Generic tacho input circuit to micro...

[Eclipze]

I've already searched this well covered topic, but couldn't find an answer :? .

I want to pulse measure the tacho for cars & motorbikes, however are unsure about a suitable generic interface.

What different types of signals can I expect?
12V, 5V, low side of coil, high side of coil, max voltage, negative voltage?

I'm thinking a resistor divider, 10k/10k, with a 1nF filter & zener clamp, followed by 100k of series isolation into the micro.

Alternatively, perhaps a input series 400V diode (cathode on output), 2.2k series isolation, 1k pullup to VCC, 1nF filter, reverse voltage diode clamp to ground, and 100k series isolation into micro. This may provide the most protection and accommodate any positive voltage switching input to ground.

Apart from that, I was thinking of a transistor buffer for further isolation.

Any advice most appreciated. I don't know a lot about the different tacho pickups to expect. (speed pulse interface will be the next topic :wink: )

P.S. This looks a neat forum...

 

[Optikon]

I've already searched this well covered topic, but couldn't find an answer :? .

I want to pulse measure the tacho for cars & motorbikes, however are unsure about a suitable generic interface.

What different types of signals can I expect?
12V, 5V, low side of coil, high side of coil, max voltage, negative voltage?

I'm thinking a resistor divider, 10k/10k, with a 1nF filter & zener clamp, followed by 100k of series isolation into the micro.

Alternatively, perhaps a input series 400V diode (cathode on output), 2.2k series isolation, 1k pullup to VCC, 1nF filter, reverse voltage diode clamp to ground, and 100k series isolation into micro. This may provide the most protection and accommodate any positive voltage switching input to ground.

Apart from that, I was thinking of a transistor buffer for further isolation.

Any advice most appreciated. I don't know a lot about the different tacho pickups to expect. (speed pulse interface will be the next topic :wink: )

P.S. This looks a neat forum...

Consider using an opto coupler for real isolation. I would strongly recommend it if you are interfacing to a micro. They are very cheap, can run fast (for your application) and provide many thousands of volts of true electrical isolation.

 

[Oznog]

Easiest thing is to use a 5.1V zener on the pin (if you've for a 5V Vdd), and use a large value resistor divider. As an example, if the coil signal has a peak of +/-20V, you might use values of 100K in parallel with the zener and 200K between the coil and the pin. The other side of the coil will need to be connected to ground.

Now if the voltage on a pin exceeds Vdd or is below GND, protection diodes will conduct current to either Vdd or GND. They are very limited in how much current they will handle, and will blow if the voltage is above/below by over 0.3V. Also, current conducted to the positive rail will raise Vdd and your regulator probably won't stop it,

The resistors limit pin current to 1mA, the zener clamps the input voltage at +5.1v to -0.7v. The positive voltage probably isn't enough to engage the protection diodes at all, the negative will definitely conduct but it's limited to 1mA so it's not going to hit Gnd - 0.3V. Current going to ground won't muck with the regulator at all.

An optoisolator will work too. This should do the job reliably with these values, though.

 

[Eclipze]

I'm running a 3.3V micro. Thinking about it further, I really want to avoid any current injection into VCC, and also 3V3 zener can be a pain to work with because of the sloppy clamping leaky region being wider. I'm also concerned about input offset to ground, as if significant may not get below the micro's threshold, and if the divider is set too high to compensate, I may not pick up 12V type signals. 3.3V can be a pain to work with!

Optocoupler is certainly an option.

Starting to think of using a PNP transistor buffer. Series diode on input (cathode facing signal at connector), resistor divider to VCC (100k/100k)with PNP base-emitter on the top end of the divider. Collector to ground via a resistor, collector to micro. This system would work for any input that oscillated below ~2V, and upto the rating of the series diode (400V). The micro would also get a full 3.3V to 0.0V swing, guaranteeing input capture. Only concern may be the -ve spikes. With a transistors that had -80V rating on the base, after the resistor divider it would be -160V. I could put a diode clamp to ground at the input resistor divider, as well as a 1nF to get rid of some nasty spikes, and a cap across Vbe to filter the signal.

Hmm.... this option is starting to sound nice Can anyone see any probs with it?


If I want to