This solution seems to be the best approach. If you know where the oil will pool and can get a sensor to that location, why not put in a pickup to clear it?
The system could be designed with an orifice and some method of measuring the pressure drop across it to detect when it was drawing in oil or just sucking air. This signal would keep the evacuate pump running and inhibit engine starting.
The advantage of such a system is that it would (eventually) allow the engine to be restarted after the sensor had determined that the oil had been evacuated.
I think a few people have suggested variations on this approach already.
--
Paul Hovnanian mailto:Paul@Hovnanian.com
------------------------------------------------------------------
Bloody typical, they've gone back to metric without telling us.
Damn - I was thinking of the time constants specified in the data sheets, which are mostly for thermistors immersed in a well stirred oil bath. In air they are roughly ten times slower - see
formatting link
on page 21 - on thermal time constants - and page 22 on thermal resistance to ambient (which puts the concepts in the wrong order, if you want my opinion).
You can get a time constants of five seconds in still air, but only for very small, minimally encapsulated thermistors
Perhaps I was a bit too terse. Here's the reasoning.
The Nichrome wires (pure nickel would be better, TCR ~ 0.6%/C) have a TC that is reasonably linear with temperature, unlike thermistors. Check out the difference in resistance of a carbon bead thermistor at -55C and
+125C--it's about a factor of 1,400 (see
formatting link
It isn't at all impossible to get a circuit with that wide a range working reliably, it just requires a lot more testing than a hot wire.
The single-thermistor idea suffers from the same problem in the industrial temperature range--it could be done, but would require significant smarts in the controller to set the excitation currents, and therefore significantly more involved testing, *in the engine*. Testing a large industrial diesel over a wide temperature range isn't something to be contemplated lightly, so a design change to something with a smaller delta-R would be pretty attractive. The application under discussion is low enough in volume that the difference in engineering time is very significant, I would think. If the engines are all operating indoors, it gets somewhat easier.
The thermocouples would have to be attached to something dissipating heat, obviously--the advantage being that you get the equivalent of a nice isolated four-wire measurement, rather than something jiggling up and down with the alternator current pulses, which is what you'd have in a typical industrial sensor. Even in a bridge measurement, any nonzero difference signal will have power supply noise on it. This disadvantage is not big enough to make thermocouples the best option for small (few-degree) differences, though.
Your carbon bead thermistor is a little on the insensitive side - the resistance of the "F" material Thermometrics SC30 thermistors (up to
10k at 25C) decreases by a factor of about 2000 over the same - military - temperature range.
formatting link
The "industrial" temperature range for integrated circuits is -25C to
85C, and the commercial range is 0C to 70C.
The original poster didn't specify the temperature range that had to be coped with, which suggests that the commercial temperature range would be more appropriate. The inlet manifold of a diesel engine will get hot when the engine is running, but the problem was detecting oil present when the engine is started, and presumably not much warmer than ambient.
Coping with a wide temperature range would complicate the design in any event - it is a judgment call whether the simplification you get from using an insensitive sensor balances the difficulty of coping with a ten-fold smaller signal - and it is by no means obvious that the OP would need to design for an extended range of ambient temperatures in the first place.
and
Why "in the engine"? And what do you mean by "significant smarts"? - the dumbest PIC could handle it.
under
up
in
Not if you went to the trouble of regulating the power supply to the sensor - on occasions I've used a cheap voltage reference to bias a sensor, to avoid exactly this problem
So get rid of the power supply noise - it isn't difficult, and doesn't cost that much.
option > for small (few-degree) differences, though.
The added mechanical complexity of adding a separate heater is likely to be even more of a discouragement than the pain of amplifying the tiny voltage signal you get from a thermocouple - and note that you can AC-excite a resistive temperature sensor (or use reversing DC)to get rid of amplifier offsets and adventitious thermocouple voltages, which you can't do with a thermocouple.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.