Quote:
Originally Posted by tsk1979 I was simply referring to articles which said high speed applications have trouble with electronic valve timing.
Low speed engine do not have those problems.
That said, won't using electromagnets solve the rare earth problem to a degree?
About reliability of electronics, I would disagree. Its more of a QC issue than a design issue. Military grade electronics and chips are still manufactured and can run in adverse conditions.
Even normal chips are taped out with an aim of 125-130 degree operational parameters. However to cut costs the casings are cheap quality which cannot withstand those temperatures.
In an expensive automobile a few dollars/chip will be acceptable.
Already we have components like cam sensor and sensors in the exhaust circuit. Exhaust circuit sensors run at much higher temperatures.
The biggest drawback i could see is the magnetic valves, as you mentioned. I am not well versed with electrical engineering, so you are right, there may be many problems in those areas. However, as far as ICs are concerned, its easy in current tech to manufacture chips which can survive 200 degree C temperatures without skipping a heartbeat.
Even consumer grade hardware like GPUs touches 100-110 degree C before thermal shutdown algos kick in. |
Tanveer
I saw your profile - we are in the same business buddy !
Coming back to the topic:
Electromagnets would require a significant DC current to hold the magnetisation and double the size (you need a moving coil and a static coil) not really a good solution, but may be workable. However all said and done this may eliminate almost all advantages that you mentioned (lesser components ...)
You can disagree about what I said about the reliability, but I'll stand my ground. The tape our of 125C is for junction temperature - not ambient. Even good packages cause a difference of at least 30C between outside and inside temperature. 200C of operation is not going to happen with CMOS or CMOS like technologies - (
sorry to everyone who doesn't have a background in semiconductors) every one of the reversed biased diodes will leak so heavily, it'll be a short for all practical purposes. On top of that thermal stresses (SiO2 and Si do not expand at the same rate, neither does SiNx) will be a problem within the chip and on the board.
Military grade and Military are not the same thing. Automotive temperature range includes 125C - but those are chips that go in music systems in cars (even car music systems are more reliable than what you get for home use). Military grade has more specs that I'm sure you would have never simulated. e.g. What happen if temperature increase by 100C in 3 seconds (chips used in rocket controls), what happens if the chip is going through a shock of 10g (almost all aircraft), a shock of 50g (even industrial sensors need that).
Exhaust circuit sensors are not ICs that you refer to, they are sensors - they give you a weak electrical (usually current) signal. Thermocouples tolerate thousands of degrees C but they do not do any signal procession whatsoever. And semiconductor doesn't mean silicon. Long before there was silicon there were other materials (I'm not talking about Germaniumor GaAs - CdS, Zinc Blende and stuff was what caused semiconductors to be initially discovered and used - a lot of them are still in use, though not in ICs, in sensors, displays and what not) however these other materials are no good for signal processing and controls - which is what you need here.
Regarding your comment on GPUs, imagine the scenario that you are playing NASCAR 2011 on your xbox, trying to overtake on the game and the GPU overheats - all hell breaks lose your computer shuts down you lose to your buddy for the first time in months !
Now imagine the same experience on a highway at a modest speed of 80kmph while trying to overtake a truck. GPUs have zero reliability if you use the lingo of Automobile designers (whose stuff has zero reliability if you talk to military types).
And you haven't even consider soft error rate (I mentioned it by its real effects - alpha particle strikes) which can play havoc with any and every algorithm you can consider.
And we haven't even started discussing what happens
after you make a chip.
And all this for what? Even assuming that all this were possible for $1, what exactly do you gain? Can the transmission take the higher rpm? how about flame front inside the cylinder, heat and mass flow ...
The technology will take immense investment to develop (if it is possible which is not obvious to me) for gains which are debatable do a good extent, and there are many other (better) uses the investment money can be put - look at Tesla.