E-Field Mill




About: Electronic Engineer. Living in northern germany. Working in the R&D department of a mid-sized company manufacturing medical products.

You may already know that i'm addicted to any kind of sensor measurement applications. I always wanted to track down the fluctuations of earths magnetic field and i also was fascinated by measuring the ambient electric field of the earth that is maintained by charge separation processes taking place between the the clouds and the surface of the earth. Incidents like clear sky, rain or thunderstorm all have an dramatic impact on the electric field that surrounds us and new scientific findings are showing us that our health depends heavily on the surrounding electric fields.

So, that's the reason why i wanted to make myself a suitable measuring device for static electric fields. There is already one pretty good design existing, also called electric field mill that is widely used. This device uses an effect called Electrostatic Induction. This always happens when you're exposing a conductive material to an electric field. The field attracts or repels the free electrons in the material. If it is connected to ground(earth potential) charge carriers are flowing in or out of the material. After disconnection of the ground a charge remains on the material even if the electric field vanishes. This charge can be measured with a voltmeter. This is very roughly the principle of measuring static electric fields.

A few years ago i built a field mill according to plans and schematics i found in the internet. Principally it consists of a rotor with some kind of propeller on it. The propeller is a twin set of metal segments that are grounded. The rotor turns around a set of induction plates that are electrically covered and uncovered by the rotor. Every time they are uncovered the electrostatic induction of the ambient electric field causes a flow of charge carriers. This flow is reversed when the rotor covers again the induction plates. What you get is an alternating more or less sinusoidal current which amplitude is a representation of the strength of the measured field. This is the first flaw. You don't get a static voltage showing the field strength but have to take the amplitude of an alternating signal that has to be rectified first. The second issue is even more tedious. The field mill works pretty well in an undisturbed environment -lets say on the dark side of the moon when you're far away from power line hum and all this abundant electric mist that is penetrating our environment everywhere we are. Especially the 50Hz or 60Hz power line hum interferes directly with the desired signal. To tackle this problem the field mill uses a second set of inductions plates with another amplifier that takes the same signal with a 90° phase shift. In an additional operational amplifier both signals are subtracted from each other. Because they're out of phase a remainder of the desired signal remains and the interference, which is equal in both signals, is cancelled out theoretically. How good this works depends on the equality of the interference in both measurement circuits, the CMRR of the amplifier and on the question if the amplifier gets overdriven or not. What makes the situation even more uncomfortable is that you roughly doubled the amount of hardware just to get rid of the interference.

Last year i had an idea to overcome those problem with my own design. It's a little more work on the mechanic but simple in question of electronics. As always this is not a detailed step by step replication of complete device. i will show you the working principles on my design and you may change it in different ways and tailor it to your own needs. After showing you how to build it i will explain how it works and show you the result of my first measurements.

When i got the idea for this device i was proud to the bones but as you know arrogance is preceding any downfall. Yes, it was my own idea. I developed it on my own. But as always there was someone before me. The separation of charges by induction and amplification by using the capacitor effect was used in nearly every electrostatic generator design during the last 150 years. So there's nothing special about my design despite the fact that i was the first one who thought about applying those concepts for measuring weak electrostatic fields. I still hope one day i will be famous.

Step 1: List of Materials and Tools

The following list shows roughly which materials you will need. You may change and tailor those as much as you want.

  • Sheets of 4mm plywood
  • timber beams 10x10mm
  • 8mm aluminum tube
  • 6mm aluminum rod
  • 8mm plexiglass rod
  • 120x160mm single side copper plated PCB
  • brass or copper wire 0.2mm
  • a piece of 0.2mm copper sheet
  • solder
  • glue
  • 3mm screws and nuts
  • A 4mm test socket
  • conductive rubber tube (Inner diameter 2mm) I got mine from amazon
  • Electronic parts according to the schematic(download section)
  • A 68nF styroflex capacitor as a collector for the charges. You may change this value in wide ways.
  • A capstan motor for 6V DC. These are motors that were especially designed for disc players and tape recorders. Their rpm is regulated! You still can find them on Ebay.
  • A 6V/1A power supply.

These are the tools you need

  • Soldering iron
  • Arduino development environment on your PC/Notebook
  • USB-A to B cable
  • file or better a lathe
  • electric drill
  • small buzz saw or hand saw
  • tweezers
  • wire cutter

Step 2: Making the Mechanics

In the first picture you can see the whole design is based on two sheets of plywood 210mm x 140mm in dimension. They are mounted above each other, connected by 4 pieces of timber beams that keeps them 50mm in distance. Between both sheets the motor and the wiring is contained. The motor is mounted with two M3 screws fitting in two 3mm holes drilled through the upper plywood sheet. A sheet of PCB material is working as a shield against the ambient electric field. It is mounted 85mm above the upper plywood sheet and its inner edge is just ending about the motor shaft.

The core component of this device is a disk. It has a diameter of 110mm and is made out of single side copper coated PCB material. I used a mill to cut out a round disk of of the PCB. I also used a mill to cut the copper coating into four segments that are electrically insulated. It is also very important to cut a ring around the middle of the disk where to motor shaft will go through. Otherwise it would electrically ground the segments! On my lathe i cut a small piece of 6mm aluminum rod in a way that it takes a 3mm hole at the bottom with two rectangular 2,5mm holes that have M3 threads cut in. The other end i cut down to a small 3mm shaft to fit in the middle hole of the disk. The adapter then was super-glued to the bottom of the disk. The disk assembly could then be screwed to the motor shaft.

Then you see another important component. A segment of the size of the ones on the disk, made out of 0,2mm copper sheet This segment is mounted on two sheets of plywood. When the disk is mounted this segment is very narrowly under the rotating disk. the distance is just about 1mm. It's important to keep this distance as small as possible!

The next important things are the ground whisker and the charge pick-up. Both are made of aluminum tube and rods with cut in threads to mount them all together. You can do any type of variation you like here. You just need something conductive running over the surface of the disk. For the whiskers i tried a lot materials. Most of them were damaging the disk segments after a while. Finally i found a hint in a book about electrostatic devices. Use conductive rubber tubing! It's not damaging the copper coating and wears and wears...

The ground whisker is placed on a location in a way that it loses contact to the underlying disk segment when it is starting to uncover the ground plate. The charge pick-up is placed such that it takes the segment in the middle when it's at maximum distance from the ground plate. See that the charge pick-up is mounted on a piece of plexiglass rod. This is important because we need a good insulation here. Otherwise we would have a loss of charges!

Then you see that the 4mm test socket is placed in the "basement" of the assembly. I provided this connection because i was not sure if i would need a real "ground" connection or not. Under normal conditions we are dealing with such low currents that we have an intrinsic grounding anyway. But maybe there will be a test setup in the future where we might need it, who knows?

Step 3: The Wiring

Now you have to electrically interconnect everything so that it works properly. Use the brass wire and solder together the following parts.

  • The 4mm test plug
  • The ground whisker
  • The shield
  • one wire of the charge collect capacitor

Solder the 2nd wire of the capacitor to the charge pick-up.

Step 4: Making the Electronics

Follow the schematic to place the electronic components on a piece of perfboard. I soldered pin headers to the edges of the board to connect it with the Arduino Uno. The circuit is damned simple. The collected charge is picked-up at the capacitor and fed into an high-impedance amplifier which boosts the signal by 100. The signal is low-pass filtered and then routed into one input of the arduino's analog-to-digital converter inputs. A MOSFET is used for the Arduino to switch on/off the disk motor.

It's very important to connect the ground of the mechanic assembly to the virtual ground of the electronic circuit which is where R1/R2/C1/C2 meet! This is also the ground of the charge collecting capacitor. You can see this on the last picture in this chapter,

Step 5: The Software

There isn't much to say about the Software. It's written very straightforward. The application knows some commands to get configured properly. You can access the arduino if you have the Arduino IDE installed on your system because you need the virtual comport drivers. Then plug an USB cable to the arduino and your PC/Notebook and use a terminal program like HTerm to connect the arduino via the emulated comport with 9600 bauds, no parity and 1 stopbit and CR-LF on enter.

  • "setdate dd-mm-yy" sets the date of the RTC-module connected to the arduino
  • "settime hh:mm:ss" sets the time of the RTC-module connected to the arduino
  • "getdate" prints date and time
  • "setintervall 10...3600" Sets the sampling intervall in seconds from 10s to 1h
  • "start" starts the measurement session after syncing to the upcoming full minute
  • "sync" does the same but waits for the upcoming full hour
  • "stop" stops the measurement session

After receiving "start" or "sync" and doing the synchronization stuff the application first takes a sample to see where the zero-point or bias is. Then it starts the motor and waits 8s for the rpm to stabilize. Then the sample is taken. Generally there is a software averaging algorithm that continuously averages the samples over the last 10 samples to avoid glitches. The previously taken zero-value is now subtracted from the measurement and the result sent over the comport together with date and time of the measurement. A example of a measurement session looks like this:

03-10-18 11:00:08 -99

03-10-18 11:10:08 -95

03-10-18 11:20:08 -94

03-10-18 11:30:08 -102

03-10-18 11:40:08 -103

03-10-18 11:50:08 -101

03-10-18 12:00:08 -101

So, the measurements are shown as deflections from zero measured in digits which can be positive ore negative depending on the spatial direction of the electric flux. Of course there's a reason why i decided to format the data in columns of date, time and measurement values. This is the perfect format to visualize the data with the famous "gnuplot" program!

Step 6: How It Works

I just told you that the working principle of this device is electrostatic induction. So how it works in detail? Lets assume for moment we would be one of those segments on the disc. We are rotating at a constant speed continuously being exposed to the ambient electric field and then hiding again from the flux under the protection of the shield. Imagine we actually would get out of the shadow into the field. We would get in contact with the grounding whisker. The electric field would act on our free electrons and lets say the field would repel them. Because we are grounded there would be an amount of electrons fleeing from us and vanishing in the earth.

Losing ground

Now, while the turning of the disk continues at some point we would lose contact to the ground whisker. Now no more charge can flee from us but the way back for the charges already gone is also closed. So we are left behind with a lack of electrons. If we like it or not, we are charged now! And our charge is proportional to the strength of the electric flux.

How much charge do we have?

During the time we got exposed to the electric field we lost some electrons. How much have we lost? Well, with every electron we lost, our charge climbed up. This charge generates an rising electric field of its own between us and the ground. This field is opposite to the ambient one which generated the induction. So the loss of electrons continues up to the point where both field are equal and cancel out each other! After we lost contact with ground we still have our own electric field against the grounded plate which has earth potential. You know how we call two conductive plates with an electric field in between? This is a capacitor! We are part of charged capacitor.

We are a capacitor now!

You know the relation between charge and voltage on a capacitor? Let me tell you, it is U=Q/C where U is the voltage, Q is the charge and C the capacity. The capacity of a capacitor is reversely proportional to the distance of its plates! That means the wider the distance the lower the capacity. Now what happens while we keep on turning on the wheel with no contact to ground? We are increasing the distance to the ground plate. While we're doing this our capacity falls dramatically. Now look again at U=Q/C. If Q is constant and C is declining, what happens? Yes, the voltage is rising! This is a very clever way to amplify the voltage by just applying mechanical means. You don't need an operational amplifier, noise filtering and statistical computing here. It's just clever and plain physics that boosts our signal up to a level where signal processing with electronics just becomes a boring tasks. All the cleverness of this device relies on electrostatic induction and the capacitor effect!

What does it mean?

But what exactly did we boost in this way? Do we have more electrons now? No! Do we have more charge anyway? No! What we boosted is the ENERGY of the electrons and this is what enables us to use simpler electronic circuits and less filtering. Now we reached the aphel of our trajectory and finally the charge pick-up takes our energized electrons and collects them into the charge collector capacitor.

Immunity against interference

When you take a look at the video you'll see that despite the usual interference in my home the output signal of the device is steady and practically noise free. How is this possible? Well i think its because signal and interference are not going separate way up to the amplifier as in the classic field mill. In my design the interference affects the collected charge right from the moment on the connection to ground is lost . That means every sample is affected in some way by interference. But because this interference has no DC component as long as it's symmetrically, the interference result is always averaged-out in the charge collector capacitor. After enough disc turns and samples fed in the charge collector the average of the interference is zero. I think that's the trick!

Step 7: Testing

After some testing, debugging and improving i installed the field mill together with my old win-xp notebook in my attic and did a test run over approximated one day. The results were visualized with gnuplot. See the attached data file "e-field-data.dat" and the gnuplot configuration file "e-field.gp". To view the results just start gnuplot on your target system and type at the prompt >load "e-field.gp"

See the picture showing the results. It's quite remarkable. I started the measurement on 2018-10-03 when we had fine weather and blue sky. See that the electric field was pretty strong and negative, while we have to take care because whats "negative" and whats "positive" currently is not reasonable specified. We would need a calibration of our device to align with real physics. But anyway, you can see that over the measurement cycles the field strength went down alongside with the weather starting to deteriorate and becoming cloudy and rainy. I was somehow amazed about those findings but still have to check if these correlate with physics.

Now it's your turn. Go on and make your own electric field mill and explore the secrets of our planet on your own quest! Have fun!

Step 8: Gathering and Interpreting Data

Now as everything is (hopefully) working fine you should gather some data. I'd recommend to use a fixed place for the field mill. Otherwise the data would be hard to compare. The local field parameters may vary a lot from place to place. I configured the mill that it took one measurement value every hour. I let the mill run for about 3 months. If you take a look at the graphs presenting the gathered data of the month november 2018, december 2018 and january 2019, you see some remarkable findings.

First you can see that the field strength in november was merely positive turning into negative by the end of the month. So something general must have changed, likely according to weather. Maybe there was a reasonable temperature drop. Then the average signal stayed negative up to the end of the measurement cycle. The second thing is that there are several spikes in the signal graph indicating fast field changes lasting only some minutes. I don't think that changes in the atmosphere are responsible for that. Even the local weather comprises huge masses of gas and incorporated ions. Also clouds and raain or snow usually don't change within minutes. So i think man-made influence may have caused those sudden changes. But this is also difficult to explain. All power line sources only provide ac-voltage. That doesn't count for the dc-changes i observed. I suspect there might have been some electrical charge processes by cars passing by on the asphalt of the street in front of my flat. Thinkable would be also charge processes caused be dust carried by wind and getting in contact with the face of my house.



    • Sensors Contest

      Sensors Contest
    • Colors of the Rainbow Contest

      Colors of the Rainbow Contest
    • Pets Challenge

      Pets Challenge

    36 Discussions


    Question 5 weeks ago

    Really a brilliant design! This is the first one I've seen online that appears within my abilities to build. Thanks so much for sharing!

    I have a couple questions...Please excuse my rudimentary electrical engineering knowledge:

    1) I see your insight below on motor RPM, and all of the 6V capstan motors I see on Ebay are listed as 2400RPM motors. DO you think one of those would work?

    2) Does the charge collecting capacitor need to be a polystyrene capacitor?

    3) When you say the capacitance of this capacitor can be changed, how much is the tolerance?

    4) I'm having trouble finding "conductive rubber tubing" for a reasonable price. (I'm assuming this is carbon filled silicone? Let me know if you have any good ideas on sourcing this material.

    Again, Thanks so much! I've been wanting to build one of these for years...with these plans, I'm hoping to measure the electric field underneath some thunderstorms this summer! :)

    3 answers

    Answer 5 weeks ago

    Hello my friend and thanks for your kind comments. I really appreciate your plans on doing researches on the variations of earths electric field in thunderstorm. This is a startling project and really of the kind i wanted to fuel with my design! Now i come down to your questions.

    1) Every Motor will work that is supposed to run on a more or less constant speed. But even this requirement may be weak. I have the theory that, depending on the setup (disk size, charge collector capacitor, input current on OP-AMP) we need a minimum rpm of the motor but from a certain rpm on upwards there would be no influence of the collected charge anymore. But as long as this is theory just choose any motor with stabilized rpm you like.

    2) I first was very concerned about the quality of this capacitor and i thought it must have as less leakage current as reasonably possible. This was due to my experiences with electrostatic machines where you tediously and carefully have to collect charge carriers in a leyden jar until you gain enough of them to realize a fulminant discharge. Those setups don't forgive you the slightest design flaws that yield into leakage current. Now, here in this design my experience is that this is not that dramatic. We don't reach high voltages and this is where the leakages become nasty. So i would say you could use any capacitor apart from electrolytic ones. We have a steady stream of incoming energized charges and we always have some unavoidable leakage anyway. At least the OP-AMP swallows some nanoamperes.

    3) Generally there is no tolerance and you can make it the size you want. You just have to obey two simple rules. The smaller the capacitor is the faster it will charge-up but the "averaging" effect for interference and the the discharge from leakage currents will have more and more impact on the signal. Think of the capacitor as a part of a lowpass. The higher you make it the more your signal becomes a steady DC. If it is too small the signal becomes noisy. This leads us directly ti the second rule. The bigger the capacitor is, the better becomes you signal quality but the time to charge the capacitor also increases. At some point perhaps you would need to increase the time between motor start and taking the value from the capacitor.

    4) I used this from amazon: https://www.amazon.de/gp/product/B009L9EWPU/ref=pp...
    But you also can try different materials. Sometimes electronic components are delivered in some kind of "electric foam" to protect the against ESD. Also the small bags where the components are wrapped-up are made of plastic that has a conductible surface. You can take everything that is smooth and a little conductive. A high impedance is nor problem because we are working with very tiny currents. But you should avoid using metal whiskers. I tried copper and steel and they where ruining the plates.


    Reply 5 weeks ago

    Awesome! Thanks for such an in-depth answer. That all makes sense. I'll start ordering parts today. It may take me a while to put it all together, but I'll keep you posted! I'm hoping I'll be able to calibrate it as well so that I can have a dimensional number, but we'll see...I'll need to buy a real power supply for that!

    One last thing: A lot of DIY and commerical field mills I have seen all have grounded shielding on all sides. Your design appears to just have the shield on the top covering half the spinning plate. I'm assuming it doesn't really matter too much though, as the field mill will still be measuring the total electric field perpendicular to the plates, right?


    Reply 5 weeks ago

    This is correct. In my experiments i made the assumption that the electric flux is vertically oriented and in fact i never observed horizontal field components giving me a signal. Of course the current design would be susceptible to field components with spatial orientations between horizontal and vertical principally. To be sure you'd need to enclose the sides with shields as you described. For our purposes, i don't think this is necessary but if you do you should take care to keep at least the same distance between those new shields and the disk as it is now between disk and the upper shield. Otherwise the voltage gain by the capacitor effect would be negatively affected.


    Question 6 weeks ago

    May I ask your question.
    Please tell me how to change the measurement interval?

    1 answer

    Answer 6 weeks ago

    This is simple. Use the command "setintervall <value>" to change the sample interval as you wish in seconds. E.g enter "setintervall 60" for setting the sample interval to 1 minute. The value can be between 10 and 3600.


    Question 2 months ago

    Excuse me.May i ask question?
    What type RTC dod you use? Please tell me using RTC type.

    2 answers

    Answer 2 months ago

    Hello! Sure, the RTC is an I2C coupled module available as the standard Arduino RTC shield. It runs on an DS3231 clock chip.


    Reply 6 weeks ago

    Thank you for telling me.
    Thanks of you, I could settle these proglem.
    Thank you very much.


    Question 2 months ago

    hello.I am japanese.I am not good at English.Please allow me.

    By the way,I have question about Hterm. I installed it and conected to arduino.
    I sent arduino "setdate 26-03-19" but arduino reply"165-85-165"
    Arduino dose not reply correct time and date.

    Please tell me solving method.

    1 answer

    Answer 2 months ago

    Hello again! That really strange. May i ask what RTC-module you are using? Do you have the the CR2032 backup battery inserted in the RTC-module?


    Question 5 months ago

    Hello I would like to know what is the Red and Black element on the printed circuit board that you have noted "e field".
    Thank you in advance

    6 answers

    Reply 5 months ago

    A red and black element? You mean on one of the pictures, the video or are you refering the schematic?


    Reply 5 months ago

    Sorry, I was talking about the one in the photo in step 4, the one on the circuit board.


    Reply 5 months ago

    Ah ok, this is the potentiometer R5(22K).


    Reply 5 months ago

    Thank you. Is that a digital potentiometer?


    Reply 5 months ago

    No, this is just an ordinary potentiometer to be trimmed with a screwdriver.


    Reply 5 months ago

    Hi! Sorry but i don't know about what element you're talking about. Where is the black and red element you mentioned? Is it on one of the pictures, the schematic or the video?


    Question 8 months ago on Introduction

    RE RPM - beside being constant, is speed a concern? And if so is there any calibration required? Also, what effect does an RPM variation have on the results?

    1 answer

    Answer 7 months ago

    An excellent question! My first intention of using a stabilized rpm was that everything you do not understand properly has an impact on your setup. So i made it this way to be safe. Later i thought deeper about this issue. Generally the disk-segments are rendering a current flowing into the charge collecting capacitor. The faster you turn the disk the more current flows. On the site of the charge collector you have an unavoidable leakage current because your capacitor isn't ideal and the electronic amplifier always has some input current. So the first finding is that the current from the segment disks must be larger than the leakage current. You need a minimum rpm. The second question is what happens if we exceed this minimum rpm more or less? Obviously he charge collector will charge more or less quickly but up to what limit? Well, the upper limit is specified by the fact that every disc segment is a capacitor by itself. It takes some charge and generates a field against ground. By dragging the segment away from the grounded plate the energy of this charge gets amplified resulting in a voltage. When the segment gets in contact with the charge collector it takes this charge up to it's voltage. That means the voltage on the charge collecting capacitor can't get higher than the voltage of the segment. The conclusion is that we need some minimum rpm. Above it an increase of the rpm only results in a faster charging of the charge collecting capacitor but it doesn't change it's end voltage. So far the theory, it's unproven yet.
    A calibration is generally necessary because actually we just have digits with no dimension. I'd have to put the device between two large metal plates and apply a voltage to it. So i could calculate the field strength and the field would be more or less homogeneous. Then i could take the digits and calculate the proper transformation factor.