BADRAM.THD --- Copyright 1987 by Phil Wheeler An original compilation of Compuserve Model 100 Forum messages for use by Forum members only. One hazard of the Model 100 family is the "cold start" -- loss of everything in the machine due to some ill-natured fairy. These are recent messages on this subject: Possible causes, symptoms and recovery approaches (to get back DO files) Message range: 177294 to 177552 Dates: 12/1/88 to 12/8/88 Sb: #M100 bad RAM? Fm: Dan Fandrich 723306 To: ALL I hope someone can help with this one. My Dad bought a stock 24k M100 about two years ago so he could do some writing at the lake, in the mountains, etc. After about a month and 20k of writing later (with the machine always plugged into the wall between uses, so it's not the battery), he turned the machine on to an empty screen, the date showing January 1, 1900, the time 00:00:00. Needless to say, he now takes a large pad of paper and a pencil to the mountains. In the meantime I used it n without any problem for a while, then built an 8k memory expansion (with a 6264LP RAM and 74HC147 priority encoder, which lost bits over a period of days but that's another story) and it eventually happened again. Then a few days ago when I was trying to run the BASIC loader for DSKMGR (maybe XMD100, but it's probably not important) which had worked flawlessly for a couple of days, I got an 'FC' error. The LIST showed the 1 line program looked OK, so my assumption is that there was some kind of memory cn. Anyway, after a minute or two of trying to get it to work, ZAP -- my model 100 time machine took me and my RAM back to January 1, 1900. One final problem (make that nuisance). Once in a while I turn the machine on and the date is one year later, but the RAM is (seems?) intact. Well, there you have it. The tragic story of "My life with the M100". Luckily I can still end it on a happy note. If anyone's heard of this problem before, please let me know, because actually using the machine in its present state is like playing Russian Roulette. Fm: Tony Anderson 76703,4062 To: Dan Fandrich 72365,306 If your dad experienced a date and time change, but all files were still on the menu, all he had to do was go into BASIC and reset the calendar and clock with DATE$ and TIME$ commands. If the files were gone, and he had a clean menu, he experienced what is called a "cold start", whersystem pointers are reset indicating it's an empty machine. The files were still in place, but the system directory had been cleared. You could still have retrieved the files, had you known how. Cold starts are one of the random events in all "partitioned computers", where both programs and files are stored in RAM at the same time, and move around, depending on what else is happening. You experienced it too, but your's was more dramatic, since it happened while you were actually using the comp.. Same diagnosis; the files were still there, but the directory had been cleared. You could have recovered them. The date changing is a well-documented bug in the ROM code; a clash between the system interrupts, and usually happens when the clock routine is being updated while you are also trying to send data out the RS-232 port at higher baud rates. Would you be interested in selling your Model 100? That will make the problems "go away". Fm: Wilson Van Alst 76576,2735 To: Dan Fandrich 72365,306 I have two M100's and a T200, and what you're describing -- the way you're describing it -- has never happened to me. The 100s are known to have a bug in their calendar routine, resulting in an occasional problem with the year-component of DATE$. The quirk has been traced to the Modem/RS232 interrupt, and it occurs when you're engaged in high-speed communication with another computer, a disk drive, or an external modem. I have never a cold-start resulting from this bug. I respectfully disagree with Tony Anderson's opinion that crashes are random and apparently inevitable events. When they happen, they happen for a reason, and 99 percent of the time the problem is a software conflict. The designers of the M100 ROM very generously provided a number of "hooks" that other machine -language programmers could attach special routines to. But when one program tries to grab another one's "hook" -- or a program is removed from its runninbefore its "hooks" are disabled -- the computer will follow an unpredictable path that often ends at the North Pole. When you got that FC error from an apparently uncorrupted BASIC loader, it's an almost certain sign that the m/l program was grabbing for a hook that wasn't there. You should re-check the program document to be _sure_ your loading procedures were as specified. And if you had another m/l program running previously, you should double-check its instructions to be sure the hooks, if any, were neutalized before the program was deleted. (Some software ignores this process. Shame!) But none of this accounts for the way you describe your dad's mishap at the lake. It sound as if there was nothing in the machine except the factory firmware and his musings. If that's true, you _may_ have a defective computer. There have to be some of those lurking around in the world; but if yours turns out to be one, it would be mighty rare. Wish I had more specific help to offer, but I'm afraid you're just gonna have to live with the machine until you determine, once and for all, whether it's a lemon, or one of the very reilable and rugged M100's that thousands of people put to work without problems every day. Fm: Tony Anderson 76703,4062 To: Wilson Van Alst 76576,2735 Well, I would agree with the general principle that cold starts are not always random events. Sometimes they can be traced to machine language conflicts or other causes. But lacking such background information from the original problem description, sometimes it's easier to give a general answer, than to try to elicit exact details, and try to determine the true cause. But the fact that they occur in other partitioned comps, indicates they are not a unique problem to the Model 100 family. Did I use the word "inevitable"? Perhaps "inevitable" is a poor word in this context. Cold starts could only be "inevitable" in the sense that they can occur despite the most careful operation, most stringent attention to detail, and to the most knowledgable of users. The possibility is certainly lurking in all such machines, and we have found other causes, in addition to software hook conflicts. But as you know, it's often impossiboint an exact cause, after the fact. Particularly MUCH after the fact, where exact details are not known, and where users were not sufficiently aware of the problems to know what they had actually done that might have contributed to, or created the problem. Fm: Bill Brandon [DPTRAIN] 76701,256 To: Dan Fandrich 72365,306 Dan, could it be that there is a loose connection on the 100's PCB? A chip which had backed part-way out of its connector could give an intermittent cold start. Since I infer from your having built a memory expansion yourself that you are handy with this sort of thing, why not pop the case open and have a look? It couldn't hurt. . . You will probably have more than one cold start, though they get less common as you get more familiar with the machine. (I once cold-started mine because of the way I was holding it - with left index finger inadvertently pressing reset button - when I turned the mn. Embarassing!) Also, using various programs (even straight BASIC ones), especially m/l, will sooner or later lead to the dreaded January 1, 1900. I keep this little goodie taped to the top of my PDD for those occasions: 10 OPEN "TEMP" FOR OUTPUT AS 1 : FOR X = 32768 TO MAXRAM : PRINT #1, CHR$(PEEK(X)); : NEXT Even after an arctic start, you can type this into BASIC, run it, and recover a good bit of what was in RAM - you should get _all_ of your .DO files back. There will be lots and lots of garbagett out, but it shore beats having to redo all that golden prose. Fm: Wilson Van Alst 76576,2735 To: Tony Anderson 76703,4062 I don't think you used the word "inevitable" -- but that's the sense I got from your message: that cold-starts are something to be taken for granted -- sorta like occasional irregularity -- and there's not much you can do about them. The inference, to me, was that the M100 and similar computers are entally unreliable. I just don't believe that. As far as I'm concerned, the firmware operating system in these computers has proved itself to be as rugged as the mechanical frame that supports it -- which is plenty. Certainly the computer is vulnerable. Slippery and gooey software will do to the OS what a milkshake or a can of glue will do to your keyboard. But where you say cold starts "are not always random events", I say "are never random events." In a system as well-proved as the M100'I believe there's a preventable cause for each and every crash. Some may be too subtle for diagnosis at a general user's level of expertise (eg., speed differences among varous 3rd-party RAM chips); but none, in my opinion, is just a random occurrence. What you and I fully agree on -- there is _no_ way to know what happened to that computer at the lakeside cabin last year. Maybe the butler did it. Fm: Tony Anderson 76703,4062 To: Wilson Van Alst 76576,2735 Boy, that's the most ridiculous thing I've ever heard! The butler wasn't even NEAR the cabin. (grin) Howsomever... there have been reports of spontaneous cold starts, happening all by themselves, while the machines involved sat unattended and unused. (Maybe it was an automatic 30-day purge that we don't know about.... sort of a delayed laxative effect.) (grin) It was even suggested once that such might be attibuted to stray cosmic particles bombarding the CMOS chips. (hows THAT for esoteric?) Then there are a couple of cases of it happening during battery replacement, and/or because batteries were not making a good contact with the springs in the battery compartment. And I believe there was one case reported of it happening after the machine was passed through an airport X-ray machine. Hard to substantiate any cause for such reports that would be something within the geeral users control. Or perhaps within his realm of awareness about the machine itself. The original message in this thread seemed to say the user was a "buy it and use it" type of fellow, who was quite surprised that the machine was so unreliable that it would spontaneously cold-start, and who had not taken the time to develop a background in the machine, which would lead to precautionary use, which might have avoided the occurrence. Fm: RANDY HESS 73267,552 To: Tony Anderson 76703,4062 Leaving a machine plugged in to a "cabin"'s electric supply may have been the simple cause of the original problem since an interuption to the 120VAC for longer than a FEW minutes with the converter connected will nicely clean out even the most coddled RAM! These machines are so user friendly that we take for granted a lot of conveniences that the so-called "big-boys" dream about! The very fact that VERY unsophisticated userdo so much with these computers with hardly ANY training speaks volumes for the ruggedness of their design and quality of the programming. Fm: Tony Anderson 76703,4062 To: RANDY HESS 73267,552 Ah, but it shouldn't. In the event of an AC power outage, the internal nicads are supposed to keep the contents of the CMOS RAM intact. And theoretically, for eight days or more. Fm: Wilson Van Alst 76576,2735 To: RANDY HESS 73267,552 Aha! So the butler didn't have to be near the cabin: he just cut the power lines. Fiendish folk, those butlers. Your analysis of the cabin crash is interesting. I've never had that specific problem, but can't rightly say whether it's because my wall sockets are dependable, or because I've got a NiCad charging circuit installed -- which may keep battery power flowing even when the wall-juice is ed. Fm: Dan Fandrich 72365,306 To: Dan Fandrich 72365,306 Thanks for all the input! I'm glad to hear that mine isn't the only M100 with that date bug at least. As far as the cold starting goes, it could very well be a loose connection somewhere -- I'll have to go in and take a look around. The first time the "cold start" happened, I wrote a one-liner that did very much what Bill Brandon's did, namely take a PEEK through all the RAM to display the contents, but I got nothing intelligable (all 0's and 0ffh's if I remember correctly). I have doubts as to whether or not this was a software cold start, but rather a loose power connection. The other times I didn't even bother checking because I figured the whole RAM was gone again. Had I realized that software cold starts aren't all that unusual, I wouldn't have be so hasty the next few times and overlooked that one-liner (the last time especially, after the FC errors). I've had more thanhare of problems with m/l hooks and nonstandard software and the like. I think I've quite easily _recovered_ more data from my CP/M machine's hard disk than I've created on it. (My thanks to Peachtree Software for their Magic Wand word processor that writes to group 01F8h on the hard disk whether something's there or not...) With my luck, the M100 probably is that one-in-a-million lemon. What else could you say to a person who bought a Timex-Sinclair 1000 that overheats after 10 minutes of use, and an IBcompatible whose real time clock speeds up by a factor of 3 when it's warm? At least it took 5 years before my Commodore 64 broke down. Oh, that's right. That one was the third warranty replacement. Fm: Bill Brandon [DPTRAIN] 76701,256 To: Dan Fandrich 72365,306 Dan, have you considered the possibility that you may be telekinetic? Or afflicted by poltergeists? Fm: RANDY HESS 73267,552 To: Tony Anderson 76703,4062 I'm glad to hear that but it's happened to me morethan once and, although I've never studied a schematic, used "empirical" evidence to assume the worst. I'll try to find a schematic. Fm: Dan Fandrich 72365,306 To: Bill Brandon [DPTRAIN] 76701,256 I knew you'd say that! :->