How to design an electronic switch using a MOSFET?

[faeraki]

Hi!

I am kind of new in ''practical electronics'' and I have to design an electronic switch using a MOSFET to cause ignitions(few times per second..).
I guess that the main idea is to connect an NPN MOSFET between the power supply and the primary winding of the ignition coil.(I am going to connect a pulse generator to the gate and I can set the frequency there.)
After a small research I saw that I have to use in parallel with the MOSFET a diode and a capacitor..I don't really get why but I guess that it has to do with protecting the the transistor...
Considering the fact that I don't want to melt anything or blow up the MOSFET is there anything else that I have to do?
And how could I choose the type of the MOSFET that I need?

Any help is appreciated!:-D

 

[Tunelabguy]

re: how to design an electronic switch using a MOSFET?

The protection diode does not go in parallel with the MOSFET. It goes in parallel with the coil. But even then it might not work.

First, as for why you need a protection diode, consider that when current flows through a coil, it gathers what can be thought of as momentum, to use a water analogy. When you try to stop the current flow all at once, that momentum is going to try to overwhelm your switch and keep some current flowing for a little while. The consequence of this is that the voltage on the MOSFET will rise tremendously, probably the point of destroying the MOSFET. The proection diode going back to the power supply gives this current somewhere to go without an undo increase in damaging voltage on the switch. The current flows for a very short while and then decays to zero and stops.

But here is the thing about ignition circuits. They only work if the primary coil is interrupted suddenly. If you let the primary current decay gradually through a diode, you will not get a very high secondary voltage, which is the whole point of an ignition circuit. So you do need to interrupt the primary coil current fairly suddenly, but not too suddenly. The solution is to get a MOSFET with drain/source voltage spec that is as high as you can afford. Let's say you get one that can withstand 50 volts. And let's say the power supply is 12 volts. When the primary coil is turned off instantly, the voltage on the switch that interrupted the current (whether it is a mechanical switch or a MOSFET) will rise to a zillion volts (so to speak), if it has nowhere to go. If you use a simple diode around the primary of the coil, this current will be able to flow back up into the 12 volt power supply. The forward drop of the diode, being about 0.7 volts, means the voltage on the switch rises to only 12.7 volts. No problem for the switch, but it is a big problem for an ignition circuit. By opposing the primary current flow with only 0.7 volts, the current will decay slowly and the secondary voltage pulse will be not very high. But what if you put a 35-volt zener diode in series with the protection diode? Then the protection current would still flow when the coil is turned off, but that protection current would be opposed by 35.7 volts (35 volts from the Zener and 0.7 volts from the regular diode). This will cause the current in the primary to decay about 50 times as fast as it would have with only a regular diode, and so the secondary voltage pulse will be 50 times higher (great for making sparks!). And what about our switch? Well, it will see a voltage of 12 + 35.7 = 47.7 volts, which is less than its limit of 50 volts. So the switch will not be destroyed.

So to make an ignition primary switch you have to use a fairly high-voltage capable MOSFET and then use a protection diode and Zener diode that lets the voltage rise to almost the limit of that switch. That way you can get a nice spark from the secondary and still protect the switch on the primary.

By the way, your description of where to connect an N-channel (they are not called NPN) MOSFET is a little off. The N-channel MOSFET is for interrupting the circuit to ground. The positive power supply goes to one end of the coil, and the other end of the coil goes to the drain of the MOSFET. Then the gate can be driven as needed by a signal referenced to ground. If you must insert your switch between the power supply and the coil, where the other side of the coil is permanently grounded, then you would use a P-channel MOSFET, and the gate would have to be driven by a signal that is referenced to the positive power supply.

 

[faeraki]

re: how to design an electronic switch using a MOSFET?

Thank you very much for your help Tunelabguy !
You really made things clear!

But what about these cases that some are using capacitors between the drain and the source?

 

[Tunelabguy]

re: how to design an electronic switch using a MOSFET?

A capacitor between the drain and the source does sort of the same thing. It provides the current somewhere to go so the voltage does not rise too high on the switch. But there are some risks in relying on a capacitor. If the capacitor is too small, the voltage can still rise above the point where damage occurs. But the biggest difference is in how a capacitor works over time. A capacitor gains charge gradually as the current from the coil continues to flow. The current in the coil is not opposed by much voltage at first. Then as the capacitor gets more charged up it gradually opposes the current flow more and more. All this time spent opposing the coil current just a little bit is not producing any high voltage output in the secondary, but it is reducing the total amount of energy in the magnetic field. So when the field finally does collapse, it has less energy to put into the secondary, and the resultant voltage is not as high as it could be. By opposing the current flow mostly with a Zener diode, you start opposing the current flow as much as you can as soon as you can, before much energy can drain away from the magnetic field. This will give you a bigger spark. Now I can see a place for a capacitor. A very small capacitor can limit the voltage during the short time it takes for a Zener diode to avalanche. Diodes do not begin conducting instantly. In the very brief time before the diode begins conducting, if the MOSFET turns off too quickly, the voltage might still rise to the point of doing damage to the MOSFET. A small capacitor between the drain and source prevents that from happening. But that capacitor should only be big enough to cover the very short period of time before the protection diodes begin conducting. From that point on the protection diodes take over and the small capacitor has no more influence.

 

[dick_freebird]

re: how to design an electronic switch using a MOSFET?

Your secondary voltage will be limited in one case by the
turns ratio and the breakdown of the switch FET. The FET
will get beat up if the plug cap fails to break over and you
have a lower breakdown path than the back-referred air
gap (which is higher breakdown in operation than in free
air testing).

A snubber network can relieve the voltage stress peak
on the FET. You do not want to return primary current
to the supply as a catch-diode would, this only will steal
from delivered spark energy. You might want to look for
FETs with a very good repetitive avalanche rating and
as high a breakdown as possible, so as to tolerate more
unfortunate (mis)application.

 

[chuckey]

re: how to design an electronic switch using a MOSFET?

The standard coil is designed for a "Kettering" ( contact breaker type) coil, these rely on a capacitor across the points, so when the points open the voltage pulse is not a thin very high voltage one but part of a LC tuned circuit, which is ringing say ~100KHZ, so as to provide a "fatter" - more current in the spark. Standard cars will not run without the capacitor, which I believe is .2 MF. So lets get this cleared up, do we want the ultimate voltage spike or do we want the 10 microsec Kettering substitute?
Frank

 

[faeraki]

re: how to design an electronic switch using a MOSFET?

@Tunelabguy: I am concern about 47.7 volts. Not sure if this voltage is enough for a nice arc..
@dick_freebirdo you think that a snubber RC circuit between the drain and the source would be ok? I am too confused with the protection of the mosfet. I have seen RC snubber circuit, diode between drain and source, capacitor between the drain and the source, diode in parallel to a capacitor and to the mosfet.. and many other snubber circuits..
@chuckey: Regardless the ignitions in cars, I want to create electric arcs for a project in the lab. My issue is that I wanna use an electronic switch to have 100 arcs per second.

 

[Tunelabguy]

Re: how to design an electronic switch using a MOSFET?

@Tunelabguy: I am concern about 47.7 volts. Not sure if this voltage is enough for a nice arc..

That was just for illustration, based on a MOSFET that could only handle 50 volts. If you pick a MOSFET with a greater drain voltage spec, then you can raise the protection voltage accordingly. But don't actually go within 3 volts of the maximum, like I did in the example. You want to have some margin in case circuit components produce a slightly larger drain voltage. In any case, this is not the voltage that produces the arc. This is only the voltage on the MOSFET when the primary is turned off. The voltage that causes the arc is produced by the secondary of the ignition coil It has many more windings, and by that ratio of turns in the primary to turns in the secondary, it produces a voltage that is proportionally higher.

Regardless the ignitions in cars, I want to create electric arcs for a project in the lab. My issue is that I wanna use an electronic switch to have 100 arcs per second.

But you still have to answer chuckey's question of whether you want the arcs to be spread a little over time. As chuckey said, they do that in cars because a very short spark is not as effective in igniting the fuel as a slightly longer spark. Although, the fact that you need 100 arcs per second suggests that you don't want much temporal spreading of the arc. So perhaps a short intense high-voltage pulse is all you need. In that case you should try to stop the primary current as fast as possible, i.e. without much of a capacitor on the drain. You will have to run some tests on whatever you plan on using for an ignition coil. Do you plan on using an automotive part, or do you have some other kind of ignition coil in mind? If the sparks last too long for your laboratory use, you will have to either raise the protection voltage on the MOSFET or lower the inductance of the primary coil, or add some power-wasting resistance in series with the coil. This last method will also reduce the overall current, and therefore the strength of the spark.

 

[faeraki]

Re: how to design an electronic switch using a MOSFET?

Ok yes..I see what you mean now (please excuse my ignorance..) I wrongly thought that 47.7Volts was the voltage produced on the secondary of the ignition coil.
First of all to answer to chuckey's question what I need is short intense high-voltage pulses.I will use an automotive part.
Concerning the duration of the sparks I guess that as I need 100sparks/sec will be 10msec.
So till now I am thinking to search fora MOSFET with as high breakdown as possible and very good repetitive avalanche rating.
I have to connect a diode and a zener diode in parallel with the primary of the ignition coil which is going to be connected to the power supply and the drain of the MOSFET.
Are these right?
Apart from these, do I have to connect a diode between the drain and the source for extra protection?
Do I have to add a pull down resistor at the gate?

 

[Tunelabguy]

Re: how to design an electronic switch using a MOSFET?

Concerning the duration of the sparks I guess that as I need 100sparks/sec will be 10msec.

I don't think so... 10msec. is the time between sparks. Each individual spark will last a much shorter time.

So till now I am thinking to search fora MOSFET with as high breakdown as possible and very good repetitive avalanche rating.

If you do the protection circuit right, the MOSFET will not ever need to avalanche. Avalanche is another word for break down. MOSFETs with a good avalanche rating can break down repeatedly without damaging themselves. But if you use a protection Zener diode, the voltage on the drain will never get high enough to break down the MOSFET.

I have to connect a diode and a zener diode in parallel with the primary of the ignition coil which is going to be connected to the power supply and the drain of the MOSFET.
Are these right?

Yes. How this works is that the Zener is positioned so the the arrow points toward the drain of the MOSFET, so that when you turn the MOSFET off, the rising voltage on the drain will be going against the arrow, through the Zener. That is how Zener's work. But Zeners also behave like ordinary diodes in the forward direction. So if you did nothing else, when you want to turn on the MOSFET, the Zener would be forward biased, and the current that you wanted to flow through the coil will flow through the Zener instead. To prevent this, put a regular diode in series with the Zener with the arrow of the regular diode pointing away from the drain of the MOSFET. That way no forward bias current can flow through the Zener - only reverse avalanche current, which is what you want.

Apart from these, do I have to connect a diode between the drain and the source for extra protection?

No, there is no good that can come of that. If the arrow of the diode points from the source to the drain, it will only conduct current when the drain is negative, which is never. And if the arrow of the diode points from the drain to the source, the diode will short out the MOSFET all the time. Actually, the MOSFET already has a built-in diode that prevents the drain from getting negative with respect to the source.

Do I have to add a pull down resistor at the gate?

That depends on how you are driving the gate. But a pull-down resistor is a good safety feature. In case your driving circuit ever became disconnected from the MOSFET, the pull-down resistor would keep the MOSFET turned off.

 

[faeraki]

Thank you again Tunelabguy! Now I am looking about the type of Mosfet and diodes that I have to select for this switch.
The problem is that I don't know anything apart from my 12 V dc power supply.
I was thinking about a mosfet with Maximum Drain Source Voltage at 620 V but I don't know if this is safe or waste of power.
Moreover could you suggest any value for Rds(ON)?

Any help is appreciated!
Thanks again!

 

[dick_freebird]

You probably don't care about Rds(on), you want to
control primary current to a reasonable limit value
(through on-time) and then snap the switch "off"
as quickly as you can. The tolerable current in an
automotive ignition coil is probably a couple of amps
max, and a FET with 0.1 ohms Ron is less than a 1%
loss at 12V, 1A.

What Vds you need, depends on the output voltage
desired and the transformer turns ratio. Anything
lower than Vout/((Tsec+Tprim)/Tprim) will limit the output
and be breaking down the switch device. Or something
like that, it's an autotransformer pretty much.

But old style points ignitions worked with pretty small
gaps and I don't think more than a few hundred volts
is necessary for 10kV-range output. I've made relay
based chatter ignitors using little 5V reed relays that
weren't rated more than 250V at the contacts, and
while I never measured raw output, they would throw
more than 1/2" free air arc.

 

[faeraki]

Thank you again for your help!
While searching for the most suitable MOSFET and diode, I found this
https://www.fairchildsemi.com/ds/IS/ISL9V5036S3ST.pdf
This is an IGBT for ingition coil and I was thinking that while it includes the configuration with the diode and the zener that Tunelabguy suggested
maybe I could use this instead, connected directly to the primary of the ignition and to the pulse generator(gate).
Is there anything I am missing and I shouldn't do it that way?