Nuts and Volts - March 2018

be adjusted and should be able to handle 100V or more.

Note the two diodes D2 and D3 which clamp the input to the first op-amp. This is necessary to avoid over-stressing the high impedance input in case the op-amp goes well outside its linear region.

The circuit does some simple capacitive integration and then bleed-down between C1 and R1, followed by a comparator that can detect when the total duration of the shocks over some interval have exceeded some level. We use op-amps for all of this. They make short work of problems like this, though it could be done with transistors as well.

Again, all of this is guesswork, so you may have to make some adjustments.

Sensor Inversion

QI have a sensor that goes LOW when triggered. I would like to have it go HIGH instead. What’s the simplest way to accomplish this? Randal Mahone Keyport, NJ

ASince we’re already on the topic of sensors, this makes a nice segue. You may have seen this technique used by me already. I find that the simplest way to invert a sensor signal — or any other digital signal that doesn’t need to be very fast — is to use a simple transistor. Bipolar transistors are naturally inverting devices, as are FETs and tubes.

First, we’ll start with the trivial inverter shown in Figure 2. There is both an NPN and PNP version. This simply limits the base current to the transistor and senses the collector current with a resistor, providing the output voltage.

One thing to be aware of is that the amount of current that can be delivered is asymmetric; the transistor can source or sink much more than the resistor generally can. You can reduce the value of the resistor on the collector to increase that current, but the price is more wasted power and generally slower turn-on times for the transistor.

We can speed this up slightly using the technique shown in Figure 3. Here, we simply add a capacitor to the base resistor to allow the sensor output to force current / charge into the base and to pull it out quickly.

This can enhance both the turn-on and turn-off times by rapidly changing the charge state of the base-emitter junction.

Even when the transistor is in saturation — as it is when the input is high and the output is low — pulling the charge out rapidly will reduce the time to come out of saturation. The value can be small (in this case, 100 pF) since it doesn’t need to move that much charge to have a large effect.

Note that this still won’t be ideal. It’s the equivalent of resistor-transistor logic or RTL, and so it asymmetrically drives the output. It’s active in one direction but passive in the other, so it relies on a resistor to return to one state.

Resistors don’t provide constant current like a transistor can, so the resistor side will generally look like an exponential decay when driving some capacitance, instead of a linear decay.

Exponentials asymptotically approach a value. Fortunately, most circuits detect the low condition at a voltage higher or lower (depending on the direction) than the final voltage.

Training a Train

QThere’s a model train at our library that runs around the perimeter of the room near the ceiling in the Children’s Books area. It’s rarely ever running as the management doesn’t want it to wear out. I’m wanting to build a timer or motion sensor of some kind that will trigger the train to run when kids are there. Can you help?

Kenneth Klein Altoona, KS

AThis project could easily be an entire construction article, and that would be a bit out of the scope for the Q&A, but I’ll try to give some suggestions.

While a timer might be okay for some situations, I think that a motion sensor combined with a timer might be the most comprehensive solution. To fully implement this, I’d suggest using some type of embedded processor.