Forums » CircuitLab Support Forum

Hello, My name is Jose Guadarrama, I am trying to mimic the induction heater from the following website: http://www.rmcybernetics.com/projects/DIY_Devices/diy-induction-heater.htm

However, I don't have the circuits background necessary to understand what's happening at each component. I am trying to copy the exact same schematic to circuitlab but I can't get it to work. Also where would I connect the negative terminal from the power supply to?

Here's what I have so far: https://www.circuitlab.com/circuit/86uytm/induction-heater/

Thanks in advance.

1) Where did you get the model parameters for your diodes from?

They are not CL "standard" parts. They also seem to be different from each other.

2) Same questions about the MOSFETs.

3) You have not created a centre tapped inductor correctly. You have inserted a single inductor connecting the MOSFET drains together and nothing else. The wire you have drawn to a notional centre tap does not in fact connected to anything because there is no pin on the inductor for it to connect to.

CL does not directly support a tapped inductor.

However, you can use the simple transformer as a tapped inductor. Set the turns ratio to N=1 and then connect the dotted end of one winding to the un-dotted end of the other winding and use that node as the centre tap.

You can also use the CL centre tapped transformer to create a centre tapped inductor. Just connect each end of the untapped winding through 1G Ohms resistors to ground.

However, this just complicates things so stick with the 1:1 transformer.

You have omitted the inductor L2 in the original drawing. It is essential to make a Royer - or as this is, a Royer-like - oscillator work.

You'll need to make a guess at the inductances since the article makes no attempt to put values on them.

I'd guess making the tapped inductor about 16uH in each half of the winding (each side of the centre tap) and L2 about 100uH might get things going.

Note that inductance is proportional to N^2 where N is the number of turns. What this means is that that if you have a coil of N turns wound on a core it will have an inductance L=m*N^2 where m is a constant determined by the core material and the laws of physics.

If you then double the number of turns to 2N, your inductance increases to 4L.

It is important to understand this because if you use the CL 1:1 transformer to make your tapped inductor and you set the inductance of one half of the inductor - the primary side inductance, L_PRI - to 16uH then the total inductance will be 64uH.

Each side of the centre tap has N turns, so the total number of turns is 2N, but because the inductors are tightly coupled, the total inductance is proportional to 4N.

So be careful to remember the distinction between the inductance of each half of the winding as compared to the total inductance.

4) In almost any circuit that shows a positive supply connection and a ground connection, the negative terminal from the power supply goes to ground.

Here's a fixed version to play with:

:)

BTW:

http://en.wikipedia.org/wiki/Inductance

Also note that in the case of the project you are basing your design on:

http://www.rmcybernetics.com/projects/DIY_Devices/diy-induction-heater.htm

the coupling between the coils is not tight. It is very variable and depends on the contruction of the coils and - more importantly - on what you stuff inside them!

That will also affect the permeability of the effective core so will change the inductance.

That in turn will alter the oscillation frequency.

The rules on inductance are the same but the loose and variable coupling makes calculations of inductance very difficult.

Some links to true Royer oscillators:

http://en.wikipedia.org/wiki/Royer_oscillator

and see:

APPENDIX K

WHO WAS ROYER AND WHAT DID HE DESIGN?

in:

http://cds./docs/en/application-note/an65fa.pdf

Thank You for all your feedback signality: The reason why I wasn't using CL standard parts was because I was trying to mimic the induction heater exactly and added the parameters from the datasheets. Thanks, I am just trying to see the theoretical values for the voltage and current.

Also I noticed you made your voltage into an expression..can I ask why is that?

Other than that thanks your awesome :)

" ...and added the parameters from the datasheets. "

If only it was that simple.

Have a look at:

https://www.circuitlab.com/forums/modeling-and-simulation/topic/4jex93sw/importing-transistor-from-datasheet/

" ... you made your voltage into an expression..can I ask why is that?"

You just did.

:)

Here's a clue:

"V1 (vcc) has exponential ramped startup to kick-start oscillation."

and the answer is:

https://www.circuitlab.com/docs/expressions/

https://www.circuitlab.com/docs/circuit-elements/#behavioralsources

https://www.circuitlab.com/circuit/b24363/ramped-and-glitched-signals-and-supplies-01/

https://www.circuitlab.com/circuit/w83xna/relaxation-osc-fixed-02/

https://www.circuitlab.com/circuit/9rd742/astable-flip-flop-classical-01/

https://www.circuitlab.com/circuit/7w869q/astable-flip-flop-04/

https://www.circuitlab.com/circuit/8qsm6f/joule-thief-01/

@Signality: Thanks, the induction heater is a great example for a workaround for tapped inductors. Your text also shades some light into my inductor - darkness, love it!

I’ve tried to learn about possible starting issues for the Royer when having a current limiting power supply, but no luck in CL? ( Tried to set V2 to 11V or so, played with R_i)

However, (to me) it’s not easy to understand why and how oscillation starts in the “ideal and symmetric” simulation, anyway.

I've replaced the expression by the simple switch PSU - it works, but then I found the oscillation starts depending on CL’s "Time Step", would not expect that?

In the other example, the joule thief, the simple power switch works, too, but the issue with the time step is even worse (Ri=10 Ohm, I tried to speed up by increasing steps to 500n, it changes the frequency and more).

This user defined time steps may not be the final solution, although I can clearly see that CL adds additional steps at some “critical” points …

But back to the induction heater and starting the oscillation: Let’s assume the sim using Time Step 1u is correct (see the huge current at C1, good ! ), now, …

Um, … do you have any proposal “how to bring a small screwdriver” into the coil (e.g. after 400us)? Is it basically the simple resistor in parallel to C1?

Regards, Sancho

Hi Sancho_P,

Glad some of it's helped.

"However, (to me) it’s not easy to understand why and how oscillation starts in the “ideal and symmetric” simulation, anyway."

The algorithms CL uses are not made public so I don't know. However it's possible that even in the ideal case slight numerical errors may occur - effectively noise - which start oscillation even in what seems to be an ideal perfectly balanced circuit.

CL sometimes starts circuits which in spice never do without some external stimulus such as a power supply startup ramp or step (that's why your switched supply works), a spike or step in an otherwise steady supply or a component imbalance is the resistors, turns ratio or transistor parameters.

What does happen if you do any of those imbalancing tricks is that the oscillation tends to start much earlier and ramps up faster than without.

Given that:

https://www.circuitlab.com/forums/feature-requests/topic/7yyp49h5/feature-request-delayed-start-of-plot-time/

doesn't exist yet, you don't want to have to wait for all that time before oscillators in CL start.

As for starting with a current source: it should work: that is what L1 is doing. It makes the AC source impedance look high so at AC the oscillator is being fed from a current source.

I have a 10kV inverter design that is regulated by a switched mode supply which effectively regulates the current through the "Royer" oscillator as a means of regulating the output current through a gas discharge on the secondary side of the transformer.

About Time Steps:

Again, no visibility of CLs algorithm so very hard to know what's happening. Transient sims often seem to go off into a little world of their own busy calculating away when there seems little reason for it to do so ...

but:

https://www.circuitlab.com/circuit/653up7/damping-vs-time-step-01/

https://www.circuitlab.com/circuit/4erqjv/damping-vs-time-step-02/

https://www.circuitlab.com/circuit/y97g97/damping-vs-time-step-03/

"Um, … do you have any proposal “how to bring a small screwdriver” into the coil (e.g. after 400us)? Is it basically the simple resistor in parallel to C1?"

Need to think about that.

As I said, the screwdriver changes the permeability of the core so it changes the primary inductances, which changes the oscillator frequency.

The coupling between the primary and the secondary - the eddy current path in the screwdriver (remember the screwdriver is not physically connected to any of the circuit other than by the inductance and some small stray capacitances) - is hard to predict or calculate because it is a small peg in a big hole.

Might be possible to do as an example but the lack of parameterisation in CL makes it unclear for a more general case.

@Signality: The background of my questioning the current limited power supply and starting to oscillate (in both, simulation and reality) were some hints (comments) at the originally linked webpage that starting with a limited / -ing current supply would be difficult. I could not really believe that. So I wanted to find out if you really need a hammer (…the hinted big caps) to start oscillation or simply a “reasonable” oscillator circuit (let it be sim or real).

What you say regarding (CL) simu time steps and the absolute time between power ON and starting the oscillation makes sense, as I saw that CL’s simu with bigger time steps often just needs more time to begin to oscillate - hence more iterations until some errors cumulate as you said.

To me a fascinating experience, as I am the one who will set up his “perfect” oscillator on the breadboard just to stare in horror at smoking parts then (plain DC, no oscillation of course).

“Transient sims often seem to go off into a little world of their own busy calculating away when there seems little reason for it to do so ...”

I take it you are talking about “His Highness” because CL never does … ;-)

Regards, Sancho