photo sensor ignition

[tomcer34]

hi, I want to replace hall sensor from my ignition to for some photo sensor. Can u suggest me some fast one? More then 30kHz. Reason for replacing hall sensor it´s because it´s too slow for me.
Thank u for help!

[Benta]

A one-cylinder 2-stroke engine rotating at over 1.000.000 rpm? Or a 12-cylinder rotating at over 80.000 rpm?
 

[CaptDon]

We use hall effect sensors in turbochargers giving 8 pulses per revolution and the turbos run at a max of 28,000 rpm. Our sensors don't lag at all. Why do you feel your sensors aren't fast enough? Probably the crappy circuit surrounding your sensor maybe? We run dynamic compression on the incoming signal to keep it leveled in amplitude then hit a high gain limiter/clipping stage to form a square wave whose leading edge remains relatively accurate in phase relationship to the 8 teeth on the shaft. This whole mess is used with a pair of X/Y accelerometers to accomplish precision balancing of turbocharger compressors and turbines. It is real cool when 50 pounds of 28,000 rpm rotating mass comes apart!!! I've seen the pieces go through 1/4" steel plate! 8 pulses per revolution at 28,000 rpm is still only 3740 hertz. Why in the world do you need resolution at 30Khz??

[Terry Bites]

Rise times are in the uS or nS for hall sensors. How many Mhz do you need?

[floobydust]

Optical switches do not do well in vehicle/automotive applications- if the dirt doesn't contaminate them, the heat or moisture does.
You are very hard pressed to find slotted optical switches that are AEC-Q102 qualified. I think they aren't reliable- can't seal them due to moving parts, temperature cycles or the heat are very harsh and CTR degrades as they rapid age.
Chevy 90's LT1 Optispark was a flop for all the reasons I've listed.

I notice more automotive racing distributors are using optical pickups, which surprises me but there is much less sensitivity to EMI.
Most are primitive 8-slot like Mallory Unilite, using 5mm LED/photo-transistor.
The MSD (pic) looks totally overboard with what 1,024 slot wheel? Silly because the slop in timing chain and distributor gears makes that pointless (pun intended).

A crankshaft 60-tooth pickup is fine. Unless a one or two cylinder engine, because they have huge variations in crank velocity an ECU can make timing errors generating the next spark or fuel injection event. A V-twin is terrible, the crank slows down during a compression stroke, speeds up after ignition, but a misfire can happen and the 45° firing interval Harley riders love makes it rough to follow to estimate spark timing.

[tomcer34]

Hi, there is my schematic. It´s my own design so it´s very possible that there are few mistakes. For hall sensor I´m using 55100-3H-02-A. I couldn´t upload my schematic here, because it´s in Eagle and this page doesn´t support that format. So reason for optical sensor is that i can´t rev my engine more than 2000 RPM rignt now with this setup(with my old thyristor ignition i can go to 13500RPM). When i check with my osciloscope i can see 700Hz frequency going from hall sensor. I need high freqency for precission of igniton. 55100-3H-02-A should be capable of 10kHz but i cound managed it. btw those two imputs from lm339 are for 1 hall sensor and second for optical sensor.

[tomcer34]

it´s one cylinder two stroke engine. And max rpm is around 13500. High freqency of sensor is for precission and now max rpm   

[tomcer34]

What sensor are u usnig? I´m running 55100-3H-02-A for now. Also could u send me diagram for it? down there I send my.

[NiHaoMike]

Have you considered inductive sensors? Not much to go wrong with simple coils of wire and their low cost makes it trivial to have multiple for redundancy.

[floobydust]

I think your circuit has a few mistakes that would make it work poorly. 555 output pin 3 needs base-drive resistor for Q4, otherwise you're overloading it. I would use around 330R. Tie pin 4 to pin 8, tie unused LM339 inputs to gnd. 7805 output should have bulk cap at least 10-22uF. Usually a capacitor is across R11 for noise. R9 is 10k? R17 is low, 520R it could be 10k 555 triggers on falling edge so speed loss not important.
The Littelfuse Hall sensor 55100-3H-02-A rated 12kHz, open-collector output to ... pad 8 right.

Slotted optical switches not always fast Vishay TCPT1300 seems relatively slow 20-150usec rise/fall times.

[tomcer34]

this is my new revision. Few mistakes were corrected, now i can rew up to 5k rpm, but it´s still not enought. Do u have any suggestion where problem might be? thank all of u for ur time

[langwadt]

it´s one cylinder two stroke engine. And max rpm is around 13500. High freqency of sensor is for precission and now max rpm

one cylinder two stroke at 13500rpm is 225Hz ...

[tomcer34]

it´s one cylinder two stroke engine. And max rpm is around 13500. High freqency of sensor is for precission and now max rpm

one cylinder two stroke at 13500rpm is 225Hz ...

yes, but i need exact impuls, when piston is at the top, that´s why i need high frequency. With 225Hz it would need to be same rpm whole time and that´s not possible. I need at least 10kHz for good timing of my igniton

[tomcer34]

Have you considered inductive sensors? Not much to go wrong with simple coils of wire and their low cost makes it trivial to have multiple for redundancy.

yes, it was originaly there, but i had problems with reading and converting it to digital. There was around 60V and also negative voltage. so my tranzistors were not happy even with 100k rezisotor. Bike I´m working on is jawa babetta 210 typ 50

[tomcer34]

We use hall effect sensors in turbochargers giving 8 pulses per revolution and the turbos run at a max of 28,000 rpm. Our sensors don't lag at all. Why do you feel your sensors aren't fast enough? Probably the crappy circuit surrounding your sensor maybe? We run dynamic compression on the incoming signal to keep it leveled in amplitude then hit a high gain limiter/clipping stage to form a square wave whose leading edge remains relatively accurate in phase relationship to the 8 teeth on the shaft. This whole mess is used with a pair of X/Y accelerometers to accomplish precision balancing of turbocharger compressors and turbines. It is real cool when 50 pounds of 28,000 rpm rotating mass comes apart!!! I've seen the pieces go through 1/4" steel plate! 8 pulses per revolution at 28,000 rpm is still only 3740 hertz. Why in the world do you need resolution at 30Khz??

can u suggest or send some diagrams please. I want to improve my own designt based on someting working. On intertnet is not mutch about this topic

[Benta]

Have you considered inductive sensors? Not much to go wrong with simple coils of wire and their low cost makes it trivial to have multiple for redundancy.

yes, it was originaly there, but i had problems with reading and converting it to digital. There was around 60V and also negative voltage. so my tranzistors were not happy even with 100k rezisotor. Bike I´m working on is jawa babetta 210 typ 50

Instead of reinventing the wheel, I'd have kept that and done something with the signal conditioning instead. Just because a couple of transistors were unhappy there's no reason to go over the top. That can be solved.

[eugene]

I'm confused by the application. Here's what I am guessing:

The crankshaft position sensor outputs are connected to the inputs of the LM339 comparator (PAD8 and PAD13.)

The 555 generates a pulse of defined period that is used to trigger the ignition module.

The ignition module is connected to the open collector Q4 (PAD12.)

Is this correct? If so, how do you control the spark timing? Do you just fire at top dead center (TDC) under all conditions of throttle and RPM?

[tomcer34]

I'm confused by the application. Here's what I am guessing:

The crankshaft position sensor outputs are connected to the inputs of the LM339 comparator (PAD8 and PAD13.)

The 555 generates a pulse of defined period that is used to trigger the ignition module.

The ignition module is connected to the open collector Q4 (PAD12.)

Is this correct? If so, how do you control the spark timing? Do you just fire at top dead center (TDC) under all conditions of throttle and RPM?

basicaly yes, pad 8 is for hall effect sensor and pad 13 is for second one/optical sensor(only for experiment right now). Pad 12 is for igniton coil. right now sparks fire where is sensor possiotion on crank, but in the future it will be controled by arduino(it will shift spark timing by couple of degrees depended on the load and rpm, that part sort of work, but this doesn´t). sorry for confusing diagram/schematic but it´s made for pcb production.

[tomcer34]

Have you considered inductive sensors? Not much to go wrong with simple coils of wire and their low cost makes it trivial to have multiple for redundancy.

yes, it was originaly there, but i had problems with reading and converting it to digital. There was around 60V and also negative voltage. so my tranzistors were not happy even with 100k rezisotor. Bike I´m working on is jawa babetta 210 typ 50

I'd have kept that and done something with the signal conditioning instead.

what do u mean by that? what´s that can be achived? Do u have some example please?

[langwadt]

I'm confused by the application. Here's what I am guessing:

The crankshaft position sensor outputs are connected to the inputs of the LM339 comparator (PAD8 and PAD13.)

The 555 generates a pulse of defined period that is used to trigger the ignition module.

The ignition module is connected to the open collector Q4 (PAD12.)

Is this correct? If so, how do you control the spark timing? Do you just fire at top dead center (TDC) under all conditions of throttle and RPM?

basicaly yes, pad 8 is for hall effect sensor and pad 13 is for second one/optical sensor(only for experiment right now). Pad 12 is for igniton coil. right now sparks fire where is sensor possiotion on crank, but in the future it will be controled by arduino(it will shift spark timing by couple of degrees depended on the load and rpm, that part sort of work, but this doesn´t). sorry for confusing diagram/schematic but it´s made for pcb production.

just a bare ignition coil and that wimpy 100mA transistor?

[Benta]

what do u mean by that? what´s that can be achived? Do u have some example please?

I mean that: if your electronic circuit is so sensitive that it it will not work in an automotive application (motorbikes are part of this), you'll never get anything to work.
Spikes, noise, high voltages are standard.
The original magnetic sensor was probably the best solution, but was incompatible with your apparently very sensitive circuit.
80% of automotive (motorbike) design is dealing with an electrically very harsh environment.

[viperidae]

Does the signal disappear when you reach the 5k rpm limit you currently have, or is the ignition timing too retarded the rev higher than that?
The faster the engine spins, the more you need to advance timing, yes? Especially if your trigger signal includes the fixed dwell time.

Modern ECUs keep their own internal crank position and sync it to the crank position input. That way the ignition timing can be much more precise than the crank sensor. You'll be able to apply the correct dwell for the coils so they have enough energy to fire.

[floobydust]

It would help if we knew what engine this is going on, a moped or what and what the original ignition system was. No idea what OP is using for the magnet either. How about some pictures. Otherwise this guessing is just silly.
Simple ignition systems have irregular cam lobe and staggered pickup coils to achieve timing advance and rev limiting.

[tomcer34]

I'm confused by the application. Here's what I am guessing:

The crankshaft position sensor outputs are connected to the inputs of the LM339 comparator (PAD8 and PAD13.)

The 555 generates a pulse of defined period that is used to trigger the ignition module.

The ignition module is connected to the open collector Q4 (PAD12.)

Is this correct? If so, how do you control the spark timing? Do you just fire at top dead center (TDC) under all conditions of throttle and RPM?

basicaly yes, pad 8 is for hall effect sensor and pad 13 is for second one/optical sensor(only for experiment right now). Pad 12 is for igniton coil. right now sparks fire where is sensor possiotion on crank, but in the future it will be controled by arduino(it will shift spark timing by couple of degrees depended on the load and rpm, that part sort of work, but this doesn´t). sorry for confusing diagram/schematic but it´s made for pcb production.

just a bare ignition coil and that wimpy 100mA transistor?

Sorry, my mistake, i dont use BC547, it´s there because i dont have library for tranzistor im using, it was just same footprint. Tranzistor that is used is E19009-2

[tomcer34]

what do u mean by that? what´s that can be achived? Do u have some example please?

I mean that: if your electronic circuit is so sensitive that it it will not work in an automotive application (motorbikes are part of this), you'll never get anything to work.
Spikes, noise, high voltages are standard.
The original magnetic sensor was probably the best solution, but was incompatible with your apparently very sensitive circuit.
80% of automotive (motorbike) design is dealing with an electrically very harsh environment.

it´s not that sensitive. casing is shielded. Osciloscope didn´t show any high voltage spikes or anything, signal is clean. Problem is that pull rate of hall senosor is too low(on scope 1.5kHz right now, but datasheet says it should be 12kHz).