Centre-tapped PWM matching Xformer, will it work? / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

Centre-tapped PWM matching Xformer, will it work?

1 2 3 next

Move Thread

LAN_403

Ash Small

Sun Mar 30 2014, 10:04PM

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

I was wondering how to make use of the PWM function on a TL494 to control power into, say, the LC tank circuit of an induction heater, and came up with this:

1396216248 3414 FT1630 Untitledlc Tank Driver

The basic idea is that you synchronise the frequency of the oscillator with the resonant frequency of the tank plus workpiece, and then turn up the PWM. The TL494 allows PWM from 4% to 48%.

I'm a bit concerned about what happens when PW is >48%. I imagine that in this scenario, some current will try to flow from the tank circuit into the secondary of the matching transformer. Am I correct that, if this happens, current will be limited to the magnetising current of the core? (no current can be transferred to the centre-tapped primary if both switches are off, although I imagine some substantial voltage could be 'reflected', but as long as the breakdown voltage of the switches isn't exceeded, everything should be ok.)

Is my thinking correct, or have I missed something?

I imagine the impedance of the secondary under these conditions will be quite high, but I'm not certain about this.

Any other opinions? Is this just another 'epic fail'?

EDIT: mosfet drivers are:

]0900766b80f9e21f.pdf[/file]
or

]ixd_614.pdf[/file]

teravolt

Mon Mar 31 2014, 07:56PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

Hi Ash, I think it should work fine if you use a transformer with a core. it is kind of hard to do if you use a tesla. I had a lot problems. think you might need some diodes in series or parallel with the collectors to prevent the anti parallel diodes in the FET from being forwarded. do you plan to use a ferrite core

Ash Small

Mon Mar 31 2014, 08:37PM

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

Hi Terravolt, thanks for your reply. I will use a ferrite core if I decide to build it (one is indicated in the schematic). Also, I was considering using IGBTs which (generally, at least) don't tend to have the freewheeling diode, although your suggestion of using diodes in series with the collector'source is a good one, and I'll do this if I use mosfets, which I think generally do have the freewheeling diode (I think I've go that the right way round).

This was just a 'thought experiment' to start with, but if others think it might work then I may well build something to test the idea.

I am expecting quite a bit of reflected voltage, though, depending on turns ratio, etc, so I imagine I'll need high voltage switches.

I think ideally this circuit would want some form of feedback to keep everything 'in tune', and then just dial in more 'pulse width' to increase power, or less to decrease power. This could be usefull for some induction heating processes such as, for example, 'heat treating' where the workpiece requires maintaining at a constant temperature for extended periods, or where the workpiece has to be cooled slowly over an extended period, or even (and this is something I'm contemplating) for holding molten aluminium at a constant temperature in a vacuum induction furnace before pouring.

I can see that this wouldn't work with an 'air core transformer', but I don't see why it wouldn't work for driving the primary LC tank circuit of a 'conventional' TC, for example.

Steve Conner

Mon Mar 31 2014, 09:39PM

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

You can't drive a parallel resonant tank circuit with a voltage source inverter. The inverter's output voltage wants to be a square wave, but the tank voltage wants to be a sine wave, so efficient power transfer is impossible. You have two options:

Convert it to series resonance using a series feed transformer, or a matching inductor. The latter is called the LCLR topology. Both options have been successfully used by hobbyists, there are several well-documented builds. Richie Burnett prefers the LCLR, I prefer the series feed.

In your topology, the leakage inductance of the matching transformer may well be enough to make it run as a LCLR, especially if you realise this in advance and design it that way.

The other option is to use a current source inverter instead, like the ZVS. The DC link inductor allows the output voltage to be whatever the tank circuit wants, while constraining the current to a square wave. But note that a current source inverter can't be PWMd by its nature. They are usually controlled by varying the DC link voltage with a separate buck converter or SCR controlled rectifier.

Ash Small

Mon Mar 31 2014, 10:42PM

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

Thanks for your reply, Steve. Do I understand correctly that by LCLR you mean I may need to add inductance to form the square wave into a sine wave?

I sort of assumed that the inductance of the matching transformer would sort of do this anyway.

I think I will try building something along these lines as you and Terravolt have both implied that it's not impossible to achieve.

It will also give me some experience with the TL494, which the majority of members here were advising me to use when I first joined this forum and was still learning the basics, but I was asking lots of questions regarding astable multivibrators built from discreets.

By the way, it's good to see you back after your absence. Do you have any relevant links to series feed transformers or matching inductors?

Steve Conner

Tue Apr 01 2014, 07:37AM

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

Richie Burnett explains the LCLR

Mindchallenger aka Iamsmooth describes his series feed project

For more info Google is your friend. Also look for the "OMG Induction Heater" thread on this forum.

Push-pull inverters tend to be a bit wimpy. The leakage inductance between the two half primaries limits the power throughput. If you want more than a few hundred watts, it's really better to use a half or full bridge as shown in the links above. You can still drive this from a TL494 using a gate drive transformer.

Ash Small

Tue Apr 01 2014, 09:52AM

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

Steve Conner wrote ...

You can still drive this from a TL494 using a gate drive transformer.

Steve, while I've not yet studied the links you provided in detail, the reason I got to the 'push-pull' topology in the first place is because, as Artlav pointed out in his thread, the PWM won't get through the GDT (the GDT switches each time the TL494 switches, rather than only remaining 'on' for the duration of the 'pulse width'.

As I understand it, you can't use the PWM function of the TL494 to control/reduce power if you are using GDT's, or am I mistaken?

Steve Conner

Tue Apr 01 2014, 10:03AM

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

You are mistaken, it works fine. The majority of PC power supplies used a TL494 driving a half bridge through a GDT, with voltage regulation by PWM.

The secret is "bipolar PWM".

Irrespective of the duty cycle, the waveform stays symmetrical, so it can pass through a transformer unchanged. The other kind of PWM ("unipolar") doesn't work with transformers because the waveform contains a DC component for duty cycles other than 50%.

teravolt

Tue Apr 01 2014, 08:13PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

Steve could he use ZVS like mazzilli flyback driver to create the circulating current needed. If he got it tuned the IGBT's would not have to do as much work.

Steve Conner

Tue Apr 01 2014, 08:56PM

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

Yup. ZVS driven induction heaters exist and work fine, except for the usual touchiness of the ZVS circuit at high power levels.

You don't even need a matching inductor, you can just use the work coil as the ZVS tuned circuit. To avoid the need for a centre tap, you can use two DC link inductors and connect one to each end of the coil.

1 2 3 next

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