Arduino based four point probe

As title mentioned, my project is use Arduino to design an I-V measurement system for semiconductor wafer by using four point probe method. I used 1.02mm diameter spring test probe pogo pins for the probes. These four pins are soldered on strip board with same height and distance between each others.

Basically my idea is the arduino will output 0-4mA current from PWM digital pin (suggestion from my supervisor), and measure the voltage from the analog pin. The current increase by 0.25mA each time and back to 0 when exceed 4mA. I done this by using PWM theory and for loop operation. Here is some coding part for the voltage measurement (just want to show how i measure the voltage).

double readingV = analogRead(A0);
double Volts= ((readingV*5)/ 1024);
Serial.print(Volts);

From the picture given, my idea was connect the I+ pin to the digital pin which output the current, I- connect to ground, V+ connect to analog pin. I am not sure about V-, i connected that pin to ground also. The I+ pin ables to output desired current value (measured with multimeter, some offset value occurred such as 0.23mA current, but it is increasing until 4mA ). However, I did not get desired voltage value from analog pin when i tested my system with a standard reference wafer, the voltage obtained was not increasing as current outputs increases (sometime i got 0V from analog pin).

So i want to ask, regardless of the project coding, the wafer, or any mistakes I made, is it the connection of the probes wrong? since from the figure the V- is connected to voltmeter negative, not the ground. However, I have tried testing by connecting the two voltage pins with digital multimeter, I+ to PWM digital pins and I- connect to ground, still cant get desired voltage value from multimeter. Or the analog pin itself would not provide such accurate voltage value? Should I add something or external circuit for the voltage measurement?

My main focus will be get the correct resistance first before resistivity. Sorry for my broken english :smiley: I am appreciate if someone experienced to this four point probes can lead me in correct way. Thank You.

four point probe.gif

Image from Original Post so we don’t have to download it. See this Simple Image Guide

c584f0ac24d53b2092bc8d77a90ad38fe36aceda.gif

…R

Basically my idea is the arduino will output 0-4mA current from PWM digital pin

It is not possible to control the current from an Arduino pin.

PWM merely switches the pin between HIGH and LOW. The current drawn while the pin is HIGH depends on the load.

I guess you’re trying to measure (or calculate) resistance?

Basically my idea is the arduino will output 0-4mA current from PWM digital pin (suggestion from my supervisor)

[u]PWM[/u] is NOT analog and the Arduino (by itself) is NOT a current-source… It’s a voltage-source and (as-usual) the current depends on resistance (and voltage).*

However, I did not get desired voltage value from analog pin when i tested my system with a standard reference wafer, the voltage obtained was not increasing as current outputs increases (sometime i got 0V from analog pin).

And this sounds like an industrial or production application… In general, measurement instruments in “serious applications” should be calibrated & verified. There’s nothing wrong with using the Arduino in an industrial application but check with your QA department before using any “homemade” instrumentation.

  • Most “things” in electronics are constant-voltage (or controlled-voltage) sources. For example, the power outlet in your home has the same voltage if you plug-in a small low-wattage lamp or a toaster, or if nothing si plugged-in… The voltage is still there. The Arduino puts-out 0V or 5V (as long as you don’t overload it). Or, your stereo puts-out (about) the same voltage if you disconnect the speakers. In the case of audio, of course the voltage isn’t constant because it depends on the loudness of the music, but it’s “classified” as a constant voltage source because it’s independent of the load (within reason and with a load resistance that’s in-spec.)

There are constant current (or controlled current) power supplies for LEDs. The voltage self-adjusts to give you the correct current… If you disconnect the load, of course you don’t get any current but you get whatever maximum voltage is available as it “tries” to maintain current. If you short it out, the proper current will flow (at nearly zero volts) and unlike a regular power supply, it won’t burn out if you short it.

DVDdoug:
I guess you're trying to measure (or calculate) resistance?
[u]PWM[/u] is NOT analog and the Arduino (by itself) is NOT a current-source... It's a voltage-source and (as-usual) the current depends on resistance (and voltage).

And this sounds like an industrial or production application... In general, measurement instruments in "serious applications" should be calibrated & verified. There's nothing wrong with using the Arduino in an industrial application but check with your QA department before using any "homemade" instrumentation.

ten years ago we used to build the devices. Now we just repair and calibrate them. Back then they were selling for $15,000. The only ones we see now are from universities. So i suspect this is a university lab project.

Paul

From the picture given, my idea was connect the I+ pin to the digital pin which output the current, I- connect to ground, V+ connect to analog pin. I am not sure about V-, i connected that pin to ground also.

Look what that does to your circuit, basically it shorts things out. The current source must be isolated from the voltage reading. Having the two connected to one Arduino means it can not be isolated and so it will not work.

So you need an isolated constant current supply controlled by an optically isolated PWM signal if you want to use one Arduino. However as the Arduino’s voltage resolution is only about 4mV I am not sure if it will be enough for a semiconductor wafer.

Grumpy_Mike:
Look what that does to your circuit, basically it shorts things out. The current source must be isolated from the voltage reading. Having the two connected to one Arduino means it can not be isolated and so it will not work.

So you need an isolated constant current supply controlled by an optically isolated PWM signal if you want to use one Arduino. However as the Arduino’s voltage resolution is only about 4mV I am not sure if it will be enough for a semiconductor wafer.

Hi Mike, thank for the reply, i think i am wrong for connecting the two part with just one arduino. But how if i use 2 arduino uno, one for the current source, and the other one for voltage measurement? Or else any suggestion for the current supply i can apply to my project? As i mentioned my system will go for analog current, not constant, as i have to follow my sv suggestion. Since this is just an university project, i would give a try and see what happens next. But i still have to ask, does the current must be constant for the measurement? :slight_smile:

Yes paul this is my final year project for university, not industry purpose. So is ok if the result is not as accurate as the pro instrument used in lab, just like more of concept project. Thanks for the replies above. Appreciated :slight_smile:

But how if i use 2 arduino uno, one for the current source, and the other one for voltage measurement?

The the two must be powered by independent power supplies, that is you can’t power them both from the same computer.

But i still have to ask, does the current must be constant for the measurement?

Yes, if you are doing this with just a voltage you must adjust it so the current is the same for each measurement. I am not sure what resistivity your wafers are going to have but you can calculate the approximate values of current and voltage if you make an estimate of the expected resistivity.

project for university, not industry purpose. So is ok if the result is not as accurate as the pro instrument used in lab

As it is a University project then it is important you quantify the errors in the final instrument.

Oddly enough in 1974 my final year project was an automated system for measurement of resistivity and Hall coefficient by the van der Pauw method. This is a four point probe arranged in a clover leaf pattern. This was in the days before micro controllers and the whole thing was done with TTL logic gates, which included driving relays to switch the voltage and current between the pins and mechanically lowering, raising and rotating of the specimen in a magnetic field.

Hi Mike, let say if i were using 2 arduino for that, is it correct if i connect the I+ and I- to the digital pin and ground of the first arduino, and then the V+ and the V- connect to the analog pin and ground of second arduino?

Two Arduinos could still be connected through the supply or USB connection.

If you supply the current pins with smoothed PWM, then you could try to read the two voltage pins with two analogue inputs. Voltages can be derived with a bit of maths.
Leo..

One day last week I walked by the area of the plant where we work on the devices that measure semiconductor wafers. I see one of the boards uses a bunch of reed relays to keep the circuits isolated. Might be useful components for your design.

Paul

Using 2 Arduinos doesn't really solve the basic problem of using PWM as a current source. The first post already found one of the problems. PWM is either on or off and sometimes the analogRead() landed on an "off".

I expect there is a chip available to do this. The MAX31865 does exactly this job for temperature sensors. It would probably work for this role, if there isn't another more-specialised chip available.

There are better processor chips too. Many non-Arduino chips have differential analog inputs so the V- probe goes to an input pin and not ground.

I'm also planning to implement a 4-pt probe with an arduino. I would think using a separate IC to provide a controllable current source is the way to go and maybe use an op-amp or inst. amp to get a single-ended differential voltage signal to send to the arduino's ADC. playing with the current magnitude and the amplifier gain should allow me to get pretty good measurements.