Lissajous, Osciloscope.

[Shelschocker]

     First of all, sorry for my english as this is not my first language, and also sorry for any inconsistance as this is my first post here on eevblog or any forum what so ever.
Let me first explain what am I doing, and then what confused me about. I am a first year mechatronics engineering student doing a component tester using Lissajous patterns. Using the arduino to generate a PWM signal and then filtering so I can have a sine wave signal, then passing to an class ab amplifier, in the simulation I didn't use the amplifier,  and than to the circuit || component  that I'm going to test.
     I first start to use the osciloscope in the university yesterday to analyze if the signals that I was generating were right. But when I use the probe of the osciloscope to measure the voltage across the led I get a sine wave that is in the images, going to negative to positive, I was expecting a "rectified?" or half wave to be the signal, but what I got was a full sine wave with a brief cut, from 2.24V to -2.43V, when I measured with the multmeter I got 1.75V, I don't know if I'm doing something fundamentaly wrong using the osciloscope as I know little about how to use it.
     All the circuits I used are in the images, using the arduino I measured the voltage drop and current through the 1ohm resistor and ploted the graph and get what I was expecting for a diode. I'm confused about the readings I was getting on the oscilloscope, I ask for some teachers on my university but as my course if foccused on mechanics I didnt get answers from them. If anyone have an insite that could be helpfull I would be gratefull.

[Towz]

Nothing wrong there, only that you should expect the rectified voltage after the diode/LED. The voltage from the anode to ground will be alternating but from cathode to ground will be half wave rectified, so if you mesure across the 1 ohm resistor (hence, the current) you will see the rectified wave form you expected.
The 1.75 V from the multimeter is the rms value of the sine wave (0.707*2.40 aprox.)

[Shelschocker]

Thanks, I'll be testing today at the university.

[Shelschocker]

I simulated here and I'm getting nothing on the cathode to ground. Channel C of the oscilloscope.

[oldway]

First, wellcome to the forum....

The I/V curves of electronic components is used not only for repair but also for the pre test of electronics circuits in the production line.
It is particularly useful because it allows testing the components even in circuits without the need to power on the circuit.
This avoids the risk of destroying components.

The simplest systems use the 50 or 60 hz network at reduced voltage.
These circuits are usually called "octopus test component".

Some analog scopes have this function of component tester (eg Hameg oscilloscopes)

More sophisticated devices with memory enabling comparison between a good and a bad circuit were developed by the firm HUNTRON.

http://www.huntron.com/products/trackers.htm

You will find a lot of informations on this site.

[Shelschocker]

Thanks, yes, actually is because I started to work on a lab in the university that provides the components for the projects of the students, and my professor wanted to make a circuit like this to see if the component that the student give back is actually working.

I generate a 60hz pwm with arduino, as the quadratic wave is a sum of all odd harmonics, I used a band pass filter for 60Hz to  generate a sine wave, I tried to make a Wien bridge oscillator but failed, so the out of the filter I passed through a class AB amplifier, as the current was low.

Now I'm focusing on get the current and the voltage through the component under test, I used the analog pins of the arduino to get the voltage through the 100ohms resistor to measure the voltage and through the 1ohm resistor to know the current, so I plotted the graph and I got the analog signature of the LED.

So far I got the LED, diode, and capacitor shorted, and a led ruined right.

But my doubt actually was on the use of the osciloscope, as I started to use him this week, I didnt know what I was doing wrong in the circuit, as I mentioned  I was getting a full sine wave but Towz helped me with that.

But yet the problem is even if I measure the from the cathod I'll be getting nothing, and I couldn't understand that.

[oldway]

A component tester is very simple.

You don't need to generate a 60Hz sine wave, you get it for free from the mains.

http://pe2bz.philpem.me.uk/Power/-%20LV/Site-900-WallXformers/octopus.html

EDIT: more informations about the component tester here:

http://www.qsl.net/kd7rem/pdf/octopus.pdf

[alsetalokin4017]

So the Vdrop across the 100ohm resistor is ...what? If only the scopes were set to some reasonable V/div values and cursor positions we might be able to tell. This would then allow us to determine the current through the system when the LED is on. (By subtracting the LED anode Von  from the peak value before the 100R.)

When you use the scope to measure "from the cathode" of the LED in the circuit above, you are actually measuring the voltage drop across the 1 ohm resistor. This is converted to a current value by I=V/R, where V is the scope reading in volts and R is the one ohm value.  Say you are putting 20 mA through the LED when it is on (that is, when the voltage on the Anode is at 2.24 V and the LED is conducting). This means your scope signal at the Cathode will be half-wave rectified and varying from 0 V to 20 mV. (V=IR).
 
So to see this on the scope, the channel's V/div setting has to be set to a very sensitive value. In the scopeshot immediately above, if I am not mistaken the Cathode channel is set to 0.35 v/div, that is, 350 mV/div. So a 20 mV signal would not be visible.  Not only that but according to the Anode trace the LED isn't on for very long in the cycle.

The simulated scopes shown above all appear to have the fine V/div knobs set at some strange values, so the graticule markings are nearly useless to gauge the amplitudes of the signals visually. This results in, for example, a signal that "looks" like it is of greater amplitude than another one when it is really not, even though the _coarse_ V/div settings are the same... since the fine setting knobs are at different values.

[alsetalokin4017]

For the purposes of the stockroom checking returned components, what is needed is a circuit that uses the same components to make some simple signal that can be monitored on a scope, or even by blinking LEDs or making a DMM reading. The technician simply removes the matching component from the reference circuit and replaces it with the returned or suspect components. If the circuit behaves the same as before, the component is OK. If not... well, do the math. How many such circuits do you need? Depends on the coursework of course. If today's lab is a 555 astable oscillator, you breadboard it up and you can check every component in the circuit.

For example, to test a 555, you make an astable oscillator circuit using some caps and resistors and of course a working 555, and you monitor its output on the scope and input current on a DMM. Swap the 555s and see if the output waveform and current consumption are the same, within limits.

[alsetalokin4017]

But what does any of this have to do with Lissajous patterns?

[chris_leyson]

But what does any of this have to do with Lissajous patterns?

Because the scope is used in X-Y mode and not Y-T mode and maybe your perception or understanding of what a Lissajous pattern should look like is maybe based on Sci-Fi films.
 
From a historical point of view Lissajous figures or curves were also used for frequency comparison before practical frequency counters could be built. Later on they appeared in many Sci-Fi Films with X and Y driven at different frequencies but harmonically related by an integer ratio A/B but with one generator slightly offset in frequency so you would get a "rotating" pattern usually displayed on a TEK scope. Spirograph used cycloid and trochoid and probably other curves to generate similar patterns.

Welcome to the Forum Shelshoker, good to see someone investigate analogue signature analysis, it is a very useful tool for fault finding. "Fundamentals of Signature Analysis" is a good read http://www.huntron.com/sales-support/pdf/ASA-paper-extract.pdf as well as the Wiki page on Lissajous patterns. If you use a linear test load like an R-C or R-L or just C or L then you will get an ellipse displayed and that is a perfectly valid Lissajous curve, see the Wiki page for an animation. Your design thinking is good because you can "drive" the component under test at any frequency up to the bandwith of you drive amplifier and so you are not restricted to 50Hz or 60Hz. 10kHz or 20kHz would be better for testing smaller L's or C's so you have built in some flexibility.

Post a circuit diagram or sketch that would make it easier to see your measurement problems.

[Shelschocker]

A component tester is very simple.

You don't need to generate a 60Hz sine wave, you get it for free from the mains.

http://pe2bz.philpem.me.uk/Power/-%20LV/Site-900-WallXformers/octopus.html

EDIT: more informations about the component tester here:

http://www.qsl.net/kd7rem/pdf/octopus.pdf

Yes, I know that with a osciloscope it's relatively easy to make this circuit, but here where I live the price of the osciloscope it's about 3000 R$, you could say that's about 3000 U$ in comparison. So the idea it's really to not use the scope and use only the arduino to get the price down so you could make some and leave it to students use. For an example in the lab that I work we have 2 oscilloscopes and sometimes there are 6 projects going on at the same time, so this circuit would help to debugg some circuits as well.

About using the main the problem it's that my professor wanted to generate a range of waves, and with the main I get just 60Hz.

But even then I looked up at the page you send there are usefull informations, thanks.

[Shelschocker]

So the Vdrop across the 100ohm resistor is ...what? If only the scopes were set to some reasonable V/div values and cursor positions we might be able to tell. This would then allow us to determine the current through the system when the LED is on. (By subtracting the LED anode Von  from the peak value before the 100R.)

When you use the scope to measure "from the cathode" of the LED in the circuit above, you are actually measuring the voltage drop across the 1 ohm resistor. This is converted to a current value by I=V/R, where V is the scope reading in volts and R is the one ohm value.  Say you are putting 20 mA through the LED when it is on (that is, when the voltage on the Anode is at 2.24 V and the LED is conducting). This means your scope signal at the Cathode will be half-wave rectified and varying from 0 V to 20 mV. (V=IR).
 
So to see this on the scope, the channel's V/div setting has to be set to a very sensitive value. In the scopeshot immediately above, if I am not mistaken the Cathode channel is set to 0.35 v/div, that is, 350 mV/div. So a 20 mV signal would not be visible.  Not only that but according to the Anode trace the LED isn't on for very long in the cycle.

The simulated scopes shown above all appear to have the fine V/div knobs set at some strange values, so the graticule markings are nearly useless to gauge the amplitudes of the signals visually. This results in, for example, a signal that "looks" like it is of greater amplitude than another one when it is really not, even though the _coarse_ V/div settings are the same... since the fine setting knobs are at different values.

If I could put "say thanks" there I would put more times hahah, your explanation on the scale of the mV helped a lot, the 100ohm resistor is just to limit the current through the LED, I put a 10k ohm resistor and change the scale and then I got the half bridge on the LED, now I can run some calculations and understand it better, but it helped a lot, thanks.

[Shelschocker]

For the purposes of the stockroom checking returned components, what is needed is a circuit that uses the same components to make some simple signal that can be monitored on a scope, or even by blinking LEDs or making a DMM reading. The technician simply removes the matching component from the reference circuit and replaces it with the returned or suspect components. If the circuit behaves the same as before, the component is OK. If not... well, do the math. How many such circuits do you need? Depends on the coursework of course. If today's lab is a 555 astable oscillator, you breadboard it up and you can check every component in the circuit.

For example, to test a 555, you make an astable oscillator circuit using some caps and resistors and of course a working 555, and you monitor its output on the scope and input current on a DMM. Swap the 555s and see if the output waveform and current consumption are the same, within limits.

Yes, but it's quite a range of components, and actually it's not just the stockroom, is the lab for the students of the mechatronics engineering use to make their projects, as after the second/third year every semester have a class that it's a project, and then as we have a lack of oscilloscopes this circuit would also be helpfull to debugg some simpler circuits, and also have the advantage that we can make a certain quantity of them and let to the students use.

[Shelschocker]

But what does any of this have to do with Lissajous patterns?

Because the scope is used in X-Y mode and not Y-T mode and maybe your perception or understanding of what a Lissajous pattern should look like is maybe based on Sci-Fi films.
 
From a historical point of view Lissajous figures or curves were also used for frequency comparison before practical frequency counters could be built. Later on they appeared in many Sci-Fi Films with X and Y driven at different frequencies but harmonically related by an integer ratio A/B but with one generator slightly offset in frequency so you would get a "rotating" pattern usually displayed on a TEK scope. Spirograph used cycloid and trochoid and probably other curves to generate similar patterns.

Welcome to the Forum Shelshoker, good to see someone investigate analogue signature analysis, it is a very useful tool for fault finding. "Fundamentals of Signature Analysis" is a good read http://www.huntron.com/sales-support/pdf/ASA-paper-extract.pdf as well as the Wiki page on Lissajous patterns. If you use a linear test load like an R-C or R-L or just C or L then you will get an ellipse displayed and that is a perfectly valid Lissajous curve, see the Wiki page for an animation. Your design thinking is good because you can "drive" the component under test at any frequency up to the bandwith of you drive amplifier and so you are not restricted to 50Hz or 60Hz. 10kHz or 20kHz would be better for testing smaller L's or C's so you have built in some flexibility.

Post a circuit diagram or sketch that would make it easier to see your measurement problems.

I saw the the theory behind the Lissajous patterns, as my understand goes, what make this plot of signal assinature is the difference betwen phases angles in voltage and current sine waves, if there's no difference in the phase angle, you have a line, that's the resistor signature. But that's the limit of my theory behind it so far. I downloaded the material you recomended thanks.
About the restriction in the frequency, the only is the sampling frequency of the arduino uno that is theoretically of 1kHz/sec so by the Nyquist theorem theoretically I can sample the signals I got till 500Hz, but my professor ordered an Arduino Due that has 15kHz/sec , so we can use a range of frequencies, also if  in the future I needed a more range of frequencies I could use a FPGA, but that's out of reach for me now.

[oldway]

...
Yes, I know that with a osciloscope it's relatively easy to make this circuit, but here where I live the price of the osciloscope it's about 3000 R$, you could say that's about 3000 U$ in comparison.

Change is 3.26 R$/1 US$, 3000 R$ is not 3000 US$ 

For x / y, DSO are not that good but you can use the cheapest old 1 channel analog scope you can find on "mercado livre"....

... my professor wanted to make a circuit like this to see if the component that the student give back is actually working.

No need of several frequencies for this, only 50 or 60 Hz is enough.

... Using the arduino to generate a PWM signal and then filtering so I can have a sine wave signal,...

This may not work for a broadband of frequencies.

So the idea it's really to not use the scope and use only the arduino to get the price down so you could make some and leave it to students use.

Ok, that's the same concept as HUNTRON Tracker 2800.

[Shelschocker]

...
Yes, I know that with a osciloscope it's relatively easy to make this circuit, but here where I live the price of the osciloscope it's about 3000 R$, you could say that's about 3000 U$ in comparison.

Change is 3.26 R$/1 US$, 3000 R$ is not 3000 US$ 

For x / y, DSO are not that good but you can use the cheapest old 1 channel analog scope you can find on "mercado livre"....

Actually what I mean is a person who earn 500 R$ here will earn 500 U$ in the US, so If you can buy an oscilloscope in the US for 500 U$, the same scope here is about 3000. 

I'm gonna to take a look at this huntron thanks

[oldway]

From Hameg HM 605 manual, pg 24 :

[setq]

That section always bothered me. There are a few missing

[oldway]

That's not right: a blown up component is an horizontal line and a shorted component, a vertical line.

[setq]

That's not right: a blown up component is an horizontal line and a shorted component, a vertical line.

Doh!  Was half asleep when I did that. You are indeed correct. Voltage sweeps, current does nothing if it's open circuit.

Have corrected image.

The really stupid thing is that I built a curve tracer and should have seen that.

[Shelschocker]

Hey guys, thanks for all previous answers, I know got the figures for the components, now I'm doing a PCB for the project, but something that my professsor wanted me to do that I couldn't is a current limitter, for example, you would chose 1mA, 10mA, 50mA max, but I couldn't make it, If someone could just give me a direction would be nice. Thanks. For those who are interested, this figures I got from the arduino, so I save it and plot on Excel, but now I made a python code that plot the figures from the data of the serial port. The figures I got now a better resolution also.

[glicos]

Here's a project with same purpose but instead, use the soundcard of your PC and software (attached). You might get an idea with this.