How the Tach Amp works
 

First of all here are two images of the schematic I drew:

Small: Small
Large: BIG!

What the Tach Amp does is to take in the signal from the sensor that detects the pin on the crank pulley and produces a 3 mS pulse to the tachometer every time the ping goes past the sensor.

How it does it:

There are three parts to the circuit.
1. Voltage Conditioning
2. Sensor interface.
3. Amplifier

Voltage Conditioning

The 14V nominal voltage is first run through a 32 Ohm resistor to limit the maximum current the amplifier will draw. That voltage is filtered by a 100 microfarad capacitor and fed to a 360 Ohm resistor and then to the cathode of a 5.6 V zener diode. The 360 Ohm resistor and zener diode maintain a set voltage at the point where the two come together. This regulated voltage is fed to the LM1815.

Sensor Interface

The sensor is called a Variable Reluctance Sensor, which is nothing more than a small electromagnet. When a chunk of steel (the pin on the crank pulley for instance) passes through this magnetic field, the voltage across the sensor changes slightly. The 169 Ohm resistor attached to the Emitter of the transistor sets the maximum current that can flow through the transistor and then to the sensor. The Transistor itself and the 830 Ohm resistor from its base lead to ground set up a constant current source to reduce noise from the sensor (if the current flowing through the sensor isn't constant, you would end up with lots of noise and voltage fluctuations into the Amplifier). The 2.2 microfarad capacitor blocks the DC voltage from the input to the amplifier, but allows the pulses from the sensor to get through.

To condition the signal before it enters the amplifier, it is low-pass filtered by the circuit made up of the two 18K resistors and the 0.0033 microfarad capacitor. This filter cuts out any high frequency noise that may cause the tachometer to produce bogus readings.

Amplifier.

The LM1815 Amplifier takes the signal, amplifies it and compares it to a threshold it sets based on the average level of the incoming signal. The threshold is maintained by the 4.7 microfarad capacitor attached to pin 7. The 1.2 M resistor slowly bleeds off the charge from the capacitor, without it the threshold would just keep climbing - if it were too small a value the threshold would never get off the ground and the signal to the tachometer would be too noisy. The 4.7K resistor between pins 5 and 8 sets the initial threshold. When the voltage rises above the threshold, the LM1815 arms itself so that when the signal falls back below the threshold it can trigger an output pulse.

To make sure the signal to the tachometer is of a set pulse width, the LM1815 incorporates a timer that is set to produce a 3 millisecond pulse via the time constant set up by the Resistor/Capacitor pair made up of the 470K resistor and 0.01 microfarad capacitor attached to pin 14.

Since the output from 12 can only pull to ground, the 27K resistor from pin 12 to 8 pulls the output pin high unless the LM1815 is actively pulling it low.

Pretty good for something that fits into less than 1.5 inches square!

In other words, its an analog computer.

Thta is some really nice work!!! We needed this badly!

So, it might be possible to either breadboard a replacement with OTC components or, with some effort, layout and manufacture a PCB that would fit into the existing amp housing and offer it as a DIY kit.

I think I'll probably breadboard it into a small plastic box, connect it with some stock-looking wiring to the socket and hide it all in the engine compartment somewhere.

Thanks again!

Huh???

I can turn on a light switch, but I didn't catch any of that! So, when y'all get the "fix"....I want one! That dang tach is the only thing that doesn't work (after I get the fuel lines back in).

Have you tried the old cigarette butt trick?

Okay, so I put a tach in my cluster, got a tach amp (already had the connector it goes into) ran a wire from the the #3 pin to the back of the tach, gave the tach +12v, and it doesn't work. (#2 and #6 pins already had gnd and +12v, and 7, 8, and 9 already go to the sensor)

Is there a way to test the tach amp output at #3? Because if its putting out the correct...whatever...then I've got a bad tach.

Thanks,
-Geoff

nice job
 

looks impressive!

My greatest deficiency (sp?) is electronics. I don't know the first step of how to use an ohm (sp?) meter.....

Very nice, Pokey. I'm going to take a crack at this using a simple 8-bit microcontroller. Since it already has comparators, voltage references, timers, and A/D's built in, it should use about half the parts.

Regarding the output pulse - is it logic high (each pulse is 5V for 3ms) or logic low (each pulse is 0V for 3ms)? Any chance you can post scope traces of the waveforms?

I'm working on it! Since my scope is an ancient analog model, I'll have to use my camera to take pictures.

Are you certain the sensor is working? There have been cases where the cable that goes from the sensor near the crank pulley to the tach module. Its a shielded cable. And a pulse generator can be connected to the tach input to see if that is working. Or you can shotgun the parts off a working car (maybe easier).
I put a tach in my '79 300TD with good success. I had to find a source of 12V for the module, that was the only thing that threw me at the beginning.

OldPokey, your schematic is good! I recall from school that is a Schmidt trigger. Now if someone makes a surface mount board to replace the short lived VDO module I'll bet they can sell a lot of them, starting here at the Shopforum!

I've got at least 10 tach amps in my garage. It doesn't matter if understand how it works... if it does not I pull it out and put in another one. ;)

Some measurements:

Signal out of the amplifier is normally at 0V, when the pin on the crank pulley is detected, output goes to 6V for 3 mS, then back to 0.

The pulse from the sensor is approximately 400 mV and looks like a triangle.

I will take my lashup into work and get some proper captures with the nice stuff we have there...

John,

Thanks, very helpful. Signal out should be clamped to the zener voltage (5.6V).

400mV is not much signal to work with. No wonder they were so cautious about noise. Writing some blanking intervals and filtering into the software should take care of triggering on noise, but I'll need to work carefully around the low signal level.....

Thanks again.....

OP-AMP of some kind to get the 400 mV up to something reasonable, then employ a better noise-filtering algorithm in the software of whatever you're doing?

Might be easier than trying to get something useful out of the 400 mV by itself.

Here is a capture from an oscilloscope. The yellow line is the input, and the green is the output.

Notice that the volts per division for the input is 100 mV (or 0.1V) and for the output is 2V. The input pulse is only about 200 mV, and it rides on the voltage differential across the tach sensor. Also note that the output pulse does not begin until the input pulse has dropped to below the threshold set internally to the LM1815.