Need Help Troubleshooting a mosfett

Ok, I need you to show me a circuit. There’s no way I can visualize all the pots and stuff. But here’s some stuff about MOSFETs: they have insulated gates (no current flow at all), and because of the way they work they’re gates basically act like capacitors (in the low nF range). So, if you apply voltage to the gate, then it won’t draw any current. If you even entirely disconnect the gate so it’s floating, then it would theoretically remain on as it’s gate capacitance will have nowhere to discharge since the gate is insulated from the rest of the transistor. MOSFET stands for metal-oxide-semiconductor field-effect transistor. The metal-oxide-semiconductor means that there’s a layer of metal, which is the gate, then there’s a thin oxide layer, which is an insulator, and a semiconductor (usually silicon) after that. Since it’s a field-effect transistor (FET) it need no current; it acts by the electric field produced by the gate. However, the MOS arrangement, which is conductor on both sides of a very thin insulating layer (one side is semiconductor, but can be considered conductor here), is basically a capacitor. In fact, it’s the exact same arrangement as a capacitor. So, if you charge the gate capacitor of a MOSFET it will theoretically stay charged indefinitely keeping the MOSFET on until a pull down/discharge resistor is put in place. So, unlike a BJT, you can’t just put a switch and current limiting resistor in series with the gate. A better idea would be a switch and a pull down resistor, or a SPDT switch for activation/deactivation.

Ok here it is: http://everycircuit.com/circuit/5291919063384064

Ok, so you tried to turn of the MOSFET by adjusting the input voltage via the first pot, right? If so, did you turn it the right direction? Did you decrease input resistance with pot 2 or leave it at 1MOhm? If you left it at 1MOhm it would’ve taken a while for the gate to discharge because of the gate capacitance. Because of that, it would’ve seemed as if the MOSFET was broken and wouldn’t turn off when in fact it was working perfectly. If the problem persists, measure the gate-source voltage with a multimeter to see if there’s somehow still voltage keeping it on. If there isn’t any voltage between the gate and source and the LED remains lit then there’s a problem with the MOSFET. Either that or it’s actually depletion mode instead of enhancement mode, but that shouldn’t be the case because you bought enhancement mode MOSFETs. If it only turns off with a negative gate-source bias then it’s depletion mode.

Actually, I only used the 1st pot to adjust the threshold gate voltage that I wanted to bias it with (which was about 2V) then I didn't touch it after that. I was then using the 2nd pot to adjust the amount of resistance between the gate of the mosfett and the middle of pot1... (I was slowly increasing the resistance of pot2 to see if it would have any affect on the mosfett: which was already initially on)... which, it appeared at first that it did not have any affect on the mosfett because the led seemed to have stayed lit at full brightness, and the voltage at the drain was the same, but then I went to turn it off, and that's when I became unsure. I expected the mosfett to remain on or unaffected during the test, but I had no idea what to think when it stayed on cause I even tried grounding the gate after that, and it still remained on. It wasn't until I grounded the gate and tied the source to ground that it actually cut-off.

By the way, don't mosfetts have regions like a transistor, like ohmic saturation, and cut-off I think are some, there might be one more?... so how would I make a voltage follower, would that be using ohmic region? And then what about a simple amplifier circuit, wouldn't that use sat./cut? ...How do you put the mosfett into those different regions/ states?

Yes, there are regions and stuff, but let’s not worry about that yet. Why didn’t you have the source grounded? I thought source was connected to ground and the gate was pulling current from the LED turning it on. If the MOSFET doesn’t turn off when the gate is hooked up to the source then there is a problem. Else it’s likely some confusion/error that makes it seem broken when it’s not.

I dont think the gate was pulling current from the led, (do you mean the drain possibly?).. I was feeding the gate with a positive voltage from some pots, I think I grounded the source, but I honestly can't remember exactly if I grounded source and connected drain to cathode of led, or connected cathode of led to source and grounded drain, but wont a mosfet conduct in either direction anyhow?

Sorry it’s been a while but I forget to check up on EC comments/circuits. Anyway, yes, I meant drain not gate. Silly error. Also, yes, MOSFETs conduct either direction but the conductivity/insulation is based on the gate-source voltage regardless of how the MOSFET is configured.

Dont worry about it man, I hardly ever check the community myself... I'm guilty too. Alright, so does that mean that the mosfet could conduct better/worse in one direction than the other depending on the gate voltage? Also, does the voltage at the gate depend on the voltage at the source too (like a bjt has a voltage drop)? And can it drive a floating load... is there a relationship between current through source/drain and the gate voltage?

You shouldn’t talk about gate voltage alone unless source voltage is known (or assumed to be 0). MOSFETs are driven by gate-source voltage difference. If the gate is charged with one more volt than the source then it’ll conduct a certain amount. If it’s charged with two more volts at gate than source, then it’ll conduct more. Conductivity here should be independent of direction (I’ll need to experiment with that in EC to make sure) and only controlled by gate-source voltage.

Ok, it seems MOSFETs act more like BJTs when conducting in the forward direction and more like voltage controlled resistors in the reverse direction. When the drain-source voltage is less than approximately half the gate-source voltage or negative, its resistance is pretty constant. When it exceeds the gate-source voltage it becomes a constant-current source type device. When it’s between the gate-source voltage and half the gate-source voltage the current is less dependent on voltage than a resistor, but still not totally a constant current source. So, in a sense it conducts better in reverse than in forward direction, but if a constant current device is what you seek then forward conduction is best.

However, when conducting in reverse make sure the driver circuit of the MOSFET is supplying gate voltages higher than the source voltage (or simply grounded directly into the source).

Ok thank you for all that, will have to do some of my own experimenting too on that.

Ok. Have fun!

Hey maybe you can help me on something. Trying to create a square wave oscillator, designing an oscillator is something I've been hung up on for a while, I've got the idea that basically a transistor is biasing itself on (when in cut-off), but when it turns on (in saturation) it reverse biases itself, thus turning itself off then back on over and over again, with the time between the transition of the two states being determined by the size of one resistor and capacitor. Now most of the time I try this, in a couple different designs even, it will rise up in voltage til it hits a specific voltage level and it flattens out at that, never even a damped oscillation... and the sometimes I'll get lucky and oscillations will occur for a short period, and sometimes I can play with the values and/or sim speed and actually make it work, but I still never feel like I fully get it. Take this oscillation circuit for example: http://everycircuit.com/circuit/5509828155015168 it works pretty flawlessly with a relay, cause relays don't have any states in between fully on and fully off, but if you try this same circuit with a transistor as a switch I stead of a relay, failure. It reaches equilibrium, and stays there at one flare voltage level. I'm constantly having to fight the resistive properties of a transistor with almost every oscillator design I try, I just wanna figure out how to design an oscillator that doesn't get stuck like that. How would I make that circuit from the above link work with a transistor for example?

Ok, I haven’t looked at the link yet. I will tomorrow (currently going to bed right now), but you see you have to use the transistor’s attempt to reach equilibrium to make it oscillate. Basically, put a fixed delay in the feedback. If the input goes a little high, the output tries to compensate by going a little low, but there’s a delay so it keeps going lower as there’s no feedback to stop it. But then the feedback kicks in, and the reverse happens. This goes on like that for eternity as you now have an oscillator. This delay in the feedback can be for example a RC low-pass filter, but it needs at least three capacitors or else there simply won’t be enough delay for it to work. However, the RC filter will attenuate the signal significantly, so your transistor (or set of transistors, such as a darlington pair or op-amp) needs to have enough gain to produce a sufficient positive feedback loop. It basically works because if you put an AC signal it phase shifts it 180 degrees, but a right filter will also phase shift it 180 degrees for a total of 360 degrees transforming the negative feedback into a double negative which is a positive feedback. But only for that one frequency. This frequency is determined by the RC low-pass filter. You have to have at least three capacitors because that’s what’s needed for 180 degrees phase shift at a specific frequency. The frequency can be controlled by the values of the components. You can use other filters too and it’ll work fine if designed with enough gain (just needs 180 degrees phase shift at a specific frequency). A nice filter good for when your transistor has little gain is an LC low-pass filter. Just two capacitors, both grounded, with an inductor connecting their ungrounded ends. The two ends of the inductor are the two ends of the filter. At the resonant frequency of the filter it’ll have 180 degrees phase shift. A little lower frequency and very little phase shift, a little higher frequency and too much phase shift. That way it gives a single very specific frequency at which it lets the oscillator oscillate. Since there are no resistors and it’s a resonant tank circuit it takes very little gain to get it to oscillate if made right. Should be easy to do.

Dang, I’m writing the equivalent of a novel here.

Ok, I got all that, thank you! ...but what about a relaxation oscillator, hurz has told me before, that relaxation oscillators dont follow the same rules as a harmonic one that needs proper phase shift, he's said that in a relaxation oscillator, you only need enough of a delay and some gain, no phase shift, prob because nothing is sinusoidal, and that makes me think of something else, a harmonic is positive feedback, relaxation must be negative?!?

Thanks for taking the time to right your novels!

A relaxation oscillator is basically a 555 oscillator. Yes, it has negative feedback, but instead of a delay like an RC low-pass for 180 degree phase shift to provide positive feedback for only a single frequency, it has hysteresis. Let’s say you have a circuit that if its input voltage drops below threshold A, it makes its output rise. Then, if the input rises above threshold B, it makes its output drop. If you connect the output to the input with a resistor and capacitor to give it a delay between the positive and negative thresholds then it will oscillate at a specific frequency indefinitely, as there simply is no equilibrium state. The most common example of this is the 555. However, you can easily create the circuit with a single op-amp, a few resistors, and a capacitor instead of all the transistors and stuff inside a 555. You can even create something that resembles a relaxation oscillator (not sure if it counts as one or not but I’ve made it) that uses no more than two transistors and one capacitor.

Ok, I looked at the link and it doesn’t quite seem like a relaxation or harmonic oscillator. It’s just, well, a relay oscillator I guess. But they’re easy to make oscillate as they have no equilibrium (just add negative feedback, and perhaps a capacitor to control frequency), and you could easily turn it into a true relaxation oscillator as the relay turn off threshold is below its turn on threshold.

You mean like this: http://everycircuit.com/circuit/4826182943244288

Yeah. Exactly. It’ll even work without the cap due to the coil’s inductance, but it’ll go really fast. Even without inductance it’ll work because the relay can only go so fast. Else it’ll just go light speed creating an explosion involving relativistic energies making a giant crater from the world’s most compact high yield bomb.

Haha, were did you get that hypothesis?

Well, I was thinking that if the relay had no delay whatsoever then it must have to travel at light speed to work, but then the little switch in it would be like a tiny relativistic meteor hitting Earth at 99.99% light speed creating a giant crater. Did you know that an iron atom that hit Earth at 9999999999999999999998% of the speed of light (exactly that much I counted the nines) contained enough energy to light a 25W LED for an entire two seconds? Just from the kinetic energy of a single atom. And even though there are more stars in the universe than grains of sand on all the world’s beaches and oceans, there are more atoms in a grain of sand than stars in the universe. And there are billions of stars per galaxy, thousands of galaxies per galaxy cluster, thousands of galaxy clusters per galaxy supercluster, and millions of galaxy super clusters in the universe. So, that’s a lot of atoms in a grain of sand. And if each had enough energy to power a 25W LED for two whole seconds... we may have to say goodbye to Earth. No, not Earth, the whole dang solar system. So, yeah. Big boom. Aliens will think it’s just another of many gamma-ray bursts which are known for their immense power and no one knows how they get they’re power. But now I do. It’s because of little ultra-relativistic relays invented by aliens only to destroy those aliens and their little relativistic relays.

Haha, well I bet that little relay would crash and burn before it even hits speed of light, perhaps a little explosion inside that tiny box, nothing strong enough to support a contact moving that fast, it would probably melt before it gets anywhere close the speed of light

Probably. And the magnetic field required to generate sufficient acceleration to speed it up to light speed in the tiny space provided would literally rip atoms apart. And destroy the Earth. And cause the sun to explode upon interacting with it’s magnetic field. Again, same result. Solar system no longer exists.

Ca- boom... 💥

Yep!