Sensorless BLDC motor control with Arduino – DIY ESC

DIY ESC with Arduino UNO

This topic shows how to build a sensorless brushless DC (BLDC) motor controller or simply an ESC (Electronic Speed Controller) with an Arduino UNO board.

There are two types of brushless DC motors: sensored and sensorless. Sensored BLDC motor has built-in 3 hall effect sensors, these sensors detect the rotor position of the BLDC motor. Controlling a sensored BLDC motor is easy since we know the rotor position like what was done in the project below:
Sensored brushless DC motor control with Arduino

The commutation of the sensored BLDC motor is done according to the hall effect sensors state.

Sensorless BLDC motor doesn’t have any sensor to detect its rotor position, its commutation is based on the BEMF (Back Electromotive Force) produced in the stator windings.

The main advantage of sensorless BLDC motor control is lower system cost and the main disadvantage is the motor must be moving at minimum rate to produce sufficient BEMF to be sensed.

How it works:
When the BLDC motor rotates, each winding (3 windings) generates BEMF opposes the main voltage. The 3 generated BEMF signals are 120° out of phase which is the same as the hall effect sensor signals. The figure below shows the relationship between the hall effect signals and the BEMF signals:

bldc hall sensor and bemf signals

As shown in the figure above, the BEMF signals are not synchronized with the hall effect sensor signals (phase shift of 30°). In every energizing sequence, two windings are energized (one connected to positive and the other to negative) and the third winding is left open (floating). The floating winding is used to detect the zero crossing, thus, the combination of all 3 zero cross over point are used to generate the energizing sequence. Totally we’ve 6 events:
Phase A zero crossing: from high to low and from low to high
Phase B zero crossing: from high to low and from low to high
Phase C zero crossing: from high to low and from low to high

How to detect the zero crossing event:
The easiest way to detect the zero crossing events is by using comparators. The comparator has 3 main terminals: 2 inputs (positive and negative) and an output. Comparator output is logic high if the positive voltage is greater than the negative voltage, and logic low if the positive voltage is lower than the negative voltage.
Basically 3 comparators are needed for this project, connections are done as shown in the figure below (example for phase B). Each phase requires a similar circuit.

bldc bemf signal detection

The virtual natural point is the same for all the 3 comparators, it is generated using 3 resistors. When the BEMF generated in the floating (open) winding crosses the zero point towards positive side, the comparator output makes a transition from low-to-high. When the BEMF generated in the floating winding crosses the zero point towards negative side, the comparator output makes a transition from high-to-low. By having three such comparator circuits, one on each of the phases gives three digital signals corresponding to the BEMF signal in the windings. The combination of these three signals is used to derive the commutation sequence.

Hardware Required:

  • Arduino UNO board
  • Brushless DC (BLDC) motor
  • 6 x 06N03LA N-type mosfet (or equivalent)  – datasheet
  • 3 x IR2104S (IR2104) gate driver IC  – datasheet
  • 6 x 33k ohm resistor
  • 3 x 10k ohm resistor
  • 6 x 100 ohm resistor
  • 3 x IN4148 diode
  • 3 x 10uF capacitor
  • 3 x 2.2uF capacitor
  • 2 x pushbutton
  • 12V source
  • Breadboard
  • Jumper wires

Sensorless BLDC motor control with Arduino circuit:
Project circuit schematic is shown below.

Arduino sensorless BLDC motor controller DIY ESC circuit


Note that all grounded terminals are connected together.

In the circuit there are 2 pushbuttons, one is used to increase BLDC motor speed and the 2nd one is used to decrease it.

The first three 33k (connected to motor phases) and the three 10k resistors are used as voltage dividers, because we can not supply the microcontroller with 12V, the other three 33k resistors generate the virtual natural point. The virtual natural point is connected to Arduino pin 6.

The Arduino UNO board is based on the ATmega328P microcontroller which has one analog comparator. The positive input of this comparator is on Arduino uno pin 6 (AIN0) and the negative input can be pin 7 (AIN1), A0 (ADC0), A1 (ADC1), A2 (ADC2), A3 (ADC3), A4 (ADC4) or A5 (ADC5). So I connected the virtual natural point to the positive pin of the analog comparator (pin 6), phase A BEMF to pin 7 (AIN1), phase B BEMF to pin A2 and phase C BEMF to pin A3. Each time the comparator compares the virtual point with the BEMF of one phase (this is done in the software). This minimizes the hardware needed and simplifies the circuit.

The IR2104S chips are used to control high side and low side mosfets of each phase. The switching between the high side and the low side is done according to the control lines IN and SD. The figure below shows input and output timing diagram:

ir2104s timing

The SD lines of the three IR2104S are connected to pins 11, 10 and 9 respectively for phase A, phase B and phase C. The Arduino UNO can generate PWM signals on that pins where only high side mosfets are PWMed.

Sensorless BLDC motor control with Arduino code:
The code below does not use any BLDC motor library.

As mentioned above, Arduino pins 9, 10 and 11 can generate PWM signals where pin 9 and pin 10 are related with Timer1 module (OC1A and OC1B) and pin 11 is related with Timer2 module (OC2A). Both Timer modules are configured to generate a PWM signal with a frequency of about 31KHz and a resolution of 8 bits. The duty cycles of the PWM signals are updated when a pushbutton is pressed (speed up or speed down) by writing to their registers (OCR1A, OCR1B and OCR2A).

The analog comparator compares the positive input AIN0 (Arduino pin 6) with the negative input which can be AIN1 (pin 7), ADC2 (pin A2) or ADC3 (pin A3). When the positive pin voltage is higher than the negative pin voltage, the output of the analog comparator ACO is set, and when the positive pin voltage is lower than the negative pin voltage, ACO is cleared.
In this project I used the analog comparator interrupt and I used its interrupt on rising (transition from low to high) and interrupt on falling (transition from high to low), this makes the zero crossing events interrupt the microcontroller.

To fully understand the code, please read the ATmega328 datasheet!

Sensorless BLDC motor control with Arduino video:
The video below shows how project is working.

Microchip AN970 document


      1. In the video there is only 5 capacitors, but the circuit shows 6 capacitors and the arduino uno default timer2 frequency is ~61 khz, but you explained and coded default prescale, can u please provide the full report with the virtual ground resistor design. Please help me i am on my major project..

        1. In my circuit it should be 7 capacitors (the 6 capacitors shown in the schematic diagram and one between (+) and (-) of the 12V source). The biggest capacitor in the video is for 12V source. It could be done just as shown in the schematic diagram.
          The frequency of Timer2 PWM signal is about 31KHz because when it is ON its register TCCR2A = 0x81 which means WGM20 bit = 1 (PWM phase correct mode) and therefore: freq = 16000000/(N*510) = 31372Hz where N is the prescaler = 1 (TCCR2A = 1), you can check it with a real oscilloscope or even by Proteus simulation. All data are provided in the Atmega328P datasheet.

  1. Thank you, and what you used as power source Lipo battery or power supply. And the design of virual ground and back emf

    1. I used 12V power supply for both the Arduino and the 3-phase bridge, a battery (Li-ion, Li-po ..) can be used for powering both of them.

      1. I also used 12v power supply, but when i see the voltage across power supply dropping to 5v or lower when turning on motor and motor vibrates and try to turn but not rotating. Also tested with lipo battery but motor condition is same and i am using STP105N3LL mosfet. Could you please verify the program and also the circuit diagram,

    1. It’s A2212 1000KV, it’s well used in quadcopters, even PC cd-rom (dvd-rom) sensorless BLDC motor (spindle motor) works fine.

    1. It’s the analog comparator interrupt service routine (ISR), the analog comparator is used to detect zero crossing events and at any event the analog comparator interrupts the microcontroller in order to update the commutation state.

      1. Thanks for the quick reply! I understand that this is an interrupt service routine and that is updating commutation state. I just do not understand why “i” is decremented by 1 in “for” loop when ACSR.ACO is “1” or “0” and why loop is counting to 10.

        I suppose that these “if” statements make decisions regarding the rising or falling period of BEMF, because ACSR.ACO is in these cases going from 1 to 0 (rising period of bemf) and from 0 to 1 (falling period of bemf), but can not figure out why “i” is decremented by 1 and why is counting to 10.

  2. Hi! I’m having issues understanding the commutation code:
    1 – ex:
    void BH_AL(){
    PORTB = 0x08; // 0000 1000 >> PB3 >> PIN 11 ON
    PORTD &= ~0x28; // 0010 1000 CLEAR >> PD5, PD3 >> PIN 5 OFF, PIN 3 OFF
    PORTD |= 0x10; // 0001 0000 SET >> PD4 ON >> PIN 4 ON
    TCCR2A = 0; // Turn pin 10 (OC1B) PWM ON (pin 9 & pin 11 OFF)
    TCCR1A = 0x21; // 0010 0001 >> CLEAR OC1B ON COMPARE, PWM PHASE CORRECT 8 BIT >> PB2 >> PIN 10

    Why did we set pin PB3 when we clear it ?

    I am trying to convert this code to run on my hardware, the gate drivers have separate input for High and Low side FETs, I think I shall be able to do that with inverted output PWMs, any comments here ?

    1. See the post below, it’s a little bit different, I used IR2101 instead of IR2104 gate driver. The IR2101 has separated gate control inputs for high side and low side mosfets.

    1. You can try it by your self but doing that may cause some problems to the analog comparator because it’s used for zero crossing detection and the analog comparator share the same multiplexer with the analog-to-digital converter (ADC).

  3. It can work your design with the following engine I have .. N5065 1820W 320KV Outrunner Brushless Motor For Electric Skate Board DIY Kit New

  4. Hello. Nice work thank you! Do you think, is it possible to check that engine is started after start procedure? If engine will not start correctly or its blocked. I tried to check bemf but when engine is fixed, the comparator goes crazy…

    1. As you say, when stationary, there’s no BEMF signal at all, and since there’s no (defined) hysteresis on the comparator it’ll just be triggering on noise. The traditional means of detecting operation is a stall counter – increment a variable every millisecond or so using a timer interrupt, and resetting every time a correctly sequential commutation is detected i.e. rotation rather than oscillation. If there’s no correct sequence in say half a second then disable – or start a restart routine.

  5. Hello, very good job.
    It’s works perfect, i try it in my project and need to make some changes( revers, startup by button, current restart(when its high, and higher speed by make the pwm is higher) and therer some problems…..
    For example- try to make the revers by button- wrote some lines of code to stup the motor and use bldc move in another way- but the motor after perfect work stops after i press the button-and start work crazy……
    Can you help me with code( can and want to simmelar pay for it)
    can you connect with me by e-mail(for example)
    Best regards, Igor

  6. Good job! But if i use a power supply higher than 12v, what resistor should i use for the virtual point and voltage divider? If i dont make any change, can i damage my microcontroller?

    1. Of course you may damage your Arduino microcontroller. Make sure that the input voltage doesn’t exceed 5V (for the ATmega328P)

      1. congratulations for your project, I have seen this same circuit all the same in another project, the only difference is the output uses the N-type mosfet IRF840 and fed with 400 volts, the voltage divider of the BFEM the resistances are of the same value as samples in the circuit, nomas that changes to 5 watts, my question is if there is higher bfem with 400 volts the divider has the same resistance values but at 5 watts, but the working voltage is 400 volts there is no risk of damaging the arduino .
        How to use this circuit that you propose with 400 volts with the IRF840, and resistance values that must be put in the divider, thanks for your contributions

        1. Increasing the voltage without changing voltage divider resistors may damage the Arduino. Project the circuit diagram is for 12V BLDC motor (just use voltage divider equation).
          The mosfets 06N03LA also doesn’t work with 400V, you may find that in its datasheet.

          1. thanks for your comments, if the 06N03LA is for 25v Vds, for the irf840 it is 8 A, 500 Volts would be to calculate the voltage of the divider and the value of the resistors, I need to know the CFEM, thanks again, if you have more data to be able to operate it with irf840 at 400 volts, I thank you in advance

  7. We have 3 capasitors (values are:10microFarad) and we finished the circut.But us second condenser(in the middle ) blowed up 2 times and we don’t know main problem.Us energy source is city network and 12VDC converter.Can you help us ? Thanks…i

    1. a copper strip which passes from the 12-volt Arduino area into
      the motor area and it is felt that there is the potential for a back-EMF voltage spike to be fed back along that copper track and damage the IR2104 chip or even the Arduino board itself. It is suggested that the introduction of a fast Schottky diode in those links would block that possibility.

  8. i am using this ESC circuit to run a compressor, its working fine but the flow rate is not up to the mark so i want to increase the speed of the compressor, to do so should i try a higher voltage supply or can the speed increased by the code itself? if so then please help me with the code to increase the speed of the compressor.

    1. The is no thing to with the code to increase the maximum speed of your motor, the code give the maximum speed when the PWM duty cycle = 255.
      Increasing the voltage may increase the speed (you should use your motor nominal voltage), but be careful, with this circuit diagram you shouldn’t exceed 20V, unless you change the back emf resistors.

  9. Circuit is working but speed is less how can I get a maximum speed is it possible to increase pwm max duty above the 255
    Instead of 06N03LA mosfet I am usingP55NFO6 mosfet is it responsible for low speed

    1. The duty cycle of the PWM signal may be vary from 0 to 100%, so the motor reaches its maximum speed when the duty cycle = 100% which is represented in the code by 255.
      The maximum speed depends on type of motor (max speed, voltage …).
      The P55NF06 should work without any problem.

          1. void AH_BL(){
            PORTB = 0x04;
            PORTD &= ~0x18;
            PORTD |= 0x20;
            TCCR1A = 0; // Turn pin 11 (OC2A) PWM ON (pin 9 & pin 10 OFF)
            TCCR2A = 0x81; //
            void AH_CL(){
            PORTB = 0x02;
            PORTD &= ~0x18;
            PORTD |= 0x20;
            TCCR1A = 0; // Turn pin 11 (OC2A) PWM ON (pin 9 & pin 10 OFF)
            TCCR2A = 0x81; //

            can you please explain the logic behind the above code.

  10. I have used
    1) IRF3205 n-channel MOSFET instead of 06N03LA N-type mosfet
    2) IR2102 gate driver IC instead of IR2104S (IR2104) gate driver IC
    I simluate the schematic in proteus using the above mentioned components.
    I don’t get any error message.
    At the same time I don’t get any outpput.BLDC motor refuse to rotate.
    Is there any problem of using differnet MOSFETs and ICs ?PLease help me to get my output.This is my final year project.
    Thanks in advance…

    1. 1) Although it may work, but the IRF3205 is not a good choice for this application (slow).
      2) You can’t use the IR2102 directly without doing some modifications to the application code (if you want to use it you’ve to make some modifications to the code, read the datasheet of the two devices!).
      I don’t know if you can simulate this application using Proteus, it’s better to do it in real hardware circuit.

      1. I have comopleted this project using the components
        Driver chip – IR2104S
        N-Channel MOSFET – P55NF06
        BLDC motor is as same as your’s

        The problem is the motor doesn’t start automaticaly.
        I switched on the power supply and rotate manually by fingers with full speed settings using speed up pish button.After only my motor rotates.
        There is no prblem with the speed control.But how can I start the motor automatically.What is the problem behind this?
        please help me to find a solution for this.
        Thanks in advance.

  11. would you please reply to my previous question…I am struggling with it for more than a week…
    Please help me to complete my project…..

    1. I think you’ve to make some modifications to motor starting code:

      while(i > 100) {


      just do some experiments!

  12. Sir, while uploading the program,it shows some undeclared identifiers, they are
    – bldc_move
    – AH_BL
    – AH_CL
    – BH_AL
    – BH_CL

    This is a part of my academic project……please help.
    Thanks in advance.

  13. Hello. Great work.
    Question, and request I Got a Samsung 3 phase 36 pol BLCD motor.
    Would like to set the speed to 3000RPM. And display it on 4 digit display( working onit), by counting the Zero crossing. Can U please help me by the Zero Counting with i can work and later display?
    Kindly Regards

    1. This is the project that you mention the samgsun 36 poles engine, the output IRF840 mosfet is used, powered by 400 volts, the whole scheme is the same as the one shown here, also the code is the same, the difference is only in the output of the mosfet that uses IRF840 alimenentado with 400 volts, and the divider of voltage of the CFME the resistance equal value nomas that are of 5 watts, and this is what it was consulting if it will have the problem of greater contrainduccion in the divider and damage the arduino , I’m going to have to perform tests with 12 volts, small motor, change the output mosfet with a higher voltage and larger motor, and change the value of the voltage divider, since the samsung 36 poles engine project is all same as the one that shows simple projects, as I have mentioned the only changes that are the mosfet, higher voltage, and higher watts in the CFME divider resistors, there is someone who wants to comment or suggest something, I do not know what damage the arduino, or the best one,
      Thanks to Simple Projects, for their orientations and illustrative videos

      1. hello alex!

        i am on exactly the same set up as you are talking about…i use a chopper board GTZ 1000 from ali express to supply the power side of the 3 phases boards,360 volt dc,i use also 6 IRF840 for this purpose;…i did severals mistakes by assembling and creating the board that manage those 3 phases motors…i will receive this week 20 more IRF840 because i burned up the first 6 i used this week to try ….

        i started this project by patrick kelly video on you tube,but i undestand now that patrick kelly was inspired probably by this site and this page (here is the link of patrick video talking about our subject ),i am not at the moment,abble to say if it will work because it miss the 6 transistor on my board,but what i can say ,is that patrick kelly did some mistake on the shematic he showed and that you can eventually upload on internet (check the video and go to upload the documentary in link with it and tell me what you think about it)….my motor is a direct drive from a washing machine LG,the goal is to supply it perfectly with the board i made,if you want,lets exchange in private,tell me if you want to have my email adress,i think it could be constructiv for each others to share our experiences and knowledge about it…

        hoppefully reading you soon mate…

        1. Thanks nico
          good that you are also in this idea and project to feed the motor with 36 poles with more voltage, I have the same intensity and I am working on it, if your mail interests me to have a faster communication and comment more
          so we can detail more
          if we manage to operate with more voltage and with these more economic schemes, since I have been reviewing the pwm drivers and the VFD are high cost, we will save
          I wait for your mail and thanks for contributing

        2. hello Nico again and review the schematic information again, and if I corrected it where it indicates that there was the problem in the voltage divider from where the feedback is taken, I change the value of the resistors with those values I think is more correct, and also makes the recommendation to avoid a cfem with the incorporation of the fast Schottky diode, we must be careful I see in corrected pdf scheme that in the illustrations of connections did not make the changes, in the scheme of construction of the plate with the components there if made the change, it was right that the divisor was wrong, so good that corrected, since I was discussing this situation to see if anyone could help, maybe Patrick Kelly out there read me the comment hahaha, well I already corrected
          regards we are in contact

          1. the friend Wunsch Fee comments below that the arrangement is not enough that the program can not keep the mosfet open, we will try it with voltages 100 to 300 …, well one more comment is given that you can not have more revolutions of 2000 rpm for the conditions of the engine, with 1000 to 2000 rpm are enough for my projects
            see you soon

        3. friend Nico can not find chopper board GTZ 1000 from ali express to supply the power, can you provide me the link please

  14. Hello,
    We are trying the same circuit but we are using :
    1) IRF540 MOSFET’s
    Rest the entire circuit is same as yours. But the issue is, the motor runs only once and then, starts vibrating. Even when we press the switches, the sound of vibration is changing but motor doesn’t move. Any comments, what may be wrong?

  15. Its not enough. A shkotty diod need to beplaced in the Back FM way, otherwise it eill damage the Half wave drive, and the Arduido to. .. I working on another circuit feed with Sinus wave form generated by Arduino. The earlier mentiond code dos not hold the Mofet oppen, insted its gives out small peek’s.. +. The waweformis decrease as the Half wave is operats,and changing electric phase.

    1. Wunsch Fee very interesting what you mentioned, and of the circuit that you are building you could share it since you have it, thanks for your comments

  16. Maandish Shing
    Its tray’s to rotat in the wrong direction. Because of the Fix linear wave generation Its just workingin one direction. There sould be a code, if no BEMF detekted in a among of time, contoler should stop eweryting, and restart “Start up” process.

    1. i think if it rotate the wrong direction,just try to reversed 2 phases on 3,because it is what we are doing into buildings when an asynchronous motor does not rotate as we would….let me know what do you think about this idea 😉

  17. Alex.

    I readed the projekt. Ther are some fault in tha circuit. They have “repaired it”, because of the ~380v BEMF and This is the reason why the Capacitor is blowed up as in one of the post mentioned earlier. Still the Wave form is not efficient/sufficient for a High Voltage Unit. This code can’t hold the the Mosfet open. Insted its makes a PMW( many small peek’s) output.. witch is not efficient for this type of motor, es it is a induction motor.. elekto magnetik fiel is indicated.. and the magnet on the rotor is respondin of the EMF( pull push) .

    1. Wunsch Fee, thank you very much for the information, I will review the data, if you have something to help us in that project, some novelty, some circuit that makes the motor samsug of 36 poles to 400 volts work, I thank you for informing us, thanks for share

  18. friend Nico can not find chopper board GTZ 1000 from ali express to supply the power, can you provide me the link please

    1. here is the chopper board, you have to put a bridge rectifier to the output you ll choose in order to get the voltages claimed by the commercial …this board give a maximum voltage of 153 vac 20KHZ,by using a bridge rectifier,as you break the hyper frequency ,the voltage get pretty high….

      -is somebody here abble to explain me this phenomene? the fact of breaking hyper frequency built for high voltage??

      i use this board according to patrick kelly specifications,my direct drive motor is abble to handdle 310VDC,but i have a supposition by extrapolation: even my DD motor is preview for a maximum of 310VDC,am sure it can support 400 vdc,why? simply because the particularity of this set up is the capacity to run the motor with hight voltage/low amps instead of low votage/hight amps,then ,no amps means basiquely no heat,if no heat no problem with the copper vanish of coil,it wont burn out ,could somebody confirm my theory?

      at this time of my project, the board i built looks okay,all stable while supply by the 368 VDC,for the power side,12 volt for the command side,5 volt for the arduino,all is stable,no burning wires,but at the moment IRF840 are not mounted on the circuitry because i burned up the last 6 i had with the first board i built,am suppose to receive 20 IRF 840 more while this week,then i will try again……

      1. by the way i have a question: i have a minus for each voltages (12VDC,5VDC,368VDC), on the shematic of simple project the whole set up is in 12vdc,so of course all minus can be connect all togther,but…..whats about a multi voltages set up as mine baased on patrick kelly? does it be dangerous to loop minus of each voltages together? at the moment they are all separated,and i think it is better….? throught?

        about the shottky diode you guys are talking about in order do not let high voltages damage the microcontroler (in my case a IR2104),yes you have to install a shottky diode,i use a SCS320AHGC9 rated at 650 volts,20 amps and 115 watt,the maximum that patrick kelly propose into the second video he did to mentionned this missing from the first video….it has to be placed after the condensator to the line 6 of the IR2104 or any kind of microcontroler equivalent….as it is placed on the positiv of the high voltage for the power side board,it protected from the hight voltages return the microcontroller…..

      2. thanks Nico for the data, this is inverter at high frequency and must be rectified,
        next comments, if the invesor is 12 volts, and the output 368 vcd, the current estimate 2 amperes, it gives 736 watts at the output, as we have to consume the same in the battery we have, 736/12 = 61 amperes, and if we do not generate more than 736, to be able to return these and the profit that is proposed to generate with the generator, all well theoretically.
        my first test will be to feed it with 120 AC, and then rectify it, you get 169.2 vdc. and assuming that the same is consumed, it should be minus 2 amp * 169.2 = 338.4 watts
        I mention these calculations so I want to avoid the high consumption of the battery, it is true is half power.
        there are more I want to comment and your doubts, my friend, I continue later

  19. some link about patrick kelly,i have to mentionned that i mixed the shematic of simple project and patrick kelly because,patrick kelly power side (the right side of the board drawed) was not exact,by finding out “simple project” i re ajust in the good way the minus side of the high voltage power side….


    SECOND VIDEO OF PATRICK KELLY (correctiv one about shottky diodes)

    guys i have good hope that with several brains to understand this circuitry,we ll get the point fa sure! 😉 😉

  20. THE ORIGIN OF THIS PROJECT IS HERE WITH GERARD MORIN,but be vigilant: gerard set up used 3 hall sensors effects …basiquelly,patrick kelly seemed to upgrade gerarfg concept by proposing a electronic gestion sensor less,here is the main difference between both set up,for those who did not know…. 🙂

  21. THIS IS THE NEXT VIDEO OF THE FIRST I PUT BEFORE: gerard looped up the system for 11 hours…
    ps: please noticed the amazing torque of this direct drive motor into the precedent video….gerard says ,600 watt in for 10KW of mechanical power out,witche for me makes completly sens because this electronic set up allow this kind of motor to work half fully with lentz law:

    -phase 1 + phase 2 ON /phase 3 OFF = when coils of phase 1 ans 2 are pulsed on it become repulsiv for magnets while phase 3 witche is OFF become attractiv for magnets….

    now transpose this first line to the 3 phases and it is what is suppose to do the code….

    1. you have to check the video well what Gerard Morin showed, I do not see anything special, the motor powered by the big battery, connected to the inverter to move the motor at a speed of 98 rpm (low power consumption), the generator that charges the battery small, which feeds the 1200 lumens bulb, these bulbs are low consumption (led).
      Gerard measures cd of the batteries in AC, this is an error, the charger that charges the battery accordingly.
      I do not know exactly where the 160 volt source of the generator is, and I did not see the small battery charger.
      If the battery is charging it may be a long time, but the battery that is connected to the inverse without charging will be discharged.
      the comments of the video is not available hurts, so you can not clarify things.
      With all that I comment if I make a mistake, make the comments, what I say is with all respect to Mr. Morin, but do not be fooled when things are not right.

  22. in others words,when 12 coils that are composing a phase are OFF,the rotor use only the attractive intrinsec power of magnet,this energy is for free,thats why this torque and thats why this tiny consumption in the configuration of high voltage/low amperes…..
    actually this all i guess and i think am not that bad ,right?

    1. contributing with Nico
      If this is how it works, two phase in repulsion and one that takes advantage of the attraction, and so it goes alternating phases, what you have the low consumption with respect to the power you have is; it is the frequency, and the higher voltage supply, there is efficiency, when turning on and off the mosfet.

      the proposed speed of 3600 rpm has not been obtained, the maximum that has been seen in the videos is around 2000 rpm, it is said that the conditions and characteristics of the engine was not built for high speed, which have achieved 2000 rpm , they achieve it with special and expensive drivers.

      but we are still friends, we have to seek and achieve efficiency.

      As for the video that Gerard shows, he used an Anhmeim controller BSCKB1-120081 with a voltage of 160 vcd output. it seems to me that the speed it reaches is 1600 rpm.
      Thomas Buie is improving the program and showing the system project as the pdf file shows.

  23. am even sure we could run even definitly better in terme of consuptions by this :

    -phase 1 ON /phase 2 OFF / phase 3 OFF

    why? because my guessing is “if we have 2 phases OFF then we have 2 times more attraction,so 2 time more a free intrinsec attractiv energy used between hearts coils and rotor magnets ?

    at least instead of only one phase OFF and 2 phases ON,we manage 2 phases OFF and only 1 phase ON……

    what you guys think about this idea? am i completly fucked? haha

    thanks for reading me 😉

  24. To turn on a phase can not, it has to be in pair.
    since the phases are connected in star configuration.

    What they present in the document is in draft is to say what is experiencing Thomas Buie, there is a video where they show their actual operation, but it is good that you are working on it, like the video present here
    In the video Thomas Buie details the challenges and advances he has with the program and his intention to achieve what is shown in the document, he really sees a guy very dedicated to the work, he also says that he needs resources and asks for support and donations from money, if really for these projects it takes time, dedication and efforts, and resources.
    we will follow the advances and contributions
    see you soon

    1. i think you re right about thomas attitude,what i understand about all of that (the approach of thomas meanly),is that at the moment,nobody can properly show a direct drive moteur of washing machine running perfectly by arduino actual code so,i guess there is probably multiples adjustments to realise on the code,to make it efficient enough to run such this kind of motor,right?

      if simple project could interact with us about ours comments,it would be awesome….

      1. ALEX, thomas showed the osciloscope screen and we can clearly see that waves are not well formed,he even tells that “it is not what he expected and something is wrongs with thoses actuals waves” ,i think we should go back to fundamentals in order to crate 2 perfect half waves ,because even thomas seems not abble at the moment,to re created such perfects half waves ?

  25. friend Nico, be very careful when you operate the inverter the high frequency and the voltage can kill you, take your precautions.

  26. Hi Nico
    You mention that you have already assembled the project and that you only wait for the IRF840.
    I consider that the program that presents here simple projects, is already practical.
    1.- Start the project as it is presented here, with the voltages output that proposes 12 volts that are applied to the mosfet. in your case they would be the mosfet IRF840
    2.-Test it with a bruless guy engine as the video shows it. And if everything works, it’s fine.
    3.- Increase the voltage to where you propose Simple projects.
    4.-change the engine and put the 36 coils and feed the output of the mosfet to about 48 to 60 volts.
    5. If the motor moves and it is working, raise the voltage to about 80-100 volts.
    6.- Depending on the results, and if they are good, increase the output voltage of the mosfet to about 169 cd. (rectify the 120 AC, you get the 169 CD.) If you can check the revolutions in each voltage increase
    7.- And see the results, if they are favorable, if you want to increase the voltage
    What I am saying is that if the Simple Projects project is already functional, it has to be coupled to the 36 poles motor.
    The project of Thomas and Gerard of the pdf, as we commented already it was corrected the system of protection and the divider of voltage.
    performing the steps would have a valuation that is obtained in each voltage increase, it is a matter of proving it.

  27. Here my email address to wunschfee[at]
    For me the mentioned ciruit above did not worked. With the IRF840 mofet. I also tried a faster, IRF530 100v mosfet.
    I used a 0.8A 90volt power suply. For testing my gol is a 650v/1100v motor driver. Becouse on higher voltage takes les amper.
    Gerald Moris usesa external Motor river cirkuit witch Hi has opend, and its a Frequency generator Motor drive, not a PWM drive. The. The built in 4 cabel twinn Tacho signal generator is working fine in my setup.
    I cust recived a NAND chips to trie out.

  28. wunschfee thanks for sharing
    but the link is no longer available
    can give more details if it worked in some way with 100 volts, and show the video
    see you soon

  29. I’m whaiting for my 4CH oscilloscope & function/Arb. waveform generator. I mad successful contakt with Nico. Till the Oscilloscope not arrives, i can’t chek the malfunction of my cirkuit.

  30. Very nice guide. From what I see, only one PWM should be required, since all three PWM Generators are outputing the same frequency and same duty cycle, so only one PWM pin could simplify both hardware and software.(Correct me when I am wrong)

    It would be interesting to see the specific operation to not only control the speed, but also control it’s direction. Good project anyways!

  31. i run this code with few modifications on a Samsung BLDC drum like gerard morin did in his videos. as i did more research on the circuit i noticed that the duty-cycle went down automaticaly to the min-duty-cycle that i set in the code. the motor starts automaticaly with a duty-cycle of 240 when the voltage is set to 90V or more. then when i increase the voltage to 200V the drum automatically ‘corrects’ its given duty-cycle and decreases it. can you give me a hint why this is happening? i did this testing only on the big BLDC drum and have not tried another motor. if you read this and like to comment i would be very happy 🙂 thx and big regards from austria.

    1. The duty cycle is set by the Arduino microcontroller and it’s not related to motor supply voltage.
      Since the resolution of the PWM signal is 8-bit, the duty cycle may vary between 0 and 255. If duty cycle is zero the motor will stop. Using a min duty cycle prevents the motor from stopping, you can test which minimum duty is adequate for your application.
      Also, the duty cycle is controlled from the 2 push buttons, after your motor starts try decreasing the duty cycle!

      1. it behaves exactly as you mentioned and can be clearly seen in the serial monitor if we do the necessary addition to the code. what we are also trying is to integrate the PWM library so that we can use more frequencies to play with. but it seems to be difficult from my point of view.

        1. There is no need for the PWM library! you can change the frequency with Timer1 and Timer2 registers (TCCR1A, TCCR1B, TCCR2A and TCCR2B), ATmega328P datasheet shows how to do that.

          1. yes but there only a few options. i would like to run the drum with for example 120kHz. if i use 0x01 for TCCR1A the frequenzy is preset to 31kHz. don’t understand how to achieve this.

  32. Hey i tried this circuit with the IR2101 and 2104, the code and everything works well. I’m able to run an old alternator as bldc even. Now i want to scale it up a bit and use TO-220’s. could you recommend some fets for a 24v system? i was looking for mosfets with the same gate charge, but still need a recommendation, cheers from AUS

  33. If i’m only using the two buttons to control the speed, is there any advantage of using the lm339 over the internal comparators? And should I tie the rest of the analog in pins to gnd for noise immunity? cheers

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.