Microwave transformer inrush current solution?

That would dramatically shorten life of cathode. It will lose emission.

Use SMPS with PWM regulation. That would allow for much smaller transformer, too.

Q: Why it is impossible to have sex in Red Square in Moscow ? A: Because every bystander idiot would be trying to give his invaluable advice.

Vladimir Vassilevsky DSP and Mixed Signal Designs

I presume that they're using a mechanical relay to switch the primary of the transformer, and that simply switching to a slow(er)-blow fuse or circuit breaker is not an option for some reason.

High inrush current is a function of the magnetization of the core from the previous cycle and the point in the AC cycle when transformer is energized.

You could use a commercial product such as these ones:-

Hello Vladimir,

Vladimir Vassilevsky wrote:

That is very interesting to know. Can I assume that this also happens when switching the transformer including magnetron element off and on? Because that's what these people are doing at the moment, and they may be interested to hear that they can prolong the life of their magnetron elements.

If this only happens in my scenario (so turning the filament off and on with HV present), then why is that?

...

You mean switching just the HV off and on, instead of the whole magnetron element, and keeping the cathode filament powered-up?

One problem is of course that they already have a complete set-up, transformers and all, and don't want to replace most of the hardware. But I'll certainly look for a HV switching solution.

Anyway, thank you for your advice already,

Best regards,

Richard Rasker

Keeping cathodes always at optimum temperature is one of the ABC truths of vacuum electronics. There are several degradation mechanisms:

• active layer cracking when heating up and cooling down

• ion bombardment when HV is applied and cathode is cool

• blowing up hot spots on cathode surface when emitting too much current from cold cathode.

That would be ideal. However I don't have numbers how magnetron life compares to useful life of microwave itself.

Vladimir Vassilevsky DSP and Mixed Signal Designs

On a sunny day (Wed, 26 Jun 2013 12:21:50 -0500) it happened Vladimir Vassilevsky wrote in :

Vladimir, in normal magnetron / microwave ovens the heater winding is on the same transformer core as the the HV winding, so switched at the same time. Magnetrons normally last many years (> 10).

yes, they make/made HV reed relays for this. Old microwaves had these to switch the HV side from the transformer on and off for the defrost cycle. I don't recall if the AC or side or the doubled up DC was switched though. The transformer and heater coils were powered up the entire time, so you never got crazy inrush current problems like you're facing.

same transformer core

In home/office use, 10 years translates to some 2000 hours of operation and 20000 on/off cycles. Not much.

Vladimir Vassilevsky DSP and Mixed Signal Designs

same transformer core

I've wondered about how that can work...maybe not so big a problem at kilowatt power levels. I can assure you that it is very bad at megawatt power levels! Magnetron life can be shortened to approximately zero in a hurry if HV is applied when ther cathode is not at full temperature.

...

They're already on the slowest circuit breakers feasible, and 99 out of a

100 times things switch on without tripping. But this still means that thigs go wrong every 15 minutes or so.

I know. Remanence combined with same direction field from new inrush current = saturated core = no more induction, just DC resistance (0.5 ohms). Hence the triac solution, with switch-on at mains voltage zero-crossing, and switch-off at current zero-crossing (plus perhaps a small extra timing factor to allow for magnetic hysteresis). Now if only those triacs didn't have this >1 volt voltage drop ...

Thanks, I'll contact these people for some pricing information (although experience tells me that these things won't be cheap).

Best regards,

Richard Rasker

These things will be running 8 hours daily, with an estimated 500 up to 2500 on/off cycles. Indeed a bit of a waste if they needs replacing every two weeks or so.

I'll look for switchmode (and switchable) HV generators with separate, continuous filament power.

Thanks for all the help.

Best regards,

Richard Rasker

try a suitable sized incandesent light bulb across the on/off switch/relay. When the switch opens the filament will be cold so the current through t he xformer will continue and will gradually reduce as the filament heats. In other words you demagnatize the core each time it shuts off so that it won't be able to saturate at the next turn on.

Mark

Is the dissipation in a triac really that much of a problem?

lots of stuff run of SSRs like:

-Lasse

How about timing the relay so it switches at the zero crossing? Switch on when the voltage is zero and switch off when the current is zero. With a constant voltage the time a relay needs to close the contacts should be pretty constant. With a microcontroller and volt / current sensing you could make it self-adjusting.

--
Failure does not prove something is impossible, failure simply 
indicates you are not using the right tools... 
nico@nctdevpuntnl (punt=.) 
--------------------------------------------------------------

On a sunny day (Wed, 26 Jun 2013 19:22:24 GMT) it happened snipped-for-privacy@puntnl.niks (Nico Coesel) wrote in :

I was thinking you need some thing like a big flywheel, motor that will work as generator for a few periods... Or a big LC circcuit.

Mechanical relays are too inaccurate for this kind of timing -- typically up to 10 ms (1/2 mains period) activation and switch-off time, with quite a bit of tolerance (20% easily), not to mention a bounce time of at least half the activation time. And indeed this would require a controller with software, current and voltage sensing, a low-voltage supply ... For now. it's probably easiest and cheapest to follow Spehro Pefhany's advice and simply buy a couple of commercial transformer relays, especially designed for the job. For the next generation of magnetron drivers, I'll try and find switchable HV supplies with a constant cathode filament voltage.

But thanks for thinking along anyway :-)

Best regards,

Richard Rasker

Your problem is saturation of the transformer. It is not precisely "inrush current", assuming this is the typical microwave where the tube itself is the rectifier, and there are no capacitors or DC anywhere.

If the power switch (either relay or SSR) turns on at the voltage zero crossing, then the transformer experiences a complete half-wave of magnetization in one direction. In the steady state, the transformer is fully magnetized at the voltage zero crossing, and spends a half cycle de-magnetizing before it starts to magnetize in that direction. An SSR rigged to turn on at the voltage PEAK would nearly solve this problem, the only little bit left would be the remanent field left at the last turn-off.

Switching the HV would work, but I doubt you could find a relay that could take this abuse for very long. Switching the filament might not turn the tube off, old radars often ramped the cathode heater off as RF heating provided enough to keep it hot. Anyway, applying the HV to a cold tube for long might cause other problems.

Jon

Nope, that GUARANTEES saturation every time. You want it to turn on at the voltage peak, as contrary as that sounds, so the transformer only gets a quarter cycle of magnetization in one direction.

Jon

Does the magnitron have a torroid transformer? There are gapped torroids just for this problem.

You could also try bleeding off the core to reset it during the OFF time. This may be the easist to try. Once the core is reset, you dont have to worry which AC cycle you start on. Likey the core is not reset and starting again on the same AC cycle saturates the transformer and pops the fuse.

Cheers

"Richard Rasker"

** Switching on at a zero crossing creates the LARGEST possible *magnetic* inrush surges with transformers - this is so because the AC wave takes a whole cycle to average to zero volts.

For minimum magnetic surge current, you switch on at a voltage peak - this way the AC wave averages to zero in a mere half cycle.

To make a transformer that does not surge with random switch on points, simply wind the primary for double the AC supply voltage - this has the same effect as using the transformer at double the usual supply frequency.

NB: In cases where the transformer is loaded with rectifiers and large capacitor bank, surge current due to charging the caps dominates if you switch on at AC supply peaks. This is not the case with the OP's question.

... Phil