In Reply to: Re: Dunce question posted by andrew on 02/23/03 at 6:04 PM:
Hello Andrew. The gear ratios you have listed are a bit odd for simple counting. I would suggest that you use a 100 pulse-per-rev optical encoder, which, for one rev of the input shaft would provide counts of 346, 221, 140, 100, and 371 respectively for each of the 5 gears. These numbers would then be divided after capture by a 1:10 divider, then by ignoring the "ones" digit, there would be left 5 separate and discrete numbers of 34,22,14,10, and 37. Ignoring the "ones" takes care of any bobbling in the count due to minor differences caused by phasing errors (the same gear tooth is not always meshed every time a shift occurs). A decoder for these abbreviated numbers would then provide a separate output for each gear meshed, and you could use this for whatever sort of display LEDs, or whatever other indication you might prefer. Although the encoder would be running backwards for "Reverse", the encoder would still be providing a "37" output, and no additional directional switching is required.
"Neutral" would, of course, not provide any count, since the output shaft is not rotating. Likewise, if the clutch was enabled, regardless of gear selected, the gear selection indicated could be erroneous, as the output shaft could be coasting due to being unlocked from the input shaft. I need to think about this a bit in case there is a way around this.
The idea of using the ignition pulse to provide a timing "window" depends on the ignition system in use. The traditional Kettering design has a centrifugal spark "advance/retard" mechanism which, when the engine accelerates, causes the ignition pulse to arrive sooner. This would have the same effect as shortening the "count window" -- and lengthening it on deceleration, which would cause the decoded gear-counts to be incorrect and the indicators could flash randomly.. Once the relative speeds have stabilised, however, the correct selection would show again. There are probably a number of other problem areas likely to be experienced without trying to introduce one deliberately, so I think it would be prudent to have a separate fixed input pickup. Incidentally, for a 4 stroke 6 cylinder motor there are usually only 3 ignition pulses per crankshaft rev.
If the ignition is some form of electronic system then it will be necessary to explore this before commenting. If the motor uses an Electronic Control Module then there will certainly be some form of reluctor/pickup unit installed to provide the key shaft position indication to the processor. This would be a good place to investigate whether this signal could be "tapped". Professional mechanics tend, with probable good reason, to distrust any non-commercial attachments wired into original-equipment harnesses, and it is the first thing which is snipped out if any motor problems are experienced regardless of diagnosis.
Returning to the output encoder installation, this would be best installed on a small bracket and driven from wherever on the output shaft looks practical. There is no mechanical loading with it and it is usual to drive it with a light toothed drive-belt. I recall seeing advertised a few years back a special "split" toothed gearwheel which came apart in two halves -- allowing it to be snapped around an existing shaft without dismantling anything. I seem to recall that you had to order the right size to suit the shaft diameter. Maybe an Internet search would show something up.
As you can see, the problem areas are not electronic. I suggest that the method of implementing both the input and output mechanical mods need to be addressed before getting into the rest of the design. In the meantime, I will ponder on the "neutral" and "coasting" situations and try to work something out.
Let me know what is what on the ignition system when you have time.