Thread Below--

Mike Hyatt lists devices with imbeds (no software in this question)

The list goes from appliances to Xerox machines. (nuclear plants) Albert Eistein just named Man of the Century.

Question: Is it a truth that the engineers that manufactured the devices known as imbedded chips honestly do not KNOW if the device that they manufactured has an aspect of its engineering that will seize or compromise its operable ability when a certain date is reached.

I'm not an enginner or scientist, I find it a little difficult to believe that you can manufacture a machine and not know its stress points or a least be able to test the stress points. And come up with an absolute. YEA or NaY??

Where is Good housekeeping when you need them???

-- d----- (dciinc@aol.com), December 27, 1999

Answers

d,

I'm no engineer or expert either. However, I can tell you that you need to understand the difference between an embedded "chips" and an embedded "systems" which could include a number of different chips, software, etc. or customized boxes built specifically for their respective applications in industry, etc.

We're talking BILLIONS of chips and numerous points of vulnerability, or not.

Mike

===================================================================

-- Michael Taylor (mtdesign3@aol.com), December 27, 1999.

Clearly y'all hain't never done no software/firmware engineerin'. They use off the shelf chips and just use the functions they want to use, regardless of the internal housekeeping being done by the chip/ware itself. (Or so a OEM guy once tol' me, over a few adult beverages.....)

Night train

-- jes a confused ol footballer (nighttr@in.lane), December 27, 1999.

Michael,

I will assume they CAN know about the chips but not the systems the chips are in?? Still seems a stretch to manufacturer devices and install them in a system and not be able to concur absolutes of stresses involved.

-- d----- (dciinc@aol.com), December 27, 1999.

Also you need to realize that a co. may use a part from another co. in their device. Then another co. may buy that device to use in their own. They in turn sell this product to another co. that uses it to operate a machine or something. Meanwhile the first producer is bought by a larger co. the second producer went bankrupt. The third no longer produces these products anymore, and the workers at the last plant are on strike. The whole thing is just a mess.

Lots of Co.s can't cut through all this B.S. and even if they could they couldn't find out which part will fail because its impossible to track down. So they are going to wait until the part stops working (fix on failure) and replace then. The problem is there isn't a large enough inventory for everyone to run out and buy a new Cutler Hammer switch for their applaication. And even if they could their isn't a guareenmtee it will work in their applaication and the electricians who will be needed to install will be at other plants trying to troubleshoot the same thing.

This embedded chip thingy is much larger than most understand.

-- Johnny (jljtm@bellsouth.com), December 27, 1999.

Many on the list are not a problem! When equipment designers develop an application or device they rarely develop all the electronics to support that device. They use off the shelf stuff that's available. Many PLC's, controllers etc. may have RTC'S (real time clocks or secondary counters) that may not be part of the design or needed to function properly. However in many cases these clocks are active and not visible or re-setable (in a normal fashion). It is these types of devices that are the great unknown of Y2K. Some are going to fail (00 may reset the clock and function normally or in some cases a failure signal to shut down the device). That is why many on this forum have prepared to the hilt...Nobody wants catastrophic failure (except a few of the chem-trailers) it's just better to be prepared than face the consequences.

Doomer

-- Polly-Morphic Doomer (Greenem31@aol.com), December 27, 1999.

I my wanderings I have heard that Texas Instruments had something like a 97% error rate in their clean rooms (at one point). Also, as a software developer, I can tell you that it is easier than you think to build something more complicated than you can fully understand. How many people can say how many files they have put on their PCs hard drive and exactly which folders they are in? Theses projects are like Legos. You just keep adding and adding. If care is not taken, you may end up with one big mess. One that almost no one understands.

-- Reporter (reporter_atlarge@hotmail.com), December 27, 1999.

Many of the embedded chips aren't calendar date sensitive. Approx. 80% of the embedded chips don't give a flip what calendar date it is, they function based on days in a week, days in a cycle, etc.

The 20% that are calendar date sensitive should have been replaced, reprogrammed, or bypassed before the software was touch.

One would assume that this was one of the first things done with all those billions of dollars spent fixing Y2K.

You can program until the "Cows Come Home" but if the software you are working on deals with calendar date sensitive embedded chips and they aren't made compliant first then you just wasted you time.

'Have A Fantastic Day'.

-- KWT (ktuttle@netbird.com), December 27, 1999.

problem is that you design a chip, but not necessarily the code that is burned on to it. And even if you did, frequently changes are made to the code during a production run. So chips made before such-and- such a date had one version of the code and after that date the prom had a different (to some degree) instruction set burned into it. Now that happens ALL the time. Some production runs of chips for embedded systems had hundreds of code changes over the course of the production run. And no changes to any aspect of labeling were made to the chip so even sitting side by side, you cannot tell which version of the code does which chip run.

And then you do not know who used your chip in what and how.

Basic conclusion is that....we be hosed.

pliney (feeling a whole lot older this week and it is only monday) the younger.

-- pliney the younger (pliney@vallier.com), December 27, 1999.

Reporter,

Unless of course You are Koskinen,Clinton,Richardson,etc.

-- d----- (dciinc@aol.com), December 27, 1999.

KWT I agree with most of what you are saying.... However, most industries in the US have not remediated embedded chips (flashed ROMS, EPROMS etc.) because of the difficulty and expense in tracking down these buggers. Corporate America has decided to fix on failure when it comes to embedded technology. The spin from the Govt. assumes that manufactures compliance statements are gospel. Microsoft should be a good example to us all about compliance statements. (although most of the code problems are equivalent to the early Pentium slipped decimal problem). The embedded chip issue is truly the great unknown with respect to 01-01-00 problems. 2-5% of failures in critical infrastructure will bring the systems down, maybe one piece at a time but the hand writing is on the wall. The oil and water scenario alone should make things interesting by mid Feb. or March.

Don't wannta be but.... Doomer.

-- Polly-Morphic Doomer (greenem31@aol.com), December 27, 1999.

Systems theory always encouraged designers to treat, say, counter chips as "Black Boxes" that count, and to pay no attention to what went on inside.

Well, ALMOST ALL counter chips were made with an Epoch Date embedded at time of manufacture...

Now the Black Box fallout is about to infect the world. LOL!

-- Z (Z@Z.Z), December 27, 1999.

--d-----

I don't know much about embedded's with the exception of what I have read. This is a link to one of G.Ns posts, were Beach explains what the secondary clock is, and how it may affect embedded's. Hope it help's, I haven't heard anything concerning this issue since North posted this in April of 99. http://www.garynorth.com/y2k/detail_.cfm/4399

-- ~***~ (~***~@earth.ebe), December 27, 1999.

I posted this link a few months ago here, but it may be helpful to post it again. This is a "literary" description of how an embedded chip came about, not a literal one. But I think it helps non-technical people to get a clearer picture of how it could be that no one knows what these chips might do or how they might act.

http://www.drivezero.com/herbal/framec/lampoilc.html

-- Katherine Carr (livineden@yahoo.com), December 27, 1999.

Ms. Carr,

-------------------Thank-YOU------------------

-- d----- (dciinc@aol.com), December 27, 1999.