Designing Encoder/Tach Circuit for Digital Display and CPU Interface 3

[DanBo1221]

Mechanical guy tackling electronics here. I'm looking to design and fab a consumer grade personal-use wind turbine with portability and cost being main design constraints. My main goal is to have a charge controller box that displays voltage from the DC motor, as well as the motor shaft RPM, on the top of the controller box. I've also toyed around with the idea of having USB interface with the controller to record data onto a laptop for efficiency measurements for various blade designs. What's the most effective means of monitoring motor shaft speed and displaying it on a box some far distance away from the motor in the wind turbine nacelle, and having the option open to record that data onto a PC? Colleagues have suggested using resistors to calibrate the voltage to correspond with RPM, but I'm not sure that's the most effective/accurate means. Would an encoder be better for this task? Outside my territory here, so any and all advice is appreciated!

Deisng criteria:
Perm DC Brushless motor, .38" shaft diameter, 5" diametrical area behind the motor for monitoring equipment/cabling
Calculated RPM speeds 0-300 RPM (furling device will be used as braking method)

 

[LiteYear]

I wonder if you're over-complicating the design? The Arduino route certainly gives you a lot of flexibility, but as you're discovering, you need to do a fair bit of work yourself to decode the input and drive a display. Nothing unusual for a Elec/Comp Engineer but a bit of a steep curve for a first timer.

I propose that the volts->RPM route is highly error-prone because as others have said, the terminal volts will depend on load as well as RPM. The pulse count is a much more accurate way to go and isolates the measurement from your power conversion circuitry. Doing pulse count in Arduino is a lot trickier than voltage mapping. You can do it, but it'll take a little head scratching.

The good news is that pulse counting is exactly what a tacho does! If you buy any standard tacho (eg. an automotive tacho) it'll have a simple configuration for pulses per revolution. Then you just feed the pulse train in and it'll spit out RPM! There are analog and digital versions in every shape and size.

The question then is, how do you get a suitable pulse train out? I would first look to see if the signal from the Hall Effect sensor in the BLDC is available - that will be as close as you'll get to a ready-to-use signal. If that fails, then the voltage from one of the coils will do - you'll probably just have to threshold it to convert from the phase waveform to a pulse train. A simple Schmitt trigger ought to do the trick. Just be careful to consider the range of voltages that the generator will produce - a high input impedance Zener circuit will help clip the larger voltages.

And finally, to get motor volts, a voltmeter will do. Again, the display comes for free!

 

[DanBo1221]

LiteYear the tacho is definitely a more simple solution and it has been brought up. Like you said, the Arduino route may be easy for electrical/comp guys, but I'm considering this a good way broaden my own horizons away from composites and solid modeling! The display may be a problem at first, but I'm am heavily considering purchasing a $10 preconfigured LCD shield that is HD44780 compatible and uses the Anduino LiquidCrystal library. I'm not sure the program will be as easy as Anduino and Sparkfun claim it will be, but I'll try anyway. Anyone try this method yet?

 

[OperaHouse]

It must have been 30 years ago I was on a plane back from europe. A French woman got in this heated argument with a passenger. All at once she just ended it saying, "You will die stupid." At the time I thought this was the lamest insult I ever heard. I still think about that comment every day. There is no worse fate. We all have to learn something every day or we will die stupid. We didn't choose to go to the moon because it was easy.

I haven't used it but pulsein is likely to be a little unstable. Average a number of them as I have shown and it should be a repeatable number. Afrer all you are likely only interested in the 200-300rpm range. I would transformer to couple one of the phases and detect any positive signal. This would just be a simple LM339 with the output tied to the UNO pin and a 1K pullup to 5V. Feed the input signal through a 300K resistor to one of the inputs of the 339. Limit the voltage by back to back diodes to common where the other input is tied. The rest of the 339 will find other uses in this project.

 

[DanBo1221]

This would just be a simple LM339 with the output tied to the UNO pin and a 1K pullup to 5V. Feed the input signal through a 300K resistor to one of the inputs of the 339. Limit the voltage by back to back diodes to common where the other input is tied. The rest of the 339 will find other uses in this project.

Have some questions:
Are you saying a 1K pullup to the LM339 Vcc? Also, not understanding what you meant to do with the diodes.

 

[MikeHalloran]

Do not assume that VCC of a LM339 will be or must be 5V.
They are often used with +/- 15V supplies.
The open collector output makes them useful for interfacing between analog and digital circuits.

This

http://my.ece.ucsb.edu/bobsclass/2C/tutorials/App notes/an-74.pdf

should get you started on understanding some of the things the 339 can do.


Mike Halloran
Pembroke Pines, FL, USA

 

[OperaHouse]

Power the 339 from the common and 5V of the UNO. One of the input pins is connected to common. It really doesn't matter which input. The other input has two diodes in parallel, facing in opposite directions. This limits the maximum voltage to + - .6V. the 300K limits the current by providing a voltage drop. The trandformer must be an isolation type. A small wall wart 120 to 9-16V would be ok. Hope you will post your results on fieldlines.com