Eliminate Huge Spark when Connect ESC to battery

[LittleGiant]

Thanks

OK...thanks for the tips on using flyback diode protector. Will try to add it next time I visit Sim Lim Tower because I do not have any diode component at home


EDIT: The IRFZ44N is called "HEXFET". I think it already has a schottky diode build into it. I can't confirm yet. Need to dig up more info.

SH

Ehh you are using mosfets only on the positive wire right?

[Super-Hornet]

I assume u asking me where I connect the MOSFET right?
On the -ve side or the low side. This is because of N-Channel. In other word, battery +ve goes to ESC, ESC -ve goes to MOSFET Drain and MOSFET Source goes to ground.

SH

[LittleGiant]

I assume u asking me where I connect the MOSFET right?
On the -ve side or the low side. This is because of N-Channel. In other word, battery +ve goes to ESC, ESC -ve goes to MOSFET Drain and MOSFET Source goes to ground.

SH

I tot you were using pmos. Yes nmos is better at draining and usually nmos have much higher rated currents. Another question is why are you using a microprocessor when u can make a delay switch using a resistor and a capacitor. Future expansion for more functions?

[Super-Hornet]

I wanted to use P-channel type due to I want to have common ground...but can't find a good high current P-Channel in Sim Lim Tower. Therefore, boh pien but to use N-Channel. I was thinking that, my ESC already has 5Amp BEC and therefore, no need another BEC. Or, i will use another battery for the external BEC...thus, it does not matter if I use N-Channel (no common ground) instead. Thus, I stick to N-Channel.

I use microcontroller due to I want it to be fanciful or sophisticated and also future functions..as what u suspect.

Yes, I can just use a high resistor with capacitor as a delay also. Just change the top module instead.

The ideal of this design is consist of 3 parts:
1. Driver/Solid State switch
2. Controller
3. Power for the Controller.

To have high current, u design the driver with high current. To have different delay or method of controlling the Driver, u change the controller. The controller required power..thus the power supply module.

The Controller can be either microcontroller, resistor+capacitor or even a tiny switch.
The power supply module can be either a LM7805 or LM317 or even something higher (like, maybe for Adrianli's boat where it required 50V. LM7805 cannot handle 50V)

SH

[Super-Hornet]

If u do not want to have fanciful microcontroller controller, U can put all "Driver" and "Controller" (Resistor + Capacitor) in one board. The size can reduce even further.

The power supply model is coming from the battery directly.

SH

[LittleGiant]

I wanted to use P-channel type due to I want to have common ground...but can't find a good high current P-Channel in Sim Lim Tower. Therefore, boh pien but to use N-Channel. I was thinking that, my ESC already has 5Amp BEC and therefore, no need another BEC. Or, i will use another battery for the external BEC...thus, it does not matter if I use N-Channel (no common ground) instead. Thus, I stick to N-Channel.

I use microcontroller due to I want it to be fanciful or sophisticated and also future functions..as what u suspect.

Yes, I can just use a high resistor with capacitor as a delay also. Just change the top module instead.

The ideal of this design is consist of 3 parts:
1. Driver/Solid State switch
2. Controller
3. Power for the Controller.

To have high current, u design the driver with high current. To have different delay or method of controlling the Driver, u change the controller. The controller required power..thus the power supply module.

The Controller can be either microcontroller, resistor+capacitor or even a tiny switch.
The power supply module can be either a LM7805 or LM317 or even something higher (like, maybe for Adrianli's boat where it required 50V. LM7805 cannot handle 50V)

SH

Since you are using nmos on the -ve terminal, you don't need diodes anymore. Good smart design to have it modular. Will be waiting for more updates.

[Super-Hornet]

OK.. here is the proposed design for using resistor+capacitor to create delay for the controller.

{ Note that the capacitor mentioned in this post is about the capacitor to create delay...not the ESC capacitor }

Initially thinking of using resistor and capacitor only (schematic #1, resistor charge up capacitor and the capacitor will trigger MOSFET) but notice that there is a disadvantage of it. When battery is disconnected, the capacitor take time to discharge. Because of that, if you plug in battery, after a while, ESC turn on. Then u unplug battery and plug back again (without waiting for capacitor to discharge), the MOSFET is still turn ON due to capacitor still has some charge, spark will appeared. That mean, when u unplug battery, u have to wait for capacitor to discharge 1st.

This does not happen when using microcontroller because I can program the microcontroller to initialized the control pin to be logic "0" (low) 1st.

Because of the slow discharge of capacitor, a method to discharge the capacitor fast (when it sense battery disconnected) is required.

Therefore, the 2nd schematic show a revised method. The diode is to prevent voltage from the capacitor to flow back to the common +ve and the P-Channel MOSFET is use to short the capacitor +ve and -ve pin together when the Vgs sense low voltage (or 0volt). When battery is connected, the MOSFET will not turn on because the Vgs is in high voltage...thus the capacitor can charge up. When battery is remove, the P-Channel MOSFET gate is tie to ground and therefore, the MOSFET is turn on and this allow the capacitor to discharge in faster time

Oh... The P-Channel MOSFET of IRFD9120 is available in Sim Lim Tower. It is a 4 pin DIP packing type.

Manage to find a some "Char Par Lang" diode at home and test it on a breadboard just now. The capacitor do get discharge when battery is removed. Next is to develop the controller board.

SH

[LittleGiant]

OK.. here is the proposed design for using resistor+capacitor to create delay for the controller.

{ Note that the capacitor mentioned in this post is about the capacitor to create delay...not the ESC capacitor }

Initially thinking of using resistor and capacitor only (schematic #1, resistor charge up capacitor and the capacitor will trigger MOSFET) but notice that there is a disadvantage of it. When battery is disconnected, the capacitor take time to discharge. Because of that, if you plug in battery, after a while, ESC turn on. Then u unplug battery and plug back again (without waiting for capacitor to discharge), the MOSFET is still turn ON due to capacitor still has some charge, spark will appeared. That mean, when u unplug battery, u have to wait for capacitor to discharge 1st.

This does not happen when using microcontroller because I can program the microcontroller to initialized the control pin to be logic "0" (low) 1st.

Because of the slow discharge of capacitor, a method to discharge the capacitor fast (when it sense battery disconnected) is required.

Therefore, the 2nd schematic show a revised method. The diode is to prevent voltage from the capacitor to flow back to the common +ve and the P-Channel MOSFET is use to short the capacitor +ve and -ve pin together when the Vgs sense low voltage (or 0volt). When battery is connected, the MOSFET will not turn on because the Vgs is in high voltage...thus the capacitor can charge up. When battery is remove, the P-Channel MOSFET gate is tie to ground and therefore, the MOSFET is turn on and this allow the capacitor to discharge in faster time

Oh... The P-Channel MOSFET of IRFD9120 is available in Sim Lim Tower. It is a 4 pin DIP packing type.

Manage to find a some "Char Par Lang" diode at home and test it on a breadboard just now. The capacitor do get discharge when battery is removed. Next is to develop the controller board.

SH

You are driving the pmos gate directly using battery source. You don't need a resistor? Not sure about this but i think a resistor a better protect the gate from emf.

[Super-Hornet]

Well, initially I did put a resistor between +ve and the PMOS's gate...but after doing testing on breadboard, it can work with or without the resistor.

But u are right about putting the resistor there...but for another reason that I just think about. Without the resistor (or direct connect o +ve), the gate will register whatever voltage the battery is. If lets say the battery is 40V or higher, the gate may have problem.

SH

[LittleGiant]

Well, initially I did put a resistor between +ve and the PMOS's gate...but after doing testing on breadboard, it can work with or without the resistor.

But u are right about putting the resistor there...but for another reason that I just think about. Without the resistor (or direct connect o +ve), the gate will register whatever voltage the battery is. If lets say the battery is 40V or higher, the gate may have problem.

SH

How's the latest version? Got pics? Testing and results?

[Super-Hornet]

Oh. Sorry.. was busy trying other things (Navigations light controller and Boost Converter).

Oh! I notice that if I use 5V to trigger the IRFZ44N (that is, I using PIC with power supply from LM7805), the resistance between Drain Source is slightly higher (and thus hotter) than if I use 6V.

With that, I modify my Spark Eliminator on my bo-105 and my OH-6D (TRex450 with 6S1P) using LM7806 regulator instead. Now, the heat sink of the MOSFET is much cooling.

I haven't solder the resistors+capacitor+PMOS+Diode into a board to try but it should be working based on the breadboard testing.

By comparison between:
1. resistors+capacitor+PMOS+Diode
2. PIC + LM7806
3. PIC + bridge resistor (to divide the i/p voltage to 6V for the PIC)

The #2 and #3 is the best in terms of component required. The #3 method, due to bridge resistor using 1/4 watt resistors, it is not recommended to solder any LED indicator.

I try to take some pictures of the current Spark Eliminator I'm using in my bo-105 and OH-6D. So far it is working perfectly.

SH

[LittleGiant]

Oh. Sorry.. was busy trying other things (Navigations light controller and Boost Converter).

Oh! I notice that if I use 5V to trigger the IRFZ44N (that is, I using PIC with power supply from LM7805), the resistance between Drain Source is slightly higher (and thus hotter) than if I use 6V.

With that, I modify my Spark Eliminator on my bo-105 and my OH-6D (TRex450 with 6S1P) using LM7806 regulator instead. Now, the heat sink of the MOSFET is much cooling.

I haven't solder the resistors+capacitor+PMOS+Diode into a board to try but it should be working based on the breadboard testing.

By comparison between:
1. resistors+capacitor+PMOS+Diode
2. PIC + LM7806
3. PIC + bridge resistor (to divide the i/p voltage to 6V for the PIC)

The #2 and #3 is the best in terms of component required. The #3 method, due to bridge resistor using 1/4 watt resistors, it is not recommended to solder any LED indicator.

I try to take some pictures of the current Spark Eliminator I'm using in my bo-105 and OH-6D. So far it is working perfectly.

SH

For nmos higher gate voltage means lower channel resistance. Check the specsheet. There should be a recommended value. If you are using a resistor for the gate then just reduce the resistance value.

[Super-Hornet]

No resistor. Direct drive. Unless I read wrongly on the spec, the Vgs voltage is ranging from 4.5V to 15V... So, 5V should be OK. But If I not wrong, the different is like in 0.00x ohm differences.

Even though it is slight 0.00x Ohms different, it will help reduce the waste wattage to heat.

SH

[khatiban]

No resistor. Direct drive. Unless I read wrongly on the spec, the Vgs voltage is ranging from 4.5V to 15V... So, 5V should be OK. But If I not wrong, the different is like in 0.00x ohm differences.

Even though it is slight 0.00x Ohms different, it will help reduce the waste wattage to heat.

SH

Vgs of 5V and 6V is actually quite a big difference
(fig.1 on pg3 of datasheet). see the difference in Vds at
Vgs=5V, Id=50A
Vgs=6V, Id=50A
I think its quite hard to reach the maximum performance if Vgs is 5V
by the way this is the datasheet I referred to, let me know if I referred to the wrong MOSFET:

http://www.irf.com/product-info/datasheets/data/irfz44n.pdf

[Super-Hornet]

Yup, Same MOSFET. Yes, I was looking at the Fig 1 graph. That is why I change from 5V to 6V. I can't use anything higher than 6V because it might damage the PIC if I supply more than 6V. I could use something to have different voltage but that will increase the component counts.

Thus, for PIC, using 6V is the best. Using Resistor+capacitor+Diode+PMOS, u can change the voltage divider and capacitor to have voltage higher than 6V.

SH