ZVS Debugging

I want to discuss methods that I had used to debug my ZVS curcuit. The reason for the need to debug it came from the mosfets overheating quickly with and without a load. Having read up on other peoples stories and requests for help and discoveries of my own, I am writing the post to help anyone else that may run into similar problems and also get help to understand why my solution worked in my case.

Below is a circuit diagram of what I am using at the time of writing this:



(Note: Incase of any confusion, above 1 - drain, 2 - gate and 3 - source)

As you can see, some of the parameters are unknown, This is due to them coming from old computer power supplied and not having a LCR meter... but that doesn't stop this debugging guide.

From the basics, if we are experiencing FET's getting hot we generally check the following:
⦁ MOSFETS
⦁ You are using a MOSFET that can handle the input power and the output power. (Using a circuit simulator like ltSpice can help find the rought values of what is going in and out of your FETs). It is suggested to aim for a FET that can handle 4x the amound of voltage you hope to have running through it (and also current too).
⦁ You are using MOSFET's with a low Ron value (a good value is under 50mOhms), For mine i went to the extreme and the IRLB3034pbf has an Ron of 1.7mOhm.
⦁ DIODES
⦁ Making sure they are fast with a low voltage drop, schotty diodes are good for this. I am using SB5100 schotty diodes here, they have a voltage drop of about 55mV which is fine in my circuit, but lower would be better. (I've not tried other types of diodes in the circuit but feel free to add to this to help others)
⦁ TRANSFORMER
⦁ Make sure that the primary windings are even and firmly in place (Even the slightest differnence can change the performance of the circuit)
⦁ If you are using this for High Voltage, making sure there is a paper fin gap in the core can prevent it getting satuated during ciecuit operation

Now is all the above seems to be ok, we can move onto looking into other factors that have an impact on the circuit but not many tutorials or blogs explain it well (with respect to technical details and I probably won't either, but I have a theory on some parts that make sense, feel free to expand on it if you have any input).

GATE RESISTORS:
The resistors at the gate of the FET are used from what I have read, to control how fast they can turn on and off, the higher the resistance and slower it is and vise versa.
A lot of posts and tutorials on ZVS say to use a default of 470ohm and that is what I had Done, but my FETs where getting hot and fast.
This make me thing about the circuit as a whole, we have the ZVS part and we also have the Tank circuit, the tank circuit is tunes to a certain frequency dependant on the values of L and C chosen, and the FETs are limited to the value of R chosen of switching...
My theory then was that, what is the tank circuit was trying to operate at a frequency that the FETs could not provide due to the chosen resistance? would that cause it to overheat? Well I tested that theory but changing the 470 ohn resistors to 220, the results spoke for themselves, the FET's still heated up over time but it took a lot longer (HURRAY), but was I truely happy with that, no! I wanted a ZVS that didn't heat up even slightly when in operation like so many tutorials and blogs claimed theirs didn't. so I lowered the resistance to 100 ohms and well, What can I say the heating issue seemed to disappear (Danced around the room and also burned my finger a few times... arcs are hot).
With this success and the proof that it seems to hold some validation to my theory, what now? well, we need some proof of this other than just 'Well, it works doesn't it?!', Unfortunately I do not have the experience to come to a proper conclusion about this theory other than Maybe, The Fets capacitance and resistance value making an RC limit to the frequency?

Also having measure frequency across the two drains on my oscilliscope and calculated the RC frequency, there should not in theory been a problem as the RC frequency of the Fets was higher than the actual output frequency, so what could of been causing the overheating?

Does anyone have any idea's to what would be causing the overheating and how lowering the gate resistance actually solved the problem?

I hope this post might help people in the future that are having trouble with ZVS circuits and overheating mosfets and that someone with more knowledge than myself might be able to explain what is more likely the cause of that problem and too how my solution seemed to fix it.

Kind Regards
David

 

Further debugging my ZVS as I was still having issues with the mosfets overheating when an arc was being produced, I found posts for finding out how much heat should be generated from the data sheet, I think the following is correct:

C/W = Id*Id*R(on)* R0ja

In my circuit when arcing it was drawing only about 3-4amps (and the fets heated up quickly... hardly able to touch after around 10 seconds) and when not only around 30mA, and from the above calculation the fets should only be producing (best case scenario) between 1.116 - 1.984 degrees from the data sheet for the IRLB3034PBF, with a R0ja = 62 and a R(on) = 2mA at 4.5v (I'm not using heat sinks yet), My thought's on the matter was that because I was only supplying 3.7v, that the fet's where not fully being turned on and as such producing more heat, only problem I have at present, is that if I supply more than 3.7v to the high voltage transformer (i'm using from a small kit bought on amazon) it arcs in the coil of the secondary (gotten through 3 so far... oops).

Having thought about the above, I decided to test the idea and using a separate power source for the gates of the mosfets and the drain and source both with common ground and also changed the circuit slightly as I though having separate pull down resistors would help, it now looks as follows: (On a side not I notices that when I only apply power to the gate - ground then circuit still oscillates without the drain to source additional power, just at a low voltage of about 1 - 1.5 volts when applying about 4.5v to the gate)



and though now (figures crossed) the mosfets seem to behave even when arcing, seeming to help with regards the heating problem (still burned though another transformer - probably because it's now getting the full 3.7 volts to the transformer rather than being split with the gates? I'll try more winding on the primary when i get round to fixing it, so that the output is a bit less and also planning on casting the transformer in resin to increase it's breakdown resistance (sorry if my terminology is wrong)). I'm left wandering how best to create this circuit from only one power supply (i'm looking at portability, and also trying to keep it using discrete components for now as I'm wanting a deeper understanding of how the circuit works before using power isolating ICs).

My questions here are:
1: am I right in thinking that the problem was more than likely caused by not providing enough voltage to the gate and as such causing overheating or am I totally missing something simple.

2: where my calculation correct for the above data sheet temperature calculation? (I'm kind of new to working with data sheets, so been reading and watching a lot about them, and think I am on the right path)

3: If I wanted to use just on power supply without using an IC, would a resistive voltage divider work here or would the current demand of the drain be too much for such a divider (High power resistors would probably be required i'm guessing, already using 1W resistors)?

Thank you for any input you may have.
Regards
David

 

Just a Quick update from what I have found, Found a nice video on youtube Why We Must Always Use a Gate Resistor, this video helped answer a question I could never really find an answer too, which was how do we choose the resistor for the gate, to which this video enlightened me that if you look at the datasheet and the on/off-rise/fall times it presents a Rg value to use for the mosfet at a specific Vgs, I've never been a fan of tutorials or diy articles that specify a value for a component but fail to say why they chose that specific value and I needed to find something on that matter for this project, so there we go (Obviously in some cases it's self explanatory but not in this case... I think anyway).

Not heard anything back from anyone else my questions above (scrap no 3, if anything I think looking into the idea of a forward converter or flyback converted would be the way to go for isolation... but I'll look into that later) so I'll keep digging and share what I find.... on that note I wanted to also share that when picking a Mosfet for this type of project look at the capacitance values in the data sheet, as too the on/off-rise/fall times and try get them as low as possible to increase it switching ability.

(Having tested the lower resistances as per the tests of the datasheet, I've found that even 2w resistors heat up quickly at very low values around 2.4Ω, not blown up anything yet but getting better understanding overtime)

(My aim with this project is to make either a high voltage power supply or maybe an induction heater or both)

more later
Regards
David