Classic Computer Magazine Archive ANTIC VOL. 2, NO. 12 / MARCH 1984



A cautionary tale

by David and Sandy Small

Our "Nightmare Mission" series has now stretched over four issues of ANTIC (October 1983, November 1983, December 1983 and February 1384). In this series, we've covered some of the techniques and tools we used to develop a game cartridge in much less time than is usually required.

But, so far, this process has seemed too easy, too effortless. We've passed along some tips and shared some of our experiences and problems with you, but we've failed to mention the disaster that befell us on the day before our finished project was due to be delivered. You'll recall that if our game was not ready by a certain date, we would not be paid for our frantic efforts. This is the story of the final stage of our nightmare mission.


It wa's early Sunday night. Our contract called for the game to be delivered to Washington, D.C., on Monday. The program was completed, and we were getting ready to burn it into EPROMs. This involves taking the program front disk and copying it into EPROM memory chips (Eraseable-Programmable-Read-Only-Memories). These are then plugged into an Atari cartridge to form a cartridge-based game program.

You'll recall (from last month's column) that we wanted to use two 27128 16K EPROMs to create a 32K cartridge. We had plugged in an MPC EPROM burner via joystick ports 3 and 4, and had connected it to the two 27128's. We also had a fast, homebrew disk drive of our own design hooked in via the Operating System slot, a Bit-3 80-column board in the back slot, and a RAM-PAGE 32K board in slot 3. This gave us a total of 48K in the machine.


When it was time to burn the first EPROM, we powered up the Atari and our disk drives, loaded the EPROMMER program for the MPC, and reached down to plug in the MPC's power cord. Just then, there was this blinding flash! Sparks flew across the room! Was the world coming to an end?


After a few seconds, we opened our eyes. Computer reflexes took over: We yanked the plug out of the wall. The smell of burnt bakelite was in the air, and we knew that some circuits had burned out.

We looked at the EPROM burner. Smoke was drifting up from it. Its two capacitors had been blown apart (this is common when an electrolytic capacitor is suddenly overwhelmed). We followed the MPC's cord to the Atari's front ports: The front-port traces on the plugs had been fried right off the circuit board. All that was left was burn marks.

With a shudder, we realized that those traces led directly to the Atari's PIA chip, a sensitive component that needs to be treated carefully at all costs, and from there directly into the system data and address busses - which lead in all directions. Things didn't look good.

Saying a prayer for the soul of our new machine, we turned the power back on. This was an automatic reflex; the TV screen showed static, which indicated that the Atari was turned off. But the power was already on!

We had seen this happen once before, when a memory board had shorted out. Hoping that only the memory boards were affected, we stripped them out and replaced them. Still nothing. So we opened up the Atari to find . . . carnage. There were chips with the middles blown out of them, diodes with smoking innards, and pieces of LS logic scattered throughout the inside of the machine.


The bad news was not yet over. We turned to the Bit-3 ($350) and RAM- PAGE ($200) boards, and tried them in a different Atari. When we plugged them in, the machine quit working. Both boards were ruined.

To top things off, our disk drive system no longer worked (we sell these for $1500). We added up the damages, and found that the mysterious flash of light had cost us at least $3000. And, at that moment, we very nearly gave up on our game project. It was midnight. The project was due the next day. And our equipment had just been destroyed.

But we didn't quit.

Someday, something like this will happen to you. You, too, will face the wall, as we did, in a cloud of smoke, with a contract about to be lost, and discouragement in the air all around you. But you just can't quit. Being part of this field means that one day you will receive a baptism by fire. When it happens, you have to grit your teeth and do whatever has to be done. If everyone could do it, where would that leave us?


We still had one working Atari and an 810. We called Dave Mann, our local users' group president, hoping that he wouid still be awake. He was, and he offered to loan us his EPROM burner. By the time we got it, though, it was after 1 a.m. At that point, we started, once again, to burn the EPROMs (fortunately, we had a spare -- our first one had been cooked in the MPC).


While the EPROMs burned, which took several hours, we looked over our damaged equipment to try to determine the cause of the accident. We soon found it: a ground loop.

Several 110-volt lines enter the MPC EPROM burner and end on the circuit board. The unit's case fits tightly over the board, so tightly, in fact, that in this case one of the 1l0-volt wires actually touched the case. This electrified the case, and the ground, of the MPC machine. Ordinarily, you wouldn't notice that this had happened, because the Atari "floats" (i.e, isn't electrically grounded to its power supply). Unfortunately, our Atari was grounded, via the 3-wire disk drive units that our disk system ran. When we'd plugged in the MPC, 110 volts of A.C. had shot through the + 5-volt D.C. power supply for the MPC, Atari and disk drive unit.

Voltage measures units of pressure, and 110 volts exert much greater pressure than five volts. The insides of the integrated-circuit chip are designed to withstand five to ten volts of pressure. But not 110 volts. We were unable to find a single chip in any of our damaged units that had not burned out; all of them had blown apart internally under the force of the 1l0-volt current. And some of them had heated up so violently that they'd actually exploded, and blown the lid off the circuit base.


Could this happen to you! Do you ever hook up anything to your Atari that does not use one of the standard Atari power supplies (which come with the special connector)? Anything that plugs into an outlet is suspect. Examples include: Percom disk drives, the ATR-8000, the California Microlink 8" disk drive setup, and of course the MPC EPROM burner. If any one of these is defective, and provides a ground outlet for the 110-volt current, your Atari could be in danger.


To protect yourself and your equipment, you should check the grounds on these machines using a 110-volt A.C. voltmeter. They should register O as compared to the wall ground. Plug in one device at a time and check it for 110-volt leakage. Also, remember that the bypass circuits used to prevent this sort of problem are often blown out by a nearby power surge or lightning strike. After such an occurrence, it would be a very good idea to check your equipment for ground safety before turning it on.

In our case, our Atari was destroyed. If we had been caught between a 1l0-volt current and the ground, we could have been killed. It is certainly worth the investment to get yourself a voltmeter and be safe.


If you don't know much about electricity, and don't know how to check the grounds, your local users' group or dealer can probably help you make sure your equipment is safe. We've also seen: pamphlets on the subject at several local electronics stores. And a voltmeter need not cost more than $10 or so.


Finally, at about 5 a.m., we finished burning the EPROMs. We plugged them into the cartridges for the first time, and found that, of course, none none them worked. We shortly discovered the writing mistakes that had caused the problem, and soon we had all four cartridges running. By then it was 6:30. We rushed to the airport, caught a 7:30 flight to Washington, D.C., and hand-deliveied the cartridges to the software representative who met us there. This, at last, was: the end of our nightmare mission. And at the end of twenty-eight hours without sleep, it came none too soon.


After the trauma was over, we remembered that, some time back, we had purchased a special computer insurance package. Computers are not covered by the typical homeowners' policy, so we had looked for and found a policy designed to cover our machine (it's protected by Columbia National General agency, 88 E. Broad St., Columbus, Ohio 43215).

Most likely, your home computer (and all of your software) is not currently insured. If anything -- from theft to fire to a ground loop - happens to it, you're out of luck.

Our insurance policy covers hardware and software that is lost -- for any reason. So the day after we turned in our project, we called our insurance company and arranged to have them send a local adjustor to survey the damages. They promptly took care of this, and just as promptly settled our claim.

Professional software developers cannot afford not to have their computer equipment and software insured. We were very lucky to have had insurance; we could just as easily have had to completely write off the value of our damaged equipment and start from scratch. Insurance is not expensive, it can be comprehensive, and it can save you in a pinch. We highly recommend that you look into it.

We wrote this article for two reasons: To show that our "nightmare mission'' actually had some nightmarish aspects, and to warn you not to let the same kind of thing happen to you. The computer experience is not always an easy one; in our years of dealing with computers and computerists, we've heard many tales of woe.

Remember our warnings. But remember, also, that our story had a happy ending after all: Our insurance paid for the damages and our game program was finished on time. And best of all, we survived the ordeal.

David and Sandy Small are professional programmers who work extensively with Atari computers and Atari-compatible peripherals and software to produce commercial software for the Atari. In Systems Guide, they share discoveries, insights, experiences and secrets of professional programming that should be of interest to others who are at or near their level of practice.