative lead, as seen here:
0DE4C1DD-FF72-4B75-8D2B-6C9A6880F051.jpeg

Once I was done, I reinstalled the transformer assembly, power supply board, and heat sink back into the cabinet. I turned on the power with everything else disconnected because I wanted to quickly test the voltages being produced by the rebuilt power supply. To my dismay, LED #2 didn't light up on the power supply board, indicating I wasn't getting 12V. Before rebuilding this, I was getting all three LEDs to light up, so I knew I had made a mistake. I confirmed with my DMM that 12V was missing, so I started troubleshooting. Eventually I realized that I didn't have continuity between F3 and BR2, which surprised me since I always meticulously double-check my solder joints by testing for continuity — thanks, OCD! I remembered this joint in particular as being a pain to test since the trace connecting BR2 to F3 is on the parts side, so I had to have my DMM leads on opposite sides of the board. Taking a closer look now, the pad on the solder side was lifted, and new solder hadn't flowed through the via to the parts side pad connected to the trace. I'm guessing this was damaged when I removed the old BR2. The legs were a very tight fit in the through holes, so I had to snip them off individually and it was still tricky to remove them. Installing the new BR2 was very snug as well, and it was even trickier to now get solder to flow to the top side pad without removing the part. After a few magical incantations (i.e., lots of shouting & cussing), however, I finally got a good solder joint, continuity tested ok, so I fired it up, and my 12V was back... phew! 😅

Sprayed Triple Thick on Bezel

I also wanted to address the severely flaking paint on the back of the bezel. I figured I would spray it with triple thick to keep it from getting any worse, and maybe touch up the paint by hand if I really couldn't stand it. Generally speaking, I prefer an imperfect original piece over a perfect reproduction, so I'm hoping I can make this work.

Here's an up-close look at some of the flaking (it's all over, but most severe with the red & blue paint):

D028923A-94F1-426E-80A8-B3DD8CF1773B.jpeg

I carefully masked off the viewing area and instruction card. As I was about to spray on the triple thick, I reread the directions on the can and noticed it said "Do not use in high humidity... If extremely humid, do not spray." So I checked the weather, and it said it was very humid... and would be for several more days. I also googled how long you could leave painters tape on a surface, and saw recommendations against leaving it on for more than a day or so. Yay. I then peeled off all the masking, waiting a few days until the humidity dropped back down to normal, and then carefully masked off everything again.

3F352F9C-F895-46CE-BAC7-04AB9FAFF4C9.jpeg

Per the instructions, I applied a "FULL WET COAT," and then a second coat about 2 minutes later. As it dried, I became a little concerned because about half of it was looking cloudy. So, again per the directions on the can, I reapplied "after 30 minutes to restore gloss." That didn't seem to help at first, but by the next morning, everything was looking nice and clear.

Recapped Monitor Chassis

I also used Peter's cap kit to recap my K4900, and while I had the chassis out, I replaced the bottle cap transistors, B+ filter cap, and the axial fuse, too. Here are some notes from this step:
  • The vertical output section looked like it had seen some heat.
  • I accidentally cut J45 while removing the glue holding C311, C507, and C701 in place, so I replaced it with the excess lead trimmed off of a new cap.
  • C366 had a partially lifted pad.
  • R503 had partially melted the insulation of the blue wire connecting point 51 to the other white resistor mounted on the frame.
  • Both bottle cap transistors were very rusty.
  • The B+ filter cap had a partially burnt pad.
I couldn't tell if this chassis had been worked on before, but either these jumpers, etc. were factory mods or old operator repairs (I left them alone):

1E0868C7-EC6A-4DD7-8681-15BF3E93C79C.jpeg

Another thing I noticed was that two of the old caps had values that didn't match the new ones in Peter's kit. C308 was 3.3uF and 313 was 47uF, while the replacements were 2.2uF and 100uF respectively. There was a note about C308 in the cap kit instructions, indicating that a rare version of the K4900 chassis needed a different value for this cap if installing the replacement from the kit resulted in a fold-over issue. I did a bit more research and learned that this was indeed an early -023 version of the K4900 chassis. This had been discussed a few times before (see links to forum threads below), so I went ahead and ordered some additional new caps with values that matched what the -023 chassis needs.
I then centered all the pots, reinstalled the chassis in the monitor frame, and connected everything back together. I was feeling lucky, so I turned on the machine with the game boards disconnected to see if I could at least get the monitor to come up. I was happily greeted with some HV buzz and neck glow, and once I turned up the screen pot on the flyback just a smidge, got some good raster on the tube:

3E358447-FE35-4EA0-80AB-88AD97CB85CF.jpeg

It was getting late, so I didn't bother balancing out the color pots at this point, but I did think it was kinda cool that the recapped monitor first came up with a hue of blue that matched the marquee. 😎

Next up this weekend, I'm planning to cap kit the sound board, and then replace all of the remaining IDCs with new trifurcon pins and headers. At that point, I should be ready to fire up the game to test it out for the first time. 🤞
 
Another thing I noticed was that two of the old caps had values that didn't match the new ones in Peter's kit. C308 was 3.3uF and 313 was 47uF, while the replacements were 2.2uF and 100uF respectively. There was a note about C308 in the cap kit instructions, indicating that a rare version of the K4900 chassis needed a different value for this cap if installing the replacement from the kit resulted in a fold-over issue. I did a bit more research and learned that this was indeed an early -023 version of the K4900 chassis. This had been discussed a few times before (see links to forum threads below), so I went ahead and ordered some additional new caps with values that matched what the -023 chassis needs.

AWESOME JOB!!!

BTW: the C308 has a note in the instructions about it and the C313 should be the 100uf as an upgrade.
thank you for your support.
 
I've been able to spend a lot of time on this project over the weekend, and it's definitely paying off. Like I mentioned earlier, this is the first Williams game I've worked on, and to my surprise, everything seems to be going really smoothly. Here's a recap of what I've taken care of over the past two days...

Recapped Sound Board

The final cap kit I wanted to install in this machine was for the sound board, which I again got from @security0001 via arcadepartsandrepair.com. No real problems with this one either. In terms of the old components, I noted that C15 and C26 were bulging, and C24 had a cold solder joint on the parts side.

One small thing that tripped me up for a hot second was that the final cap in Peter's kit was listed as C38 in the instructions, but I couldn't actually find C38 on the included cap map. The one remaining original electrolytic cap on the board corresponded to one marked C36 on the cap map, which I confirmed by looking at the original drawing set. I also confirmed that the cap value for C36 (both the original cap and Williams' specifications) was the same as the replacement in the cap kit, so I went ahead and installed it there. Maybe this is a board revision thing? My sound board was marked 1C-2001-146-6 and under C27 it also said 24118BVO REV. F. Or maybe there are low-res scans of the drawing set floating around that make the 6 in C36 look like an 8? I did find a forum thread where this cap was also referred to as C38, so it could be a common confusion. In any case, it turned out just fine with that replacement cap installed at C36.

Replaced all Insulation Displacement Connectors and Headers

This was probably overkill, but given how painless it was to replace these on my power supply board, and how rough several of the IDCs looked, and how many people have had problems with these in the past, I got into the zone and spent an afternoon replacing every damned IDC and corresponding header in this cabinet. Every. Last. One.

That's 3 on the CPU board, 3 on the ROM board, 2 on the interface board, 4 on the sound board, and 1 on the monitor chassis, in addition to the 4 others I already replaced on the power supply board. They all got new connectors with trifurcon female pins and new headers with square male pins. I also meticulously double-checked my work by testing continuity on both ends of the wire for each crimp job, and testing continuity on the boards for each of the new header pins. Hopefully, I'll never have to worry about any of these ever again.

Here are some brief notes from that process:
  • Connector 3J2 on the interface board could now be easily installed upside down since there's no key nor a retaining clip on the new headers, so I used a sharpie marker to indicate the top and bottom sides of the connector to be sure.
  • Similarly, 3J2 could be accidentally swapped with 3J3, so I marked them as right and left as well. I was also worried about the video signal connector with the new headers on the monitor chassis, so I also marked that connector to indicate the correct orientation.
  • While all of the other PCBs had the rectangular retaining holes on the connectors facing outward towards the edge and away from the middle of the board, and Peter's cap kit for the power supply board even had a note recommending to keep it like this, I noticed that the sound board originally had them facing in. This presented quite a dilemma for my OCD: keep it original or make it "right?" I ended up deciding to flip them around and make the rectangular retaining holes on the new sound board connectors face outward towards the edge like the rest.
  • However, I did not flip it around for the video input connector on the monitor chassis, which also originally had the rectangular retaining holes facing inwards towards the middle of the PCB. I figured that if I ever replaced the monitor for any reason, and if the new monitor still had the original header pins with the plastic retaining clip, then the new connector wouldn't be able to be installed properly given the offset openings on the connector. So, I kept the original orientation for the new video signal connector.

I had the most trouble replacing the headers on the CPU board, which were the last ones I did. It was late, I was getting tired, and I probably needed a new tip for my Hakko FR-301 desoldering tool. In particular, I wound up completely destroying the parts-side solder pad for header pin 7 of 1J3. By "destroyed," I mean the pad came off entirely and even pulled out the via along with it. Fortunately, this is an unused pin (i.e., N/C) on the monitor control connector, so I didn't sweat it. 😎

Another thing that gave me some trouble was replacing connector 2J2 at the ROM board, which required me to convert the IDC looped terminal pins to the new crimped trifurcon pins. The old insulation displacement connectors made it easy to make these loops, and apparently this technique was quite popular among pinball manufacturers. Here's what that originally looked like here:

81BC74F2-3CF4-47F9-B2AB-137D1655220F.jpeg

I found a bunch of pinball repair/restoration discussions (e.g., link, link) on how best to approach this. At first, I tried to crimp two wires into the same pin to recreate the loop. Tried and failed, that is. It was a tight enough fit just to get a single one of the thicker gauge power wires crimped into the Molex .156 trifurcon pins to begin with, and this was a +5VDC line that had to be delivered to two separate header pins on the ROM board. There was just no way I was going be to able to crimp two wires into one of these pins, and I gave up after multiple failed attempts.

I ended up stripping off a bit of the insulation in the middle of the existing wire and soldering in a T-splice to add a new wire for the second pin. It was a pain in the asteroids and looks pretty sloppy (see for yourself below), but continuity tested out fine at both pins, and I got the expected +5VDC at both as well when power was applied. So... whatever. 🙃

1269B3C4-82E2-4DD1-82BB-29CB381EA99A.jpeg

In retrospect, I should have made a simple Y-splice forked connection to attach two new wires to the old one. I guess I was frustrated and not thinking straight as I worried about running out of length with the original harness wire.

One last thing I noticed was some really weird green oxidation or residue on connectors 3J3 and 3J2. I had no idea what it was, but I didn't want to leave it there, and it was a contributing factor to my decision to just go HAM and replace all the old connectors and header pins.

1DD2DAC6-700D-41B6-82BA-D9EF713A4F9F.jpeg
 
Reassembly and Test

After all that, I decided it was finally time to put everything back together and fully power-on this machine for the first time in 20 years (allegedly).

There were two quick things I noticed during re-assembly:
  1. This CPU board is marked as 5770-09656-00 Rev. B and it has the two-wire BS.BA inversion factory mod. It also has a small daughter card installed at position 2G, which is apparently "a little adaptor board to use a 74LS368 in place of a 74LS367." I left it as-is.
  2. I was planning to replace the original battery holder with a new 3-cell AA battery holder, but when I went to look for it, I realized I hadn't actually ordered one. I thought about doing the CR2032 coin cell battery mod since I did have the parts on hand for that, but decided against it since I want to keep this looking as original as possible. Maybe I'll pick up a new AA battery holder the next time I order parts, or maybe I'll just give the original a deep cleaning. In the meantime, I put 3 new lithium AA batteries into the original holder. The previous owner was smart enough to pull the old batteries before putting the cabinet in a storage shed for 20(?) years, so there wasn't any acid damage — just a tiny bit of corrosion on the terminals that I quickly cleaned up with some isopropyl alcohol. Don't blame me — it was 3:00am and I wanted to see this game in action! 😂
At that point, I had everything wired up and ready to go. I hooked up my DMM to measure voltage at the CPU board, plugged in the power cord, flipped on the main switch, and pulled out the interlock while keeping my fingers crossed. The machine instantly sprang to life with nothing but reassuring sounds, and here's what I saw from my vantage point at the back of the cabinet:

D168F245-4898-41B6-9E4E-7AE889B18B67.jpeg

I was incredibly relieved to see all three LEDs lit on the power supply board, a zero showing on the ROM board segmented display, and a voltage reading just shy of -4.9VDC on the CPU board. I almost couldn't believe my eyes.

I walked around to the front of the machine and saw some blurry text on the screen. I had to crank the focus knob on the flyback all the way to make it nice & sharp (I'm not sure if I should be concerned about that?), and then I was able to make out this message:
HIGH SCORE TABLE RESET
BOOKKEEPING TOTALS CLEARED
ADJUSTMENT FAILURE
RESTORE FACTORY SETTINGS
BY OPENING FRONT DOOR AND
TURNING GAME OFF AND ON

I assumed that was because the game had been sitting with no batteries for two decades, so I did what it said and then turned it back on. It was glorious:


I played a quick game and entered my high score, then turned it off again to replace the coin door bulbs and make sure that the high scores and settings were being saved. When I turned it back on again, it displayed the rug pattern and then the "INITIAL TESTS INDICATE: OPERATIONAL" message.

Everything seemed to be working great. I was ecstatic and I wanted to spend some time dialing in the monitor, etc. but like I already said, it was past 3:00am by now so I called it a night.

There are quite a few more things I still want to do to this cabinet — like installing new leg levelers, addressing the slightly chewed-up bottoms of the side panels, and investigating why light is bleeding out from behind the marquee and illuminating the control panel to the point where it washes out the bulbs underneath the translucent buttons — but I'm out of the woods now with this project. It hadn't been turned on since the previous owner put it into storage years ago, and neither of us wanted to power it on until it had been gone through. Even after that, there were so many things that could have gone wrong (e.g., making a mistake in wiring up the new connectors), so I consider myself quite lucky that this machine fired right up on the first try!
 
I've been super busy with kids stuff and a new job, but I'm still slowly making progress on this project. Lately, I've focused on taking care of little odds & ends, like replacing the leg levelers, installing new locks, and glueing & clamping the bottom of the cabinet sides to address some slight expansion/splitting of the plywood. I also made a new upper back door, which I think came out really well. I cut a 2x4 sheet of 1/2" MDF to size, cut the rabbet joint (my first ever attempt) on my table saw, spray-painted it satin black, drilled & installed a new lock, and attached an original diagram sheet that I found on eBay. It's not perfect, but I'm quite happy with it.

I followed the two posts from @Sinibun on this thread for all my measurements: https://forums.arcade-museum.com/threads/measurements-for-joust-upper-back-door.300482/

8A7435C0-B9E0-418A-9475-8F61348ED56C.jpeg

F24AEDBC-D33E-451C-93D8-91170DCFCC50.jpeg
 
Back
Top Bottom