[Prev][Next][Index][Thread]

Re: Capacitor Quality Factors


From: 	D.C. Cox[SMTP:DR.RESONANCE-at-next-wave-dot-net]
Sent: 	Friday, November 28, 1997 12:54 PM
To: 	Tesla List
Subject: 	Re: Capacitor Quality Factors

to: Fr. Tom

The biggest factor effecting cap Q quality here is the choice of materials.
 Aluminum is commonly used because it is widely available and cheap. 
Copper is the best choice and works much better than alum. at these low RF
frequencies but copper is more expensive and usually not worth the trouble.
 Good design with alum. will give you a very acceptable Q factor.  Most
builders destroy the Q factor in their rotary spark gap designs and
interconnecting wires.  Use only copper like welding cable and use only
brass (never steel or alum) in the RSG design.  People do break these rules
but it really hurts their peak current levels.  With the large plate
surface area cap Q factor is usually so high that it is not an important
design parameter unless you are running peak currents of 3,000 amps or
higher (cap values of 0.1 or higher typically).

90 mils is .090 inches and is good up to 12 kv operating.  60 mils (.060")
works best at 7-10 kv and 120 mils is usually preferred at 15 kv.  Again as
you suggested, keep total design values to less than 9 kv per sheet (never
exceed operating value) to prevent undesireable dielectric to plate loss
interactions. 

Hope this helps out.

DR.RESONANCE-at-next-wave-dot-net
 

----------
> From: Tesla List <tesla-at-pupman-dot-com>
> To: 'Tesla List' <tesla-at-poodle.pupman-dot-com>
> Subject: Re: Capacitor Quality Factors
> Date: Wednesday, November 26, 1997 6:27 PM
> 
> 
> From: 	Thomas McGahee[SMTP:tom_mcgahee-at-sigmais-dot-com]
> Sent: 	Wednesday, November 26, 1997 1:23 PM
> To: 	tesla-at-pupman-dot-com
> Cc: 	pacster-at-ibm-dot-net
> Subject: 	Capacitor Quality Factors
> 
> Tesla List members,
> Every once in a while I respond to something on the Tesla-2 List
> that I think would also be useful for the members of Chip's
> Tesla list. So here is a post dealing with capacitors.
> 
> My apologies to those who already got a copy because they are members
> of both lists.
> 
> Fr. Tom McGahee
> 
> ----------
> > From: Tesla-2 <tesla-2-at-emachine-dot-com>
> > To: tom_mcgahee-at-sigmais-dot-com
> > Subject: How does one make high Q 
> > Date: Wednesday, November 26, 1997 5:29 AM
> > 
> > From: pacster-at-ibm-dot-net
> > Date: Wed, 26 Nov 1997 10:29:37 GMT
> > Subject: How does one make high Q caps?
> > 
> > Hello !
> > 
> > Finally!, my exams are over for the year, so now I can get coiling
again (I
> > definitely prefer coiling to exams). I was reading some posts on
capacitors
> > and it was said that for a 15KV cap. the polyethelene should be around
> > 90mils thick. Coming from a 'S.I. units' country I'm a little confused,
is
> > that thickness 24.5/1000 x 90 = 2.28 mm? If so then its no wonder that
my
> > caps cook after only one minute of coil operating, I was using 0.45 mm
of
> > poly between plates, oops!
> > 
> > What is the recommended thickness for 7.5, 9, 10, 12 and 15 KV poly
> capacitors?
> > 
> > I am hoping to build high Q caps from now on and after a fair bit of
reading
> > apparently these can improve the capacitor Q:
> > 
> > 1. Clean all sides of the poly with meth. spirits.
> Keeps the crud out.
> > 2. Round all foil edges. 
> This cuts down on corona.
> > 3. Keep the plates as tight as possible .
> Increases capacitance and decreases chance of bubbles.
> > 4. Add lay sheets of foil between the layers of poly.
> This enhances charge distribution across dielectric layers.
> > 5. Add sheets of oil soaked Craft paper next to each live foil plate.
> This cuts down on air bubbles.
> I add the kraft paper between all surfaces. The idea is to reduce
> air bubbles, and these can form anywhere. It may be overkill,
> but I like the idea of oil being wicked up between ALL surfaces.
> > 6. Provide the biggest terminal ends for each plate.
> Maximizes current transfer to the outside world.
> By the way, I personally do not like the use of "tabs" on the 
> capacitor plates. I prefer to use the entire edge of the plate
> as my connection. This requires that the plates extend out
> opposite ends and uses a bit more aluminum foil (whoopie-doo!),
> but it maximizes current path and makes the cap plates easier
> to build. (Making tabs is a pain in the you-know-what).
> > 7. Put the caps in series.
> Above 10KV you really can't make a single cap that will stand up
> to the corona. Over time the dielectric fails, especially at the
> ends. Making many caps and putting them in series makes it
> so each cap has a low enough voltage that corona is not a major
> problem.
> > 8. Submerge the cap in light mineral oil.
> The light mineral oil penetrates more readily between layers than the
> heavy mineral oils. The oil helps in the exclusion of air, which
> in the presence of HV RF results in spot heating and nasty corona.
> > 9. Use equal numbers of + and - plates.
> Otherwise one set of plates is being wasted :(
> > 10. When using caps in series they should be the same individual part
> > capacitances.
> Very important! The capacitors in series will share voltage in 
> inverse proportion to their relative capacitance. In simpler terms,
> the smallest capacitor will get the biggest voltage. Zap!
> > 11. Vacuum pressurising the caps.
> Not necessary with most flat cap designs. In fact, a dangerous thing
> to attempt if your case has flat sides! The rolled cap designs in
> cylindrical PVC pipe take the vacuum well, however.
> > 12. Keep all dust and dirt out of the cap.
> Goes right back to #1. Don't let crud get inside EVER! It becomes
> a site for corona generation and mucks up the oil.
> > 
> > What is the most useful in improving the Q and what has little affect
on the
> > Q, is there anything else that improves the Q? 
> 
> OK, my own personal thoughts on factors affecting the Quality of the
capacitor.
> In groups, with like items together.
> Relative scale of 1-5, where higher is better:
> 
> First the things that directly affect capacitance :)
> 
> (5) Size of plates: larger is better
> (5) Number of plates: more is better
> (5) Distance between plates: shorter distance is better
> (5) Dielectric material: higher dielectric constant is better
> (5) Dissipation factor of dielectric: lower is better
> (5) Breakdown voltage of dielectric: higher is better
> 
> Things that make it work better and live longer
> 
> (5) Use multiple thin dielectric sheets between plates
> (2) Use inactive floating plates between multiple dielectric sheets
> 
> (3) Low inductance: flat caps are better than rolled caps here
> (4) Low resistance internal connections: larger surface connections are
better
> (3) Low resistance external connectors: larger is better
> 
> (5) Exclusion of air by using oil (reduction of corona)
> (5) Exclusion of air by using vacuum or compression
> (4) Exclusion of air by allowing it to escape (vertical flat plates best)
> (3) Exclusion of air by using Kraft paper around active plates
> (2) Exclusion of air by using Kraft paper around inactive floating plates
> 
> (5) Reduction of corona by using multiple series-connected capacitors
> (5) Reduction of corona by allowing adequate extra poly around end
overlap
> (5) Reduction of corona by allowing adequate extra poly around top/bottom
> (2) Reduction of corona by using rounded edges on plates
> (3) Reduction of corona by using adequate container insulation
> (3) Reduction of corona by excluding small foreign matter
> (2) Reduction of corona by excluding large foreign matter
> (1) Obsessive-compulsive cleanliness of all surfaces and materials
> 
> Other factors
> 
> (3) Size (smaller is better)
> (3) Weight (lighter is better)
> (2) Portability (the more portable the better. No messy oil spills)
> 
> (1) Aesthetics (the way *You* think it looks) 
> (0) Aesthetics (the way *Other* coilers think it looks)
> (-5) Aesthetics (the way *NON* coilers think it looks)
>                 
> 
> > Normally I roll my caps, but
> > I think that if I should be using a thickness 2.3mm of poly this could
be a
> > bit of a task. But perhaps I could make ten 10nF caps and parallel them
up
> > to 100nF and then series four sets of them down to 25nF?
> > 
> > I would also like to take a look at my coil on the scope when it is
running.
> > To do this can I just connect a piece of wire to the 10x probe and
dangle
> > the wire in the air about 10 to 15 feet away from my coil 
> 
> Yes.
> 
> > (do I need to
> > connect the scope earth to anything ?),
> 
> Yes. Earth. Mr. Ground. The Ultimate Reference point.
> 
> > and then fire up my coil, with
> > hopefully my new high Q caps (so that I can watch it for more than a
minute
> > at a time). Is this O.K. or is there a safer and better method?
> > 
> 
> Safest method: Don't build a coil. Watch TV instead.
> Better method: Fire it up and enjoy the sparks. Tweak for maximum Fun
Factor.
> 
> > Thanks all for your time everyone!
> > 
> > Mark
> > 
> Mark,
> I hope the above comments are helpful.
> Fr. Tom McGahee 
> > 
> > * Send list Replies to: Tesla-2-at-emachine-dot-com
> > 
> > * To UNSUBSCRIBE from this list, put the command UNSUBSCRIBE TESLA-2
> >   in the body of your message to: Listserv-at-Emachine-dot-com
> > 
> 
>