– Clip 1 Introduction – Clip 2 Course Description – Clip 3 Test Parameters – Clip 4 Test Equipment – Clip 5 Overview – Clip 6 Sine Wave – Clip 7 Sine Wave to Square Wave – Clip 8 Clipped sine Wave – Clip 9 Square Wave without driving power – Clip 10 First run – Clip 11 First performance overview – Clip 12 Square wave falling Edge – Clip 13 Recording with Laptop – Clip 14 First power analysis – Clip 15 First performance result – Clip 16 Current Requirement – Clip 17 Benchmark Result – Clip 18 In circuit current – Clip 19 Driver Current Requirement – Clip 20 Chart Overview – Clip 21 Second Benchmark – Clip 22 Increased Driver Voltage – Clip 23 Increased current Requirement – Clip 24 Further increased driver Voltage – Clip 25 IRFP450 chart – Clip 26 Calculation 1 – Clip 27 Conclusion – Clip 28 End of Part 3
Clip 1
Clip 1
This is the third part of the Back EMF charging of a battery based on a power MOSFET, IRFP450. It has 500 Volt 14 A, 190 Watt power rating. It is ideal for power switching applications up to 1 MHz. I will explain the implications to use a power MOSFET in an analogue circuit and how we can compensate for that. At the end a conclusion is provided
Clip 2
Clip2
Back EMF charging via IRFP450 Part 3
– Analyze the performance of the power MOSFET IRFP450
– Investigate the signal conditions to drive a power MOSFET
– Measure the charging rate and compare with the 60 Watt charger
Clip 3
Clip 3
Test Parameters
1. Voltage source 1 is a Benchtop Power Supply
2. Voltage source 2 are two batteries with 12 Volt and 8 AH
3. Bedini SSG circuit with IRFP450
4, 2 Bifilar solenoids to drive the wheel
5 Motor wheel made of Nylon 6 with 3 kg or 6 pound rotor
6. Ferrite (Ceramic) magnets replacement in the wheel
Clip 4
Clip 4
Test Equipment Part 3
Tools
– Keithley 2110 Bench Digital Multimeter 5.5 Handheld DMM’s
– Agilent InfiniiVision 7000 Series oscilloscope MSO7014B
– Agilent InfiniVision 3000 Series oscilloscope DSO3014A
– Tenma DMM
– UNI-T current clamp
– HP Laptop
– Bench Top Power Supply 20 Volt 5 A
– Bench Top Power supply 120 volt 1 A
Clip 5
Clip 5
I will use in this video as a Bedini SG circuit the power MOSFET IRFP450. I try to make it work. I will show what it means to use this chip. I demonstrate the implication to use it and what is required for as signal. How we can modify the signal and what we have to look out for.
Clip 6
Clip 6
The first thing you need to know is that a power MOSFET can not be driven by a sine wave. It also has a gate which is driven via capacitive coupling. The trigger coil is producing a sine wave but not a clean one, it is based on the fast moving rotor over the solenoids. This is more a pulse and therefore much smaller than a standard sine wave. We look here for the pulse width which does not apply for a sine wave. The MOSFET need to be triggered via a sharp signal with fast rising and falling edges as in a square wave. The analogue signal would permanently switch on the MOSFET and damage it. Operation is in this way not possible. The solution is to convert the sine wave into a square wave. That is one of the jobs of a MOSFET driver. I use mainly the TC4421/22/23/24 series. It converts the input signal into a digital signal and outputs based on the input voltage up to 15 Volt to the gate. The sine wave on the scope will now be converted into a digital signal.
Clip 7
Clip 7
In my first test I use my commonly used TC4421/22 driver. I have the driver on the board already included. I will explain the driver circuit in a minute because I have configured it into a special way, which I will explain a little bit later. We come into the driver via the sine wave, which we simulate as the trigger coil. I have the power supply for the driver connected but I do not need to start it. There is a residue voltage already existing of around 0.5 volt and this is enough to start the driver. Let me show you what happens when I activate the signal. You can see that the sine wave is now converted into a square wave of the same frequency. What happens is that we have on the sine wave also rising and falling edges or curves. The driver has a threshold which is based on the supplied voltage. That means that a large part of the sine wave is ignored. Only when the threshold level is reached the schmitt trigger of the driver is setting the signal high. Once the wave is going past 90 degree, reaching the switch off threshold the square wave stops and is setting the signal low. Here only a small part of the sine wave is contributing to the signal width or pulse width. This also depends on the input voltage. Obviously it does set the threshold higher, the higher the voltage. We end up not achieving a 50% pulse width or duty cycle. I can tell you the exact value which is 39.7% duty cycle. The sharper the sine wave the shorter the pulse width or duty cycle. The implication now is that the signal from the trigger coil does not give us the full bandwidth of the signal to the power MOSFET and therefore will not be efficient to drive the rotor.
Clip 8
Clip 8
I will show you now what I have done as modification on the input side of the driver. The front bit or the left side contains the driver. The signal comes in here. The input side of the driver is very sensitive. The trigger coil produces high voltage transients, which the driver cannot handle. Yes, it has a zener diode on the input but it cannot handle large currents. I have therefore added another zener diode , plus a resistor to ground. On the oscilloscope you see that the bottom of the sine wave is cut off. You might think it has to be the opposite way around. The spike appears on the opposite side of the coil. Not on the plus side but on the minus. You have a couple of options to reduce the voltage. One way would be to use a voltage divider. That is elegant and will not cost you much current. The zener is consuming current but I am not concerned in my example because it comes from the trigger coils which has enough power. I have now started the power supply for the driver. Remember we had our square wave inside the sine wave. Let us see now what happens to our signal. If I measure now the signal from the driver to the gate I get a signal like that. Our pulse width is now much smaller. My first guess is, it is around the 25% mark. That would be a very short pulse to drive the power MOSFET. I activate the second probe and you can read now the pulse width. I am correct we have a duty cycle of 25.6%. The frequency here is for both signals identical with 500Hz. We have only a fraction of the energy available to drive the rotor. Under this condition I have not much room to drive the motor efficient. I have to make a lot of tweaks to make that working in an analogue circuit.
Clip 9
Clip 9
I want to show you briefly what happens when I switch off the power supply to the driver. The voltage falls and the pulse width increases back to the 39% we had before. The low signal would not be high enough to drive the MOSFET.
Clip 10
Clip 10
I have now the MOSFET connected in the system and running. I use here my secondary power supply for tuning purpose. I feed the driver with 5.1 Volt and 12 Volt from the secondary power supply. On the scope you can see the signal to the gate does not exceed more than 15 Volt but it is the amplitude and we have to divided it by two. I use now voice over because of the very high noise of the motor. It is in fact much more silent as before but I move the camera to the motor and it picks up the sound to a much higher value than I did anticipate. On the circuit you see a blinking component. This is a GDT (Gas Discharge Tube) or voltage surge arrester of 400 Volt. The voltage reaches the level above and it flashes. This is a controlled spark gap if you like, without the usual fallout. We see on the left side all the rheostats . I use two 200 Ohm with sliders, I use in full value and the rotational rheostat of 500 Ohm. What I wanted to demonstrate is that the higher the resistance the higher the brightness of the spark and the more power it has but it slows down the wheel until it stops. That is the down side to deal with in this application. But we see the potential and can utilize that later. We have a voltage reading of 550 volt which does reach over 600 Volt as top limit. The connection to the battery is not made here.
Clip 11
Clip 11
I have the battery now connected and removed the secondary power supply. Now the performance is completely different. I can almost switch off the power to the driver but I can not use more than 5 Volt input power, otherwise the wheel will not run anymore because the pulse width will become too short. For my test I use only 4.8 Volt. The charging performance is very high. Now with the battery connected I still have 140-160 Volt left. We remember that it was not above 10 Volt before, when we tested the transistors. Note, also the voltage on my driver input. It does not exceed 4.7 Volt. My Zener is a 3V5, 100 mA I use. I will now connect a current probe and read the values.
Clip 12
Clip 12
The driving battery has at the moment 12.187 Volt fluctuating around this value. We have charged the battery now to 12.279 Volt and rising. I have not recorded that yet but will do that a little bit later. On the scope you can see now current and power readings. We have 3.5 A and between 65-84 Watt, over the 120 Volt. That is calculated on an amplitude and not as RMS. RMS is not relevant here because we achieve the charge via a pulse and not via continuous power. That would be far too low to do that. What is nice to see is that the square wave produces after the switch off the very sharp transient. That is how it works. A transistor is not able to do that because the sine wave is continuous achieving only low values. As Tesla said you need high and sharp, short pulses to excite the field. The battery charge reads 12.29 Volt and we around the 10 minute charging time mark for 200 mA
Clip 13
Clip 13
I have increased now the resistance to the driver. That has an performance gain. We see now 4A and power around the 90-100 Watt level. The RPM is reduced slightly. We charge now faster with 12.340 Volt on the charging battery. Achieved in 2 more minutes. However to keep it running is a high balance act. The MOSFET is very sensitive. Ideally we would need to adjust the trigger signal either delayed or advanced switching. The transistor is much easier to handle. The MOSFET requires fine tuning. The manual turn of the wheel has to be high in order for the gate to switch on the MOSFET.
Clip 14
Clip 14
I have now everything connected and prepared for the test. The driving battery is fully charged and reads 13. 586 Volt. From the previous test we have and residue charge left on the battery of 12.147 Volt. That is now my base voltage for the charging measurement and time. I will record now all details and put it at the end on the board.
Clip 15
Clip 15
The motor is now running and I record the performance. I measure close to 4 A and 160 Volt, calculated around 83 Watt. The wheel frequency or RPM is at 110 -115 Hz at the moment and increasing. I have increased the resistance to the driver to close to 1 KOhm. The higher the resistance the higher the power gain, visible on the scope as the white line. We have charged to 12.272 volt at moment, fast increasing. The voltage on the driving battery is increasing slowly as well.
Clip 16
Clip 16
I take now a current measurement and read 0.656 A. I will take this value together with the voltage of 12.328 Volt of the driving battery as base value on the board.
Clip 17
Clip 17
The RPM of the wheel is marked on the sope with 120 Hz. This is the frequency I will use for the calculation at 12.333 Volt.
Clip 18
Clip 18
I measure now the current which circulates within the circuit as only 0.395A. Some of the energy is lost as heat. The MOSFET is getting very hot.
Clip 19
Clip 19
As additional value I measure the current requirement of the driver. My current clamp is not very precise but I get a reading of 1 MA. Correction to what I said in the video. That is a value of 4.8 mW
Clip 20
Clip 20
This is a first glance of the performance chart. I will ignore the beginning of the recording and select a section in the middle. This is a more conservative view but overall more realistic.
Clip 22
Clip 22
I have increased the driver voltage now to 4.9 Volt and see a dramatic change in performance. Remember I did show that the increased driver voltage will lead to a shorter pulse width. We achieve higher current flow from the MOSFET and sharper pulses to the coils. The downside is that the shorter duty cycle is in the middle of the trigger coil signal and not at the beginning. It therefore is dragging now the RPM of the wheel down. On the scope you see now twice as much current and power. I stay corrected, I did not mean the comparison to a carburetor but to a spark distributor, where you adjust the degree of delay for pre- or late ignition. The battery charging rate is now very high. I will show that on the chart later on but will not take that into consideration for the calculation. The driving battery is losing much faster energy.
Clip 23
Clip 23
I like to show you what impact 100 mV on the driver has. I read 1.306 A. That is as well twice as much as we had before.
Clip 24
Clip 24
To take that a step further I increased an additional 100 mV on the driver. We read 5 Volt now and measure on the scope a power of over 200 Watt. The driving battery is losing now very fast energy it is going below 12 Volt. The RPM is going down and when increased further the wheel will stop. There is no power anymore in the coils at the right time to repel the magnets of the rotor. I will take that to 12.6 Volt and stop my calculation for our performance review.
Clip 25
Clip 25
The chart is showing now the measurements. I take the conservative value for the board but have for comparison also marked the fast charging rat to the 12.6 Volt. I charge at the high power level 150 mV in 15 Minutes. That is close to a 250% improvement. Note that has to increase even higher in order to fully charge the battery.
Clip 26
Clip 26
We integrate our measurements into the formula
V = 12.33 Volt Driving battery
I = 0.656 A
P= V * I = 12.33 Volt * 0.656 A = 8.1 WattRPM = HZ/8*60 = 120/8*60 = 900
100 mV charge in 26 Minutes * 15 = 390 Minutes or 6.5 Hours
Power requirement = 8.1 Watt * 6.5 h = 53 WattFrom 60 Watt battery charger, charged to 13.5 Volt in 39 Minutes and 12 Seconds
Power consumption is 39.12 WattThe charge performance is only 74% of the standard charger
Clip 27
Clip 27
First Conclusion
The IRFP450 power MOSFET does have the potential to outshine the
competition but is not able to do so based on the restriction to digital
signals. The conversion of the audio signal from the trigger coil is
challenging and would require additional circuitry to optimize that. The
performance is therefore low and not recommended to use for the Bedini
SG circuit. Where it shines is in the use of a motor drive controller where
precision is required. That is a stepper motor as used for a CNC router,
3D printer or inkjet printer. The Bedini SG circuit does at best create an
initial charge but fails to charge a battery full. The wheel is only a controller
or timing device without additional benefit.
Clip 28
Clip 28
End of Part 3
Power Benchmark Test
Harnessing Back EMF