using 100 capacitors in parallel for high current...

[Glyph]

Some microwave ovens have a 1uf capacitors rated at 2000v in their klystrons cavity circuit. If that would work then they should not be to hard to get hold of.

the magnetron in a home microwave is too integrated for me to successfully pull apart its capacitor. The voltage doubling capacitor from what i know runs at 50/60hz and not suitable for induction heating.

But i've never seen a Klystron capacitor. Although to be honest i've never seen a klystron. I will look into it though, if this fits the bill i'll use that.

Would a klystron capacitor operate at far above kilohertz levels? sounds like it would run at gigahertz. Would it still be suitable for a kilohertz induction heater?

Have you thought about immersing them in transformer oil? I found this interesting thread on the topic when searching for "dielectric oil".

If that oil isn't carcinogenic or toxic then it sounds like a good solution. is immersion all that's needed or do i need to pump it through a cooling system too? if i can find a way to pump it (sounds like it would be highly viscous) then i might consider it. Although if its too expensive and/or bad for human health i'd have to think otherwise.

 

[Roff]

If that oil isn't carcinogenic or toxic then it sounds like a good solution. is immersion all that's needed or do i need to pump it through a cooling system too? if i can find a way to pump it (sounds like it would be highly viscous) then i might consider it. Although if its too expensive and/or bad for human health i'd have to think otherwise.

I don't know any more than you do, but you can do a little Google research. This stuff looks pretty good, but I have no idea where you would get it. I would probably look at motor oils.
I think the pumping question depends on how much surface area your oil container has, and how much air moves past it.

 

[crutschow]

Would a klystron capacitor operate at far above kilohertz levels? sounds like it would run at gigahertz. Would it still be suitable for a kilohertz induction heater?

The capacitor is likely used to filter the high voltage DC power supply to the klystron and thus is probably not capable of handling high AC currents. The klystron is a cavity oscillator, not an LC oscillator so has no capacitors in the RF circuit.

 

[ronsimpson]

I like WIMA polypropylene capacitors. I have used Panasonic polypropylene capacitors. Illinois capacitor are OK.
https://www.electro-tech-online.com/custompdfs/2008/04/WIMA_MKP_10.pdf
If you look at the data sheet for the MKP10 (MKP4, FKP4, FKP1) capacitors and other pulse caps from WIMA you will see graphs of AC voltage and frequency for the capacitors. With a little math you can get from AC voltage and frequency to current.

If you are into math, search for more information on dissipation factor and ESR for capacitors. WIMA and Panasonic have application notes for determining maximum current in a capacitor at set frequency. These notes may be slanted for CRT monitor applications where the current flow is a saw wave (for “S” capacitor)(or 20% duty cycle saw tooth for fly-back cap). There are also application notes on using capacitors in resonate power supplies. There are ways of getting from RMS current to heat in a capacitor. In your case the current is sign wave 100% duty so the math will be simple.

Most high current capacitors should not be used at their maximum voltage rating when under high current stress. (I know this the hard way) Say you put 100Vp-p @100khz in a capacitor and get 10C rise. The pk voltage is only 50 volts. If you use a 100 volt cap that is only 50% of the rating. If you apply a DC voltage on top of the AC, under low stress applications the AC+DC peak could safely get to 100 volts. Under high stress application the PK voltage should be de-rated some. WIMA seem to have this problem more than Panasonic.

I have paralleled up capacitors to increase current capacity. Leave some space between them for air flow. They will get warm in your application. Use large copper traces or planes to carry the current.

https://www.wima.com/EN/pulseselection.htm
Selection of Capacitors for Pulse Applications. This might help.

Capacitors have a “current rating” much like resistors have a current rating. That comes form power loss, heat rise, ESR*current, dissipation factor.

What voltage are you thinking of?

 

[Rolf]

The capacitor is likely used to filter the high voltage DC power supply to the klystron and thus is probably not capable of handling high AC currents. The klystron is a cavity oscillator, not an LC oscillator so has no capacitors in the RF circuit.

You are correct, the cap is in the voltage doubler circuit for the magnetron.
But the the higher the frequency the lower the capacitive reaction. So I would guess the limitation is the internal heat rise. Soaking it in oil for cooling would help but it might come down to the duty cycle, it might not be able to withstand 100% or even 30%.

 

[Glyph]

I don't know any more than you do, but you can do a little Google research. This stuff looks pretty good, but I have no idea where you would get it. I would probably look at motor oils.
I think the pumping question depends on how much surface area your oil container has, and how much air moves past it.

Thanks for your help. I'll look into the oils. At first i'll probably run the system in low power mode using forced air cooling and gradually crank up the power while monitoring temperature of the capacitors. If they get too high for comfort i'll switch to oil cooling.

Thanks for the idea, i would never have thought of it.

I like WIMA polypropylene capacitors. I have used Panasonic polypropylene capacitors. Illinois capacitor are OK.
https://www.electro-tech-online.com/custompdfs/2008/04/WIMA_MKP_10-1.pdf
If you look at the data sheet for the MKP10 (MKP4, FKP4, FKP1) capacitors and other pulse caps from WIMA you will see graphs of AC voltage and frequency for the capacitors. With a little math you can get from AC voltage and frequency to current.

If you are into math, search for more information on dissipation factor and ESR for capacitors. WIMA and Panasonic have application notes for determining maximum current in a capacitor at set frequency. These notes may be slanted for CRT monitor applications where the current flow is a saw wave (for “S” capacitor)(or 20% duty cycle saw tooth for fly-back cap). There are also application notes on using capacitors in resonate power supplies. There are ways of getting from RMS current to heat in a capacitor. In your case the current is sign wave 100% duty so the math will be simple.

Most high current capacitors should not be used at their maximum voltage rating when under high current stress. (I know this the hard way) Say you put 100Vp-p @100khz in a capacitor and get 10C rise. The pk voltage is only 50 volts. If you use a 100 volt cap that is only 50% of the rating. If you apply a DC voltage on top of the AC, under low stress applications the AC+DC peak could safely get to 100 volts. Under high stress application the PK voltage should be de-rated some. WIMA seem to have this problem more than Panasonic.

I have paralleled up capacitors to increase current capacity. Leave some space between them for air flow. They will get warm in your application. Use large copper traces or planes to carry the current.

https://www.wima.com/EN/pulseselection.htm
Selection of Capacitors for Pulse Applications. This might help.

Capacitors have a “current rating” much like resistors have a current rating. That comes form power loss, heat rise, ESR*current, dissipation factor.

What voltage are you thinking of?

excellent resources, thanks alot!

After some searching i'll think i'll go with illinois capacitor as they seem to have the best ESR for the capacitance and voltage levels i'm going for.

The calculated voltage as well as simulated voltage at peak load comes out to about 600-800volt. I intend to incorporate an overvoltage protection circuit that will block further power input if the voltages reach 1kV. The capacitors I'm thinking of using from illinois capacitor are rated for 3kV. I think that should be adequate to handle most spikes or other unforeseen voltage events.

To deal with heat i intend to mount the capacitors on 1/4" thick copper plates with heat sinking and forced air cooling on the capacitors. If it still gets too hot i'll use oil cooling as suggested before.


The total unit (if my math is correct and the specifications right) will have a working voltage of 3kV, ripple current handling capability of 200A rms, and a "typical" ESR of 0.00046 ohms or 0.46 milliohms.

Voltage max during use will be set by the electronics to 1kV.

 

[Oznog]

OK, first step, make up a company name.
Seriously. You can get a lot further with just a company name. People will take you seriously. No one will ever check up on whether your company exists.

Give them a call, explain your needs and for god's sake don't be shy about saying what it is. People like that have a high chance of annoying the saleperson. Say "We have a project involving an induction furnace operating at 120khz and calcs show the capacitor bank will need to handle roughly 100 amps RMS at that frequency".

Chances are you'd be able to get them to sell you what you need- IF you have the money of course.

There's a lot of merit to many small caps. The heat dissipation of one cap at 100 amps RMS is a big issue even if the ESR is very small.

However, you need to pay CLOSE attention to the capacitor inductance at 120khz, regardless of whether we're talking about one large cap or many small caps. It sounds like this circuit will be low voltage, high current, right? At 120khz even a small reactive impedance could throw your circuit out of tune or simply prevent high currents from happening.

 

[Glyph]

Oznog: I already addressed the company question back in one of my posts.


As for tuning the circuit i'm not too worried about inductance. The circuit has a feedback mechanism that finds the right frequency through the phase lag of voltage applied to the matching inductor versus the voltage detected in the capacitor. The inductance of the tank circuit actually changes quite a bit upon the insertion of a metal workpiece and so a feedback mechanism HAS to be implemented for this to work at all. The capacitor inductance itself from a mathematical standpoint can be lumped together with the inductance of the output coil and treated as one large inductance.

The feedback mechanism should handle any capacitor inductance that i get.

 

[ronsimpson]

Do I understand right, you are having problems getting high current capacitors? Where are you located?

If you knew what I have been doing last month…..
I got a semi load of power supplies I designed years ago. Scrapped them out for $0.84/ pound. There were thousands of high current capacitors. I needed the money and no one wanted the parts.

I also sold old stock for about $.03/cap. I am going to see the store I sold the capacitors to and I will see what they have. They probably want $1.00 each now.

I have some of these. Maybe a bushel of them.
.0047uf 1000vac 1250vdc Matsushita (very good capacitor)
.022 630V MKP10
6500PF 1500V FKP1
3300PF 2000 KP1

 

[Glyph]

getting lots of small value, high current capacitors isn't the big problem, its getting the single REALLY high current capacitor (200amps rms) thats the problem.

So i'm trying to string together the much more readily available small capacitors to equal the performance of the large one.

 

[Boncuk]

Nice looking caps.

Although their capacitance is too high for my application by a factor about a hundred million.

I need a small value cap (1uF) that can handle huge currents. this is because the cap is used in a resonant tank circuit who frequency depends on the size of the capacitor and inductor. too high a capacitance, and the frequency becomes way too low for my application of induction heating.

nice looking caps though.

How do you imagine to draw huge currents of a tiny capacitor? Remember: The capacity of 1F allows a current flow of 1A for 1 second. Vice versa you must pump 1A into the cap very fast to achieve that current flow.

May be you should try another approach. A nuclear power plant or something similar?

 

[Roff]

How do you imagine to draw huge currents of a tiny capacitor? Remember: The capacity of 1F allows a current flow of 1A for 1 second. Vice versa you must pump 1A into the cap very fast to achieve that current flow.

May be you should try another approach. A nuclear power plant or something similar?

Did you read the entire thread?

The problem is with ESR. For example, 300 amps through 1 milliohm will dissipate 90 watts.

Glyph, you have probably already considered voltage, but just in case you haven't:
1uF has a reactance of 1.3 ohms at 120kHz. At 300 amps RMS, that's 390V RMS, or 560V peak. Keep that in mind when you're choosing capacitors.

We're talking about sine waves here.

 

[Rolf]

getting lots of small value, high current capacitors isn't the big problem, its getting the single REALLY high current capacitor (200amps rms) thats the problem.

So i'm trying to string together the much more readily available small capacitors to equal the performance of the large one.

Have you contacted this manufacturer?
**broken link removed**
https://www.vishay.com/company/contacts/

 

[Glyph]

yeah i did.

Vishay is a huge company with many branches and subdivisions, the particular division that makes induction heating capacitors is "Vishay ESTA".

ESTA is a highly specialized division and infact its so specialized that almost no distributor of Vishay parts distributes parts from that particular division. the one that i did find only sold in bulk. I'm still tracking down a distributor and still making calls. The fun part is that while many of them are helpful knowledgable people i eventually encounter someone along the chain that didn't even know they HAD an ESTA division and so communication has been slow, but its going.

The problem with being a huge multinational corporation is that you're a huge multinational corporation.

unfortuantely many of their capacitors are designed for medium frequency application (<100kHz) and are unusable in my setup because my inductor would then become unreasonably big.

things are going forward but i'm going to assume for now that i won't be able to find a single unit distributor of ESTA parts.

 

[Rolf]

Also.......

yeah i did.
{snip}
unfortuantely many of their capacitors are designed for medium frequency application (<100kHz) and are unusable in my setup because my inductor would then become unreasonably big.

things are going forward but i'm going to assume for now that i won't be able to find a single unit distributor of ESTA parts.

There should be a supply of spare parts available for cooking stoves now, that some of them have gone high tech. But I don't know what frequency they operate on. These capacitors must certainly have a large safety margin and it should be possible to operate them on a higher frequency at lower power level. Is your operating duty cycle going to be close to 100%?

 

[Glyph]

Its a great idea to use replacement parts for induction cooktops as those capacitors can handle the current.

Induction cooktops run at 20kHz-50kHz and thus the capacitors might not handle my application.

but that being said, I did look into it. So far the only available replacement capacitors come integrated with the circuit board. cost of the complete assembly is twice the cost of me putting together the 100 caps or buying the pure capacitor from the before mentioned companies.

Nonetheless its still a good idea and if i could get my hands on a broken stove that still had a good circuit board then i might consider it.

If its rated for 500 amps then i might be able to push my 200 amps at higher frequency if i could cool them properly.


As for duty cycle, i'd need to know the specific capacitor specs to know if i'm getting too close to their maximum.

 

[ronsimpson]

Here is another bad idea:

I know the construction differences between high current and low current capacitors.
Also between low ESR and high ESR, low ESL and high ESL.
Basted on the similarity of the above capacitors it is likely a very high current capacitor will work fine at high frequency even if it was not built for high frequencies.
Once upon a time I got a phase correction capacitor from the power company. They had 6.5kV and 12kV capacitors. I do not know the current rating but the wires going to the caps were 1” diameter. I know they were not built for 100khz and not characterized for high frequency but that may be because no one ever tried it. If I had one of these capacitors today I could test ESR and ESL.

What did I do with this cap? I built a 3,000 volt 6 amp supply.