questions about magnets

[Dr_Doggy]

i think id rather stay away from mains and keep it on a nice 15amp powerbar, but just so im clear i have redone your pic with a few details,

did i connect the supply correctly?
I should use 20 fets to make the 5 series hbridges, what is happening there at the (BASE)?
what frequency and voltage should i calculate lc3 for? should i just match LC2 accordingly?



im relieved that its such a simple circuit, the only one i could find had all these weird things like plls and such, btw cool design for the work coil.

i still cant help but feel that there is a certain frequency i should be putting out at the work coil, does it really matter? is ther a range i should aim for?

 

[johansen]

there is no certain frequency.

the work coil is typically 2-10 turns of copper pipe, with 1 to 10 litres per minute of water flowing through it to keep it cool.

typical frequencies are 10-200Khz.
typical tank circuit currents are 50-1000 amps
typical tank voltages are 50-1000 volts.
typical Volts times Amps flowing in the tank circuit is 10,000 to 1 million.

the work coil is typically 80% of the inductance of the tank "inductor" which resonates with the tank capacitor.
this frequency can be calculated but the insertion of a piece of aluminum, a piece of steel into the work coil changes the inductance, and the frequency.

if you want 2 kw of heat delivered into a block of iron you will need about 300$ worth of parts.

the transformer coils connected to the gates of the mosfets are known as "gate drive transformers"
and you need to build an inverter to drive that transformer.
which can typically be done for about 5$ in parts.

take a look at DRSSTC tesla coil circuitry for ideas.
DRSSTC tesla coil inverters and induction heaters are about equal when it comes to the demand placed on the components.

 

[Dr_Doggy]

i have been noticing that it does look like a tesla coil circuit,

about that gate drive and inverter though, if there are 5 h-bridges in that circuit, do i use 2 transformers for every h-bridge or is that one transformer with lots of coils?
, and the inverter, what kind of output should it have, ie 110v 60hz?

im used to hbridges enough, just not so much seeing them in series like that...

 

[johansen]

one transformer per H bridge is the way to do it.

each H bridge has its own local capacitor btw. this is most important.
you need low inductance caps on the supply side.

the reason to use many in series is because mosfets are cheaper per unit of power in lower voltages.
if you want to run one H bridge you would need to use 500 volt mosfets, and the on resistance follows the voltage squared...

 

[Dr_Doggy]

k, i think i got it,

but once again, this looks like a coil circuit, and has high power demands like coil circuit, i wonder if i could step up and use a tc circuit with sparkgap switching, using a output of 24kv that gets me 75ma, think that would work?

 

[johansen]

I'm unfamiliar with a coil circuit.

Induction heaters are more than just components with wires.
A 2 kw induction heater will have a minimum of 20 pounds worth of hardware and electronics and cost you ~300$ in parts, plus 150$ worth of test equipment

i'd figure 50% efficiency for a spark gap induction heater, and that's on a really good day. (efficiency of the spark gap inverter alone)
then figure half of the power is lost in the work coil again.

modern induction heaters are transformer coupled and operate at 90+% efficiency, most of the power lost is lost to the I^2r losses in the work coil.
btw, a 1/4th inch diameter copper pipe at 120Khz has the same resistance as a 10 awg wire. so when you have 300 amps flowing through it you need to figure 90 watts per foot of pipe ends up in the cooling water.

 

[Dr_Doggy]

ok, maybe i wont use the spark gap then,

so far i got the rectifier to 120v 35a, with a smoothing cap of 100uf, for supply,
and 20 of the MOSFET N-CH 60V 50A for my bridges, closest i could find with that part number,
the workcoil will be similar to 1uh, like this one;
https://www.dw-inductionheating.com/wp-content/uploads/2011/10/levitation-melting-300x143.jpg


so just a few parts left;

what values/types of caps should i use for the work coil?
and those local series capacitors?

also about those transformers,
i see that one in the middle thats all connected up, but what are the specs for it?
and same thing with those 5 h-bridge transformers how should i spec them, and where does the drive coil attach?
i dont mind ordering proper ones over fumbling with winding my own, since all i know is transformers are complex things aswell..

 

[johansen]

The work coil resonant capacitor should be made from a whole lot of small capacitors in parallel.
if you google around you can find lots of examples how people have built them.

i build my induction heaters from .068uF 250vac capacitors, typically 100 or more in parallel. i may have bought all of the ebay supply though.
another option is to buy capacitors from https://www.celem.com/

read this guy's tutorial: https://www.mindchallenger.com/inductionheater/
and **broken link removed**

The local supply capacitors connected across the H bridges need to be rated for the rms current they will have to carry, to be on the safe side just figure they need to be rated for the same as the dc current flowing into the supply. Their parasitic inductance needs to be kept very low as well.

to drive the gates of the mosfet you need gate drive transformers, if you keep the frequency high you can use twisted pair ethernet wire and ferrite beads for a core, figure something on the order of a cubic inch in size.

the main transformers can be anything from ferrite yoke cores to powdered iron to ferrite toroids such as **broken link removed**
look around you may find something cheaper.
you do not need 5 cores but you will need 5 insulated primaries.

don't neglect the parasitic inductance of the transformers and you capacitor when calculating the resonant capacitance needed.

1uF resonates with 1uH at 159Khz iirc.

 

[Dr_Doggy]

ok i will aim for 1-10 uf on my tank caps and double the voltage going in.\
but how much capacitance should i use for those series caps again?


and back to that main transformer, and my verry first question, wouldn t 12v@200amps be ok on the workcoil, or would 1200volts@2amps be better? and what inducance should i aim it for, is there a minimum number of turns i can use on each coil?

also about that bridge transformer, \
just to get it straight, i get a single toroid, and feed each bridge with a coil from that toroid , of how many turns??? , then another coil of how many turns that connects to a second toroid that wraps around the workcoil, with how many turns ?


i just want to be sure, i have read a few pages now, and i am getting the jist of it, but not totally1

 

[johansen]

you need to read those tutorials to get an understanding of how this works.

realistically at 100 Khz, tank circuit currents are 100 amps or more, and 100 volts ac across the tank capacitor.

the transformers need to step the voltage down and they all pass the current by the turns ratio.
for the first iteration i would wind 10/20/40:1 transformers, it is very handy to be able to change the turns ratio while you are using the induction heater.

the number of turns can be calculated from the area of the ferrite cores, and the safe flux levels. figure 100 to 150mT.
note that this is frequency dependant to so it has to be calculated for the lowest practical frequency you intend to run it at.
transformers are not calculated from inductance, the inductance doesn't even matter, what matters is that the core doesn't saturate, and you don't have excessive leakage inductance causing excessive lagging power factor in the inverters.

practically you only need a single turn for the work coil side of it. simply wind a toroidal transformer and slip it over the half inch copper pipe you would be using to build the tank circuit. note that you can use more than one core, you can stack them together.
here is a photo of one way to do it: **broken link removed**
and **broken link removed**

the gate drive transformers are self explanitory, google that term and you'll find out how to build them.

 

[Dr_Doggy]

OK1, so i found this capacitor https://www.digikey.ca/product-detail/en/B32794D2106K/495-4135-ND/2182153

which is 10 uf, and matched voltage as yours aswell, but i noticed you mentioned using multi smaller ones, should i aswell?

also what about those series capacitor values!!

i think though im ready to get started, im sure ill bbl with lots more ??'s,, thnx for all the help for now though, all i know is this is going to be rad

 

[johansen]

do you really think that capacitor can handle 100 amps?

it would explode after 10uS

 

[Dr_Doggy]

an thats why i need a million little ones, k!

 

[Dr_Doggy]

ok, so im just worried about these caps still pls confirm that they will be ok,

tank = 399-5864-ND x 100

can you please pick exact capacitors for me though, for both the series caps and rectifier capacitor,
i was trying to get it but, i fear my values arent matching properly,

 

[johansen]

those should be fine but you can probably find the same thing for about half the price somewhere else.

the capacitors that need to go across each H bridge need to be low inductance capacitors. they can be anything or a combination of whatever capacitors you can get your hands on.
electrolytic, film, whatever. 400uF should be enough.

 

[Dr_Doggy]

UM k,are these going to be ok?

P5313-ND

i figure 4 per bank of 5 stages so 20 of those? i am just worried because they have low ripple currents...

also im not sure how to identify a low inductance capacitor.



EDIT: I found the low inductance,, will these work??
https://www.digikey.ca/product-detail/en/FK22X5R1E226M/445-8475-ND/2815405
having 5 stages at 400uf per stage, which is 80uf total, is that going to be ok??

also i just thought, these series caps only needed to be rated for 60v/5*2 =
24volts each since 120v rectified is 60v?

 

[johansen]

if you're on the 120vac line then that gets rectified to 180vdc. I though you said you were running off a 240vac line, which is what you need to be running off of to get a 2Kw machine. (thus drawing about 3 to 4 kw of power.. try that at 120 vac.)

there are lots of options for the capacitors. for bulk energy storage you can use 200 or 400 volt electrolytic capacitors from dead computer power supplies. ebay is a good place to look as well. across each H bridge however you need the low inductance capacitors. search ebay for 2.2uf 100v . a bag of 400 of them is 10$ plus shipping.

50 per h bridge should be plenty, and you can use the rest of them for the ac capacitors you need in series with the output of the H bridges, which i forgot to draw in the schematic.

 

[Dr_Doggy]

nope, 2kv, = 120v@15amps on the breaker.. is this still going to work for me?

 

[johansen]

2KW into a piece of metal is going to require minimum 3KW from the outlet, so that's 25 amps average, but the high current pulses drawn by a capacitive input filter will turn 25 amps into 40RMS or more. You can dump the bulk energy storage capacitors, but then you're going to have a power output that varies from 0 to ~1.5 times the average power draw. this voltage variation on the bus makes trouble shooting more difficult, it makes H bridge conduction losses higher, tank capacitor losses higher and the work piece may vibrate at 120Hz.

you don't have to have pure dc fed into the induction heater but it needs to be low ripple.
figure 50 joules worth of bulk energy storage at the rectified line, they of course need to handle the ripple current, which is about 20-30 amps.