Ceramic HV capacitor as primary capacitor / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Ceramic HV capacitor as primary capacitor

1 2 next

Move Thread

LAN_403

Flyglas

Mon Sept 03 2018, 09:49AM

Registered Member #62305 Joined: Fri May 11 2018, 03:39PM
Location: Germany
Posts: 6

Hello,

does anybody has experience with ceramic high voltage capacitors as primary resonant capacitor?
At the moment I am searching for door knob style mica capacitors for my qcw coil project,
but my ebay search was not very fruitful.

Example of HV ceramic capacitor: https://psearch.en.murata.com/capacitor/lineup/dhs/#a02

greetings
Flyglas

Hazmatt_(The Underdog)

Mon Sept 03 2018, 03:53PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

I only use doorknob capacitors for TC's because they are nearly indestructible.

Mica capacitors are ONLY good for CW power, not pulsed. Mica capacitors in TC use can short due to non-uniformity of their mica dielectric and cause a "hot spot" which will eventually cause early failure of the mica and will short your charging system.

I lost a small transformer due to mica caps, and I know a guy who killed his PT with mica caps, he still wants to buy back his other PT from me as they are exceedingly rare now.

I only run the ceramic TDK caps because you can abuse them and they will survive, its a massive chunk of Strontium Titanate ceramic, and they work fine.

Coil with TDK caps. They are behind the rotary.

Flyglas

Tue Sept 04 2018, 05:46PM

Registered Member #62305 Joined: Fri May 11 2018, 03:39PM
Location: Germany
Posts: 6

Thank you very much Hazmatt.

I did not know mica capacitors tend to build up hot spots. This is a good information.

I want to build a QCW coil and mica capacitors would be perfect for this.
It is very hard to find a suitable mica capacitors without paying a fortune. Because of that i will either try to use the door knob ceramic capacitors or building a classic MMC.

For my coil i only need around 12nF as primary capacitor. At the moment my plan is to arrange 11 assamblies of 20 parallel MKP capacitors (6,8n/1600V Kemet R76 series with lead spacing 15mm) in series. --> 12.36nF/17.6kV with a current handling capabiltiy of 40 A(rms) at 400 kHz

I can buy this typ of capacitor for only EUR 0.2 a piece.

Hazmatt_(The Underdog)

Wed Sept 05 2018, 12:48AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

You'll find one eventually, they always turn up.

Like this one, he's like 25 lbs of capacitor!

teravolt

Thu Sept 06 2018, 04:34AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

mica transmitter caps are exhalent but for a SGTC Maxwell caps are good to. door knob and Russian caps are good if they are pulse rated. Any cap should have a 200% voltage rating ie. 15kvac should use a 30kv cap and have a low esr, esl. and high di/dt rating. you also may balance that with cost mica is grate but expensive Any how this is how I would approach.

MRMILSTAR

Thu Sept 06 2018, 02:28PM

Registered Member #62119 Joined: Sun Feb 04 2018, 04:59AM
Location: Cedar Rapids, Iowa
Posts: 136

I second the Maxwell caps. That is what I use on my large SGTC. They are low ESR and this is what they are made for. Just be certain that you stay within the specs:

* Maximum voltage reversal
* Maximum voltage
* Maximum PPS (pulses per second)

My coil uses 6 of the Maxwell model 37667 (30 nF, max 35 KV, max 20% VR, max 500 PPS) in a series-parallel configuration to yield a 45 nF capacitance with max 70 KV. You certainly wouldn't buy them new because they are horribly expensive but they can be purchased used for a reasonable price. There are only two disadvantages. One is that, as used units, you don't know how much life is left in them but I have never had a problem with failure. The other is that it may be hard to find multiple identical Maxwells.

It is much simpler to connect a small number (6 for my coil) of Maxwell capacitors together instead of 100 or more for an MMC array. If one fails it is also a lot easier to isolate the bad capacitor.

Two of the Maxwell model 37667 in series would give you capacitance of 15 nF which is close to your 12 nF. It would also be rated for 70 kV at up to 500 PPS.

Hydron

Thu Sept 06 2018, 05:32PM

Registered Member #30656 Joined: Tue Jul 30 2013, 02:40AM
Location: UK
Posts: 208

Flyglas wrote ...

For my coil i only need around 12nF as primary capacitor. At the moment my plan is to arrange 11 assamblies of 20 parallel MKP capacitors (6,8n/1600V Kemet R76 series with lead spacing 15mm) in series. --> 12.36nF/17.6kV with a current handling capabiltiy of 40 A(rms) at 400 kHz

I can buy this typ of capacitor for only EUR 0.2 a piece.

This is very similar to what I have done - I used 256x of these capacitors for a 8s32p 24.8nF/12.8kV MMC:

(price has gone up a little since, thanks Brexit!). These caps give impressive RMS/Peak current and voltage ratings (even are rated at higher than nominal voltage for a limited duty cycle and time period during their life - perfect for a tesla coil application!) - from memory I think it ends up at ~60A RMS / 600A pk (at ~350kHz), more than I need.

Your design sounds pretty sensible, and the use of film caps also similar to what other QCWs have successfully used. Depending on what peak current you're aiming for, you may not need such a high DC voltage rating for yours (though the cheap caps make it mostly academic).

As for mine, 24.8nF is a pretty large value for a QCW MMC - I'm experimenting with a lower impedance tank circuit than is usually used in a QCW coil. It has not been pushed anywhere near it's limits yet, but seems to work well so far. I have mounted mine on PCBs, with 32 caps per board and 2x series SMD bleed resistors (1206 high voltage rated ones) underneath each cap. If you're interested in the design files then send me a PM.

Finally on a completely different topic, are you trying to reverse engineer an Weller WX series iron on another forum? I recognise the username Flyglas.

Flyglas

Fri Sept 07 2018, 04:13PM

Registered Member #62305 Joined: Fri May 11 2018, 03:39PM
Location: Germany
Posts: 6

MRMILSTAR wrote ...

Two of the Maxwell model 37667 in series would give you capacitance of 15 nF which is close to your 12 nF. It would also be rated for 70 kV at up to 500 PPS.

Thank you. I will check the Maxwell model 37667. Perhaps I am lucky and can by two of them for the suggested 2s configuration.
You are completely right with the argument of finding defect capacitors in a big MMC.

Hydron wrote ...

Your design sounds pretty sensible, and the use of film caps also similar to what other QCWs have successfully used. Depending on what peak current you're aiming for, you may not need such a high DC voltage rating for yours (though the cheap caps make it mostly academic).

Thanks for the feedback on my 11s20p configuraton. The current rating 60 A RMS (with 32p) and 40 A RMS (with 20p) is also plausible. Around 2 A RMS per capacitor. It was not my goal to have a 11s configuration. The number results from the current rating, the capacitor value (i only get 6.8n/1600V on the cheap) and the desired final capacitance.

I will also order PCBs to connect the capacitors. My idea is to design a round PCB and mount the capacitors along the circumference (with the bottom to the center of the circle). So there is only one potential on every PCB. At the end I will have a big rod out of capacitors. This assembly is completely symmetrical and i hope to achieve an even current distribution. Perhaps this is not really necessary.

Hydron wrote ...

Finally on a completely different topic, are you trying to reverse engineer an Weller WX series iron on another forum? I recognise the username Flyglas.

Yes, that's me. I have build a solder station replacement to use a WMRP soldering handpiece

During this time i contributed to this thread

I visted this link today and the thread is now much longer. You printed my hand drawn WXMP schematic :)

MRMILSTAR

Fri Sept 07 2018, 07:29PM

Registered Member #62119 Joined: Sun Feb 04 2018, 04:59AM
Location: Cedar Rapids, Iowa
Posts: 136

Probably not as cheap as you can find on E-Bay, when you can find them, but here is a source for used Maxwell 37667 capacitors. $250 for two would probably compare favorably with the cost of an equivalently rated MMC array.

I see that you are located in Germany so shipping may be prohibitive.

Flyglas

Fri Sept 07 2018, 07:40PM

Registered Member #62305 Joined: Fri May 11 2018, 03:39PM
Location: Germany
Posts: 6

MRMILSTAR wrote ...

Probably not as cheap as you can find on E-Bay, when you can find them, but here is a source for used Maxwell 37667 capacitors. $250 for two would probably compare favorably with the cost of an equivalent MMC array.

Thank you. They are currently out of stock. With the cheap MKP caps (220x0,2 EUR=44 EUR) the MMC will be incredible cheap.
I think at first I will try my configuration described in an earlier post.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob