Re: Microwave Oven Inverter
Conundrum, Wed Oct 21 2009, 05:26PM

i have a broken one here, iirc its the same as a MOT.

the problem seems to be that they overheat really easily and fail, so finding broken ones may or may not be useful.

the transformer is multicore primary around 20 turns, secondary 4*pie wound with around 20ga wire.

on mine the secondary failed near the ground tapoff...

regards, -A

Re: Microwave Oven Inverter
quicksilver, Fri Oct 23 2009, 11:54PM

Harry:
I have worked with some but just don't have the background you have to get in-depth with them.
On another working unit I set up an old HV wand and checked it out. It came from a rather new Sanyo I believe and put out 1500 @ 250ma. I used a carefully set stand of clear poly blocks and was quite cautious. It WAS pretty damn hot!There are two switching transistors on HEAVY heat sinks that get very hot. This unit was not pulled from a dysfunctional oven, I therefore believe that they run pretty hot as is. They are T220s (big ones) along with some very useful caps. I may just start gathering up quite a few of these as they contain some damn fine HV components.

Basically I was seeing 1500 @ 250ma: but this was a SMALL home unit. I have access to big restaurant units and would love to see what they have as I saw the unit itself once and it was substantially "beefier" that the residential type.

I pulled the flyback-type tx-former and played with it also. It's an odd design. After pulling the Lutz wire on one unit I experimented with the coil; I have seen this design in monitors and TV's but they seemed to be used as a choke. I used it in place of a AC flyback driven with a simple 12V HF unit and gotta small spark. The driver was of simplest design. Attempting to measure the output was a loss and I only used a cheap DIMM as I think that thing was toasted.

NOTE: I live in the USA therefore I'm putting in 120VAC, etc - The EU stuff may be much hotter.The current did not stay neatly at one figure but altered with time / temp more so than would expect.

Re: Microwave Oven Inverter
Proud Mary, Sat Oct 24 2009, 09:22AM

Thanks for that, Mr Quick. Very helpful.

Re: Microwave Oven Inverter
GeordieBoy, Sat Oct 24 2009, 11:13AM

Hi Harry, I think the reason you might not have got many replies to this thread is that the way these things work is quite complex.

As far as I know they take in the mains and rectify it, but don't smooth it (because that would cause poor power factor on the mains input side.)

They then chop this up at some high frequency to put it through a ferrite power transformer. This HF is also switched on and off with adjustable duty ratio at a few kHz to achieve the different power settings (0%, 20%, 40%, 60%, 80%, 100% or whatever is on the control panel.) This chopping on and off to achieve the different power settings is just a faster version of the usual on off buzzing that old microwaves used to do. Low frequency burst-fire like that is now difficult to implement without exceeding flicker noise limits for CE approval.

This HF signal is off-course modulated by the rectified mains voltage envelope too because of the lack of smoothing!

The HF output from the transformer is rectified using the HF equivalent of the diode-capacitor voltage-doubler that you normally see in a line-frequency microwave oven power supply. This doubles the voltage but leaves an output with lots of HF ripple on it. This is what is fed to the magnitron.

So... The microwave power output by the magnetron is 2.4GHz amplitude modulated by the HF switching waveform, modulated by the power control signal, modulated by the rectified mains ripple!

This complicated waveshape is probably quite hard to analyse or measure accurately by most amateurs. That's probably why you haven't got lots of simple V and I figures in response to your question.

-Richie,

Re: Microwave Oven Inverter
Proud Mary, Sat Oct 24 2009, 11:55AM

Thanks for you helpful insights, Geordie.

D'you think this complex OP waveform could be made into useful DC by a good old-fashioned L(swinging)CLC filter? (My intention to use one to supply HT for power valves?

Re: Microwave Oven Inverter
Wolfram, Sat Oct 24 2009, 12:37PM

Harry, is your inverter from a Panasonic oven? These seem to be the most common ones around here at least.

I have found one document with some information on them, here . Do you know anything about the topology used here Richie? It looks somewhat like a resonant single-switch converter with the top IGBT acting like an active voltage clamp or something along those lines. Sadly, these inverters seem to be of the constant-power type, and the control electronics will shut it off if the load does not draw the current that it expects. Modifications to the control part is not easy either, as everything seems to be controlled by an unmarked custom IC.

Richie, do you think it would be hard to make a new control section for it, so that it could be used to power linear amplifiers, voltage multipliers, VTTCs and the like? What would happen if I applied a constant frequency and duty cycle signal to the gate of the lower switch?


Anders M.

Re: Microwave Oven Inverter
Proud Mary, Sat Oct 24 2009, 02:32PM


That's the one, Anders!

The circuit diagram looks helpful, thankyou - but look what economists they seem to be - paralleling all those resistors near the magnetron rather than choose a more expensive single one of the right value (or maybe it is to keep the board profile lower for some reason, or spread the heat over a greaert area of the board?)

And to Richie - You're our man here, so what d'you think of Anders suggestion, please?

Re: Microwave Oven Inverter
GeordieBoy, Sat Oct 24 2009, 07:45PM

I had not seen that paper before and haven't actually worked on these myself. However, that design looks like an active-clamp resonant flyback converter. (The biggest problem with classic flyback converters at high powers is dealing with energy trapped in primary leakage inductance, and this is likely still quite innefficient!)

The bottom switch turns on and charges up the magnetising inductance of the transformer. Then, when the bottom switch turns off, the top switch turns on and the magnetising inductance forms a resonant circuit with the capacitor in series with the top switch. The natural behaviour of the circuit will be constant power because a fixed on-time of the bottom switch will ramp up the magnetising current to a fixed value storing a given number of joules on each switching cycle. This will be modulated by the mains ripple though of course!

These things are very application specific. A given power supply will most likely be designed around the nuances and quirks of a particular magnetron, and to interface to a given controller. The microwave oven market is also fiercely competitive, so they will be designed down to the absolute minimum cost possible to just manage the intended job.

Microwave manufacturers happily made appliances with big 50/60Hz iron transformers, half-wave voltage doublers and clunky burst-fire power control for many years. It seems nobody cared about the poor power factor, harmonic distortion, and pulsating load they presented to the mains supply. But now emmisions standards like flicker noise and harmonics have made them look elsewhere. The real driving force behind these new inverter driven ovens is standard compliance and safety, not new features, added value or better performance. And the much greater complexity and number of components certainly doesn't help reliability!

If it was me, I think I would just take the transformer and possibly the rectifier diodes and design a new full-bridge converter around those. That would be much more suited to running a linear amp, and I think hobbyists have done this before. Conventional voltage-mode control with current-limiting would give good line and load regulation.

Trying to modify this cheap supply to behave totally differently to how it was designed is probably going to take much more work, than starting from scratch on a more appropriate one.

-Richie,

Re: Microwave Oven Inverter
Proud Mary, Sat Oct 24 2009, 08:48PM

I have a couple of dozen 240 to 1kV neon sign inverters, and decided to do with them as Richie has suggested for the MOT inverter- toss out all the application specific parts and just save the output section. In this case, an opto-coupler links the one to the other in the mddle of the board, which invites simply sawing the board in half at that point, but I don't see such a simple option with this Panasonic unit.

Re: Microwave Oven Inverter
Herr Zapp, Sun Oct 25 2009, 12:46AM

Anders, Harry -

Note that there is a typo in Panasonic's schematic: the caps in the voltage doubler (C704 & C705) should be 8200pF, not 8200uF.

Also, can anyone provide a clear description of the function of Q702 and its associated components?

Regards,
Herr zapp

Re: Microwave Oven Inverter
Proud Mary, Sun Oct 25 2009, 01:00AM


I think Q702 is configured as a flyback or snubber diode. I'd guess that all those parallel groups of resistors of equal value are there either because it's cheaper to have two or three in parallel to achieve a given value of one single one of higher wattage, or to distribute the heat over a larger area of the board.

Re: Microwave Oven Inverter
Herr Zapp, Sun Oct 25 2009, 05:09AM

Harry -

As noted in the pdf document at Ander's link, this "appears" to be used as a some sort of a freewheeling circuit. But IGBT Q702 has an integral body diode, which would appear to be all that's required for a freewheeling function. I should have more clearly asked, is IGBT Q702 actually being "turned on" periodically through its gate? If so, exactly what happens when it conducts? And, when does Q702's body diode conduct?

It would interesting to know what the turn-on voltage is for the series-connected Zeners ZD703 & ZD704.

Regards,
Herr Zapp

Re: Microwave Oven Inverter
Proud Mary, Sun Oct 25 2009, 05:34AM


Hr Zapp: As with the paralled resistors being used for, I speculate, the sake of some economy, to achieve a lower value of higher wattage, I think the series zeners are in series simply to achieve some higher (more expensive) value. Considering these cheese-paring economy measures ("made to a price") it may be that the body diode is a completely irrelevant, redundant circuit element, there only because the diode came with the cheapest MOSFET that Panasonic could buy in bulk.

However, we have already seen that there are errors of notation in the circuit diagram, and it is not impossible that the 'series' zeners are in fact zeners-back-to-back - where we know of one error, there are probably others that we havn't yet spotted.

Re: Microwave Oven Inverter
Steve Conner, Sun Oct 25 2009, 07:15PM

I discussed this with Anders a while back. I believe the topology is basically a single-switch converter, and the top IGBT is some kind of lossless snubber.

The MO inverter is so optimized for its application that I'd personally be scared to try and use it for anything else, like running a valve PA. I remember Ralph Hartwell built a high-voltage, high power SMPS to power a Heathkit 1kW linear amp, and I'd make something based on his design from scratch, sooner than try to modify a MO inverter.

The only thing I might change about Ralph's design is to use a SLR or current-fed topology, to do away with the high-voltage output filter choke and its troublesome resonances.

Re: Microwave Oven Inverter
Zeus, Tue Nov 10 2009, 10:24PM

The Vout starts at 6Kv and drops to 1.5 to 2Kv.
I was told this by a MO repairman.
I'm in Australia which is 230V 50Hz