Go Down

Topic: High Voltage AC Sensor To Arduino (Read 1 time) previous topic - next topic

wvmarle

Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Mk1_Oz

If you put the tachogenerator signal through a series resistor through a pair of back-to-back diodes you'll
get a nice 1.2V peak to peak signal whose voltage is relatively insensitive to the speed.  Try 10k or so.
You blokes amaze me!  Nice.

This solution will give a ramp from zero volts up to the clip level.  How will an Arduino react to this on a digital pin?  Will it read the value as low until it reaches some value over 2.5v at which time it will set to high?  Not sure if the Arduino (Uno or Mega) needs a definite 0 or 5v on a digital pin or whether a slightly varying voltage, as provided by your soltion, will still work?  If your method is suitable then I guess I just vary the resistor size to clip the voltage at between 2.5 and 5v?

If a square wave output is better for the Arduino then I could incorporate both ideas.  I could use your clipping circuit to limit my AC voltage (currently up to say +/-50v) to something that a Schmitt Trigger could handle (I was led to believe above that a ST would not be able to handle 50v).  That output could then be processed by a ST to convert it to a square wave at trigger points set by me by varying the resistor values.  Does that sound reasonable?

A couple fo specific questions:

1. What type of diode for the clipping?  Do diodes come in different sizes as do resistors and caps? (sorry...I am very new...)

2. To make a ST do I just buy a ST chip and add the resistors to make my own complete unit?  There has been mention of starting with an op-amp but I have only found bare bones chips not nice complete boards.  Any links to parts gratefully accepted lol

wvmarle

This solution will give a ramp from zero volts up to the clip level.  How will an Arduino react to this on a digital pin?
At the 0.7V level it will not react. It needs 2.5-3V to detect high (see data sheet for details on this, if you're interested). This signal is best passed into a Schmidt trigger to make it a 0/+5V signal.

The resistor value is purely to limit current, as a diode has a fixed voltage drop that varies little with current. So the actual resistor value doesn't change how much voltage you get out, it could in principle be any value but low values mean higher current. You want to keep this current low to not affect the sensor, and to prevent heat from building up and so. That's why 10k, but a higher value or a bit lower will do as well. Whatever you have on hand.

Quote
1. What type of diode for the clipping?  Do diodes come in different sizes as do resistors and caps? (sorry...I am very new...)
Different sizes - for power/breakdown (reverse) voltage rating.
Different values - not so much (well, there are zener and Schottky diodes - look them up - which have different applications).
Voltage drop over diodes is pretty much the same for all of them, 0.6-0.7V.
A standard 1N4002 will do, or a higher value (1N4003..7) The 2 goes to 100V, the 7 to 1000V reverse breakdown. At 10k resistor value you get just a few mA so no need for high power rating.

Quote
2. To make a ST do I just buy a ST chip and add the resistors to make my own complete unit?  There has been mention of starting with an op-amp but I have only found bare bones chips not nice complete boards.  Any links to parts gratefully accepted lol
OpAmps usually come in DIP or other solderable package. DIP is very nice for experimenting - do get a solderless breadboard and a set of jumper wires. Makes experimenting a lot easier. Then get some protoboards to build finished projects, so you can properly build it into something.

A Schmidt trigger is normally based on an OpAmp indeed. Very easy to build, just a few external components.
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Mk1_Oz

Cool.  I think I know enought o get started and do some damage!!!

Is there any decent (and free) electronics simulation software available that could assist me before I blow up components for real?  I saw one called 123D Circuits.io but know nothing about it.

Mk1_Oz

Been playing with circuits all arvo.  Is it possible to have a non-inverting asymmetric schmitt trigger using an op-amp?  I cannot find a reference to this, only to an inverting asymmetric.

Mk1_Oz

Firstly, deepest apologies for posting three items in a row.  Don't mean to spam my own thread but I keep having questions as I have been sitting here for around 6 hours straight nutting this out!

I have designed the attached circuit on some software called CircuitLab (a random find).  It is based on the help provided in this thread so far.  Rather than clip the negative part of the sine wave, I have clipped off the positive.  This is solely because I need teh square wave to be positive (high) and this is the only way I could get it to work.  I didn't have much luck with non-inverting schmitt triggers.  It works........almost!

I have used a clipping circuit that outputs to an inverting schmitt trigger.  To be honest my resistor values were complete guesses.  I have outputed what the software shows as a simulation based on a 1Hz sine wave that has 8v peak to peak (this simulates the approximate idle speed of the engine on which I will be measuring the RPM).  I have run it with a 100v peak to peak and it output the same (that's good).  Not sure if simulation software highlights if components are overstressed or not.  The clipping diode shown has an output of 3.9v whcih still clips at the low idle speed.

The issues;
  • the square wave output is just over 5v (5.089v).  This may be an issue for the Arduino but I do not know how to reduce it.  My only thought is to add another zener diode on the output to clip the voltage to 5v or just under.
  • the square wave goes high when the sine wave is a fraction under 0v (around -0.002v) which is fine.  However, it then goes low at just over 0v (+0.002v) which is not fine.  The actual sine wave I will receive is not continual i.e. it will have a positive peak, negative peak then 0v for a period of time before another set of peaks.  Having the square wave go low at over 0v means it will stay positive until the next set of peaks arrives.  I guess this isn't technically an issue as I will be telling the Arduino to read the rising edge but it just feels messy.  I would rather the square wave goes low at the same voltage point as it went high i.e. just under 0v.  Is that possible?


I have tried many times to calculate the trigger points but I never seem to get it to match the simulation (I am of course trusting the simulation is working correctly!!!!!!!)

MorganS

1. No problem at all.

2. You need some hysteresis. Small variations around zero should not trigger a change. The Schmitt trigger should do that. If it isn't then there is something wrong.
"The problem is in the code you didn't post."

wvmarle

1. Fine for Arduino, it's rated for up to 5.5V. But your output voltage can never be more than the Vcc provided to the OpAmp. If you see anything above 5V (I assume you set Vcc to 5V) then that's a rounding error in CircuitLab's math.

2. Some hysteresis is good indeed. You may even want a bit more than that - your 0V is never going to be exactly 0V (theres LOTS of RF noise in cars, especially older ones). As long as your signal is symmetrical, hysteresis will simply delay your square wave a little compared to the input sine wave. The wavelength remains the same.

Do look a bit into hysteresis, you can easily find some graphs that show you how it works, and how it won't change the duration of peaks for symmetrical signals.
Quality of answers is related to the quality of questions. Good questions will get good answers. Useless answers are a sign of a poor question.

Mk1_Oz

Changed a few things....

I have now clipped both the +ve and the -ve of the sine wave and added a bias voltage to both.

The switching points are still 0v (set high) and just over 0v (set low) and I just cannot get this to change for some reason.  I have read alot about hysterisis and think I understand the theory....putting it into proactice is harder for me :(  Feel free to change my circuit if you can work out how to make that change!!!

I bought parts this arvo so will see if I can produce the magic smoke tomorrow :(:(

Rx7man

I know this is an old topic, but there's a chip especially designed for your purpose, look up the Maxim MAX9924, it can operate on signals from millivolts up to 100's of volts and adapts automatically and has zero-crossing detection so you have accurate timing points

Did you ever get your thing going?

Mk1_Oz

Thanks for the reply.

I played around for quite some time but ultimately dropped this part of the project.  I wanted to try to use this particular sensor as it was what I had but since using hall effect sensors for other things I will just use one of them if ever I get back to the project.

Go Up