Induction heater from Instructables, or what determines the current draw? / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Induction heater from Instructables, or what determines the current draw?

first previous 3 4 5 6 next

Move Thread

LAN_403

Artlav

Mon Mar 31 2014, 09:17PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

Putting everything together:

The larger the load, the better is the heat transfer. At 1KW input the aluminium can is smoking at 200*C in 5 seconds, while everything else stays fairly cold, without any forced or liquid cooling.

As with an aluminium chunk, having a big piece of aluminium in the coil raises the frequency quite a bit, from 65KHz to 80KHz in case of a cola can.
I haven't planned for that, but it should be within the handling ability of the IGBTs.

I still wonder about power control.
Powering the thing from a variac makes ominous 50Hz hum from my 1KW one, and i don't fancy getting a bigger one.
Powering it directly from the mains would need extreme care in tuning to avoid overload - that implies always running out of tune, which AFAIK is far from ideal for the IGBTs.
I do have a 10A breaker on the board, so hopefully nothing will blow up too hard.

What alternatives are there to a variac?
Would a triac power regulator work for such an application?
These are rather small for a rating of a few KW.

Question is, what would triac-dimmed output look like rectified?
AFAIK, it won't be anything linear in voltage, with a lot more ripple instead?

What other power limiting methods are out there?

Artlav

Tue Apr 01 2014, 03:12PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

I'm fairly new here, is it ok to just write down build logs in here like that?

Anyway, measured the efficiency.
An aluminium can with 300g of tap water was heated from 33.5 to 37 *C in 10 seconds.
That adds to 4400J of energy, or 440W of heating power.
Plus 40J for heating 14g of aluminium.
Input was 144VDC at 6.4A, 920W.
That gives efficiency of about 48%.

Is 50% efficiency a good figure for an induction heater?
If so, no wonder they need water cooling of the coil...

Steve Conner

Tue Apr 01 2014, 04:21PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Artlav wrote ...

It draws 5A unloaded, 4A with a chunk of aluminium (which heats up eagerly now), and 2A with a steel bolt.
So completely opposite to a ZVS heater, which consume nothing unloaded and a lot loaded.
Is that normal?

Another peculiarity - with a big chunk of steel it looses power, but keeps the frequency.
With a large chunk of aluminium it keeps the power, but the frequency goes up.

This is all normal behaviour for a series resonant or LCLR matched induction heater. They do the exact opposite of the ZVS which is a parallel resonant, current-fed circuit. If you're building a high powered heater, you'll find that the current draw tends to infinity when the workpiece is removed from the coil, and some sort of current limiting may well be required to stop it blowing up.

Also your earlier statement about dead time not propagating through a GDT is false.

Artlav

Tue Apr 01 2014, 05:30PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

Steve Conner wrote ...
some sort of current limiting may well be required to stop it blowing up.

Would a circuit breaker at the input be sufficient?
Or are we talking about something way too fast here?
AFAIK, linearly interpolating the idle current to full working voltage gives 30-40A, well within IGBT handling capabilities, especially for the short time it would take for the breaker to trip.

Steve Conner wrote ...
Also your earlier statement about dead time not propagating through a GDT is false.

Hm?
In what way?

With the gates being driven like in the circuit from first post, i haven't observed any effect of the dead time on them.
There is a notch where it starts, and it moves around as i turn the POT, but the gates don't discharge or change in any useful way.

Steve Conner

Tue Apr 01 2014, 06:24PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

It's false in so far as plenty of switchmode power supplies successfully use PWM drive through a GDT, including about half of the PC power supplies ever made.

The problem is probably with the Instructables circuit. In particular, if you use a full bridge to supply gate drive to another larger bridge, the PWM may not make it past the first bridge. All four devices turn off during the dead time, allowing the second GDT to float. Really two of them should stay on to clamp the second GDT to zero voltage.

Artlav

Tue Apr 01 2014, 06:34PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

Ah, so it will be more correct to say that it does not propagate in this particular circuit.

Fortunately, it does not seem to change the point relative to the question asked back then - you can't use dead time POT to control the power in this circuit.

TwirlyWhirly555

Tue Apr 01 2014, 07:59PM

Registered Member #4104 Joined: Fri Sept 23 2011, 06:54PM
Location: Uk .
Posts: 122

Artlav wrote ...

Steve Conner wrote ...
some sort of current limiting may well be required to stop it blowing up.

When the work piece is in the work coil it dampens the resonance of the tank circuit so current draw is lowered the more the resonance factor is dampened , Its one reason a PLL is used sometimes in a induction heater , when the work coil is loaded the resonance of the tank circuit will change and to maintain full power throughput resonance must be followed .

In series resonance induction heaters when the work coil is unloaded the resonance effect will cause the impedance of the tank to be stupidly low so the current drawn from the supply will sky rocket , same thing happens as the work piece heats up to the curie point ( current will increase but not as fast )

With your the circuit you are using , power will be controlled mainly by the fact that you can manually tune the driver , but as the resonance of the tank circuit drifts if its not followed power throughput will lower .

I built a PLL series induction heater without current limiting ( not the best idea :P ) , when trying to levitate a piece of AL bar it fell out the work coil and with 300V DC on the half bridge the current exceeded the IGBT ratings in seconds and bang xD . It may be not a problem with big IGBT bricks , but its best to avoid the situation

Steve Conner

Tue Apr 01 2014, 09:01PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes, exactly the same thing happened to me except it was an aluminium bike chainring bolt rather than a piece of bar.

My driver had current limiting, but I was so eager to see levitation that I turned it up too high. The bolt levitated for a few seconds, then it melted and dropped out of the work coil, at which point my bridge exploded.

Depending on the size and thermal capacity of the IGBTs, slow-acting protection might be fine.

Artlav

Tue Apr 01 2014, 09:16PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

Hm, that is unfortunate. We just confirmed that dead time, and thus turn-off, do not propagate in this circuit.

So, if i use the tl494's inbuild capabilities to limit current at the input, what would really happen is gdt slowly decaying, driving the IGBTs across linear region a couple of times during overcurrent condition.
Not. Good.

I guess that's a big disadvantage of this circuit.

What sort of current limiting did you use?
Any ideas?

IamSmooth

Wed Apr 02 2014, 12:17AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Oh, I remember watching the current spike when the loads were removed.

First, if you plan on doing levitation and melting metals you should coat your coil with a thermal insulation so it will not short if the metal spills on it.

I put ultrafast breakers on my DC input to the inverter. This was there only to protect the board from shorting; the switches would be toast.

I found the best way to solve the current overshoot was using a microprocessor to monitor the current. If it went too high it would detune the circuit by increasing the frequency above resonance. Of course, this gets into a much more complicated circuit.

first previous 3 4 5 6 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob