I found there's not enough stories about how people fix their machines, and my kids are currently playing nicely in another room, so I figured I'd type this up, to help further the hobby for other people.  This turned out quite a bit longer than I expected. Sorry!  But it has what I believe is good advice about fixing machines, especially if you're a first timer.

WHY?

Well, I got tired of shopping for mint condition games.  They're expensive, for one thing, because you're competing with people who can't fix broken machines, so demand is much greater.  I also get a lot of enjoyment out of diagnosing and fixing problems as well.  Being educated as an Electrical Engineer, I was hankering to buy a machine that needed to be fixed.  I used to enjoy tracing circuits and running scopes etc, and I hadn't done it in a while, so I searched on EBay and found this great game.  I was more into Asteroids than Asteroids Deluxe as a kid, but I decided to give this one a go.  I got it for just under $200 on Ebay.  It was listed as "having sound, but no picture."  I had to drive an hour to pick it up, and this machine is massive.  Luckily I have a minivan.  Now this machine was obviously in an arcade the whole time.  It was filthy, but in good shape.  It has at least 24k games on it (it has mechanical counters on the coin slots that can't be easily reset), and there was definitely no picture.  I also love the vector graphics games, like Asteroids, Tempest, Battlezone, Star Wars and Empire Strikes Back.  I'm currently jonesing for a Star Wars cockpit version.

I sat down after the kids were in bed that night in a chair next to the machine with the manuals (it came with the manuals, so it saved me some downloading, at least at first).  I was there all night.  Sleeping mostly (kids have that effect).  I didn't know the situation with this machine, I'd bought it from a non-technical guy who had bought it at an auction with the intent of getting someone to fix it, but he couldn't find a guy to take on the job.  He'd been told that nobody fixes Vector Monitors anymore, and it sat in his garage for years. 

So I started from step one

I tested the power supply.  Before I even turned the machine on, I pulled all the plugs off the power supply and tested them with a multi-meter to make sure they weren't going to overload the board or the monitor.  Nope, everything checks out.  So I plugged the board in without the monitor attached and powered up. 

There were a couple things I wasn't expecting about this machine.  Unlike the original Asteroids, this one has a backdrop.  It's supposed to work by reflecting the monitor on a partially reflective mirror with graphics behind it.  It makes a cool backdrop.  I had seen inside the machine with the power off, and I assumed that the dusty artwork would need to be cleaned before the effect work.  I was wrong!  What I thought had said "backlight" on the circuit diagram had said "blacklight".  So when I powered it up the first time, the background was glowing and jumped right out at you.  All the blemishes scratches and flaws were not illuminated, and the backdrop looked great and 3d as well.  Definitely a cool effect.  The marquee lit up as well.  Then the game pcb powered up and you could hear it playing.

The problem is in the monitor  

Here is a picture inside the monitor showing the yoke connecting to the back of the CRT and the metal HV cage.  Not picture are the circuit board that control all this stuff.

I pulled the monitor out of the case and looked at it.  Definitely a scary thing to work with.  I had no idea what was wrong with the thing.  And CRT tubes are huge capacitors.  They can store tens of thousands of volts.  So the thing has to be discharged before it's safe to work with it.   I read up on the subject, found all the info I could in the internet.  You will hear about the "screwdriver" method, basically where you take a screwdriver, connect it to ground and touch the anode.  It causes an impressive spark and discharges the CRT.  My friend Al has a good description on his page:  link.  The problem is, on the vector monitors, they have a diode on the anode wire (the wire that connects the high voltage transformer to the CRT).  This diode is fragile, and WILL blow if you try to do it this way.  You need to discharge the monitor slowly through at least 1 meg of resistance.

The Probe

I decided to take the safest route, and use a high voltage probe.  My reasoning on this was that I had no idea what was wrong with this monitor.  It might be in the high voltage circuit, and regular multi-meters don't go past 1000 Volts.  And we're talking about voltages 20-30 times higher than that.  How can I trace a circuit problem if I can't even measure the voltages?  Plus, it's definitely the safest way to discharge a monitor, and I plan to keep living a while longer.  Unfortunately,  these probes are $120-$150 brand new.  That's not an economical way to do things.  So where do you go when you want something expensive for cheap?  EBay.  Again.  So I searched and searched.  I found a probe called the "Fluke 80k-40" HV probe.  Exactly what I needed.  It has two banana plugs and an alligator clip.  It connects to a regular 10 MegaOhm multi-meter and divides the voltage by 1000.  So when your meter reads 10 volts, it means there's 10,000 volts at the tip.  Pretty nifty.  But these things, even though $120 new, were religiously going for $50 on Ebay.   That's still too much.  Then I got hit by a stroke of luck.  I found one called the Beckman HV-211.  It looked EXACTLY like the Fluke probes.  This one was for 22MOhm multi-meter, which are pretty rare now.  But at the voltages we were talking about, the tolerance was still around 2%.  And the price was $2, no reserve and $8 shipping.  With only 2 hours left.  Could it be?  I placed a bid and waited 2 hours, and won!  $10 for a $150 probe!  It took a week to get, so I just put everything on hiatus until then. 

The probe arrived and worked great!  I played around with it until I knew what I was doing.  It's basically a resistor network, and you can check to make sure it works by measuring resistances between specific contact points (so you don't have to have a 30k voltage).

Time to attack the anode

The first thing you need to to when working with a CRT tube of any time is discharge the CRT itself.  It can hold potentials of 30k or more.  The best way to do this is to use a high voltage probe to discharge it from the anode wire.  The anode wire is connected on one end to the high voltage transformer, and the other end is connected to the CRT, under a suction cup.

So I slipped it under the anode suction cup and measured the voltage.  Nothing!  All this waiting, and the stupid thing had no charge.  Better safe than sorry, I guess. So I disconnected the anode wire from the CRT, connected the metal plug on the CRT where the anode had been to chassis ground (the flat wire with the springs) and went back to the manual for the monitor.  How do you disconnect the anode from the CRT?  It's actually a spring clip that you can squeeze and pull out.  A lot of people use their screwdriver or probe to depress one end, then the other.  I recommend this way because the CRT can build up charge again, which is why you connect it to ground, to keep it safe.  Here's a picture of the spring clip here:

The vector monitors have a nifty circuit called a "spot killer".  If there's no signal from the board, or the high voltage circuits are messed up, this "spot killer" circuit cuts on and kills the picture on the screen, and lights a red diode.  Otherwise the CRT tube could be permanently burned.  When I plugged the monitor into the machine, my spot killer light was NOT on (it always blinks for a split second when you power it up, but it should turn right off again), so this meant the circuit was getting current from the high voltage board and a signal from the PCB.  Hmmm.  On top of that, I could hear what's described as a "squeaking" noise.  Or "vector chatter".  Basically is a tiny, faint clattering sound that tells you that the machine is trying to draw the vectors on the screen.  So the vectors were drawing, and the spot killer was off....and no high voltage on the anode wire.  My attention turned to the high voltage transformer, or the diode on the anode wire.  I hoped it wasn't the transformer, because that would be an expensive piece to replace, plus it's custom for the monitor, so no replacement parts could be found.  I ignored the transformer and concentrated on the diode.

The Diode

This is one mother diode.  About 1.5 inches long with two huge rubber sockets on each end.  It serves as a rectifier on the current coming out of the transformer, turning the AC into DC so it can charge the CRT.   Picture here:

 The CRT wasn't charging, but all the other voltages coming from the transformer were checking out right on the high voltage board.  So I fiddled with the diode.  I pulled it out of the sockets and put it back in.  Then I hooked it back up to the CRT, reconnected the machine to the CRT (Never work with a plugged in monitor) and fired that puppy up.  I heard a noise that gave me goose bumps.  There was a very faint static sound.  Like the sound that's made when a piece of paper sticks to a television from static electricity.  I turned the monitor around and there was a picture!  Woo hoo!  Doin' the happy dance! Ya mon!  I actually took a picture of this moment.  This is the Asteroids Deluxe test grid:

I quickly put the monitor back in the machine and played my first game!  Then my wife played a game, and then the phone rang and I left the machine.  When I came back, the picture was gone again.  Damn.  I pulled the monitor out again, and fiddled with the diode.  The connection was bad somehow, but I couldn't figure it out.  The two rubber sockets seemed okay, but just for kicks I decided to solder the diode in place directly to it's leads to see what happens.  To my shock, the one lead fell right out of the diode!  And I had found my problem!  All I had to do was replace this diode.  By the way, one hint on working with the HV cage. It gets it's ground from the monitor chassis, so if you unscrew it from the monitor and lay it flat to work with it, make sure part of it is still connected to the monitor some how.  In the above picture I have it sitting on my rubber mat, but also partially sitting on the metal rim of the monitor.

The Diode Replacement

One snafu though.  This diode, the VARO H1802  is so old, it's replacements (NTE527A or SK7333) are discontinued as well.  This was a bummer.  I know several guys who have the exact diode I neede, but they didn't want to be in the same predicament I was in, and they wouldn't sell me one.  Then I found the NTE505 diode.  It seemed to have all the correct ratings.  And only $5!  It is also becoming  a scarce diode, but places still sell it.  I got mine at http:\\www.circuitspecialists.com.  They only had 6, so I bought three for $20 including the shipping.  The "footprint" is a little different.  Instead of wires sticking out of the ends, the diode has metal end caps.  It was obviously made for a socket that I don't have, so I soldered the anode wires directly to it, and covered the whole connection with blue RTV Silicon ($3 a tube).  Why?  Because otherwise the 30k voltage from the CRT could arc over the diode and burn out the circuit.  The RTV is an excellent insulator.  But make sure you thoroughly cover the connections.  The diode is actually rated differently depending on what medium it's in.  It's rated at half the current if it's in air than in Silicone. This is because of the high voltage arcing.    The diode barely fit in the cage.  Here's a picture of the ugly but functional mess:

I turned off the lights in the room and tested for arcing.  None.  Then I powered up. Viola!  Same static sound, nice sharp picture.  Enough static force to hold a piece of paper to the screen (my new test for a charged CRT).  I reassembled the monitor and put it back in the machine.  It's been working ever since (3 months so far, with several week long tests).  So for the Electrohome G05, the NTE505 is a good replacement for the VARO H1802.  I can now say this with experience.