Magnetic pickup - amplitude or frequency

T

Thread Starter

tech77

Hi,

I would like to measure speed of generator using magnetic pickup. Range of speed is 300 rpm to 3500 rpm and min AC voltage would be 0.5V.Hence I have to interface magnetic pickup sensor with microcontroller. In this regard I require few i/ps :

a) As the speed changes does the o/p of magnetic pickup changes in amplitude or in frequency.To find out the rpm should I count the pulse or measure the amplitude of the signal.

To my understanding it should be function of frequency. Correct if it’s wrong.

Speed = 60 X Hz (to be measured) / No of Rotor Teeth


b) The signal o/p of mag pickup sensor would be sinusoidal with 50 % duty cycle ??

c) suggest a interfacing circuit with micro.

Regds,
 
You would do well to review the documentation for the microcontroller you are considering connecting the speed pick-up to. If it's capable of using a passive, magnetic speed pick-up it will likely have programming and configuration information in the manual (or Help file) that is specific to the microcontroller.

In general, most "microcontrollers" these days seems to use frequency as the feedback from a passive, magnetic speed pick-up. It was more common in decades past to use voltage (as it is generally proportional to speed, though not as precise as frequency).

As for the formula you posted, you neglected to say whether or not you were trying to convert to revolutions/second or revolutions/minute, or engineering units you wanted or needed the speed to be indicated in.

And, the signal output of a pick-up would be depend somewhat on the shape of the teeth on the toothed wheel. A "good" microcontroller will sense zero-crossings and be independent of wave-form/shape when sensing feedback.
 
B

Bruce Durdle

Since the magnetic pickup works by generating an emf through the rate of change of flux linked into a coil, both the amplitude and the frequency of the output will be proportional to the speed. At low speeds, there may not be sufficient voltage induced to give a measurable signal against noise, so even if you use the frequency you need to be aware of the voltage dependence.

Bruce.
 
Z

Zacharia, Tomy

Depending on how flexible your requirements are (cost, hardware choice, mounting) a proximity switch type of device would be better for monitoring. P&F manufactures such devices.

In a magnetic probe (Variable reluctance) the voltage and frequency both change with respect to speed. Speed is normally 'picked off' a gear or a notch in a shaft. Plus there will be artifacts (small pulses on account of gear tooth shape etc). your signal conditioning circuitry will need to eliminate these. These nuisance signals also change in amplitude with respect to speed.

A proximity switch in comparison gives a clean square wave for your circuit to measure. The IS models (these are actually cheap) work with even 6-9 volts. Just measure the frequency and multiply by 60 to get RPM.

Regards,

Tomy Zacharia
 
>a) As the speed changes does the o/p of
>magnetic pickup changes in amplitude or
>in frequency.To find out the rpm should
>I count the pulse or measure the
>amplitude of the signal.

There are many rotation speed sensors with rectangular pulses on its outputs, with frequency proportional to rpm. The amplitude is constant.

Here is some of them:
http://www.sensorsportal.com/HTML/SENSORS/RotationSens_Manuf.htm

>To my understanding it should be
>function of frequency. Correct if it's wrong.

> Speed = 60 X Hz (to be
>measured) / No of Rotor Teeth

Yes, it is correct and you will get results in rpm.

>b) The signal o/p of mag pickup sensor
>would be sinusoidal with 50 % duty cycle??

Normally, it should be rectangular pulses with 50 % duty cycle.

>c) suggest a interfacing circuit with micro.

Many mentioned sensors can be directly connected to microcontroller. If you do not want to deal with the methods of frequency (rpm) measurements you can use the specialized IC UFDC-1 with rpm measurement mode, controlled by microcontroller:

http://www.sensorsportal.com/HTML/E-SHOP/PRODUCTS_4/UFDC_1.htm
http://www.sensorsportal.com/HTML/E-SHOP/PRODUCTS_4/Evaluation_board.htm

In this case you will get excellent metrological performances and forget about methods of measurements.

Good luck.
 
D

Dr. R. Erlich

Hi Tech77,

The following is a links to an industrial standard boards that will produce 0-10V (or 0 to -10V) DC from almost any Magnetic-Pickup signals.

http://www.eidusa.com/Interface_Boards_FREQ_VOLT_W_Zero_and_Span.htm

http://www.eidusa.com/Interface_Boards_FREQ_VOLT_W_Direction_Zero_and_Span.htm

http://www.eidusa.com/Interface_Boards.htm

The TRANSDUCER-FVC-DZS, OEM, Frequency to Voltage Converter (FVC) board is designed to convert a continuous series of digital pulses (frequency) into an analog voltage. The board is ideal for a close loop control of a motor shaft. The input is designed to accept low level (as low as +/- 60 mV P-P) pulses from magnetic pick-ups or high-level square wave signals from a shaft encoder (as high as +/- 40 V P-P). Model depended; the output voltage is positive (0 to 10 VDC) or negative (0 to 10 VDC) with a magnitude proportional to the input frequency signal.

The TRANSDUCER-FVC-DZS Module delivers analog 0 to +10 VDC or 0 to -10 VDC outputs proportional to input pulse rate (frequency). Where it accepts variable pulse rate inputs from a variety of sensors including magnetic pick-ups, with maximum linearity. It the ideal board for supplying velocity feedback in closed loop speed control systems.

The Board operates on the frequency content of a sinusoidal, triangular, or square waveform.

Good luck with your project,
Ron
 
Dear Ron !

What is a reason to convert frequency to voltage ? Such devices will introduce an additional error of measurement. The best solution from metrological point of view is directly to measure the frequency.
 
D

Dr. R. Erlich

> What is a reason to convert frequency to voltage ? Such devices will
> introduce an additional error of measurement. The best solution from
> metrological point of view is directly to measure the frequency.

Syurish (?) Thank you for your question,

The original inquiry was referring to changes in amplitude by Tech77. Those could occur from various reasons such as strong magnetic field at the point of measurement. Consequently, we don’t know really what is being measured. Since the signal can be as low as 0.5 volt in some cases the noise of up to two volt can be summed up to the signal. This noise can be positive or negative (Bi-Polar) therefore matching impedance and at least two poles Low Past Filter is usually required in industrialized applications. For your question from the 0 to 10 volts, I would go further and convert it immediately from 4-20 mA as close to the magnetic pickup as possible. Now the controller or PLC can read this signal at minimum error as far as a mile away. Most micro-controller read single ended single. Those signals measure against ground, or in some cases, floating voltage, i.e., various virtual-ground. To reduce error even further more I would suggest a board with differential input. The boards that I refer in my previous comment are off the shelf industrial frequency to voltage that has been tested against many magnetic pickups existing on the market. These boards address all of the above. The linear signal can be used as direct feedback to compensate changes in velocity without waiting for microprocessor to boot up (Hence, working in industrial applications, where I come from).
 
syurish,

I think Ron's point is that this dude is trying to do something crude and to get it done easily you could simply convert that signal externally to analog (which is easily understood and read in most controllers)... vs. trying to write some hacked up code and count pulses on a digital input... depending on the controller, you're talking about interrupts and who knows what... I think Ron offered a quick and easy alternative...
 
Dear friend,

Please need a help about magnetic pickup unit sensor. Am confused about the sensor output. Is it is square or sin wave output?
How can test the Magnetic pickup unit sensor? how can i view the output?

What will be the minimum voltage of the output?

I need to interface to micro-controller.

Your answer will be appreciated.

Thank you.
 
Ramana,

There is a lot of good and applicable information in the previous posts to this thread. Many times, manufacturers post a lot of very useful generic technical information about their products trying to demonstrate how theirs is better than others. Don't neglect such information.

Most manufacturers produce/provide data sheets for the types of sensors they produce, and those data sheets have a lot of useful information about outputs and sensitivity, etc. Again, don't neglect such information.

To briefly answer your questions:

Generally, depending on the angle of the teeth under the sensor the output is a square wave, but can can some smooth edges, and some can be very close to a smooth sine wave.

You can use an oscilloscope to view the output if you want to see it visually. You can use a True AC RMS multi-meter to measure the output voltage, and many such multi-meters also can measure frequency.

Voltage depends on a couple of things, usually: speed, and gap (the distance between the face of the pick-up and the toothed wheel). The sensor manufacturer's data sheet will usually have graphs and recommendations for expected RPM ranges and gaps and outputs (voltages, specifically--since frequency is directly related to the speed of the toothed wheel).

Most micro-controllers will have data sheets for the inputs and can describe what type of signals are possible. Some detect rising edges; other trailing edges; some detect zero volt crossings of the waveform (sine; square; saw-tooth) (the most accurate).

Some of the links in the above responses still work and can provide some data sheets for different types of sensors.

Finally, if you have an idea what you need it's very helpful to contact vendors/sales representatives in your area who can help you choose the proper sensor(s) for your application. Also, the micro-controller sales representative or technical support group can usually recommend a particular sensor they've have good experience with--or some to stay away from....

Hope this helps!
 
hi,

Thanks for your response.

That was a good suggestion.
Am little confused with the sensor. The RPM sensor am using is Diesel Generator RPM sensor, where it is like Variable reluctance type. That was an old sensor where am not getting any technical data. But i have observed the output at the site, @1500RPM, Voltage o/p= 12.5 VAC, frequency 3.3KHz and @750RPM voltage o/p=10 VAC, frequency=1.7KHz.

The basic working principle is similar like Current transformer (CT) but output voltage range may be difference.

My question is can i use the

http://www.crmagnetics.com/pdf/ancrct-4.pdf

CT application for RPM sensor to inteface with controller.Where RPM sensor output pins are placed at the CT.

Will this work?
I am unable test it practically in my lab. i want to visit site for testing. Please suggest me any circuit with thge values i have givem if the CT application not applicable.

Thanks
[email protected]
 
ramana,

From the information provided, I can't tell if you are trying to replace a failed or failing speed sensor that is connected to an existing control system on a diesel generator, or if you are trying to design and apply a new diesel generator control system using a micro-controller. Or if you are trying to monitor diesel engine speed with a micro-controller for some additional purpose or function. Perhaps if you could help us understand the entire "project" it would be more helpful to us to try to help you.

This is the first time I've encounter a variable reluctance device; or at least it's the first time I've heard a MPU (Magnetic Pick-UP; speed sensing device--passive or active) described as a variable reluctance device.

A quick search of the World Wide Web for "variable reluctance speed sensors" using my preferred search engine revealed this result:

http://www.motionsensors.com/speed-sensors/variable-reluctance-speed-sensors.shtml

There were several others, but this one seemed to me to be the most descriptive and concise.

Based on the information provided, I would say a typical passive speed pick-up (with two leads, and possibly a earth/ground lead) would suffice--but the answer really depends on the input circuitry the device will be connected to. Some magnetic speed pick-ups are active--meaning that they require a power source (for example, 12 or 24 VDC) to produce a signal. Most new, digital control systems I'm familiar with these days require a frequency input above a certain voltage level, say 5 VAC RMS. Some speed sensing circuits use a zero-based input signal, say 0-5 VDC, but still at some frequency.

I don't believe the CT device you are considering will work with the input circuitry, but that's a SWAG (Scientific Wild-Arsed Guess) based on the existing device in use.

From the information in other search results it would seem that while the variable reluctance sensors are inexpensive the input circuitry sometimes has to be a little more "costly" (complicated). Without understanding the device you want to connect the speed sensor to (the micro-controller) it's very difficult to say what will or won't work. And, just to be clear--I'm not an electronics expert. I try to read manufacturer's data sheets and application instructions to glean the information I need to be able to apply (use) a particular sensor with a particular monitor/control system. If I have questions, I contact the manufacturer of the device or monitor/control system or one of their manufacturing representatives who is usually able to answer the question(s) or get answers from the "factory."

So, again, you indicated you want to connect the speed sensor to a micro-controller. You need to understand what types of sensor outputs the micro-controller's input card/circuitry are capable of working with. The data sheets/application information in the manual of the micro-controller or the micro-controller's website should describe the type(s) of input signal(s) the particular input is capable of dealing with (accepting). Then you need to determine the sensor which you can purchase or use that will work with the micro-controller's input.

Possibly if you could provide some information about the manufacturer of the micro-controller and/or the particular input card you will be connecting the speed sensor to then someone here could assist with some information/ideas about a particular sensor, or type of sensor.

I'm suggesting you may be going about this in the wrong way--by trying to choose a sensor and then seeing if that will work with the micro-controller. I suggest that by understanding the micro-controller's input capability that you can then choose a sensor that will work with the micro-controller. That's usually how things like this are done.

From the information provided, I'm a little confused. Let's say the existing speed sensor is mounted over the toothed starter fly-wheel the diesel engine's starter turns to rotate the engine during starting. If the engine is spinning at 1500 RPM and the existing sensor output frequency is 3300 Hz, that would translate into (1500 / 60) 25 RPS, and to arrive at a frequency of 3300 HZ at 25 RPS the toothed wheel would have (3300 / 25) 132 teeth. At 750 RPM (12.5 RPS), the number of teeth required to produce 1700 Hz would be 136. So, I suggest there is some measurement inaccuracy--because the number of teeth on the starter fly-wheel isn't variable, it's fixed. A 132 toothed wheel would produce 1650 Hz at 750 RPM (12.5 RPS).

Again, most digital control system speed input circuits just require the voltage from a MPU speed sensor to be above some minimum level, say, 5 VDC. The digital control system isn't really concerned with the voltage magnitude--other than it must be above some minimum value to be recognized--it's really interested in frequency. So, if the voltage value varies as speed varies, as long as the voltage is above some minimum value the digital control system's input card/circuitry will detect and measure frequency.

And, the output voltage of a MPU is related to the speed of the toothed wheel the sensor is monitoring AND the distance between the toothed wheel's perimeter and the face of the MPU. And, that "gap" is usually adjusted to match the sensor manufacturer's specifications to produce a particular voltage (range, really) over the expected operating speed of the device and it's toothed wheel.

Please write back to let us know how you fare!
 
hi,

Explanation was fair enough.

Let me explain clearly.
Am trying to build a product which monitors the sensors readings and displays it through Micro-controller PIC32.

So, i have worked on all the sensors. But, need to built a circuitry for RPM sensor. Am unable find out the exact datsheet for technical information. Becuase am monitoring the unknown rpm sensor.

But i have observed the the outputs of the RPM sensor in Diesel generator which the observations are
@1500RPM Voltage-12.5 VAC and frequency: 3.3kHZ
@750RPM Voltage- 10 VAC and frquency:1.5 KHz.

The hardware circuitry is my part to built where to monitor the speed of the sensor.

By observing the ouputs at the field, i have assumed the sensor is type VARIABLE RELUCTANCE.

So, please suggest to built the circuitry.

Thanks
 
Ramana,

When I searched for "variable reluctance speed sensors" (double quotation marks included in the search field) I found several examples of circuits in more than one of the search results.

I'm not familiar with PIC32 controllers, but your mission is to devise a circuit that converts frequency (usually a zero-crossing waveform, closely approximating a sine wave) into counts that the controller can convert to RPM with the proper scaling.

You will need to know the number of teeth on the wheel.

Lastly, if you're trying to read that speed sensor "in parallel" with the diesel engine controller, you will need to devise a circuit that does not load the sensor output such that neither control system can read the speed sensor's output.

If that's not the exact type of circuit, than it's something similar. Again, I suggest you use your preferred Internet search engine to find some examples of what you're trying to do, based on what kind of inputs the PIC32 can accept. You could, for example, convert the frequency into some scaled 4-20 mA output, say 0-1600 RPM for example, and then scale the PIC32's mA input similarly to display RPM.

Best of luck!
 
The dependency of RPM with Frequency in a pick up can't be different from linear, due to the fixed amount of teeth. If you are experiencing different rates of rpm->frequency on the pick up, there is without any doubt an error anywhere.
 
Top