H bridge MOSFET driver connections

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
Hello and thank you very much for reading my post and trying to help.

I am making a 3KW induction heater, using this resonant oscillator circuit:
https://drive.google.com/file/d/1-qKvyTg_fSOSFhRa2nMBaJ3y8fZ00i-p/view

However, I would like to remove the current gate circuitry and control the MOSFET gates using a driver circuit which can be frequency controlled by an MCU (I do some AVR Microcontroller programming).

I found a High speed Half bridge MOSFET driver:
https://datasheets.maximintegrated.com/en/ds/MAX15018-MAX15019.pdf

and thought it would fit the bill. However the H bridge uses the concept of High side and low side, and the induction heater circuit is designed as common drain. So I can't see ( don't have enough knowledge...) how to make connections from the Half Bridge driver to the MOSFET gates. Perhaps it can't be done with MOSFETs in common drain config?

My knowledge of this kind of power circuit is not all that good, my main area of expertise is in using sensors, digital signal processing and stepper motor control as a hobbyist.

I've done quite a lot of reading around power MOSFETS and I understand that if the gate switching is not controlled properly he MOSFETS are like;y to be destroyed.

Many thanks for any advice on an Interface between and MCU and the MOSFETS.

Paul
 

Ian0

Joined Aug 7, 2020
9,667
Your MOSFETs are common source, not common drain. Any low-side driver such as TC4427 will do the job.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
Thank you very much for this, much appreciated. I read the key points of the Microchip datasheet and as you say, this device does the job. Saved me a lot of time.
 

shortbus

Joined Sep 30, 2009
10,045
I am making a 3KW induction heater, using this resonant oscillator circuit:
https://drive.google.com/file/d/1-qKvyTg_fSOSFhRa2nMBaJ3y8fZ00i-p/view

However, I would like to remove the current gate circuitry and control the MOSFET gates using a driver circuit which can be frequency controlled by an MCU
I'm no expert but have a question about this induction heater circuit. By using an outside source for frequency control will the circuit actually do what your after, induction heating?

I built a one to loosen rusted bolts, using one of the ZVS modules that many use for the heaters. Works well for my application. But it's frequency is set by the caps and didn't know it could be changed with an outside frequency control. Seeing the cap bank in your circuit got me wondering, is using an outside frequency control from a MCU or anything else going to get you the results your after?

I know the industrial induction melters and heaters have adjustable frequency but don't think it's done the way your trying.
 

Ian0

Joined Aug 7, 2020
9,667
I also thought it was a resonant circuit, but I thought that varying the frequency might be an interesting way of controlling it - like an LLC resonant power supply.
I also suspect that taking it to the capacitance side of resonance would be a BAD thing to do, and those transistors might get really hot really fast.

Some of the internet circuits say that L2 and L3 should be wound on ferrite toroids. Ferrite will saturate immediately with a DC current through it. It should say iron powder.


I built a one to loosen rusted bolts,
Off topic here - but is it really good for rusted bolts??
 

shortbus

Joined Sep 30, 2009
10,045
Off topic here - but is it really good for rusted bolts??
Bolts was a mistake to say, should have said rusted nuts. I have an old lawn tractor that is being restored, it has the original paint that I want to keep. Doing this is known as "keeping it in it's work clothes". So needed something other than an oxy/acet torch to heat things up. While the induction heater isn't enough really to get some of the bolts/nuts loose by it's self, if I use both it and penetrating oil it does work real good. Where the penetrating oil wasn't up to the job on it's own.

My home made one was a lot less expensive to make than the cost of the ones they sell.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
Thanks for all the replies. I'll check out the inductor core material to ensure it's iron powder rather than ferrite. Regards using a frequency generated by an external oscillator (MCU in my case) the frequency applied to the MOSFET gates will need to be that which causes the LC tank to resonate. The reason for using a MCU is that the coil resonant frequency changes when an object is placed into the work coil (as I understand it from the research I've done) and it can be tuned by varying the gate switching frequency to optimise heating peformance. I also appreciate there will be easier ways to do this without an MCU - perhaps a 555 timer and pot in conjuction with the low side driver Ian pointed to. However, if I get it working, a further mod will be to try to introduce something like a phase locked loop to keep the circuit locked onto the resonant frequency without user intervention. We'll see...

Ian, I wasn't sure what you meant by " taking it to the capacitance side of resonance " Coud you possibly elaborate on that?

Good to hear about the effect on nuts....

Just for info, here's the link to the Youtube build video by Anton at Schematix:

and here's a link to a more sophisticated and more powerful induction heater project, which includes the idea of frequency control via MCU, section 3 has the MCU driver stuff it's a tough read because the site navigation is linear:

https://www.instructables.com/12KW-Induction-Heater/

thanks all,
Paul
 

Ian0

Joined Aug 7, 2020
9,667
If that's a picture of the inductors you have, then they are definitely iron powder. Yellow/white is a Micrometals type #26. I don't know what frequency at runs at but if it's >100kHz, then type #52 (blue/green) might make a small improvement.

At the resonant peak the load looks like a pure resistance, as the reactances of the inductance and capacitance cancel out.
At frequencies above the resonant peak, the inductance dominates, and the MOSFETs are in a zero-voltage switching mode, so there is virtually no switching loss.
Below the resonant peak the capacitance dominates, and every time the MOSFET turns on there is a voltage across the capacitor which it shorts out, and so there is lots of switching loss.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
That's good info Ian. So what I take from that is to ensure that on startup, the output from the MCU into the mosfet gate driver needs to be close to the resonant frequency and probably during running, don't let the frequency fall below resonance. I think it would be really useful to have an oscilloscope for this project, I'll have to look around, see what I can find. I checked the inductor delivery note, it says iron powder and they are light green, so that looks good, thanks.
 

Ian0

Joined Aug 7, 2020
9,667
The LLC resonant converter ICs start at their maximum frequency, furthest from resonance where they will produce the minimum output power, then ramp the frequency down.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
ah, good to know. Just thinking about a sensible way to connect the MOSFET gate driver into the induction circuit, my thought is to connect one output to the junction of R8/R9 and the other to the junction of R10/R11. I also think there's no need then for the circuitry which prevously caused the oscillation, so R1, R3, C15, C16, R2, R4, D4, R5, R6, D3, D1 and D2 can all go.
So I'll keep D6, L1, L2.
The above seems logical to me, but I'm no expert and in particular, not sure if I should keep the zeners D1 and D2

I know its a big ask, but what do you think about those proposed mods?
link to circuit - https://drive.google.com/file/d/1-qKvyTg_fSOSFhRa2nMBaJ3y8fZ00i-p/view

thanks.
 

Ian0

Joined Aug 7, 2020
9,667
D6 is just for reverse polarity protection. L1 and L2 are necessary to supply the power, so they all stay.
R1,R3,C15,C16,R2,R4,D3 and D4 all go.
D1, D2, R5, R6 can stay, but D1/D2 voltage should be higher than the driver supply voltage. Nothing wrong with protecting the gate from overvoltage, and making sure that the device is off if the driver loses power.
R2 and R4 should be replaced by a small resistor say 10Ω, depending on the size of the MOSFET.
The driver should then drive the unconnected ends of R2 and R4.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
Brill - thanks Ian, much appreciated. I'll give this a go in the next few weeks (PSUs arriving this weekend). I'll post back with hopefully positive results.
 

shortbus

Joined Sep 30, 2009
10,045
and here's a link to a more sophisticated and more powerful induction heater project, which includes the idea of frequency control via MCU,
I think you missed my point in my question, or I didn't really make the point. Just changing a frequency input isn't , I think going to work. The frequency is tied largely to the capacitor bank, so you also need a way of switching the different capacitors for the frequency ranges.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
Thanks for your reply Shortbus. My knowledge is limited on this (40 years ago I did my Physics degree and haven't touched a resonant oscillator since :) ). I tend to agree with you actually, the circuit I feel should be self resonating. However, from the research I've done, it seems that when a workpiece is placed in the work coil, the resonant frequency changes - that makes sense to me as it's like inserting a core into an inductor - the inductance will change and therefore so will the resonant frquency (as per resonant frequency formula). It seems that when the workpiece is inserted, an equilibrium is reached and further heating of the workpiece can be achieved by tweaking the frequency of operation. Some commercial products achieve this by providing a frequency adjustment Pot (and associated circuitry) for the user to tweak, and some use a phase locked loop to automate the process.

Ultimately it would be great to understand why the frequency tweak works, but I haven't found out why as yet. I suspect that basically, as the work coil induces eddy currents into the workpiece which causes heating, there's a link between frequency and heating effect.

Thanks all for your interest and help with this.
 

Ian0

Joined Aug 7, 2020
9,667
It would seem that in the circuit you posted, it uses a phase-locked loop to keep it at resonance, because the phase of the resonant tank will be in phase with the drive at that frequency. At the inductive side of resonance voltage will lead current and at the capacitive side, current will lead voltage.
If you want to keep it at the maxmimum power you need to keep it at resonance, but if you needed to reduce the power, taking it down the curve on the inductive side would work.

One thing that would change the inductance is if the material being heated was ferromagnetic and its temperature was taken above its Curie point.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
I found this theory paper on induction heating. It seems the ability to change the resonant frequency of operation as the temperature changes is a key feature of induction heater circuits. There are other variables too such as the size of the material and the reference depth (the dept at which 86% of the heating takes place),
http://www.elegron.pl/files/TheoryHeating.pdf

This article:
http://www.richieburnett.co.uk/indheat.html

Tells you all you need to know about theory and practice of induction heating. In a nutshell, the critical performance factor (as you have identified) is that the circuit operates at the resonant frequency. The resonant frequency changes as a workpiece is inserted into the work coil and the impedance matching inductor's function is to keep the frequency locked on resonance.

The link above also talks about power management in the inverter stage and varying the MOSFET gate frequency can be a way of doing this, but essentially what you're doing is detuning the circuit out of resonance to reduce the power consumption - perhaps because you want to reduce the rate of heating of the workpiece, but as Ian mentioned, the article explains the importance of detuning on the high frequency side to avoid damage to the MOSFETS.

So in conclusion, it has been very interesting and you folks have given me leads to focus my research. My conclusion is that the circuit is best left as is as it looks like it 'self tunes'. If it doesn't get a workpiece hot enough to hammer it into shape on an anvil, it's a more powerful circuit that's required, not a change in frequency of operation.

Thanks again for all the interest and help,
Paul
 

shortbus

Joined Sep 30, 2009
10,045
If it doesn't get a workpiece hot enough to hammer it into shape on an anvil, it's a more powerful circuit that's required, not a change in frequency of operation.
If your using it to heat for forging have you seen this guy's links? https://spaco.org/Blacksmithing/ZVSInductionHeater/1000WattZVSInductionHeaterNotes.htm It's worth your time to read all he has to say about induction circuits, mainly the ZVS type.

He is where I found out about the one I bought for my project. He spent a lot of time and money buying, testing and destroying the modules before finding the best one being offered to the public. He is a blacksmith and also a member here.
 

Thread Starter

paulskirk53

Joined Sep 8, 2021
45
Thanks, I'll take a look. The characteristics of the material change as it heats - especially iron or steel, so it seems to me from the physics of it that an induction heater that operates at one resonant frquency is only going to get so far. So I'm thinking that being able to change the resonant frequency would be a nice design idea. Perhaps SPACO might concur.

thanks all,
Paul
 

shortbus

Joined Sep 30, 2009
10,045
so it seems to me from the physics of it that an induction heater that operates at one resonant frquency is only going to get so far. So I'm thinking that being able to change the resonant frequency would be a nice design idea.
From my simple understanding you tune for that Currie point frequency or the one you are wanting reach. Then as the part starts to heat it starts out slowly but does heat until the temperature is reached. Unless you're doing this for just one size coil you are going to keep chasing your tail as even the coil shape and size has an effect on frequency and resonance.
 
Top