commgrp@silver.bacs.indiana.edu (08/04/88)
also posted on rec.autos.tech: What experiences have people had with car engines controlled by digital computers? Are they a mature technology, or are they like color television in the 1950's? Most car problems are electrical in nature; electronic engine control would seem to decrease reliability. [On the other hand, car seatbelt buzzers are the most reliable devices in the universe! How did they do that?? Must be a spinoff of SDI! :-] A friend owned a Datsun with computer-controlled engine, circa 1983. Her car randomly failed to start. Replacing the $300 computer module solved the problem for 2 weeks. After several new computers, another friend discovered that the connector was corroded. I've heard the following rumors about digital engine-control computers: 1. They contain a register which records the highest speed at which the car has been driven. This can be read by dealers, who will refuse to repair engines under warranty if the number is too high. 2. A new car's computer assumes default parameters, and sets new ones as it "learns" the car's performance over 50 miles of driving. If the car battery is disconnected, the computer forgets the learned values, and the engine runs less well for 50 miles while the computer retunes itself. (Is the highest-speed memory erased by disconnecting the battery?) 3. The federal government has funded research (thusfar unsuccessful) to discover a ray like Batman's which stops car engines, for law- enforcement purposes. Big Brother plans to require a remote-disable feature on engine computers. 4. Radio-frequency interference can jam the computer, causing various effects like engine stop or full throttle. This definitely happens with radio transmitters mounted in the car; there was a long discussion of it on rec.ham-radio last year. Manufacturer's recommendation: Don't install any 2-way radios. What happens when you are next to a semi with a 100-watt CB "linear" amplifier?? -- Frank W9MKV reid@gold.bacs.indiana.edu Cogito, ergo Spam: I think, therefore I ham.
jbn@glacier.STANFORD.EDU (John B. Nagle) (08/04/88)
In article <7200006@silver> commgrp@silver.bacs.indiana.edu writes: >What experiences have people had with car engines controlled by >digital computers? Are they a mature technology, or are they like >color television in the 1950's? Most car problems are electrical in >nature; electronic engine control would seem to decrease reliability. > >I've heard the following rumors about digital engine-control >computers: > >1. They contain a register which records the highest speed at which >the car has been driven. This can be read by dealers, who will refuse >to repair engines under warranty if the number is too high. Somebody put that in a Cadillac engine control program a few years back. A bit was set in non-volatile memory if a speed of 85mph was exceeded. Apparently this was just something some programmer stuck in, and there was a minor flap about the "invasion of privacy" thus represented. I did some work involving Ford engine control programs in the early 1980s, and at that time, there was no long-term storage in Ford's system, based on a custom chip later sold as the Intel 8061. There were vague dreams of storing info about, say, the last five or so times the engine had done something unreasonable, such as stall or miss. But nothing was implemented along those lines through '83, when I lost touch with that effort. >2. A new car's computer assumes default parameters, and sets new ones >as it "learns" the car's performance over 50 miles of driving. If the >car battery is disconnected, the computer forgets the learned values, >and the engine runs less well for 50 miles while the computer retunes >itself. (Is the highest-speed memory erased by disconnecting the >battery?) There was some question at one time as to whether such a setup would consititute a "defeat device" under the terms of the Clean Air Act. The problem is that any mechanism not exercised by the "city" and "highway" test cycles that affects engine performance that affects emissions can be illegal. So there are regulatory obstacles to long-term adjustment. >3. The federal government has funded research (thusfar unsuccessful) >to discover a ray like Batman's which stops car engines, for law- >enforcement purposes. Big Brother plans to require a remote-disable >feature on engine computers. The California Highway Patrol actually funded research in this area shortly after World War II. But I hadn't heard anything recently, not that I would be likely to. Comments? >4. Radio-frequency interference can jam the computer, causing various >effects like engine stop or full throttle. This definitely happens >with radio transmitters mounted in the car; there was a long >discussion of it on rec.ham-radio last year. Manufacturer's >recommendation: Don't install any 2-way radios. What happens when you >are next to a semi with a 100-watt CB "linear" amplifier?? I know that at Ford, they had some trouble with this, but overcame it. One of their tests involved driving by some huge radionavigation site in Michigan. The Ford systems circa '83 deliberately had limited authority, in that they could only tweak the throttle about 20%, for example (this was a bellcrank arrangement, and you could feel the throttle servo pushing back at you.). There was also a hardware backup system (referred to as "limp-home mode") which allowed driving with the computer down. In this mode, spark timing was fixed, so speed was limited, hence the name. Generally the vehicle electronics were designed to operate with up to a 30 watt radio transmitter on board, this being about as large as police or business band gear gets. Hams have special problems. If you want to run a few hundred watts in a car, this comes under the heading of "vehicle modification" and some engineering work on the vehicle electronics will probably be needed. John Nagle
fritz@mit-caf.MIT.EDU (Frederick Herrmann) (08/04/88)
In article <7200006@silver> commgrp@silver.bacs.indiana.edu writes: >I've heard the following rumors about digital engine-control >computers: >4. Radio-frequency interference can jam the computer, causing various >effects like engine stop or full throttle. This definitely happens >with radio transmitters mounted in the car; there was a long >discussion of it on rec.ham-radio last year. Manufacturer's >recommendation: Don't install any 2-way radios. What happens when you >are next to a semi with a 100-watt CB "linear" amplifier?? In '86 I was a summer student at the GM proving grounds in Milford MI, working with the electromagnetic compatibility (EMC) group. (Disclaimer: I didn't speak for GM then, and I don't now). GM does a lot of testing to try and make sure problems like this don't occur. Radio-frequency tests are performed by parking vehicles in large anechoic chambers and zapping them at various frequencies, amplitudes, and antenna configurations; and tests for mobile transmitters are performed outdoors in a specially constructed shed with no metal parts. Tests are also performed with 60 Hz magnetic fields, as might be encountered around power lines, and with electrostatic discharge simulators to make sure you don't fry the stereo when you scoot your fanny across the upholstery. The worst case for ESD is when you get out of the vehicle, and then reach back inside to flip a switch. Charge in conserved, so when the capacitance goes down the voltage goes up (Q=CV). It should be noted that the engine control unit is also a source of EMC trouble. Digital designers go for more computing power and faster clocks, but those fast rise times have significant high frequency components that can cause problems for other electronic parts. While I was there, GM was interested in using more component-level bench tests, since they are easier to perform than whole-vehicle tests in anechoic chambers, but I don't know how actively they are pursuing that idea. Frederick Herrmann fritz@caf.mit.edu
jans@tekgvs.GVS.TEK.COM (Jan Steinman) (08/05/88)
<I've heard the following rumors about digital engine-control computers: 1. They contain a register which records the highest speed at which the car has been driven.> Not without a drive-line transducer, they don't! This is possible on models with factory cruise control, otherwise it would be an added cost that probably isn't worth the return. <2. A new car's computer assumes default parameters, and sets new ones as it "learns" the car's performance over 50 miles of driving.> I don't know the specifics (Where did 50 miles come from? How does the computer "know" 50 miles -- see above.), but my '86 Ford Ranger seems to do this. For a brief period after changing from super to regular (or vice-versa) it runs a bit rough. <3. The federal government has funded research (thusfar unsuccessful) to discover a ray like Batman's which stops car engines,...> See #1. If it costs extra, there will be a big stink about getting it in. Look at what it took to get the auto industry to add pollution control and crash restraints! I think this particular bit would be costly. <4. Radio-frequency interference can jam the computer,...> I've not experienced this, even when using 100 watt HF gear into short antennas, or 30 watts of VHF into a rubber duckie. My after-market performance-computer went nuts, however! :::::: Software Productivity Technologies -- Experiment Manager Project :::::: :::::: Jan Steinman N7JDB Box 500, MS 50-383 (w)503/627-5881 :::::: :::::: jans@tekcrl.TEK.COM Beaverton, OR 97077 (h)503/657-7703 ::::::
fritz@mit-caf.MIT.EDU (Frederick Herrmann) (08/06/88)
In <1202@mit-caf.MIT.EDU> I described electromagnetic compatibility testing at GM. I received a question about my use of the term `anechoic' chamber, but was unable to reply to the sender, so am posting this reply to him and anyone else who might have been wondering: Sorry, anechoic means no-echo. The walls and ceiling of the room have cones of conductive foam all over. The cones are quite pointy, so incident waves should reflect into the next cone instead of back into the room, and after a few reflections the energy should be dissipated. The walls are heavily shielded behind the cones, so that nothing gets in and nothing gets out. Fiber optic cables are used to get signals from various instruments on the inside to computers on the outside. GM's EMC group has two such chambers (one big enough for a truck cab), and several shielded rooms (without the cones). That describes an electromagnetic anechoic chamber; acoustical anechoic chambers are similar, built with cones of a suitable material... I saw one when I toured GM's noise and vibration lab. Hope that answers your question... - Fritz fritz@mit.caf.edu Repeated disclaimer: I worked at GM for one summer, I didn't speak for them then and I don't now.