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

Thu Apr 03 2014, 05:08PM

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

Hm.
Tried it with full 310VDC input.
Same can water test with 3A input current gives only 230W of power delivered as heat.
The current is set by detuning the circuit - at 150V it was basically at resonance, while at 300V it was above resonant.

So, 300V*3A = 150V*6A = 900W
However, at 150V i get 450W of heat, but at 300V i get 250W of heat.

Why is that?

Also, 3500uF input capacitance + rectified mains = tripped breakers at random places along the line. :(
Inrush current limiting is not optional.

IamSmooth wrote ...
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.

Nice point.
No levitation with this coil, however - it's already in a wrong shape.

IamSmooth wrote ...
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.

I guess i'll dump the Instructables driver altogether eventually, so something along these lines is likely (assuming i don't blow up the IGBT bricks by then).

Artlav

Thu Apr 03 2014, 08:01PM

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

No surprises the efficiency dropped.
At 300V and 3A the bridge heats up more than even at 140V and 8A.

Without any load and the coil way out of tune, the bridge consumes 1A at 300V, 0.4A at 150V.
Almost negligible 60W before, serious 300W at full.
Accounting for that, the efficiency stays at 50%-ish level.

But that is not supposed to happen, at all.
It must be shooting through.

I need a better driver.
With dead time propagation.

Looks like it's official - the Instructables circuit isn't worth the time it took to try it.

IamSmooth

Thu Apr 03 2014, 11:49PM

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

You have to know when to hold them and when to fold them.

Artlav

Fri Apr 04 2014, 10:20AM

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

Ok, trying to use the drive stage of the Steve Ward's DRSSTC driver (

), similar to

Some problems, however.

Gates:

Zoomed at the "dead time":

Input:

Input zoomed:

Instead of dead time at zero, it hits the opposite side's threshold voltage, and stays there for the duration then rises up.
Is that sort of thing normal?
What can be causing it?

I can think of two problems - first, all is on the breadboard for the moment.
Second, i'm using IRFZ44n as N fets and IRF5305 as P fets, and the two are far from identical.
The schematic lists two fairly different FETs as well, so it's probably not a problem.

Good thing is, even with gates acting like that, the losses are down to 0.4A at 310V.
Not good yet, but much better than it was.

Artlav

Fri Apr 04 2014, 02:40PM

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

All right.
Put it on a PCB, cleaned up the wiring.
Drive looks good, no spikes on either the gates or the IGBTs themselves in any mode - much cleaner than before.
The bridge stays quite cold after 10s at 1KW instead of being almost burning-hot.
No anomalous current draw with no load at any voltage.
Nice and clean...

However, the efficiency at 300V@3A stays about where it was - it went up from 230W to 320W, which sounds like the heat formerly warming the bridge have gone to the right side, but it's still quite a bit less than 460W i'm now getting at 140V@6A.
Even worse, i can't even find this heat any more - nothing but the can in the coil is heating up, much less with 600W worth of heat.

I'm missing something obvious here, ain't i?

Artlav

Fri Apr 04 2014, 04:19PM

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

Clamp meter on AC line shows 5.7A, on DC line - 3A, DC meter shows 3A.
DC meter is located between the bridge rectifier and the input capacitor.
Multimeter in AC mode before the bridge shows 3.3A (same line as clamp meter!), and in DC mode in place of the analog one - 2.7A.
The 150V runs are through a variac, the 300V ones are direct from the mains.
I doubt the power factor is 1, so multiplying A by V would give nonsense.
And so on.

Looks like i was just juggling random numbers, and assigning meaning to their changes.

Which begs the question of just how do you measure the efficiency (or "successfulness"?) of a device like that?
Should i concentrate on optimizing the "time to glow" parameter, ignoring the input instrumentation?

Ash Small

Fri Apr 04 2014, 05:43PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Artlav wrote ...

Which begs the question of just how do you measure the efficiency (or "successfulness"?) of a device like that?
Should i concentrate on optimizing the "time to glow" parameter, ignoring the input instrumentation?

I'd measure the old 'tin of water' temperature rise, and then compare to your other readings to see if any are close.

The 'tin of water' temperature rise is a pretty accurate measure of 'power out'. If youi can get any reasonably accurate measurements for 'power in', you can (roughly) calculate overall efficiency.

I know you've tried something along these lines before, but the 'can of water' has to be the most accurate way of measuring, and comparing, power out under varying conditions. This must be, in my opinion, the best way to compare differing settings, etc. (obviously, some heat is lost over time, but it will be difficult to better this accuracy). 'Time to glow' from a standard temperature, say 20 degrees C will also give a reasonable comparison, but more heat will be lost over time using this method, so it's unlikely to be as accurate a comparison, although it will still be useful.

IamSmooth

Fri Apr 04 2014, 06:23PM

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

Ash Small wrote ...

I'd measure the old 'tin of water' temperature rise, and then compare to your other readings to see if any are close.

The 'tin of water' temperature rise is a pretty accurate measure of 'power out'. If youi can get any reasonably accurate measurements for 'power in', you can (roughly) calculate overall efficiency.

If you do this make sure you wrap the can with thermal insulation. This will significantly eliminate heat lost to the environment. Black body radiation is to the fourth power of temperature, so this can be a lot.

Artlav

Fri Apr 04 2014, 07:11PM

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

Ash Small wrote ...
The 'tin of water' temperature rise is a pretty accurate measure of 'power out'.

IamSmooth wrote ...
If you do this make sure you wrap the can with thermal insulation.

That's what i am doing right now, for the output power.
Question is, how to measure the input power, and more importantly - what sort of efficiency to expect?

I can't really optimize the design without a reliable reference, like input power.
Sure, i can crank it up till the breaker pops, note it down, crank it up just below that, and measure output.
But i'd like something a little bit more definitive.

IamSmooth

Fri Apr 04 2014, 08:10PM

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

You know the input voltage from the mains just put the current clamp over the wire to measure the RMS current

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