Inverter Microwave Oven circuit

[Carl Willis]

Does anyone have experience with the circuitry of Panasonic's "Inverter" microwave ovens? Being a high-frequency, high-voltage, high-power switching inverter, I'm thinking this magnetron supply would make a great driver for HV multipliers. The circuit appears to contain a hefty step-up transformer with the secondary wound in multiple "pie" sections separated by plastic insulation on a ferrite core. My thinking is that just the transformer is probably a very nice item to scrounge if these ovens show up on the curb.

This oven technology is rather expensive and has much greater availability in Canada and Australia than the U.S. it seems. Many of the ovens are rated at 1200 watts, with a total AC-to-microwave conversion efficiency of around 85%. So the step-up inverter is a healthy piece of equipment. Hijack two of these ovens for their circuitry, use the driver on one circuit to pump the switches on both boards out-of-phase, and you could get a balanced 2400+ watt driver for some of the large CW stacks that are selling inexpensively.

Any thoughts on this? Has it come up before?

-Carl

[Richard Hull]

Bill Connery, a friend, also noted that the modern day mini-TIG and some plasma cutters in the $250.00 range at Harbor freight are also high power inverter IGBT systems, ready to rock and roll.

Richard Hull

[Richard Hester]

I wonder what the motivation was for using a HF inverter for a microwave oven? The only driver I can think of may be overall efficiency, as the HF inverter will provide a square waveform to the magnetron, ensuring max power. The output voltage of this inverter may be a little low for Carl's purposes, since only about 2-3 kV is needed for the magnetron. A more appropriate match may be the guts from a high frequency X-ray tube head.

[Carl Willis]

Hi Richard,

The main motivation for the Inverter Microwave (a registered trademark of Panasonic actually) is to eliminate the weight and bulk of the large 60 Hz power transformer while adding continuous power control (probably through PWM).

The maggies I am familiar with need around 4 kV and get it from a "level-shifter" or voltage doubler. I'm uncertain on whether the Inverter ovens use any sort of voltage multiplier. The transformers look like they could handle a few kV, based on the photos I've seen. My particular multiplier stack uses a center-tapped HF input of up to 15 kV (7 kV each side of ground) to produce DC out at 100 kV. So my hopes are high. Even a single Inverter transformer, providing 4 kV, could drive half my full-wave CW stack and deliver fusor-usable juice at around 50 kV. Two of them, each driving half the stack, would merely decrease ripple and driver loading and increase the power capacity into the 2500 watt ballpark. I'd like to see this as a possible solution.

HF x-ray heads do have nice transformers but I question their ability to handle continuous power. The drivers are generally worthless for CW power applications because they are built for duty cycles of like 10 seconds per 5 minutes. Cheap, barely-adequate FETs like IRF540s are often seen in the dental head drivers, with a minimum of heat sinking and with heavy snubbering components that heat up a great deal.

I will be watching eBay for good, cheap Inverter microwaves on the market in my area. I do hope they live up to the potential I've identified.

-Carl

[Richard Hester]

Before you spend your money, it might be a good idea to pony up and buy a service manual for a typical inverter-driven microwave.
BTW, I cant imagine that weight is a big consideration, as you don't want your microwave to walk across the counter. I still think it's energy consumption, as the magnetron is probably not operating at optimum efficiency on the low end of the AC half cycle in a typical line frequency transformer setup. It probably draws current at low voltage without even bothering to oscillate. The same rationale is used to justify inverter driven X-ray tubes, as the voltages during the lower portion of the half cycle in a typical line frequency transformer driven tube generate mostly soft X-rays that have to be filtered out. An inverter can drive the tube at optimum potential all the time.

[Richard Hull]

Price out iron core, copper bearing 1KW transformers in today's market. Electronics and simple boost inverter systems will ultimately wipe out the 1kw and under power electronics market on classic iron transformers. This is well underway now. I see the monster microwave transformers at every single hamfest for $1.00-$5.00 each. they are a drug on the hamfest market. I recently bought a group of five MOTs at a hamfest. ($5.00 for all) I removed the spade lug wire attached 12kv, 1Amp diodes and hv caps and left the tranformers with the seller. He laughed. He still had them at $1.00 each at the end of the day. ( I already have 5-10 of these door stops in my lab.)

The above beiing said and with the handwriting on the wall, I will note that nothing will ever take the horrible, mindless abuse in any power system like an over designed iron core, copper wound transformer. This is coming from a long time transformer abuser and from a guy who realizes that a moment's inattention with solid state stuff will often leave the lab littered with pellets of silicon and shards of epoxy case material.

Richard Hull

[Richard Hester]

There is a sea change going on in the power supply industry that is being driven by energy efficiency requirements. California has more or less outlawed the use of 60Hz transformer based external adapters starting July of next year. Other countries and agencies are adopting the standards issued by the California Energy Commision. Having said that, most of the power supply companies would be squealing like stuck pigs if the astronomical rise in price of copper and iron had not made switching power supplies comparatively affordable (some squealed anyway).

I managed to find a representative (though not detailed) schematic of a Panasonic microwave oven inverter on the web.
A clever Aussie was eying one in the hopes of adapting it to supplying a high power radio transmitter. It's an interesting circuit using two IGBTs and not at all straightforward. Among other things, the inverter circuit operates in constant power mode, which presumably is the preferred method of driving the magnetron. Still ,the circuit can only be dialed back to around 40% of full power, then you have to go back to the old on/off scheme used by conventional microwaves for finer adjustment. It may be possible to gut the inverter circuit and use the components in a more suitable scheme for driving a multiplier. So far, I haven't found any repair houses on the web or on Ebay that are selling the inverter modules as replacement parts - it would be nice to have a couple of the modules to play with in my oh-so-abundant spare time.... BTW, the main driving forces were energy efficiency and better control (up to a point) of the magnetron.

[SteveHansen]

I was familiar with these for 12 volt portable (expensive)microwaves (camper and boat use) but it looks like they have gone mainstream over the past year or so. I've never played with one (although they can be purchase as spare parts from a number of mail-order houses and I'm sure will be showing up at the town dump sooner or later.
I've had great success using 750 watt inverters from Home Depot for making HV transformer drivers. Just decouple the inverter transformers from the output stages (rectify/60 Hz bridge) and connect them to a home made ferrite core step up.

[Carl Willis]

Hi Steve,

If you have any leads on which repair houses carry the inverter boards, I would appreciate it.

It's true that someone could design a better circuit for this application, but that does take some effort and the cost of parts and the time to make the board, etc. If the microwave oven idea is at all functional it would be quite useful. I think the transformers are a valuable part even if the rest of the board is not so great.

-Carl

[Richard Hester]

I was thinking much the same thing. The power supply circuit in the microwave inverter leaves a lot to be desired as a fusor driver, though the constant power feature might be interesting. However, you could burn up a lot of time and resources finding out the pros and cons of such an approach, and I assume what you want to do is light off your multiplier stack and use it to power your fusor, not dink around with power supplies for their own sake. A more straightforward approach would be to take the transformer and drive it using an IGBT half bridge. The only problem remaining is the control scheme, always a significant problem with switchers. The big boys use a resonant half bridge, which would be a pain to deal with and land you back in the "dinking with power supplies" scenario. An interesting approach would be to gin up a quasi-square wave inverter with fast shutdown in case of a fault like an arc. The HV transformer would be driven using a HF variac with a ferrite core. The big downside would be having to make your own variac, but they aren't that complicated, and a high power, high frequency variac would be pretty compact. You get the flexible control capability of a 60Hz-based system and skirt around the control issues of a closed loop HV supply.

[SteveHansen]

Here's one supplier. I found 2 or 3, these guys seem to be a bit cheaper.
http://www.amiparts.com/consumer/parts/hv_inverter.htm

[Carl Willis]

Hi Steve,

Thanks for that link. (Boy am them things expensive! Might as well buy the whole oven.) Maybe even more valuable is that "list of models" for the Kenmore ovens, which I was unaware used this technology.

Richard, I am not as knowledgable about electronics as you are, but I think the "constant power" feature may be effected through some marginal magnetic shunting of the transformer as is used in the 60 Hz equivalent. Steve Ward has found that a series-resonant primary circuit is nice to his H-bridges (he had to add inductance and capacitance in series with his transformer winding because the leakage inductance of his transformer was so small).

The transformer is probably good for 2000 V at least, at half an amp. I could probably get 40 kV out of my 8-stage multiplier with only one transformer. But I think one of the great things about these is that they probably have great potential to be reworked or potted in oil such that they can be operated with secondaries in series. The purported half-bridge on each board could be controlled from a "master" PWM controller. I hope to find out that the boards have gate drivers and gate transformers and some MOVs or other snubbing as part of the package. Potentially, a rewarding trip to the thrift stores or appliance exchange shed at the dump could yield up four of these boards on the cheap--which would pump my stack beautifully to 80+ kV at 6 kW and barely fill a small Rubbermaid oil bath. Unfortunately the home wiring is not ready for such a commitment.

-Carl

[Carl Willis]

This horse isn't dead yet so I'm going to beat it a little more.

Below is a partial schematic from a Panasonic oven service manual. A couple things are interesting about the circuit. First of all, the multiplier is not (as I had previously thought) of the same kind as that in an old-style oven. Here, the transformer charges two caps to the peak voltage of the waveform, and they are discharged in series into the maggie to give it 2Vpk. The caps are rated at 3 kVDC, 0.0082 uF each, suggesting that Vpk is maybe 2500 V. Now what is really great about this design is that the transformer secondary has to be isolated from ground to at least 2Vpk, and that if you had two identical inverters you could series the transformer secondaries even in air for the purpose of pumping a multiplier that wants more than 2500 Vpk.

-Carl

[bwsparxz]

Carl, check these links out:

http://www.repairfaq.org/sam/micfaq.htm

http://www.freepatentsonline.com/7064306.html

[Carl Willis]

I picked up some "dead" Panasonic inverter boards from the local oven repair shop (the output diodes are toasted but everything else is OK--apparently that is the weakest link).

They're disappointing in some respects. First of all, there's no half-bridge here, just one big FET or IGBT. There are no gate transformers: the switch transistor is driven directly. Almost none of the important parts have any sort of identifying numbers that would allow re-engineering of the board circuit. An unmarked control IC generates the high-frequency drive signal in response to an input that I think is supposed to be a 220 Hz TTL square wave of variable duty cycle. This input signal tells the board what power level to deliver over the range of 40-100%. Without a detailed understanding of the custom PWM controller, it is my conclusion that the inverter board as a whole is pretty un-hackable.

The transformers could be useful. Despite the large number of microwave oven models using inverters, there are only maybe two varieties of inverter board. Picking up a bunch of dead ones thus yields identical transformers, and as mentioned before, they are built such that they can be operated in series.

So it looks like I'm going to sack a few more boards for the transformers and end up making a custom H-bridge driver.

-Carl

[bwsparxz]

Carl, sorry to here the boards seem to be unhackable. Although like you said the ferrite is a keeper. I have also found 2 inverter boards to experiment with , both 1.2 KW. You could use a TL494 for the the drive signal, coilcraft has gate drive transformers as samples, and some HTNG40N60A IGBTs (600V @ 70A). Let us know what you come up with on those inverter boards.

[Carl Willis]

Hi Brian,

Thanks for the suggestion to look at Coilcraft for gate transformers. They have some decent looking ones with 1:1.5:1.5 windings for the 10-250 kHz band.

I can't locate a data sheet for the particular IGBT part number you provided. Do you know what the prefix HTNG denotes?

I've had an email conversation with Steve Ward about the HFMOTs from these inverters and pointed out that the core has an air spacer in it for some reason. He thinks that needs to be removed to operate the transformer with the SLR topology and I agree with his reasoning now that I think about it. Maybe the space was put in there to provide some leakage inductance and function like the magnetic shunts in a classic MOT? Anyway, it's bad news in an SLR circuit because it can result in high resonant current flow as the magnetizing inductance is very low.

Thanks for your suggestions. I'm watching your project very closely. What FETs are you using?

-Carl

[bwsparxz]

The IGBTs are actually FGH50N6S2D , available at digikey from $9.45 each. I have put in a order for some of these and can get the power supply running then. These are the same that Steve used. I have spoken with Steve several times about this project and he also told my about the problems with air gaps in the core. Basically using the series resonance topology- air gaps cause a no-load resonance that will surely eat up IGBTs. These HF supplies have alot of potential, Steve has gotten 15 KVAC from his in air. Surely he could push it more with the trannie in oil. Just as soon as I get the IGBTs and get results I will post here.

[Alex Aitken]

Could the air gap be there to prevent the core saturating and the resultant current spike that might also kill the drivers?