CDI schematic
Re: cool
I Sorry to get out on a techie trip here.
I may be the odd man out here, but your "techie trip" is exactly what intelligent readers need. I've long wondered why the concept of hundreds of opinions are desirable when good questions are posed. The answers from those that are not in the know will always outweigh answers from those that are, so by sheer numbers alone, information that is passed along tends to result in a general dumbing down of the reader. In many cases, the sole response that may go against popular opinion may be the only one that is correct.
My point? Keep getting "techie", it serves a good and valuable service to those that really do want to learn something instead of hoping to find others to support ill-conceived notions.
That's my opinion and I'm sticking to it! [img]/ubb/images/graemlins/grin.gif"%20alt="[/img]
Harold
Wise people talk because they have something to say. Fools talk because they have to say something.
Re: cool
Hi Jutz,
So in this example I take it to mean that the spark at present was initiated by a previous pulse, delayed to act as advance on the subsequent power strokes?
That might take more day dreaming to comprehend the logic behind a delay of this type to resolve a pending advance. Interesting to ponder this concept of past, present and future relationships in a less complex circuit.
Since this is the focus, to understand the concepts and then try to apply electronics to the application in real time. Knowing what the component can do and getting them to act in a group are very different propositions.
I really appreciate all the input to this subject. It just might help to keep me from over complicating this project.
DC
So in this example I take it to mean that the spark at present was initiated by a previous pulse, delayed to act as advance on the subsequent power strokes?
That might take more day dreaming to comprehend the logic behind a delay of this type to resolve a pending advance. Interesting to ponder this concept of past, present and future relationships in a less complex circuit.
Since this is the focus, to understand the concepts and then try to apply electronics to the application in real time. Knowing what the component can do and getting them to act in a group are very different propositions.
I really appreciate all the input to this subject. It just might help to keep me from over complicating this project.
DC
Re: cool
Not quite.
My airplane (like most piston aircraft currently in existence) has two magnetos. One magneto has fixed timing. The other magneto has an impulse coupling in its drive system which causes it to lag behind the other magneto when the engine is turning slower than a certain speed but catch up once the engine is turning faster than that speed. So you set the timing of the magnetos for running speed and then the impulse coupling retards the spark for starting. The impulse coupling is more complex than the direct drive, which is why you only have it on one magneto (don't want to add any more failure possibilities than absolutely necessary).
The circuit I envision would work similarly. Whatever ignition sensor you use would be set to trigger at the advanced position and a timer circuit would delay the spark some length of time dependant on the engine rpm. The frequency to voltage convertor would get the signal from the timing sensor and produce a voltage proportional to the engine speed which would be fed to the timer to control the spark delay. Frequency to voltage convertors take a certain amount of time to respond to changes in input frequency, however, so there would be some lag in the circuit's response. Using a separate engine speed sensor which produced more pulses per revolution could significantly reduce the lag.
So, the current spark would be initiated by the current timing pulse but the delay of the spark would be influenced by however many previous revolutions it takes the frequency to voltage convertor to respond.
My airplane (like most piston aircraft currently in existence) has two magnetos. One magneto has fixed timing. The other magneto has an impulse coupling in its drive system which causes it to lag behind the other magneto when the engine is turning slower than a certain speed but catch up once the engine is turning faster than that speed. So you set the timing of the magnetos for running speed and then the impulse coupling retards the spark for starting. The impulse coupling is more complex than the direct drive, which is why you only have it on one magneto (don't want to add any more failure possibilities than absolutely necessary).
The circuit I envision would work similarly. Whatever ignition sensor you use would be set to trigger at the advanced position and a timer circuit would delay the spark some length of time dependant on the engine rpm. The frequency to voltage convertor would get the signal from the timing sensor and produce a voltage proportional to the engine speed which would be fed to the timer to control the spark delay. Frequency to voltage convertors take a certain amount of time to respond to changes in input frequency, however, so there would be some lag in the circuit's response. Using a separate engine speed sensor which produced more pulses per revolution could significantly reduce the lag.
So, the current spark would be initiated by the current timing pulse but the delay of the spark would be influenced by however many previous revolutions it takes the frequency to voltage convertor to respond.
Re: cool
The "impulse coupling" seems to be doing essentially the same thing as an electronic differentiator, it has a "phase advance" as speed increases.
The same thing could probably be done by a sensor and shaped cam that gave a "ramp" instead of a good square pulse. Then at lower speeds, the rate of change of the ramp would be low and the trigger would simply be when the pulse reached a threshold.
At higher speeds, the "ramp rate of change" would be faster, and a differentiator (just the right size capacitor-resistor circuit, like a high pass filter) would cause the threshold to be reached earlier and earlier in rough proportion to RPM.
A micro would allow an advance based on the time taken for the last revolution, and would thus have a latency of about your two revolutions. I don't know if a frequency to voltage could be made to respond over the full RPM range and be fast enough for two revs.
It would not be critical to have the response fast, as it only has to keep up with the motor rpm increase.
The same thing could probably be done by a sensor and shaped cam that gave a "ramp" instead of a good square pulse. Then at lower speeds, the rate of change of the ramp would be low and the trigger would simply be when the pulse reached a threshold.
At higher speeds, the "ramp rate of change" would be faster, and a differentiator (just the right size capacitor-resistor circuit, like a high pass filter) would cause the threshold to be reached earlier and earlier in rough proportion to RPM.
A micro would allow an advance based on the time taken for the last revolution, and would thus have a latency of about your two revolutions. I don't know if a frequency to voltage could be made to respond over the full RPM range and be fast enough for two revs.
It would not be critical to have the response fast, as it only has to keep up with the motor rpm increase.
Re: cool
Now we are getting interestin!
When I was playing with motorcycles, the ignition units started out being pretty simple (CDI) and then began having advance and retard circuits contained therein. I thought that resistors or diodes were changed to change the advance. Resistors would make sense if it (the advance curve) were RC related.
There was an article in one of the last 3 or so issues of Strictly IC about making an engine using electonic control of the valves which were solenoid operated.
I thought some of the automotive applications picked up the timing signal off of gear teeth (flywheel) but shifted a position sensor to effectively make a retard for starting purposes. If sensors were cheap and small, this would be a low tech way to do it.
Just thinking out text.
Otto
When I was playing with motorcycles, the ignition units started out being pretty simple (CDI) and then began having advance and retard circuits contained therein. I thought that resistors or diodes were changed to change the advance. Resistors would make sense if it (the advance curve) were RC related.
There was an article in one of the last 3 or so issues of Strictly IC about making an engine using electonic control of the valves which were solenoid operated.
I thought some of the automotive applications picked up the timing signal off of gear teeth (flywheel) but shifted a position sensor to effectively make a retard for starting purposes. If sensors were cheap and small, this would be a low tech way to do it.
Just thinking out text.
Otto
Re: cool
The impulse coupling is not really a differentiator; it's more of a switch. It does change the phase, but it switches between two phase angles rather than smoothly progressing. Basically, it's only purpose is to make the engine easier to start. If your starter can spin the engine fast enough, you don't need the impulse coupling.
The centrifugal advance systems you find in the point ignition systems in older automobiles basically are mechanical differentiators and you probably could use an electronic differentiator for the purpose but the Hall effect sensors with which I'm familiar have very high gain op amps built into them with Schmidt (or some other type) feedback circuits so they give a square pulse regardless of the rate of change of magnetic field so some other type of sensor would probably be necessary with your shaped cam.
You're probably correct that the F to V could not maintain two revs latency over the entire range but most piston engines are only really useful over about one third of their range anyway. You might be able to tune it for that range and just live with more latency below that range (or you may not, I haven't tried).
Personally, I would just use a microcontroller because I'm more comfortable with programming than with analog circuit design.
The centrifugal advance systems you find in the point ignition systems in older automobiles basically are mechanical differentiators and you probably could use an electronic differentiator for the purpose but the Hall effect sensors with which I'm familiar have very high gain op amps built into them with Schmidt (or some other type) feedback circuits so they give a square pulse regardless of the rate of change of magnetic field so some other type of sensor would probably be necessary with your shaped cam.
You're probably correct that the F to V could not maintain two revs latency over the entire range but most piston engines are only really useful over about one third of their range anyway. You might be able to tune it for that range and just live with more latency below that range (or you may not, I haven't tried).
Personally, I would just use a microcontroller because I'm more comfortable with programming than with analog circuit design.
Re: cool
Now that makes sense. I think I captured a fleeting glimps of this concept in one of these posts.
I can see the peak RPM being the setpoint and a delay set through the sensor circuit. As long as there is no major swings in the differential between RPM fluctuations.
With the phase shift of reactance in AC circuits, either by capacitance or inductance, it would seem a usable feature for this application. Using this to ramp to a threshold as JT mentions with a built in delay.
Until I actually start playing with parts to get the why's from wherefores, this will remain just as using a flashlight in a black hole!
Some day all these little pieces may light up the whole solution. [img]/ubb/images/graemlins/blush.gif"%20alt="[/img]
DC
I can see the peak RPM being the setpoint and a delay set through the sensor circuit. As long as there is no major swings in the differential between RPM fluctuations.
With the phase shift of reactance in AC circuits, either by capacitance or inductance, it would seem a usable feature for this application. Using this to ramp to a threshold as JT mentions with a built in delay.
Until I actually start playing with parts to get the why's from wherefores, this will remain just as using a flashlight in a black hole!
Some day all these little pieces may light up the whole solution. [img]/ubb/images/graemlins/blush.gif"%20alt="[/img]
DC
Re: cool
Nah,
The self preservation instinct kicks in when one realisies it is time to let go of that direction of endeavor. Thankful the imaginary flashlight is cheap enough to begin a new journey into the unknown.
Could very well be to late, if nothing is learned by the first attempt. All be it starting over verifies that the previous failure may have been a waste of time and therefore leaves nothing left to modify or build upon?
Yet it can be found that there just might be another totally unrelated application for that failure to work elsewhere.
It's all part of the learning curve in making something out of nothing, adding 2+2, or grasping common sense with exposure to elementary fundamentals in physics. Where ever they my hide. [img]/ubb/images/graemlins/grin.gif"%20alt="[/img]
DC
The self preservation instinct kicks in when one realisies it is time to let go of that direction of endeavor. Thankful the imaginary flashlight is cheap enough to begin a new journey into the unknown.
Could very well be to late, if nothing is learned by the first attempt. All be it starting over verifies that the previous failure may have been a waste of time and therefore leaves nothing left to modify or build upon?
Yet it can be found that there just might be another totally unrelated application for that failure to work elsewhere.
It's all part of the learning curve in making something out of nothing, adding 2+2, or grasping common sense with exposure to elementary fundamentals in physics. Where ever they my hide. [img]/ubb/images/graemlins/grin.gif"%20alt="[/img]
DC