Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

New Induction Heater circuit with no center tap.

Status
Not open for further replies.
Does the OSC set the KHz of the Flip Flop?
Yes.

The oscillator frequency is 200Khz which you can adjust by the variable resistor. The 200KHz signal from the oscillator connects to the clock input of the flip flop which divides the clock signal by two to produce 100Khz to drive the transistor driver chip. (you can have any clock frequency you like, within reason, if 100KHz is not suitable).

The flip flop not only divides the frequency by two, but it also produces two antiphase clock signals (CLOCK and /CLOCK) which the driver chip requires.

The flip flop ensures that the clock signal clocks have an exact 1:1 mark to space ratio, with fast rising and falling edges.

Having an external clock generator ensures that the frequency of the signal in the induction coil is constant and the frequency is not pulled by items placed in the coil's magnet field.

The clock generator is dead simple to build and is well-behaved.:)

spec
 
Last edited:
Forcing the load coil to run at 100 KHz all the time does that not interfere with the resonance frequency of the RC circuit?

Does the circuit not need to run at the resonance frequency?

My mosfet circuit runs at 89KHz with no load. With a 3/8" solid steel rod in the coil frequency drops to 64KHz.
 
Forcing the load coil to run at 100 KHz all the time does that not interfere with the resonance frequency of the RC circuit?

Does the circuit not need to run at the resonance frequency?

My mosfet circuit runs at 89KHz with no load. With a 3/8" solid steel rod in the coil frequency drops to 64KHz.

In this design there is no resonant frequency and no capacitor. That is the thing about this approach: you do not rely on the coil osculating with a capacitor to generate the switching waveform. But you may have some capacitance on the coil to smooth the edges a bit.

The 100Kz frequency to drive the bridge is only notional. By experimenting you will probably find an optimum frequency to suit a particular coil and transistor type. That is one of the advantages of this circuit you can change the parameters quite easily.

spec
 
POST ISSUE 02 of 2016_11_09

Gary, it seems to me that there are two good approaches for your induction heater:
(1) Full bridge: four MOSFET/IGBT
(2) Half bridge with +- supplies: two MOSFET/IGBT

I think (2) has a lot of benefits so, for your consideration, I knocked out a quick block diagram to illustrating the approach using a single UCC21520 chip and two high power MOSFET/IGBT:

I have been a bit lavish with the 15V power supplies, and while two are not absolutely essential, they do give complete isolation between the logic and bridge. They also improve performance and make the operation of the circuit easier to understand. And PSUs are dirt cheap on the net anyway.:)

spec


This diodes still do not look right to me. With the diodes like they are it produces a pulsing DC wave. I though an induction heater worked by changing the magnetic field, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, at 100KHz?

**broken link removed**

2016_11_08_iss1_eto_induction_heater_ver1-jpg.102184
 
This diodes still do not look right to me. With the diodes like they are it produces a pulsing DC wave. I though an induction heater worked by changing the magnetic field, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, N, S, at 100KHz?

**broken link removed**

That is not correct Gary. The split transformer secondary, four diodes, and two reservoir capacitors are a very widely used power supply configuration that produces positive and negative DC voltage supply rails for the half bridge. By the way, the capacitors will need to be very large with +22V and -22V supply rails, about 1F. They would probably best be made by connecting 10 of 100mF (100,000 uF) capacitors in parallel. The reservoir capacitors would be a big cost unless you can salvage them from other equipment or if you already have them in your component spares box.:)

You are correct about the magnetic field in the coil being, N,S, N ... The half bridge does just that. For the N part of the magnetic field the top transistor in the bridge is turned on while the lower transistor is turned off. Then for the S half of the magnetic field in the coil, the top transistor is turned off and the bottom transistor is turn on. And so the cycle would continue so that you get a continuous N, S, N, S magnetic field in the coil at a frequency of 100KHz (or whatever frequency you find to be best to suit your particular coil).

I am more than happy to discus this circuit, but I would not like influence you one way or another, especially if you have doubts about this circuits suitability. After all is said and done, you have much more experience with induction heating coils than me. All my stuff is just theoretical.:)

spec
 
Last edited:
in the 12v zvs....
I found that the resonant frequency of the coil changes under load, frequencies run higher and less current is drawn when there is nothing in the coil... something like 1/4- 3/4 amp.

but under load , the res frequency goes down and current goes up... to like 5 amp... and you really hear the coils work(although it could have been my bad choke mentioned below)

and as the material gets hotter frequency rises a bit to maintain synch to the load. Because with the heat atoms become looser to spin

Also i should mention i had to put a choke on my power supply line as per the schematic, but i ended using one that wasnt big enough and allowed enough power spikes to travel back and damage my 14v 7A power supply(it now puts out 24v for some reason) ,

Also capacitors were added parallel to the work coil which also brought the current up more.

the 12v zvs was run by fast avalanche diodes cross connected to the gates ... in the Larger driver , they used a PLL, current detection coils, and optical isolation... also it had a microcontroller override (which I was told it was for a "soft start up", but did not have a chance to investigate in detail)

idk if these are problems or not for this design, just mentioning my experience to shed some light.... hope this helps a bit!
 
in the 12v zvs....
I found that the resonant frequency of the coil changes under load, frequencies run higher and less current is drawn when there is nothing in the coil... something like 1/4- 3/4 amp.

but under load , the res frequency goes down and current goes up... to like 5 amp... and you really hear the coils work(although it could have been my bad choke mentioned below)

and as the material gets hotter frequency rises a bit to maintain synch to the load. Because with the heat atoms become looser to spin

Also i should mention i had to put a choke on my power supply line as per the schematic, but i ended using one that wasnt big enough and allowed enough power spikes to travel back and damage my 14v 7A power supply(it now puts out 24v for some reason) ,

Also capacitors were added parallel to the work coil which also brought the current up more.

the 12v zvs was run by fast avalanche diodes cross connected to the gates ... in the Larger driver , they used a PLL, current detection coils, and optical isolation... also it had a microcontroller override (which I was told it was for a "soft start up", but did not have a chance to investigate in detail)

idk if these are problems or not for this design, just mentioning my experience to shed some light.... hope this helps a bit!

Yes, it does help. Thanks.

The current spikes you mention are classic symptoms of conduction overlap, where both output devices conduct at the same time. This as you say, can be limited by a choke in the supply lines which, incidentally, is inevitably required in Royer type inverters. Another way to reduce the current spikes is to use dead zone where one transistor conducts and is then turned off and before the other transistor is turned on there is a delay (dead zone). Dead zone has many benefits: more efficiency, less coil heating, greatly reduced transistor peak current and heating, less strain on the power supply. But you can overdo the dead zone so it is best to set it up for a particular coil.

The effects that you describe when an object is placed in the coil field is to be expected as energy is transferred from the coil into the object that is being heated.

I find this project fascinating and the more people with actual practical experience, as opposed to theory, that contribute to this post the better.:)

In my welder building phase, I got quite pally with the welder man at work. He taught me how to make beautiful seam welds in aluminum.

spec
 
Last edited:
in the 12v zvs....

Also i should mention i had to put a choke on my power supply line as per the schematic, but i ended using one that wasnt big enough and allowed enough power spikes to travel back and damage my 14v 7A power supply(it now puts out 24v for some reason) ,

!

I had a power supply problem when I first built my circuit too. Sometimes the capacitors blow up, sometimes the rectifier blew up. I changed the choke coils several times but in the end I decided RF was killing the 3300 electrolytic capacitors. I put a small ceramic .001 uf cap in parallel with every electrolytic capacitor that stopped the problem. The rectifier problem was the manufacture error they claimed bridge rectifier was rated 100 amps but test show they were not good above 20 amps. Power Supply has 12 caps 3300 each in parallel = about 40,000. Tests show Power Supply voltage is 21 VDC, the RF voltage is 70 volts on the same wires.
 
I had a power supply problem when I first built my circuit too. Sometimes the capacitors blow up, sometimes the rectifier blew up. I changed the choke coils several times but in the end I decided RF was killing the 3300 electrolytic capacitors. I put a small ceramic .001 uf cap in parallel with every electrolytic capacitor that stopped the problem. The rectifier problem was the manufacture error they claimed bridge rectifier was rated 100 amps but test show they were not good above 20 amps. Power Supply has 12 caps 3300 each in parallel = about 40,000. Tests show Power Supply voltage is 21 VDC, the RF voltage is 70 volts on the same wires.

With just 40,000uF toatl capacitance you would have a huge ripple voltage on the capacitors which would not be good for the capacitors or the power output of the induction heater. The capacitors neecd to be high ripple current, low ESR types. From the picture you show, I do not think the the capacitors you are using are man enough for the job, I am sorry to say. The conductors in the power supply etc need to be beefy and short too.

The rectifier diodes would also need to be beefy too. Assuming that your induction heater is around 3,000 watts, in very rough terms that would mean that at 22V supply lines the average current would be 3,000/22= 136 Amps. When you take into account all the factors, I think you will need at least 300A diodes.

It is never wise to sail close to the wind on one-off projects. You can optimize components for a production item though. Optimization is a complex task and in many circuits you can only do it empirically for certain areas.

And while I am in transmit mode only buy the sort of components you are using from a reliable source; after batteries, big capacitors are one of the biggest rip-offs.:banghead:

spec
 
Last edited:
There is a UCC21520 module available all that is required is attach 2 wires from the 15v 4a DC power supply then run 2 wires from each output to each IGBT. There is an adjustment for dead zone time. I dont see an OSC and it says nothing i see about an OSC imput.?

**broken link removed**

**broken link removed**
 
There is a UCC21520 module available all that is required is attach 2 wires from the 15v 4a DC power supply then run 2 wires from each output to each IGBT. There is an adjustment for dead zone time. I dont see an OSC and it says nothing i see about an OSC imput.?
Great board.:cool:

There is no oscillator. You would need an external oscillator like the one in post #60.

The inputs are marked, JINA and JINB at the left of the board.

spec
 
Last edited:
With just 40,000uF toat capacitance you would have a huge ripple voltage on the capacitors which would not be good for the capacitors or the power output of the induction heater.

Wouldn't metal film caps like a MKP type of something similar to them be more appropriate for this application?
 
Wouldn't metal film caps like a MKP type of something similar to them be more appropriate for this application?

Yes they would and that's something that should have been easily identified if any degree of up front design planning been implemented. :facepalm:
 
UPDATE of 2016_11_15: please see post #83

Wouldn't metal film caps like a MKP type of something similar to them be more appropriate for this application?

MKP (metal film polypropylene) would make fabulous reservoir capacitor but, by god, that would be a big expensive power supply.:) By the way, this is not a resonant design.

To have a 1V ripple voltage, say, at 100A current drain, 1F (1000,000uF) reservoir capacitors would be required and, as mentioned in post #66, this could be archived practically by 10 of 100mF industrial grade aluminum electrolytic capacitors in parallel.

Alternatively, two 12V series-connected lead-acid batteries could be used in place of the large capacitors, but that, and other approaches for the power supply, are best covered once the general architecture and bridge circuit is sorted.

As you imply though, in addition to optimizing the half bridge, the power supply for this induction heater will require careful design, and some local decoupling with MKP capacitors and ceramic capacitors would no doubt enhance the performance, especially efficiency.

spec
 
Last edited:
MKP (metal film polypropylene) would make fabulous reservoir capacitor but, by god, that would be a big expensive power supply

I don't know about expensive. I've bought some for my EDM circuit and they weren't that bad price wise. But they were from an electronics surplus place. When you want something to work and spend time blowing up unsuitable components, doesn't that add up? Buy what is correct once and your ahead of the game, or at least that's my way of looking at it. But then again I'm just a rank newbie at this stuff.

Just following along in this post and the others he's made about the same topic over time, I'm finding it hard to believe that he really made the Tesla coil he was trying to sell a few years ago. An induction heater is hard but not on the order of a Tesla coil. And if I remember correctly he claimed to have built that.
 
Last edited:
UPDATE of 2016_11_15; Please see post #83

I don't know about expensive. I've bought some for my EDM circuit and they weren't that bad price wise. But they were from an electronics surplus place. When you want something to work and spend time blowing up unsuitable components, doesn't that add up? Buy what is correct once and your ahead of the game, or at least that's my way of looking at it. But then again I'm just a rank newbie at this stuff.

Just following along in this post and the others he's made about the same topic over time, I'm finding it hard to believe that he really made the Tesla coil he was trying to sell a few years ago. An induction heater is hard but not on the order of a Tesla coil. And if I remember correctly he claimed to have built that.
Hi SB,

It's a question of practicability. To make 1F from polypropylene capacitors you would need one thousand 1,000uF polypropylene capacitors at £100 each = $100K US, for each reservoir capacitor and as you would need two reservoir capacitors (one for the plus 22V supply line and one for the minus 22V supply line), making a total of two thousand capacitors at a cost of $200KUS.:arghh:

On the other hand, 100,000uF aluminium electrolytic capacitors are $50US, and you would only need twenty, giving a total cost of $1000US.

Also, aluminium electrolytics would probably have a better performance, provided there was some, polypropylene and ceramic capacitors across the the supply lines at the right places.

Of course, you can reduce the above costs by using pulled and surplus components, but the polypropylene approach is simply impractical (and not necessary).:)

spec
 
Last edited:
In this application I do not see the need for such an extremely high power supply capacitance. Especially so given that in common commercial high current power supplies they only have a hundred to a few hundred uf per running amp at most on a 60 HZ iron core transformer based power supply.

In my books for a ~150 amp 22 VDC power supply 30 - 50,000 uf of lower ESR electrolytics backed by a few hundred uF of poly capacitors or even old metal can PFC capacitors to compensate for the HF load aspects would be a great plenty.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top