Need to detect flowing current, I've got an idea, but need some help.

[ZacMan]

G'day all.

I've got an idea for a wee project involving detecting how much current is flowing through an automotive ignition coil. When setting these up, you can vary the amount of time the system allows current to flow through the coil before it is switched off and the coil provides a spark at the sparkplug, this time is call dwell. During dwell, the current flowing in the coil builds up in a fairly linear fashion.

Currently for testing this I put a 10W 0.1R shunt resistor in-line with the power wire to the ignition coil and measure the voltage drop across it with an oscilloscope. I stumbled upon the ACS712 current sensing IC's recently, and reckon they'd be a nicer way to go, as they don't create a loss in the circuit.

I'd like to create a simple tool which goes in-line with the power wires feeding the ignition coils, and provides a visual indication that a suitable current level is being reached, and this the dwell time is long enough. My idea is to send the ignition coil power through an ACS712 Hall effect current sensing IC, and feed the output voltage from this chip into an op-amp comparator, which has a setpoint defined by a voltage divider. When this op-amp detects a higher voltage output from the ACS712 than the setpoint, it can turn on an LED.

How would I get the LED to stay on though? As these instances of the ACS712 output voltage being higher than the setpoint voltage will be very very brief. I had thought of using a mircoprocessor for the job, but its going to be a very electrically noisy area with lots of amps of current flying around, being sent into magnetic devices...

I'd be very very interested in hearing any experiences had with the ACS712 current sensing IC's, or just current level detection in general.

Cheers all.

 

[chuckey]

As the output of your voltage compararor would be a series of pulses every time the current was exceeded and nothing when the current is too low, The LED would either flicker or be off. So a simple RC network could turn the flicker into a decaying DC voltage to stop the LED flickering. The time constant must be set up at the lowest engine RPM. A more precise way would be to use a retriggerable monostable so if the pulses occur at a rate higher then its period, the output stays on, and if lower the LED would stay off.
Frank

 

[jiripolivka]

Your idea of inserting the 0.1 Ohm resistor to the feed line is correct. In the case of the input current into the ignition coil primary the problem is that the current is pulsed and that coil inductance generates high overvoltage.
The best solution is to observe the voltage across the 0.1 Ohm resistor on an oscilloscope, and from the patterns you can calculate what you need.
As the scope has its input referred to ground or coil chassis, I would connect the resistor to the ground line, so the DC source (battery) and the scope share the common ground point. Also, make sure that the "live" line to the scope input is well screened; otherwise the transients from the ignition coil will disturb the observed correct signal.

 

[ZacMan]

Cheers for the replies.

Chuckey - I've used the 555 retriggerable circuit to detect the pulses and maintain a lit LED, and it does work quite well. However, having to have one 555 circuit for each ignition coil (as I'd like the tool to be able to test six ignition channels simultaneously) eats up a ton of board space, once you account for a current sensing IC, Op-Amp, and 555 IC per channel, plus all the passives, things are getting a little out of control. I'd like to simplify things if anyone can think of a way to do so! I'm a but stumped, heh.

jiripolivka - I've been test ignition circuits like that for years, I'm trying to sort out a better way of doing it that will be much quicker and easier for another person to understand what is going on, cheers . Also, those 0.1 Ohm resistors get pretty hot when testing a high RPM's. Using the ACS712 current tranducer chips will remove the resistor loss and hopefully keep things quite a bit cooler .

Cheers.

- - - Updated - - -

Chuckey - I also tried using an RC circuit to decay the LED, and tried tuning it to get a fairly even appearance. I couldn't get satisfactory results, as I found found the output pulse from the op-amp comparator was too short to sufficiently charge the capacitor. Raising the dwell time, and thus the width of that pulse allowed the capacitor to charge, but meant you had to increase the dwell time over what the actual target is :-/.

 

[BradtheRad]

You need a light that has low resistance and creates very small voltage drop. Such as an incandescent flashlight bulb that operates on very low volts, but high amperes.

Or several flashlight bulbs in parallel. (Example, PR2 is a common type, 3V at .5 A.)

I was going to suggest a taillight or dome bulb that normally takes 1A (such as a #1157). However I suppose that will be too much resistance, probably between 4 and 12 ohms.

There is also a headlight, which passes about 3A.
Or 2 or 3 headlights in parallel, which will reduce resistance further.

 

[chuckey]

If you are testing a V6, unless the engine has got 6 individual ignition circuits with DIFFERING coils/drivers/timing references, surely just looking at one would be enough?. or if you are actually setting them up, a plug board so you can quickly move your measuring kit from cylinder to cylinder - current too high for a rotary switch?, could use relays though - nice and cheap, fire them up from a rotary switch..
Frank