Discharging microwave capacitor

[rharman]

I need to replace the magnetron in my GE microwave. Some references say just short the capacitor to discharge it and others say to do a slow(er) discharge with a 150k ohm 25w resistor.

I can't find a resistor in that size anywhere. What size (set in parallel or serial?) would be a good alternative?

Thanks.

 

[Steevo]

Microwave capacitors are even more fun than the old tube TV ones were.
I've been zapped by both, and saw a friend thrown to the floor by a 1980's microwave one.
I still can't get my hair to lay down right . . .

 

[zmaxmotorsports]

What size of cap is it?

 

[rharman]

GE Microwave HV Capacitor CH85 091uF 2100V

I had visions of lightning bolts doing a short-circuit discharge.

I'm probably overthinking this. The manual I just found mentions that it has a shunt resistor built in which should slowly discharge from non-use. Of course, they recommend shorting it with a screwdriver just to be sure.

Magnetron gets here Tuesday - already have the diode on site. Hope that fixes it.

 

[laser3kw]

in the older TV sets, we would attach a clip lead to a screw driver. You ground the clip lead and then work the tip of the screwdriver up under the hi voltage anode cap. When you were close, the hi voltage would jump to the screwdriver.
You could do the same, but make sure you are well insulated from any conductors.

 

[iron block]

Some microwave oven capacitors have built-in bleeder resistors. Check the label on the side of yours -- it may say "internal resistor", or it may show a little circuit diagram with a capacitor symbol in parallel with a resistor. (The resistor will probably be drawn as a rectangle, not a zig-zag.)

[Update - I see you posted that yours does indeed have this resistor. Good.]

So if you just leave the microwave unpowered for a good stretch, it will discharge on its own. If the internal resistor is 10 megohms, and the capacitor is about 1 microfarad, the time to discharge to a safe voltage will be on the order of several 10s of seconds. So let your microwave sit overnight for safety's sake and the cap should be plenty dead.

Of course, you will want to check to be safe. If you have a **good** voltmeter (safe to use with 2+ kV, in case something has gone horribly wrong with the cap's internal bleeder resistor), you could just measure the terminals. Otherwise, just short the terminals with a screwdriver. Wear protective glasses, again as a precaution against the internal resistor having failed.

Good for you to repair rather than junk it

 

[Richard Cranium]

Just short across it with a screw driver, Be sure you are holding the insulated handle

 

[rharman]

$200 to try the repair beats $1800 or more for a new one. It's worth the gamble.

I have a good voltmeter I can use to check it. No problem there.

Thanks everyone for the help/tips.

 

[Dragfluid]

Maybe wear some rubber gloves, too.

 

[Dragfluid]

$200 to try the repair beats $1800 or more for a new one. It's worth the gamble.

I have a good voltmeter I can use to check it. No problem there.

Thanks everyone for the help/tips.

WHAT??!!
Where in hell do you shop for microwaves??

 

[zmaxmotorsports]

GE Microwave HV Capacitor CH85 091uF 2100V

I had visions of lightning bolts doing a short-circuit discharge.

I'm probably overthinking this. The manual I just found mentions that it has a shunt resistor built in which should slowly discharge from non-use. Of course, they recommend shorting it with a screwdriver just to be sure.

Magnetron gets here Tuesday - already have the diode on site. Hope that fixes it.

Thats a big freaking incapacitator!

 

[zmaxmotorsports]

Maybe wear some rubber gloves, too.

 

[rharman]

WHAT??!!
Where in hell do you shop for microwaves??

GE Monogram Advantium built-in. Way more than just a microwave.

 

[6t7gto]

Holy ****!
http://appliances.monogram.com/ApplProducts/ADVANTIUM-OVENS

 

[pepi]

microwave capacitors are even more fun than the old tube tv ones were.
I've been zapped by both, and saw a friend thrown to the floor by a 1980's microwave one.
I still can't get my hair to lay down right . . .

lmao ...............

 

[chops101]

Don't short it with wires attached. The risk of potentially blowing any surface mount IC's and CPU's are very possible from the static pulse.

 

[AnthonyJ124]

GE Monogram Advantium built-in. Way more than just a microwave.

The wife an I actually just looked at one of these yesterday! Sounds like yours died. How long have you had it installed, do you like it?

 

[nehog]

... The manual I just found mentions that it has a shunt resistor built in which should slowly discharge from non-use...

And...

Just short across it with a screw driver, Be sure you are holding the insulated handle

Both of these statements are correct. If it sits more than a day it will be fully discharged. If more recently used, by the time you get it apart it probably will be discharged, but just short with an (old) screwdriver. Hold the plastic handle and keep your other hand in your pocket .

 

[rharman]

The wife an I actually just looked at one of these yesterday! Sounds like yours died. How long have you had it installed, do you like it?

Our appliances are all Monogram from our 1993 remodel. It's been fine. Only repair before this was to replace a light bulb on the microwave and a couple of defrost kits & switches on the refrigertaor. The Advantium looks like a great combo.

Because we spec'd all Monogram appliances in our remodel, we're kind of limited in replacement options due to sizes if we want everything to match. For example, we need a 27" microwave and oven and a 42" built in refrigerator. Plus, ours are white which is not a popular color now.

The exception to Monogram was our cooktop and downdraft. Went with Dacor. The downdraft died about 2 years ago - circuit board - can't get the part anywhere. No more Dacor for us. Wife wants a Wolf cooktop & downdraft to replace them.

Only downside to Monogram is the repair costs when needed. Our refrigerator needed a defrost kit 2 times - it was a design flaw. First one was warranty, 2nd was free parts but paid labor. They charge a small fortune just to ring your door bell and it's not even creditable to the repair cost. Plus, the labor rate is higher than other GE products.

We're probably just buying time now. 23 years is a pretty good run. Went out price shopping all new Monogram replacements yesterday. Yikes. Part of life though.

 

[Milton Shaw]

Check the door switches on that micro, on Advantium I have have seen more upper switch problems as the reason they don't heat. Those are a lot of fun to change the Mag tube on, they are buried inside the assembly.

 

[rharman]

Pretty certain it's the magnetron. It was buzzing for a while before it conked out.

At least, I hope so.....

 

[ForceFed70]

91uF really isn't all that much.

I owned a 1F 12VDC cap for many years.

Your voltage is likely higher, but it's still not a huge capacity. Any resistor in the 100+ Kohm will work very well. Figure out the voltage then do the math. If voltage is 120V then current is .0012A through a 100K ohm resistor. 1200V = .012A Also 25W is likely overkill. Yes, you'd need a resistor that big if this was a constant load but it's a load applied for only a few seconds. 10W would be more than adequate.

As mentioned tho - you don't need to do this with modern electronics anyway.

 

[Falcon67]

For those prices, it should double as a particle beam accelerator and matter transporter.

 

[LS6 Tommy]

If it didn't have a shunt resistor I'd say when you get done burning a hole in your screwdriver and most likely popping the capacitor, when you go to get the new one you can get a discharge resistor at the same time...

Tommy

 

[rharman]

For those prices, it should double as a particle beam accelerator and matter transporter.

Agreed......

 

[afbrian13]

We have a bunch of 1uf caps at work. 115vac 400hz power. Before we work on them you just ground them out. Even 5min after disconnecting power the most you'll get is a tiny little spark. Most times there's nothing, but it's not worth the chance of not discharging. I know this has been on here a few days, but figured I'd chime in since we do this all the time.

 

[rburke65]

Wha does a $1800 microwave oven do?

 

[rharman]

Everything? Repsonding to this would be helpful.....

 

[rharman]

Well...... Capacitor worries turned out to be for nothing. No spark at all.

BUT.... replaced the magnetron and microwave still no workie! Have ordered a capacitor. If that doesn't fix it, we're starting our appliance replacement with a microwave!

Got no spark at all attempting to discharge it immediately after testing the oven. Might be a power supply, but I'm hoping for the capacitor. It's only $40 compared to $275 for the power supply. I'll gamble $40 more but that is it.

 

[Dragfluid]

Wha does a $1800 microwave oven do?

Heats up leftovers DAMN fast, and then beams them over to you on the couch.

 

[nehog]

Dare I say... Proper diagnostics before replacing (expensive) parts?

 

[LS6 Tommy]

Dare I say... Proper diagnostics before replacing (expensive) parts?

C'mon, if we used proper diagnostics we wouldn't be suggesting shorting out a potentially good cap...

Tommy

 

[ForceFed70]

Dare I say... Proper diagnostics before replacing (expensive) parts?

I don't think you understand how expensive the equipment is that would be required to do a proper diagnosis.

Oscilloscope could probably be found used fairly inexpensively if you have the knowledge to work one. But an LCR meter is crazy expensive and if there is a requirement for a logic analyzer, most people would need to take out a loan to purchase.

There's a reason why nobody repairs this stuff. Between the time required, the cost of the components, and the cost of the diagnostic equipment - you're far better off just buying a new one.

 

[nehog]

I don't think you understand how expensive the equipment is that would be required to do a proper diagnosis.

Oscilloscope could probably be found used fairly inexpensively if you have the knowledge to work one. But an LCR meter is crazy expensive a...

You don't need an oscilloscope for basic diagnostics. A LCR meter, if desired is relatively cheap on eBay but is not really needed. Ears, eyes, a DVOM, and a few simple steps will eliminate virtually all parts except for the control system itself.

Start by checking basics, fuses, all switches (most common failure!) and move into the more advanced parts. Use the meter to ensure the transformer is not open (be careful, you can get a nasty inductive kick back on it...) same for the rectifiers and capactors--check with a meter.

Most failures are micro-switches for door or a fuse. The magnetron rarely fails, but can weaken with age. An abrupt failure of a magnetron is only going to happen if the filament opens, again a simple thing to test with the meter.

Disclaimer: I am an EE and understand exactly how these work.

Caution: Microwave ovens contain very high voltages, and can kill you. Never work with one energized.

 

[pedrodagr8]

I don't think you understand how expensive the equipment is that would be required to do a proper diagnosis.

Oscilloscope could probably be found used fairly inexpensively if you have the knowledge to work one. But an LCR meter is crazy expensive and if there is a requirement for a logic analyzer, most people would need to take out a loan to purchase.

There's a reason why nobody repairs this stuff. Between the time required, the cost of the components, and the cost of the diagnostic equipment - you're far better off just buying a new one.

An oscilloscope is unecessary for this job. Solid LCR meters are now relatively inexpensive. A hobbyist level meter (look under transistor tester or Mega328 tester) can be found on eBay for around $20. They are nothing to look at but they are well featured and can give you enough information to see if a capacitor is dead or not.

If you want to step-up to a pro-level LCR meter with selectably frequencies up to 100kHz then you are looking at around $80-120 for the Der EE DE-5000 depending on accessories. A far cry from the price of $400+ a few years ago. Note, that LCR (nor the tester mentioned above) are NOT designed to handle charged capacitors and certainly not 2kV charged caps.

In this case, he does seem to be going rather slip-shod about this. Not really testing much, just replacing parts and hoping that it works.

 

[rharman]

In hindsight, yes I should have done more diagnosis up front. Everything I researched, based on the systems, pointed to the magnetron - seemed to be a common point of failure. Had I opened up the box and ohmed it out, I probably would not have bought it. New capmay not fix it either but I'm not that deep into it yet. Hoping to be able to return the magnetron.

Live & learn. Well.... I'm old enough to know better already.... and I do.

 

[ForceFed70]

I guess it's been a long time. When I 1st considered buying a LCR meter (about 15 years ago) my jaw hit the floor and I quickly decided I'd never own one. Time to reconsider and look them up again I guess.

 

[pedrodagr8]

I guess it's been a long time. When I 1st considered buying a LCR meter (about 15 years ago) my jaw hit the floor and I quickly decided I'd never own one. Time to reconsider and look them up again I guess.

A few years ago, a company named Cyrustek released some ICs that make it easy for manufacturers to make really high quality LCRs for relatively affordable prices. The best one out there right now is Der EE DE-5000, it's rebadged as the IET Labs DE-5000 and DE-6000. Very high end device with read ability up to 100kHz (necessary for checking the health of Low ESR and Low Z caps). It has a variety of useful accessories, etc.

 

[iron block]

Start by checking basics, fuses, all switches (most common failure!) ...

Most failures are micro-switches for door or a fuse...

+100 for nehog. Things that move are the first suspects. Followed by things that get hot.

As an EE buddy of mine says, it's ALWAYS the connector. Might not be the case here, but I'd start looking hard at switches. Then check anyplace that connectors are soldered to a printed circuit board. Those solder joints are often prone to failure after enough heat cycling. The solder can crack in an annular ring around the socket pin and totally remove continuity. Hard to spot with the naked eye, BTW, so use a magnifier or just touch up all the joints with a big soldering iron.

Good luck.

 

[dclassical]

I don't think someone has said it and I realize that discharging the capacitor is not an issue anymore, but here is what I do. I put my DVM on Low-Z and measure the voltage across the capacitor, which will discharge it (and you will see the voltage go down).

I have also built a cable with an inline resistor like the one you mentioned (using heat shrink tubing).