More induction heating

[Alberto]

Hello. I saw some threads about this topic but since my question is different I don´t want to change the topic of the other threads.

I want to start making vacuum triodes and to heat the titanium getter I need a induction heater. I have some research but all of them have a lot of turns in the coil and to heat parts acurately I think I only neeed 3 or 4 turns.

Also I think it would be around 300 watts. For the power supply I have thought in a modified MOT to get 12 or 24V wit a FUUUUULL bridge rectifier. I don´t know if I would need a big capacitor for filter. Since a power suppky of this power if a little expensive

I saw some schematics of ZVS for induction heaters, so probably I onlye need to change the values of some components to get that amount of power and for 3-4 tusn coil, but as you can see, I have no idea.

I will appreciate any advice.

best regards

[romy]

on mine a 80mm od, 70mm long, 6x1mm, 81/2 turns coil produces around 50 khz.

these formulas should give a very rough approximation (i would look them up i a more intuitive presentation than i can do here):

(1) L = µ0 x rel. premeab. x area X (turns)^2 / length

(2) F res = 1/(2pi) X sqrt(1/(L x C)) = 1/(2pi X sqrt(l x c))

so if you add turns and lenght they tend to cancel out, but turns prevail (lower F).

(somebody correct me if im wrong.)


how large are your objects and what temp are you aiming at? i understand these devices will work up to about 100 khz and tend to let out the smoke above that.

[Alberto]

Thanks for your answer!

It is for heating getters and other parts on homemade vacuum triodes. I think the maximun diameter of the triode will be arround 20mm

So the parameters of the coil change the frequency, but it has something to to with the power of the heater? I mean if I use a scoil with to many or too few turns, can I burn the circuit?

[romy]

yes, thats what i was getting at. if you play around with coils and decrease inductance too much (e.g. 2-turns, much smaller diameter) the cirquit cannot handle the switching speed any more and burns.

(mine burned for another reasons, but thats another story.)

[Alberto]

yes, thats what i was getting at. if you play around with coils and decrease inductance too much (e.g. 2-turns, much smaller diameter) the cirquit cannot handle the switching speed any more and burns.

(mine burned for another reasons, but thats another story.)

Thank you!

My idea is making a induction heater with only 2 or 3 turns  and 200 or 300watts to heat getters. But I have no idea how to make it because all the examples on the internet are for big coils with a lot of turns

[Anders Mikkelsen]

The standard Mazzilli ZVS driver does not work well with low values of work coil inductance, for various reasons. If used with small values of resonant capacitance, the operating frequency is too high, and if used with large values of resonant capacitance, the tank circuit impedance is too low for the circuit to work well. You could potentially use an output transformer to step down the output voltage of the ZVS to a level where it can drive a coil of low impedance, but the resonant capacitor would likely need to be on the primary side of the transformer to avoid introducing too much stray inductance between the MOSFETs and resonant caps, and the transformer would consequently have to be sized to handle the full tank VARs, i.e. many times the actual processed power.

A multiturn coil with a copper flux concentrator would be worth a try, but like the transformer solution this is not something that has been tried and documented, so you would have to experiment on your own.

I've experimented with induction heaters for getter firing and neon electrode bakeout, and I found that transformer coupled series resonant or LCLR can work very well. These circuits can drive any practical work coil inductance by correct sizing of the matching transformer/inductor, but they are a bit harder to design. If you only want a few hundred watts, then a fixed frequency design can be worthwhile, with a pot to adjust the frequency. Just make sure it doesn't drop below the resonant frequency, as loss of ZVS will massively increase MOSFET losses.

[petespaco]

Alberto-
   You mention wanting to use a work coil with only 3-4 turns.  As others have said, there is a good chance that the resulting low inductance will damage the drive circuit and I agree.  The issue, as I see it, is that the high frequencies (much over about 120 kHz) force the  Mosfets in the ZVS circuits spend too much time in linear mode causing excessive heating and subsequent failure.
  So, I ask myself:  "what's the big deal about wanting only 3-4 turns for a work coil?"
Then I answer myself: "I guess it's because he wants to concentrate the power in a narrow area".
If that is true, then consider a multi-layer, multi-turn coil design.
You can see how one works here:

/>This video was  the culmination of some experiments I did a few years ago to maximize power transfer for annealing brass rifle cartridge casings. 

[Alberto]

Thnak you both for your answers!

...

So I could use a mosfet conected to a 555 oscillator and choose the frequency of the coil right?

...

Yess. I didn't realise that I could make a lot of turns in a small volume.

[romy]

i didnt know what getter heating is untill i looked it up, but imho the "internet zvs" devices provide around 50khz and thats way too low to heat up such a small object fast. also i see commercial equipment starting a 3kw. why does it have to be heated that fast, btw?

anders, could you expand a bit an the "copper flux concentrator"? it wont be out of copper, right? do you need a custom made powdered metal piece or is there a simpler way? how is it to be placed on a helical coil? iv seen them as plates together with pancake coils.

[Alberto]

i didnt know what getter heating is untill i looked it up, but imho the "internet zvs" devices provide around 50khz and thats way too low to heat up such a small object fast. also i see commercial equipment starting a 3kw. why does it have to be heated that fast, btw?

anders, could you expand a bit an the "copper flux concentrator"? it wont be out of copper, right? do you need a custom made powdered metal piece or is there a simpler way? how is it to be placed on a helical coil? iv seen them as plates together with pancake coils.

You can see it in action in 1:04:37. It only has 200 watts

[romy]

so what is in the wooden box?

(i have not the slightest idea about the stuff you are doing, but im impressed.)

[Alberto]

No no I wish that video was mine!!!

I dont´t have idea what is in the box, this is what I want to know. I need a induction heater like this one to make my own triode valves.

[petespaco]

romy- as regards "What's in the box?"---
Here is the link to that video:

/>I see that, as recently as about a month ago, someone was still replying to comments.  And new comments were still being made as recently as one day ago.
Why not go to the video ask the person?
 If you do that, you should also ask what they mean by "to get a higher frequency". 
Actually you need to know a bit more than just "what's in the box", since they mention that they have a separate 12 volt gate drive power supply.  We MAY even be seeing heat sinks behind that huge bunch of capacitors.  Are the power supply rectifiers there or are they for Mosfets?

(My use of "they" attempts to avoid trying to guess gender of the technician from the voice)

[Alberto]

Yess I was one of these guys in the coments, but I asked for the diagram of the induction heater and the guy don´t answer.

I bought a induction heater like this one



I made a coil with the same inductance as the original one but smaller. It works at 12V and with a screw or a blade of a cutter it drains 3 or 4 amps. The blade get red hot in seconds, and all ferromagnetics material get hot, but the titanium wire no and I want the induction heater for heat the titanium wire to act as a getter

[romy]

the guy in the vid mentions higher frequency, right? what does you contraption run at? fortunately titanim has hight resistivity.

[Alberto]

Ok, finally he replied!

The circuit is that

https://markobakula.wordpress.com/power-electronics/500w-royer-induction-heater/

but the capacitor bank is 3,15microfad and it runs at 350khz

[RoamingD]

Hello, I have a suggestion for the power supply, by far the cheapest power supply that will output 200-300w at 12V with no problem are old computer ATX power supplies. You can get one at a second hand market for very cheap (most don't cost more than 5-10 dollars) or even for free 
A traditional full bridge rectifier power supply will require huge capacitors that come at a big price if you don't have them already, run a simulation on https://www.circuitlab.com/ and it will plot out for you the voltage and ripple.

[romy]

check my math: 350khz/3.15µf → 6.4µh. 

[klugesmith]

Math check: what are you trying to figure? To resonate at 350 kHz with 3.15 uF, L would be less than 0.1 uH.

[petespaco]

Re: "I made a coil with the same inductance as the original one but smaller. It works at 12V and with a screw or a blade of a cutter it drains 3 or 4 amps. The blade get red hot in seconds, and all ferromagnetics material get hot, but the titanium wire no and I want the induction heater for heat the titanium wire to act as a getter"

Well, 12 volts at 4 amps is only 48 watts gross  Probably only half of that for net power to the work.  Is that induction heater rated for higher voltages?  If so, go to the maximum voltage that the thing can handle and try again.  Many of us started out experimenting with induction heaters by using car batteries in series to get higher voltages.  Even garden tractor 12 volt batteries can output lots of current for a short period of time.
  One other thought:  Maybe you could flatten the titanium wire to make it more like a thin piece of sheet metal.  I seem to remember seeing getters shaped like that about 200 years ago when I was young. <G>

[Alberto]

Thanks for your answers.

At first I´ll try the new circuit with a car battery. But usually people uses higher voltage than 12V

In the link there is another link to the web of Mads and he uses 35V. If I need more than 12V I´ll use small lead acid bateries in series rated for the amperage needed

But Yes, for a 350khz resonance, the coil should have a very low inductance, so I think I missing something. Can I decrease the value of the capoacitance whitoud damagin the circuit?

[davekni]

One other thought:  Maybe you could flatten the titanium wire to make it more like a thin piece of sheet metal.  I seem to remember seeing getters shaped like that about 200 years ago when I was young.

Yes, a plain thin non-magnetic wire will be VERY difficult to heat with induction.  Vacuum tube getters are usually in the form of a closed ring to make induction heating reasonable.  Per WikiPedia, induction getter flashing is usually at 27 or 40MHz ISM frequencies.  Getters are usually thin nickel sheet metal U-channels containing barium or other getter material and bent/welded into a closed ring.  Titanium wire might work if formed into a closed ring.  Lower (than 27MHz) frequency might work too with sufficient power.  (Lower frequency might make it more difficult to heat getter without heating other tube structures.)

[Alberto]

Thanks for your answers.

At first I´ll try the new circuit with a car battery. But usually people uses higher voltage than 12V

In the link there is another link to the web of Mads and he uses 35V. If I need more than 12V I´ll use small lead acid bateries in series rated for the amperage needed

But Yes, for a 350khz resonance, the coil should have a very low inductance, so I think I missing something. Can I decrease the value of the capoacitance whitoud damagin the circuit?

Thanks for your answer. 

People used to heat everything's in the tube to degass the metal but, what effect does the frequency has in the type of heating?

[davekni]

People used to heat everything's in the tube to degass the metal but, what effect does the frequency has in the type of heating?

To heat everything inside a tube, lower frequency (well below 27MHz) is probably better.  I was surprised to read that 27 and 40MHz were used.  In general, preferred induction heating frequency is higher for non-magnetic materials and higher for smaller objects.  However, the only possible reason that comes to mind for going all the way to 27MHz is to minimize heating of other tube structures.  At that frequency and with an induction drive coil similar in size to the getter and placed as close as possible to the getter, the getter should block much of the magnetic field from penetrating deeper into the tube.
BTW, this is based mostly on reading and physics theory.  I've only dabbled with induction heating at home and built one tiny unit for work, to heat SS parts just enough to speed curing of epoxy.  I've never built a vacuum tube.  My last tube electronics project was ~50 years ago in high-school.

[romy]

Math check: what are you trying to figure? To resonate at 350 kHz with 3.15 uF, L would be less than 0.1 uH.

thank you, thats what i was getting at. (i messed up the calculation forgeting to put in nano instead of µ and i was one decimal point off anyway.)

now im getting 65.6 nH*. first i thought such a coil was not possible. then i found a wire loop calculator (using a totally different formula). apparently 1 turn/40mm/3mm wire will produce 67.1 nh.

btw, how do you get 3.15µf with 0.27µf caps?


* L= 4pi x 10^-7 x N^2 x A / length (for empty coil)


edit: not a single "calculator" yields the same result as an other and they seem not to adhere to the formula, e.g. L is in no way inv. proportional to lenght. in literature i see stuff like this, and thats only the beginning for a foil, for a wire it gets twice as long. what is the difference?

[Alberto]

Thanks for your answers.

Davekni, Ok I´ll try differents values to change the frequency, but Can I change the frequency (changing for example the capacitance) without damagin the circuit?

romy, Yes my calcularions is the same value for the inductance, but the guy uses a 4 turn 40 mm diameter (at leats that seems to me) so I´m not sure. I have asked him about that

[romy]

lets practice (40mm/4T/30mm long): 12.56x10^-7x1x16x1.26x10^-3/3x10^-2 = 844 nH

a calculator: 610 nH

https://www.translatorscafe.com/unit-converter/en-US/calculator/coil-inductance/?D=2&Du=cm&l=1&lu=cm&N=10

[klugesmith]

Let's suppose 350 kHz is the frequency of choice.
Designer can trade L for C to facilitate the drive electronics. Here's an example:

First we figured 3.15 uF and 0.0656 uH. Characteristic impedance sqrt(L/C) is 0.144 ohms.
For the getter heating application, if we wanted 2 kVAR that would be 17 volts and 118 amps.

Could get same magnetic field, frequency, and VAR with 3x electrical scaling (51 volts, 39 amps).
C, L, and impedance would be scaled down or up by 9:    0.35 uF, 0.591 uH, 1.30 ohms.
New coil could be 3 turns in same space as the original 1 turn.
Volts per turn and ampere-turns would be same as original.

Note that tank V and I can be much higher than the DC voltage and current going into inverter.

[romy]

this "trading off" where you square the factor is clear. but is there an optimal proportion? if you can chose the power supply characteristics. how does P=I^2xR influence the choice?

 

[Alberto]

Thanks for your answers. I had incomplete information. The circuit could work at 100khz, and the gates of the mosfets work with an independent power supply of 15 volts.

To make that I sould cut in the red cross and put the 15VDC with the + in the point A and the - in the point B right? Maybe it is a silly question but I´m still learning



And the guy uses IRL540N but I have IRFP250N. Wich one is more suitable? It makes a big difference?

I really appreciate all your help. The level of the forum is very high and I have not much idea

[davekni]

To make that I sould cut in the red cross and put the 15VDC with the + in the point A and the - in the point B right?

Yes.  Turn on 15VDC supply before main power supply.  Also, reduce 470ohm gate resistor values to increase frequency.  100ohm resistors would keep power dissipation under 2W/resistor given 50% duty cycle at 15V for voltage across resistor.  Might be OK to 100kHz that way (with 100ohm gate resistors).

And the guy uses IRL540N but I have IRFP250N. Wich one is more suitable? It makes a big difference?

IRFP250N has higher gate capacitance, so requires even lower gate resistance to achieve 100kHz.  Perhaps drop to 60 ohms, which will consume almost 2W each.  IRFP250N is higher voltage, but a bit lower current capability.  Allows supply voltage up to theoretically 60V (if no startup transients).  IRL540N requires supply voltage of 30V or lower.

Have you tried heating a closed loop of titanium wire?  I expect a closed loop will be required for any success at all.  Titanium loop should be parallel to induction coil loops (coaxial to induction coil).

If you are more comfortable with vacuum tube designs, why not build a high-frequency induction heater with perhaps a pair of 6L6GC or whatever you may have easy access to?  Will likely require lower power at higher frequency (1MHz or higher).

[Alberto]

To make that I sould cut in the red cross and put the 15VDC with the + in the point A and the - in the point B right?

Yes.  Turn on 15VDC supply before main power supply.  Also, reduce 470ohm gate resistor values to increase frequency.  100ohm resistors would keep power dissipation under 2W/resistor given 50% duty cycle at 15V for voltage across resistor.  Might be OK to 100kHz that way (with 100ohm gate resistors).

And the guy uses IRL540N but I have IRFP250N. Wich one is more suitable? It makes a big difference?

IRFP250N has higher gate capacitance, so requires even lower gate resistance to achieve 100kHz.  Perhaps drop to 60 ohms, which will consume almost 2W each.  IRFP250N is higher voltage, but a bit lower current capability.  Allows supply voltage up to theoretically 60V (if no startup transients).  IRL540N requires supply voltage of 30V or lower.

Have you tried heating a closed loop of titanium wire?  I expect a closed loop will be required for any success at all.  Titanium loop should be parallel to induction coil loops (coaxial to induction coil).

If you are more comfortable with vacuum tube designs, why not build a high-frequency induction heater with perhaps a pair of 6L6GC or whatever you may have easy access to?  Will likely require lower power at higher frequency (1MHz or higher).

Thanks for your answer.

So, If I use IRFP250N i have to change the 470ohm resistors with 60ohm resistors but if I use IRL540N the resistors should be 100ohms. Right, but excuse me again for my few knowlodge, if I want to supply the main circuit with 36V, wich one would be more suitable, the IRFP or the IRL?

Thank you!

[petespaco]

Please look into IRFP260N (50 amps) instead of IRFP250N (33 amps).
Most, if not ALL of the Chinese 1000 watt to 3000 watt ZVS induction heaters use the IRFP260N, as do I.
The value of this higher current capability means, (to me anyway) that, if you select an operational frequency that is on the high side for the Mosfet, making it turn fully ON a bit slower,  you get less heating of the Mosfet die.  The higher current value comes from the lower RDSon of the IRFP260n (40 mOhms) vs the IRFP250N at 85 mOhms, doesn't it?

[Alberto]

Thanks for your answers. But another question. Is there any problem if for the 35v supply I use a switching power supply? I have heard it could be problems

davekni I forgot to say, yes I was making a loop, What I wasn´t making is putting parallel to the coil. Now gets a lot hotter! No red hoy but maybe 200 or 300 ºC.

petespaco I already have the irfp250N. If IRFP260N maes a bid difference I could buy them They are not expensive, the problem is the shipping.

[davekni]

Right, but excuse me again for my few knowlodge, if I want to supply the main circuit with 36V, wich one would be more suitable, the IRFP or the IRL?

I home my previous answer was clear:

IRL540N requires supply voltage of 30V or lower.

Is there any problem if for the 35v supply I use a switching power supply? I have heard it could be problems

With separate 15V gate supply, should be fine.  Turn on 15V supply first before 35V supply.

The value of this higher current capability means, (to me anyway) that, if you select an operational frequency that is on the high side for the Mosfet, making it turn fully ON a bit slower,  you get less heating of the Mosfet die.

petespaco I already have the irfp250N. If IRFP260N maes a bid difference I could buy them They are not expensive, the problem is the shipping

IRFP260 has ~1.7x gate capacitance of IRFP250, so will turn on that much slower (unless gate resistor value is further reduced).  ZVS circuits generally work best with fast switching.  Switching at "zero" volts avoids the typical fast voltage slew rate issues associated with fast switching of H-bridge circuits.  Partially-on FETs during switching dissipate more power because Vds is higher.  At high enough frequency this extra switching loss will exceed the reduced loss of lower Rds-on.  Would require simulation and/or testing to know precisely which part is better at specific frequencies and currents.
At least for now, I'd suggest staying with IRFP250.

[Alberto]

Right, but excuse me again for my few knowlodge, if I want to supply the main circuit with 36V, wich one would be more suitable, the IRFP or the IRL?

I home my previous answer was clear:

IRL540N requires supply voltage of 30V or lower.

Is there any problem if for the 35v supply I use a switching power supply? I have heard it could be problems

With separate 15V gate supply, should be fine.  Turn on 15V supply first before 35V supply.

The value of this higher current capability means, (to me anyway) that, if you select an operational frequency that is on the high side for the Mosfet, making it turn fully ON a bit slower,  you get less heating of the Mosfet die.

petespaco I already have the irfp250N. If IRFP260N maes a bid difference I could buy them They are not expensive, the problem is the shipping

IRFP260 has ~1.7x gate capacitance of IRFP250, so will turn on that much slower (unless gate resistor value is further reduced).  ZVS circuits generally work best with fast switching.  Switching at "zero" volts avoids the typical fast voltage slew rate issues associated with fast switching of H-bridge circuits.  Partially-on FETs during switching dissipate more power because Vds is higher.  At high enough frequency this extra switching loss will exceed the reduced loss of lower Rds-on.  Would require simulation and/or testing to know precisely which part is better at specific frequencies and currents.
At least for now, I'd suggest staying with IRFP250.

Thank you for your answer and help (and your patience haha)!!

Ok, IRFP250N and 35V volts with 2 separate power supplys 15V in the gates and 35 volts in the main. The 15V probably will be swiching, but for the 35 I think I´ll rewound the secondary of a MOT a full bridge rectifier and a bunch of capacitor in pararlell. Since the power will be higher and the value of the coltage it hasn´t to be exact, I wont need a regulator right?

[davekni]

Since the power will be higher and the value of the coltage it hasn´t to be exact, I wont need a regulator right?

Correct.  15V supply current is low, ~0.3A with 60ohm gate resistors.  A simple cheap wall-wart supply should work fine for gate power.

[Alberto]

Since the power will be higher and the value of the coltage it hasn´t to be exact, I wont need a regulator right?

Correct.  15V supply current is low, ~0.3A with 60ohm gate resistors.  A simple cheap wall-wart supply should work fine for gate power.

Thabk you, but I meant the 35V power supply. Since a new one is a little bit expensive, I was thinking rewound the seocndary of a MOT, rectify and filter with a bunch of electrolytic capacitors. But do I need a regulator or with the full bridge rectifier and the capacitors is enough?

[petespaco]

Re: "But do I need a regulator ":
  Whether you need a regulator or not depends on how much voltage (and therefore power) you can afford  to lose as the current draw increases.
I rewound the secondary of a 600 watt MOT a few years ago and found a 15 percent voltage drop from about 31 volts OC to 22 amps.  And that's just loss at that transformer, of course.
I will try to attach a chart showing my results.

[Alberto]

Thank you. At 35V the current should be around 7-10 amps. So with that values, wich voltage and capacitance I would need?

[petespaco]

I suggest that you visit a website such as this one:
https://learnabout-electronics.org/PSU/psu12.php

Or:
https://www.physics.unlv.edu/~bill/PHYS483/power_supply_info.pdf
 or simply google:
"linear power supply design"   -since that appears to be what you want to use.

[Alberto]

I suggest that you visit a website such as this one:
https://learnabout-electronics.org/PSU/psu12.php

Or:
https://www.physics.unlv.edu/~bill/PHYS483/power_supply_info.pdf
 or simply google:
"linear power supply design"   -since that appears to be what you want to use.

Thank you!

EDIT: I have read the links. They explain very well how works the filter capacitors in a power supply. I have a idea about that. My problem is that I don´t know how much ripple is admisible to power the heater. I mean, of course the bigger the capacity the better but I don´t know how much it would be a acceptable value

[davekni]

I mean, of course the bigger the capacity the better but I don´t know how much it would be a acceptable value

For your induction heating application with separate gate supply, no filter capacitor is probably best choice.  Induction coil current will go up and down at 2x line frequency.  That is the way induction cook-tops work.  I had a 1kW ZVS oscillator that ran that way too, with rectified line voltage and no filter cap.

[Alberto]

I mean, of course the bigger the capacity the better but I don´t know how much it would be a acceptable value

For your induction heating application with separate gate supply, no filter capacitor is probably best choice.  Induction coil current will go up and down at 2x line frequency.  That is the way induction cook-tops work.  I had a 1kW ZVS oscillator that ran that way too, with rectified line voltage and no filter cap.

Thank you

Nice!! So I'll rewound the secondary of a MOT untill I get 35VDC after rectifying

[davekni]

Just thought of one caution if using no filter caps on rectified transformer output:  Switch power on secondary side of transformer, either before or after rectifier.  If the transformer primary is switched off, transformer inductance will create a voltage spike.  MOTs are particularly bad in that respect, running large magnetization current, so producing large spikes when primary is disconnected.
If primary will be switched off with secondary connected, either a TVS diode or filter caps are needed to absorb the transient energy.  (The FETs may be able to handle that transient energy themselves.  However, if they are already running hot due to high load, then they have less avalanche energy capability.)

[Alberto]

Just thought of one caution if using no filter caps on rectified transformer output:  Switch power on secondary side of transformer, either before or after rectifier.  If the transformer primary is switched off, transformer inductance will create a voltage spike.  MOTs are particularly bad in that respect, running large magnetization current, so producing large spikes when primary is disconnected.
If primary will be switched off with secondary connected, either a TVS diode or filter caps are needed to absorb the transient energy.  (The FETs may be able to handle that transient energy themselves.  However, if they are already running hot due to high load, then they have less avalanche energy capability.)

Thank you!

And another (probably) silly question. I said 15v to the gates because the guy in the video uses that, and you suggest using 60ohm resistors. But I have a lot 12V and 5v powers supply. Could I use one of these? In that case, wich resistor I should use?


And about the capacitors for the oscillator I have read that i should use WIMA MKP, but I have KEMET MKP X2 SH. Is there s problem with that?

And last, could anyone recommend me a book about learning basic electronics? I know ohm's law, the behavior of pasive components and that things.  But I'd like to learn more for example to stop asking these questions haha

[davekni]

And another (probably) silly question. I said 15v to the gates because the guy in the video uses that, and you suggest using 60ohm resistors. But I have a lot 12V and 5v powers supply. Could I use one of these? In that case, wich resistor I should use?

If you have three 5v supplies, wire them in series for 15V.  That presumes they are not grounded (have 2-prong plugs, not 3-prong).
12V might be OK, not less.  Trade-offs of lower (or higher) gate supply are a bit complex.  If 12V, reduce gate resistance as much as possible while remaining under 2W, so around 40 ohms.

And about the capacitors for the oscillator I have read that i should use WIMA MKP, but I have KEMET MKP X2 SH. Is there s problem with that?

WIMA vs KEMET doesn't matter.  Different MKP capacitors are designed for different purposes.  You need high AC current capability.  X2 caps aren't great in that respect (usually).  You could try them with good fan cooling.  If some overheat and fail, then buy replacements.  Look at RMS current rating in spec's.

And last, could anyone recommend me a book about learning basic electronics? I know ohm's law, the behavior of pasive components and that things.  But I'd like to learn more for example to stop asking these questions haha

I'm not much help here, having learned electronics many decades ago.  I do recommend trying things in analog simulation.  There are some fairly simple on-line simulators, and simulation programs free to download and use.  My favorite is LTSpice.  Probably not the easiest for novice learning, however.  There is good online forum help for LTSpice.  Whatever simulator you try, start very simple.  Verify Ohm's law with a voltage source and resistor, etc.

[Alberto]

And another (probably) silly question. I said 15v to the gates because the guy in the video uses that, and you suggest using 60ohm resistors. But I have a lot 12V and 5v powers supply. Could I use one of these? In that case, wich resistor I should use?

If you have three 5v supplies, wire them in series for 15V.  That presumes they are not grounded (have 2-prong plugs, not 3-prong).
12V might be OK, not less.  Trade-offs of lower (or higher) gate supply are a bit complex.  If 12V, reduce gate resistance as much as possible while remaining under 2W, so around 40 ohms.

And about the capacitors for the oscillator I have read that i should use WIMA MKP, but I have KEMET MKP X2 SH. Is there s problem with that?

WIMA vs KEMET doesn't matter.  Different MKP capacitors are designed for different purposes.  You need high AC current capability.  X2 caps aren't great in that respect (usually).  You could try them with good fan cooling.  If some overheat and fail, then buy replacements.  Look at RMS current rating in spec's.

And last, could anyone recommend me a book about learning basic electronics? I know ohm's law, the behavior of pasive components and that things.  But I'd like to learn more for example to stop asking these questions haha

I'm not much help here, having learned electronics many decades ago.  I do recommend trying things in analog simulation.  There are some fairly simple on-line simulators, and simulation programs free to download and use.  My favorite is LTSpice.  Probably not the easiest for novice learning, however.  There is good online forum help for LTSpice.  Whatever simulator you try, start very simple.  Verify Ohm's law with a voltage source and resistor, etc.

Thank you!!

Ok, since it will be difficult to me find a problem if it appears, I´ll stick with the exact components to get the best result. I don´t have 15V supply and you say 12V is not sure to be OK, but I have a lot 19V laptop power supply. Can I use it with a bigger resistor? wich resistance?

Thank you, I think I´ll get banned from this forum for making this questions

[klugesmith]

You can take 19 V laptop PS and reduce voltage to 15 V with a single-IC voltage regulator.  Should begin with an estimate of maximum current requirement.

[davekni]

Ok, since it will be difficult to me find a problem if it appears, I´ll stick with the exact components to get the best result. I don´t have 15V supply and you say 12V is not sure to be OK, but I have a lot 19V laptop power supply. Can I use it with a bigger resistor? wich resistance?

19V laptop supplies are convenient, and free when I collect them from recycling bins at work.  I use them for many projects.
Yes, I think 19V would be a fine choice for your ZVS gate supply.  100 ohm resistors will be just under 2W each.  One more detail:  At 19V and 100 ohms, power of 12V zener diodes in circuit will be about 0.42W.  Make sure zeners you use are rated for that with some margin, at least 0.5W rating.

Thank you, I think I´ll get banned from this forum for making this questions

Don't expect you will be banned, though you are correct that this forum is generally for users farther along in learning electronics.

I do hope someone else answers your previous request for beginner book recommendations.  Good luck with your project and with learning.

[petespaco]

Re: "book to learn electronics":
I too, learned electronics a long time ago, starting in about 1948.
If I were starting out today, I think I'd get one of those "learn electronics kits"

Here's one example:
https://www.sciplus.com/electronic-playground-48491-p?

[Alberto]

Thank to all of you!

I´ll use 3w 100ohm resistors to be safe (that is what I have) and the zener are 3w.

And this is why I´m making the triodes, for me that is the best way to learn.

[romy]

yes, learning electronics:

i made several attempts in my life. first was before internet. lacking a systematic education in the field i always got stuck pretty fast. last effort was one month ago, concentrating on power supplies and resonnant cirquits. i started with maxwells equations and prepared a list of symbols (two full pages). by the time i got acquainted with messrs. laplace, naba, poynting, shin et al, their complex conjugates and dyadic products i was done. now the little "j"-devil is haunting me at night.

zct, sct, zvt, ccm, llc, svt, zvs, svt, zcvt, lcc, zcs, elt, sp, dab, pol, pr, vfc, hhc, nr, vco:  abbreviations dont make it any easier.

there is a lot of material on youtube. today im trying to wrap my mind around this:

/>

[Alberto]

yes, learning electronics:

i made several attempts in my life. first was before internet. lacking a systematic education in the field i always got stuck pretty fast. last effort was one month ago, concentrating on power supplies and resonnant cirquits. i started with maxwells equations and prepared a list of symbols (two full pages). by the time i got acquainted with messrs. laplace, naba, poynting, shin et al, their complex conjugates and dyadic products i was done. now the little "j"-devil is haunting me at night.

zct, sct, zvt, ccm, llc, svt, zvs, svt, zcvt, lcc, zcs, elt, sp, dab, pol, pr, vfc, hhc, nr, vco:  abbreviations dont make it any easier.

there is a lot of material on youtube. today im trying to wrap my mind around this:

/>

In my case is because I like, like learning english, I´m doing to travel and learn this things better, so in that case is easier

[klugesmith]

Not much of a riddle in that video by Professor S B-Y.   
In that last picture, as current increased from unsaturated to saturated,
the incremental L value went down but the stored energy still went up.
We can't figure stored energy from (I^2 * L/2).   
We can get it by integrating the input power during the whole inductor-charging ramp: I * E dt.
With optional refinement of subtracting the power loss due to wire resistance: I^2 R dt.

Kind of like figuring the energy to charge a MOSFET gate, when charge is not proportional to voltage.

[romy]

its his method of teaching. see subsequent vid.

[romy]

I suggest that you visit a website such as this one:
https://learnabout-electronics.org/PSU/psu12.php

Or:
https://www.physics.unlv.edu/~bill/PHYS483/power_supply_info.pdf
 or simply google:
"linear power supply design"   -since that appears to be what you want to use.

Thank you!

EDIT: I have read the links. They explain very well how works the filter capacitors in a power supply. I have a idea about that. My problem is that I don´t know how much ripple is admisible to power the heater. I mean, of course the bigger the capacity the better but I don´t know how much it would be a acceptable value

i have been told 3% ripple by the seller on the 1800w heater.

[Alberto]

Thank you for your answer.

Ok, it´s done





The 2 "loop" wires are conecting all the capacitors in parallel

A and B are the wires coming from the drains and fast diodes. My concern is, Does it matter in what point of the capacitor bank I connect A and B?
C goes to the 2 chokes and the induction coil

Thank you

[klugesmith]

>>Does it matter in what point of the capacitor bank I connect A and B?

Your connection of work coil to capacitor bank at point C is not ideal.
The closest capacitor will get largest share of current, and the 20th capacitor will get smallest share of current.  It really matters, if you compare inductance and high-frequency resistance of the 20-capacitor bus wire pair to inductance of the work coil.
The connection of driver, at A and B, is much less important because the current is much smaller.

There's a good diagram of this issue in instructable by imsmooth.   Or just scroll down on this page:
http://inductionheatertutorial.com/inductionheater/induction3.html

Can you build the 20-capacitor bank cordwood style (look it up) with two sheets of copper clad perf board?

[davekni]

The connection of driver, at A and B, is much less important because the current is much smaller.

I agree that it is much less important.  However, I have experienced issues if wires to drains are too long.  Can end up oscillating at a much higher frequency than intended, one defined by drain capacitance and wire inductance.  Not likely an issue for you unless wires are longer than in the images shown.

Also agree that current between coil and capacitors needs to be better balanced.  Even just adding wires between the two banks at the coil end would help greatly.  Copper (heavy foil or sheet) for cap to coil connections is even better.

[Alberto]

Fuuuuuudge....


First of all, thank you again both of you for the advice, I knew it was a good idea to ask here before conecting it.

 I´ll use it only to heat getters, maybe 30 seconds long, i don´t know if that infromation matters.

The capacitors are linked with 2,5mm diameter copper wire. Removing all of them and resoldering in other support would be a PITA because now the legs are too short. I´m not sure if I have understood well, but a possible solution would be conct the ends of the bus in the other side?



I mean, close the llops and connect the induction coil to A and B? With no too much wire to stay close?

[klugesmith]

>> I mean, close the loops and connect the induction coil to A and B? With no too much wire to stay close?

Yes, as Dave said.   That gets you most of the possible improvement, with very little work.

[Alberto]

Thank you for your answer. Ok, I link the loops and the heater works, but cant get red hot the loop of titanium wire...




veeeery disappointing, until I tried spot welding the loop, and now is HOT AS HELL BABY!!! The problem was the loop of titanium has to have a good conection in the joint.

If any of you are ever in the south of Spain and need something induction heated, or a beer, you can count with me!

The mosfets or the chockes don´t get any warm. The 100ohm resistors are 5wats (ceramic) and get hot. The capacitors and the bus a little bit hot. But the coil gets veeeery hot. First I connect the 19volts power supply and then I set the main power supply at 34 volts and when I connet it the voltage drops to 20 volts and drain 11 amps. Now I´ll try with different coils, maybe with copper tubing. If I´m right, the high frequency produces skin effect and doesn´t matter very much if the tube is hollow right?

[Alberto]

Hi again. Reading the message of the guy of the video, his induction heater drains 7 amps at 35votls. What could be the reason for mi draining 11 amps at 20 volts?

[Alberto]

I mean, of course the bigger the capacity the better but I don´t know how much it would be a acceptable value

For your induction heating application with separate gate supply, no filter capacitor is probably best choice.  Induction coil current will go up and down at 2x line frequency.  That is the way induction cook-tops work.  I had a 1kW ZVS oscillator that ran that way too, with rectified line voltage and no filter cap.

With the switching power supply it seems work well, gets very hot, but worked,

But I think this heater doesn´t like pulsating DC. I tried with a rewonded MOT. 30 turns of 1,5mm2 wire to get 22 volts after the full brigge rettifier and after 2 seconds started to drain 40 amps (i´m not sure, I disconected quickly, but I think I saw 41 amps) and noww one of the mosfet is dead.

[davekni]

But I think this heater doesn´t like pulsating DC. I tried with a rewonded MOT. 30 turns of 1,5mm2 wire to get 22 volts after the full brigge rettifier and after 2 seconds started to drain 40 amps (i´m not sure, I disconected quickly, but I think I saw 41 amps) and noww one of the mosfet is dead.

That's disappointing.  With gate supply on first, I didn't expect any issues.  Only possibility that comes to mind is that a loose connection caused the ZVS circuit to suddenly drop power, causing a voltage spike across transformer inductance.  Sounds like switching supply is safer route.
Did the spot weld of the titanium wire loop break?  That's another possibility for sudden drop in power.  Just removing the loop by hand should be plenty slow to avoid any issue.
Presuming you noticed my previous caution that power needs to be switched on the secondary 22V side, not on line power.  Any turn-off of line power to transformer while ZVS is connected to secondary would produce a likely-damaging voltage spike.  (Of course, after current went to 40A, damage was presumably already done, so it didn't matter where power was switched off then.  Only matters for any previous switching using the transformer.)

[Alberto]

But I think this heater doesn´t like pulsating DC. I tried with a rewonded MOT. 30 turns of 1,5mm2 wire to get 22 volts after the full brigge rettifier and after 2 seconds started to drain 40 amps (i´m not sure, I disconected quickly, but I think I saw 41 amps) and noww one of the mosfet is dead.

That's disappointing.  With gate supply on first, I didn't expect any issues.  Only possibility that comes to mind is that a loose connection caused the ZVS circuit to suddenly drop power, causing a voltage spike across transformer inductance.  Sounds like switching supply is safer route.
Did the spot weld of the titanium wire loop break?  That's another possibility for sudden drop in power.  Just removing the loop by hand should be plenty slow to avoid any issue.
Presuming you noticed my previous caution that power needs to be switched on the secondary 22V side, not on line power.  Any turn-off of line power to transformer while ZVS is connected to secondary would produce a likely-damaging voltage spike.  (Of course, after current went to 40A, damage was presumably already done, so it didn't matter where power was switched off then.  Only matters for any previous switching using the transformer.)

Thanks for your answer.  Yes I connect first the secondary and later the primary to the 220 AC.  Maybe that was the problem

I hardly had time to out a piece of iron to try the heater when I see the rise of current

[davekni]

Yes I connect first the secondary and later the primary to the 220 AC.  Maybe that was the problem

Unless there was contact-bounce in the switch, turn-on shouldn't be an issue.  Only turn-off on the 220VAC primary side is problematic.  Whatever the cause, sounds like switching supply is safer.  (Or enough filter capacitance after rectifier to absorb any transient energy, say >=5000uF.)  So far my AC line powered ZVS oscillator hasn't died.  However, it runs directly on rectified 120VAC.  No transformer except utility transformer feeding my neighborhood.

[Alberto]

Ok ok I'll try whith some capacitors to filter.

Thank you

[petespaco]

Regarding the addition of some filter capacitors-
  I have four 8600uf 60 volt electrolytic capacitors wired in parallel with 10 gauge wire that I can easily apply to the rectified output (up to 48 volts) of various transformer secondaries.  I use them if I think ripple is too much for whatever project I am working on.  Not necessarily very scientific, but it seems to work.   One thing to consider is the inrush current needed to charge them up upon power up.  So far I haven't had any issues, though.

[Alberto]

Thank you. Since they are cheap and the shop sell minimum 10 caps I'll go for 10 50v 4700uf in parallel

[klugesmith]

>> One thing to consider is the inrush current needed to charge them up upon power up.  So far I haven't had any issues, though.

In Alberto's case, using a rewound MOT, leaving core shunts in would ease the inrush current.

[Alberto]

Thanks for your answer. Yes I have kept the shunts

[Alberto]

>> One thing to consider is the inrush current needed to charge them up upon power up.  So far I haven't had any issues, though.

In Alberto's case, using a rewound MOT, leaving core shunts in would ease the inrush current.

.

How much could be that current? Since the transformer could deliver 41 amps maybe I'll need a bigger rectifier (the actual ones is rated for 25 amps)

[Alberto]

Regarding the addition of some filter capacitors-
  I have four 8600uf 60 volt electrolytic capacitors wired in parallel with 10 gauge wire that I can easily apply to the rectified output (up to 48 volts) of various transformer secondaries.  I use them if I think ripple is too much for whatever project I am working on.  Not necessarily very scientific, but it seems to work.   One thing to consider is the inrush current needed to charge them up upon power up.  So far I haven't had any issues, though.

I have bought 20 2200 uf 63V capacitors to attach in pararllel. Should I have any precautions? or it doesnt matter the way I make since is low voltage and DC

Thank you

[davekni]

I have bought 20 2200 uf 63V capacitors to attach in pararllel. Should I have any precautions? or it doesnt matter the way I make since is low voltage and DC

Can't think of any issues or critical requirements for line-frequency filtering.  And your 25A rectifier should have plenty-high inrush current specification.

[Alberto]

Thank you!

[romy]

this and this:

/>

/>
i wonder if your heater could drive such concentrators.

[Alberto]

this and this:

/>

/>
i wonder if your heater could drive such concentrators.

Very interesting!!

So with the piece of ferrite you can concentrate the induction effect. I have no idea.

I´ll made some experiments but with others cheap induction heaters. Thank you!

[romy]

i reread the thread and got lost a little. so you made your own cirquit using a separate 19vdc gate supply and were running it on 35vac (from transformer) and a rectifier? and then when you lowered the drive to 22v (rewound) the thing failed?

[romy]

Yes I connect first the secondary and later the primary to the 220 AC.  Maybe that was the problem

Unless there was contact-bounce in the switch, turn-on shouldn't be an issue.  Only turn-off on the 220VAC primary side is problematic.  Whatever the cause, sounds like switching supply is safer.  (Or enough filter capacitance after rectifier to absorb any transient energy, say >=5000uF.)  So far my AC line powered ZVS oscillator hasn't died.  However, it runs directly on rectified 120VAC.  No transformer except utility transformer feeding my neighborhood.

do you have this documented somewhere? i found this, but there is no follow up:

https://highvoltageforum.net/index.php?topic=1658.0

[davekni]

do you have this documented somewhere? i found this, but there is no follow up:

https://highvoltageforum.net/index.php?topic=1658.0

No, that is still a project idea/design on my list for building some day   Only simulations completed.  No actual construction.  Was based on a smaller 1kW version I have built, used to feed an HV transformer for Jacob's ladder arcs rather than induction heating:
https://highvoltageforum.net/index.php?topic=831.msg5491#msg5491
Schematic is below pictures in that first post.  Should work for induction heating too.  Induction heating is an easier load to drive compared to arcs since arcs can start and stop more suddenly.

So with the piece of ferrite you can concentrate the induction effect. I have no idea.

I used ferrite cores for a low-power (~60W) induction heating fixture at work.  Haven't seen ferrites used for high power applications.  Perhaps because brute force power is easier and/or it would be too difficult to keep ferrites cool enough.  For your getter flashing project, seems that ferrites might be useful.  Depends on geometry of your getter and how close external ferrites could get, ideally from both sides of getter.

[romy]

i was actually referring to this: "So far my AC line powered ZVS oscillator hasn't died." i couldnt find anything on that.

[Alberto]

Thank you for ypur answers.

romy Yes, I gave the gates with a 19V latptop power supply and the rest of the circuit I used a variable swiching power supply that work great. Later to give the heater its own power supply I tried with a rewound MOT only with bridge rectifier, and one of the mosfet died in 2 seconds.

davekni Thank you. Any idea about how cut ferrite core? It seems very fragile. And if for exame I have to pieces (very flat) and I put togueder, will it work like one piece?

[petespaco]

Any idea about how cut ferrite core?

How would you get a gapped  ferrite core around your titanium getter if the getter is somewhere within a vacuum tube?

[davekni]

davekni Thank you. Any idea about how cut ferrite core? It seems very fragile. And if for exame I have to pieces (very flat) and I put togueder, will it work like one piece?

Perhaps brittle is a better description than fragile.  Ferrites are ceramic.  Reasonable strength but very hard, so not great impact resistance.  Only way I've found for cutting is with a diamond wheel and patience.  I have some diamond wheels for my Dremel tool.  They work well, though slowly.  For this application, constructing a core from multiple pieces is fine.  The tiny gaps due to rough surfaces are small compared to large gap for getter.  I often build large cores from multiple smaller ones.

How would you get a gapped  ferrite core around your titanium getter if the getter is somewhere within a vacuum tube?

I'm picturing a very large ferrite core, something like a U100/57/25-3C90, perhaps with a couple large I-cores attached, or another cut U-core.  Not a small gap, but probably still more efficient than no ferrite to complete magnetic path.

[romy]

how about molding it youself? what would be more appropriate here, iron powder or iron oxide (hämatite)? some aluminum, maybe?

i read that this cirquit will not drive such a concentrator. if a reason was given, i dont remember it. what might be the problem?

→ https://www.ebay.com/itm/164879009641?

[Alberto]

Thank you for all your help!