I'm building an instrument which includes counting engine revolutions for a period of time (15-30 seconds) at the same time I am counting wheel revolutions.

I have an inductive clamp to go over the spark plug wire, and found a circuit which uses some diodes, resistors, and capacitors to get the induced voltage down where it needs to be to trigger an SCR. The SCR feeds a 555 in what I believe is monostable configuration, whose output nominally is a 5V square wave. At least, that is what the circuit diagram shows.

I breadboaded this, hooked it up to a frequency counter (in count mode), and to an old oscilloscope (has tubes inside it) I've pretty much forgotten how to use, and clamped onto a 6 cylinder engine's spark coil wire. This will give three sparks into the distributor per engine revolution. I manually timed the counts on the counter for ~30 seconds, and they seemed about right for what the tachometer showed, so the circuit is working.

However, the trace of the 555's output on the oscilloscope showed a sawtooth wave, and not a square wave. Since the counter counted, I may not need a square wave, but I'd like to know why it wasn't square.

I've attached the circuit, or think I have. The 555 is running on 5V, and what I take to be the timing components are an 18K resistor from Vcc to Discharge (pin 7), and a 0.1microF (100nF) cap from the Input+ (pin 6) to ground. Pins 6 and 7 are also joined.

For what I am doing, the engine won't run over 2,000 rpm, though my expectation is to have it idle at 800-1000 rpm. 1,000 rpm = 16.7 Hz, times 3 = 50 Hz. So the period should be 20ms At 2,000 rpm that would be 10ms. If I have this right, maybe I want a 1ms square wave? Plenty long for a counter to pick up on?

I tried to do the R/C calculations for this, but couldn't deal with fractional exponents or whatnot. I'm a cook book futzer, not an EE or anything like that.

What combination of resistor and capacitor do I want here?

 

Please number the pins on the 555; some of the labels are ambiguous.

 

I took the pins in the diagram to be:
5V = 8
DIS=7
IN+=6
IN-=2
Ground = 1
OUT=3

Missing are 5 (which I think should have a 0.01microF cap to ground?), and 4 (which should or could be connected to Vcc?). I didn't connect either of those.

I think I found the formula I wanted: 1.1RC, where R is in Kohms, C is in microF, and the result is in milliseconds. For the values listed, that would be an output pulse of 1.98milliseconds, which maybe is just fine for my purposes.

If I've got this right, that leaves the original question - why am I getting a sawtooth wave instead of a square wave? The rise on the sawtooth is much steeper than the fall, for what that might be worth.

 

Pin 3 (output) from the 555 should have a rectangular pulse. The sawtooth would be on the rc network.

 

Pin 4 MUST be connected to the supply voltage (+5V). It's the reset pin and will stop the chip if it isn't connected high.

The timer chip needs a 10uf aluminum cap and a .1 uf in parallel with each other and connected from pin 8 to pin 1, physically close to the chip so they can supply quick current to the output pin. The .01 uf on pin 5 to ground also combats glitches by holding that voltage still when sudden switching happens. Right now, you are timed at 1.98 milliseconds. You can drop the 18K to 9k to get to 1 millisecond, but capacitors are never exactly what the label says, so a 6.8K and a 5K potentiometer in series would give you some accuracy.

If this doesn't clear it up, you could try a pull down resistor on the output pin, somewhere between 1k and 10k should do it.

There is also a very intentional sawtooth wave on the timing capacitor. Be sure you haven't connected there.

I am not a real engineer, I only play one at work. When I speak, electrons flow because I've been doing this since I was working with vacuum tubes.

 

I frankly don't understand much about your circuit except the 555, and it has problems. Attached is a 555 monostable which will generate a 1mS pulse each time it is triggered by a low-going signal on pin 2.

 

Thanks to all. I'll try adding these connections and see how it does. It may take a while, as I have to take my old (large) oscope and the other stuff over to a friend's house to see if I've got what I want. My cars are all coil on plug, though I have a lead on a way to pick up them firing (drops resolution to 1/6th of what I can get from a spark coil, but still should be adequate), which is a next step.

I wondered if I was just seeing the input, and I may be, but not because I wasn't connecting the scope to pin 3 (output). I checked. Perhaps the things I didn't do caused the 555 pretty much just to pass the input through?

 

Any original electrolytic caps in your old scope may be suffering from loss of capacitance which could have reduced the response bandwidth. That could mis-shape the display of a square wave into something more like a triangular wave.

 

Any original electrolytic caps in your old scope may be suffering from loss of capacitance which could have reduced the response bandwidth. That could mis-shape the display of a square wave into something more like a triangular wave.

Also check if your 'scope has a DC measuring capability and if it does, is it switched to that? Can you give a sketch of what you see on the 'scope? If there are negative going pulses at the end of the expected pulse then you could have it AC coupled.

 

I'll check to see where the coupling switches are set. It was a good scope in its day - two channel and lots of adjustments. In the daylight the grid, even turned all the way up, was hard to see. Its main problem was that after a while the trace sank lower and lower on the display, and eventually went out of sight despite my twiddling with the position knob. But up till then the trace rose steeply (though far from square), then fell more or less uniformly until it rose again.

I have a little box I can use to feed 10, 100, and 1000 Hz square waves in at 5V, and I should be able to figure out a quick and dirty way to make those saw teeth, or just rustle up an op amp or something like that to create a sawtooth generator to check the 555 side of things.

 

No help to op, but;

I've been watching this thread, as I was puzzled by how scr bt149 would turn off.
Yet op verified trigger function.
Looking at data sheet, holding current is 5ma. Much more than current thru 10K.
So scr fires, discharging 47nf, starving scr to turn-off?

Does that seem right?

http://www.nxp.com/documents/data_sheet/BT149_SER.pdf

 

Yes. Starving an SCR by (pretty much) only giving it a capacitor to discharge is a well known method.