Microwave Oven Capacitors

[Silicon]

When I investigated a previously working Panasonic microwave oven that was blowing the internal 8A fuse I found that the 1.14uF 2100V AC capacitor had a dead short across the terminals.

The text printed on the side of the can includes the words 'SFM Resonance Capacitor'.

Is there anything special about this capacitor?
Why do microwave capacitors have such close tolerances i.e. 3% of their capacitance?

Could I temporarily connect a 0.92uF capacitor in its place to see if the rest of the microwave oven (i.e. magnetron , rectifier diode and the asymmetric rectifier) works OK?

[Peter.N.]

I think they are part of a resonant circuit with the magnetron so any change in value will effect the operating frequency. I'm sure someone will come along and confirm, or deny that.

Peter

[AC/HL]

The magnetron and the semiconductor diode form a voltage doubler. The capacitor value is probably critical to maintain the RF output power, in combination with the transformer characteristics. I think the magnetron itself defines the frequency. The circuit is deceptively simple!

[paulsherwin]

Although I don't doubt that you know what you're doing, it's risky to mess about with the HV circuitry of microwave ovens like this. They are extremely unforgiving if you make a mistake.

[G6Tanuki]

Originally Posted by AC/HL

The magnetron and the semiconductor diode form a voltage doubler. The capacitor value is probably critical to maintain the RF output power, in combination with the transformer characteristics. I think the magnetron itself defines the frequency. The circuit is deceptively simple!

The secondary of the transformer is essentially tuned to 50Hz by the capacitor; thats why its value needs to be so accurate.

[The tuned-secondary thing is both for transformer efficiency-vs-weight and to make it easier to comply with power-factor/not-causing-undue-waveform-distortion-on-the-mains regulatory issues]

[Radio Wrangler]

The transformer in a microwave cooker is designed to have a lot of leakage flux, this creates an inductive component to its output impedance (electric welders also use this trick to limit welding current) and it is this inductance which resonates with the capacitors.

A normal transformer would just give an output impedance that was a transformed scaled-up version of your mains supply impedance plus some winding losses. The magnetron oscillators have a negative resistance characteristic and some source impedance is needed to control the current.

David

[Silicon]

I have found a web site that partly explains the voltage doubling effect.

The text does not explain how the magnetron behaves in this circuit but it leaves me with the impression that the magnetron does not conduct until the voltage across it gets to a certain 'trigger' voltage.

It is certainly different to the conventional 2 diode and 2 capacitor voltage doubler circuits we are used to seeing.

I will wait for the correct part. Thanks for your replies.

[Al (astral highway)]

Originally Posted by Silicon

When I investigated a previously working Panasonic microwave oven that was blowing the internal 8A fuse I found that the 1.14uF 2100V AC capacitor had a dead short across the terminals.

The text printed on the side of the can includes the words 'SFM Resonance Capacitor'.

I've handled tens and tens of these for use in projects. They are usually 0.85uF or 1uF. Tolerance is indeed typically low, for nearly all the caps I have handled. Panasonic do seem to have distinctive practices and the capacitor you describe is an unusual value.

Before you go changing the capacitor - they're expensive - have a look and see if there is another component that looks remarkably similar to the EHT diode. It is a bidirectional diode, 2X062H

If there is, take it out of circuit and measure the leakage of the capacitor again. Be so careful: these things are literally one-touch lethal.

The bidirecitonal diode I am describing goes SC when it detects a leaky doubler capacitor, thus blowing the fuse and sparing the rectifier and transformer. They have been fitted on better quality ovens for a few years now.

The doubler circuit feeds pulsed DC around 5.9KV to the Magnetron, under load. Don't try to test it open circuit as you will risk the voltage rising too high and will blow the cap.

[G6Tanuki]

The key to mentally-modelling the circuit of the microwave-oven voltage-doubler is to imagine the magnetron itself as a diode-with-a-resistor-in-series.

There's no *physical* resistor in series with the magnetron of course but visualising it being there - as the 'load' the voltage-doubler is driving and into which the power is being driven (to emerge as microwaves...) - makes the working of the thing easier to understand as a half-wave voltage-doubler.

[Silicon]

I found an equivalent capacitor through Magsells on the Isle of Wight.
Same can dimensions but the blade terminals were 6.3mm instead of the original 4.8mm.

donberg.ie in Donegal may be able to supply a HV capacitor with the correct terminals.

The HV diode behaved like a rectifier on a breadboarded circuit but when I tried in on my Robin insulation tester it was leaky.

I tested the assymetric diode both ways at 250V, then 500V and finally 1,000V DC and it did not leak.

I fitted 6.3mm blade receptacles to the wires, replaced the HV capacitor and HV diode and the 10A plug fuse.
The microwave works now.

However I feel a bit uneasy about not checking the diodes more thoroughly.

Then again, I do not check every 1N4007 that I buy at 1,000V.

[julie_m]

See this earlier post of mine: https://vintage-radio.net/forum/show...2&postcount=14 This shows how the voltage doubling action is achieved.

Even if the diode (or any other part) fails catastrophically, the metal casing is designed to contain the damage -- and if you observe the minimum safety clearances specified in the appliance manual, it will not set fire to anything else.

[Boater Sam]

I think your insulation tester is not a DC output. Not a good test for a high voltage diode. We used to test with a 12v battery and a bulb.
If it works, the diode will be OK, they are pretty robust, polarity matters.

[Silicon]

I did a quick test on the insulation meter output.

The multimeter gave a DC reading but nothing when I changed it to AC.

[Al (astral highway)]

Originally Posted by Silicon

I tested the assymetric diode both ways at 250V, then 500V and finally 1,000V DC and it did not leak.

A good start, but not very conclusive.

The diodes have to cope with pulses close to 7KV and need to be 12KV rated, 500mA or higher.

Same deal with the caps. They may appear fine at low (1000V) voltages but go leakier any higher.

Note that they have to stand off a very high voltage indeed under no load conditions, which can be sufficient for them to go leaky almost to S/C.

[high_vacuum_house]

I presume at startup the power supply will be driving into an unloaded load and the voltages could go very high as there is no voltage regulation. This is because the magnetron is operating on the thermionic bases and the heater is wound on the same transformer as the HV winding. I presume it takes several seconds for the cathode to get up to its operating temperature whilst the HV circuitry will already be active.

Christopher Capener

[Richard]

If you measure the voltage on the magnetron it rises as the magnetron cuts in, which is the opposite of what you would expect. This as I understand it is the effect of the magnetic shunts in the transformer core.

Richard

[Boater Sam]

The whole circuit is a resonant design, transformer, capacitor and magnetron all have to be 'in tune' before any significant current is drawn. Which is why the capacitor is an expensive close value item. The diode is the simplest part, it just has to stand high voltage and current pulses.
I tried using the transformers for other purposes and they behave most peculiarly. All the ones I have seen have welded laminations forming shunts in the core.

[Radio Wrangler]

Radio amateurs keep trying to use microwave oven transformers as a cheap way of making high voltage power supplies for big power amplifiers, but they want stiff supplies which don't have significant output impedance, while these transformers look somewhat more like a current source.

Another way of looking at the voltage doubler is that a magnetron is basically a thermionic diode. The other diode in the doubler is a semiconductor one. But the magnetron-diode looks to be open circuit while its heater is cold. When the primary relay applies power, the doubler isn't doubling because it's short of a diode, and as the heater comes up to temperature, the appearance of the missing diode is matched by the appearance of the load as the magnetron starts to oscillate.

David

[Al (astral highway)]

Originally Posted by Radio Wrangler

Radio amateurs keep trying to use microwave oven transformers as a cheap way of making high voltage power supplies for big power amplifiers,

I can vouch for that, David. Even when teamed up with the load they're designed for, a magnetron, they are lossy and run hot. There is a fan on board MW ovens to deal with this but they do still dissipate a lot of heat.

Take away that ideal load and try to build a regular power supply and there's even more heat down the drain and not available for the load.

Some people remove the shunts.

For all their relative demerits, they are ludicrously expensive. I used to find mine regularly by the roadside and tested them in various ways to find out their foibles.

Now I'm done with them completely.

[steve1010uk]

I used to fix loads of these panasonic ovens, the capacitor going short circuit was a very common fault, i had a drawer full of them.

Fitting a lower value cap as a test would be fine, it would just give you less power, in the early days of microwave ovens the power level was set by switching different capacitors in circuit by means of a high voltage relay.

Steve