Transistor using HV to switch 5v

Hello, my project of doing a Arduino ignition requires a signal conditioner on the pickup coil from the engine. This coil generates a 30v to 60v ac signal once per revolution. My plan is to put a diode on the (+) line and install it to the base of a transistor, and run 5v to the emmiter, and run the collector to the imput if the Arduino.

My transistor d1071, because it's on hand, is a high power amplifier and looks to be in spec except vebo of 6v. I have been reading up and this seems to be ok.

I am unsure of when you need to add resistors to the circuit. Please advise I can figure it out with a little help. Thanks

The main consideration you need to keep in mind is Voltage. It does not need to be a power transistor if the load is only a port pin on the arduino.

The easiest way to interface in this case is with an optocoupler. A AC optocoupler exists. You just need to create a current limiting resistor value for the voltage seen on the diode. STandard open collector input would be the way you interface to the input pin... (with internal pullup enabled)

So again... your collector current will be very small so there is no need for a high power transistor.

For simplicity sake... You could use: https://www.mouser.com/ProductDetail/Broadcom-Avago/HCPL-3700/ for interfacing.

The Input side would have something like a 2W 2.7K Resistor inline on each AC leg.

That's the only one I have on hand that meets voltage spec. If I wanted to do it like that any npn transistor that meets voltage spec would be fine?

In pinmode, enable the pull up resistor and nothing else is needed on the input side from the collector to Arduino?

The (-) ac leg would just be hooked to ground?

I'll try with a transistor on the scope first and see how it behaves, then might decide on a optocoupler, thanks.

Hi,

To protect the transistor BE junction I would put the diode across the Base to Emitter with the anode at the emitter. This will provide an alternate route for the reverse current.

Is the sensor connected to some other circuitry that needs to function?

You will likely benefit from some some resistance in the base to sensor, else the transistor base will take the full shorted (almost shorted) sensor current. The sensors of this type I am familiar with do not generate very high currents but others may be different.

For a starting point I would take the lowest voltage (30V) divide by 2 and calculate a resistor that will allow 5 ma into the base. These assumptions calculates out to ~ 3k

right, I was composing a reply similar to this but noticed you covered a lot of what I wrote.

So, yeah, this is an instance where you want to consider

  • You are connecting AC to an DC device
  • You want to Limit Voltage to the base pin

Consider that you want to rectify the AC signal to DC. The transistor is not suited to perform that step.

I'll try it out as suggested, I think I have enough components on hand to do that.

Where would I run the (-) leg of the AC?

Thanks this helps a lot

So i didn't have enough components to try it out, so a order to mouser for 2 optocouplers and 4 3k 2 watt resistors, along with some other random stuff i thought i might need. I drew up a circuit that i would like to bread board and scope the output of the Vo on the optocoupler, so i can move on with the project.

Question:

Please review the Circuit to make sure i have everything right.

Notes:

This is on a golf cart, power is cut every time the pedal is released, problem?

Charging system can exceed 18-20 volts if out of adjustment, so a dc-dc buck converter is used for regulated 12v power.

Gm HEI module looks for voltage to drop to zero on W/G and then breaks the C or the negative coil connection to fire the spark plug. B is powered with the pedal switch.

Success! Had to change the ac leg resistors to 1k 1/4 watt as cranking ac voltage was around 12v and the 3k was limiting the low window to nothing.

Now I have two low pulses instead of one high so I will change my code for that, and I think I can use the counting between pulses to figure rpm instead of counting a whole revolution, which should work better. Thanks!