Issacsutt

modified 4 years ago

1MHz 10Ω impedance MOSFET Driver

409

04:34:46

A 1MHz sharp Square Wave oscillator that can drive a 10Ω load.  I tested it on a breadboard using the following transistors: 2N3904, 2N3906, and IRFZ44N, and it works just as you see in the simulation.  I plan on using it to hopefully drive a homemade high voltage transformer for producing some arks.  So far, when driving a 1uH inductor (simulating a small primary coil), it produces up to 68v spikes, but unfortunately gets quite hot pretty quickly (although I wasn’t using a heat sink, but I’d rather not have to rely on one anyway).  The challenge now is really gonna be making a transformer that with a large enough secondary coil that’s fast enough for this and is coupled well enough with the small primary, and in the future, if it works and I’m satisfied, then I’ll work on a more efficient design that can remain on indefinitely without getting hot, instead of having to turn everything off after only 20-30sec or so.  

**If anyone has any experience with making high voltage high frequency step up transformers, please let me know in the comments

**Edit:  I noticed it was actually oscillating at more like half a MHz in the simulation, so I had to adjust some values, however it did indeed oscillate at 1MHz in real life, but I was using inductors with values: 10uH and 100uH

published 4 years ago

jason9

4 years ago

You could go for a tesla coil style transformer. Judging by the number of youtube videos with tesla coils, it seems to be pretty successful. Also I responded to your comment a day or two ago over on my inverter circuit.

jason9

4 years ago

Also, regarding the heat, have you actually measured the voltage across the MOSFET while it’s on? It could be that it’s only partially on and as such has a significant voltage drop across it while flowing the current. Supposing 500mA and 1V that would be half a watt.

Issacsutt

4 years ago

I know, I saw your reply, sorry for the wait again, I’m currently typing up a response. I’ve been experimenting and designing several projects all day for the past week, and was a bit too obsessed with it I guess, I don’t even eat anything from the time I wake up till like 2 or 3pm some days cause I refuse to rest till I get them to work properly, I gotta work on that though, it’s a bad habit.

Issacsutt

4 years ago

As for the mosfet, I’ll check the gate voltage and let you know, but I know the simple push-pull buffer was putting out a clean square wave at full amplitude, however, the higher the frequency gets, the lower the values of every single resistor in the circuit have to be, and I think that’s because of the parasitics of the resistors, cause I can’t even read a voltage at the end of a floating 10k if the frequency is too high, so I don’t think it’s due to the Miller affect. (anywhere beyond 1-10MHz I think 10k’s don’t even work), however it could be just because I’m using a breadboard and perhaps the parasitic capacitance is the cause of it, I don’t know yet, but it drives me nuts. It could even be the probe/scope instead, I’m not sure: And the problem gets much worse the higher you go.

jason9

4 years ago

I was thinking to measure the drain to source voltage, not the gate to source voltage, since that’s what actually determines how much power is dissipated.

Issacsutt

4 years ago

Oh yeah, that’s a good point, thanks. I’ll check them both just in case

faceblast

4 years ago

with a push pull stage I can get good switching up to 100-150kHz on irfp250 power fets. driving the gate at 12V uses 1A to hard on/off against the gate and miller capacitance. voltage boost circuit to get the gate on voltage to 15-18V helps a lot. put diode clamps on it to prevent damage to the gate. the power fets will take an incredible amount of abuse across the drain-source.

faceblast

4 years ago

breadboards arent great for drive stages, the board is a huge capacitor at these frequencies, and the tracks won't handle the current

Issacsutt

4 years ago

It already has a push-pull bjt stage, are you referring to a push-pull out of 2 mosfets? As far as driving the gate at 100’s of kHz and above, your definitely right about the Miller effect, and it gets so much harder and nearly impossible once your into the MHz region, you either have to use very low value resistors, or inductors at that point, but then it just draws so much current everywhere

faceblast

4 years ago

yeah hard switching these things needs huge amounts of current.

faceblast

4 years ago

I like your oscillator though

faceblast

4 years ago

for HV transformers, old tv line output transformers are fun, but delicate, and real easy to damage around 40-50W input, and rarely have AC output. Car ignition coils will take a lot more abuse

Issacsutt

4 years ago

It’s a challenge for sure at those high frequencies, I’ve always wondered if it’s much easier for tech companies because they’re just using much faster RF transistors, or if maybe their designs are just that different.

faceblast

4 years ago

line output transformers are also prone to throwing big currents back through the primary side if you're fooling about with sparks, big enough to destroy the protection mosorbs or melt MOVs.

Issacsutt

4 years ago

Thank you! …Fun fact, at frequencies below 1MHz it can easily still oscillate just as well with a 50k-100k resistor, (I was able to power if for for a really long time with just a 100/1000uF cap)…. Also, if you instead tie one of the bases to the resistor and use a high enough ratio of the two size inductors, then you can even turn it into a boost convertor that puts out easily 20-40v or more, probably the simplest one you’ll ever see in my opinion, it’ll also run at even below a supply voltage of 1-2v. I’ve been thinking about possibly posting it.

jason9

4 years ago

Please do post it. Sounds interesting.

Issacsutt

4 years ago

Ok I’ll be sure to post it later today then!

Issacsutt

4 years ago

Just posted it…. http://everycircuit.com/circuit/5797489372037120

add comment in editor