[Home][2017 Index] Re: [TCML] **External Email** Re: Utility PFC Caps [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] **External Email** Re: Utility PFC Caps

Hi David and all,

Those are some excellent questions - I had to do a bit of digging to try to answer them!

Utility HV PFC or Shunt caps are the unsung heroes of the capacitor world. They're designed to provide reliable 24/7 service for over 40 years, from -40C (-50F) to +46C (115F), all the while shrugging off lightning and switching transients, abnormal line voltages and harmonics, continuous 100% voltage reversals, while suffering through periodic onslaughts of ice, snow, blazing sun, or pounding rain.

Based upon these challenging requirements, its not surprising that HV PFC caps are seriously overbuilt. The dielectric systems within the individual capacitor rolls operate at a significantly lower voltage stress than typical pulse capacitors. Modern HV PFC's typically use two or three layers of hazy biaxially oriented polypropylene (BOPP) film sandwiched between aluminum foil. The edges if the foil are typically rolled or have rounded laser-cut edges to minimize edge corona. By design, the voltage stress across each individual capacitor roll is kept relatively low - between 1.8 and 2.4 kV - to insure operation well below the corona inception point. For comparison, the dielectric stress in the rolls of a typical Maxwell HV pulse capacitor is 3 to 4 times higher!

To get a better feel for design margins, we need to refer to PFC cap manufacturing test and type testing and application information. This information can be found in IEEE/ANSI Standard 18 or IEC 60871, and IEEE Std 1036. From these documents the following pearls of wisdom were gleaned. The following relate to manufacturing tests, endurance and ageing tests, and operating and transient requirements:

In all the following, Vf is the faceplate RMS voltage rating of the PFC cap.

Max Volts AC between terminals:  2.0*Vf for 10 seconds

Max Volts DC between terminals:  4.0*Vf (IEC) for 10 seconds
                                 4.3*Vf (IEEE Std 18) for 10 seconds

Transient Overvoltage endurance: 2.0*1.414*Vf = 2.83*Vf (IEEE Std 18)

Ageing (Type Test)             1.4*Vf (1000 hrs) or 1.25*Vf (3000 hrs)

BIL (Terminals to Case):         Typically 85 - 150 kV (1.2/50 us)
(Note: BIL does not apply to single-bushing caps)

Continuous repetitive AC peak voltage: 1.2*1.414*Vf = 1.7*Vf (including harmonics, IEEE Std 18, section 5.3)

Switching transients: 2*1.414*Vf = 2.83*Vf = 2 times nominal peak with no degradation in service life (IEEE 1036, section

Finally, the specifications for motor (or surge) protection capacitors are even more rigorous. So the above limitations for PFC caps can can be also be used as guidelines for your recently acquired motor protection caps. (IEEE 18-2002, section 11.3).

Taking the above info, and adding a bit of "Kentucky windage", I would make the relatively conservative recommendations:

1. For TC or other high-Q ringing pulse-discharge applications, you can apply an RMS voltage that is 1.2*Vf - 1.4*Vf. Reduce voltage if any case bulging is seen, since this is a sign that partial discharges and gas generation is occurring.

2. For HV DC power supply applications (such as the DC energy storage cap for a resonant TC system), you can apply up to 2*Vf... and perhaps as much as 2.5*Vf for short-term experiments as long as no oscillatory discharges are involved.

3. For low duty cycle, multi-kA ringing high-energy discharges, I would limit maximum DC charging voltage to no more than 1.2*Vf to 1.5*Vf. This should minimize the risk of corona damage during rapid voltage reversals. Most foil-film PFC capacitors will easily handle peak discharge currents of 25 - 40 kA. Higher-ESR self-healing metalized film caps, not so much.

HV PFC and surge/protection caps have lots of design margin, and they should tolerate considerable abuse. You can draw your own conclusions now that you know the applicable manufacturing test and design standards. As with any high voltage capacitor, overvolting can severely reduce capacitor lifetime, sometimes quite catastrophically. And, if you feel lucky, you can push the limits even further. Obviously YMMV... :)

Good luck and play safely,


David Rieben wrote:
HI Bert, all,

While the subject of electrical utility PFC (and protective) capacitors is still fresh on my mind, I was wanting to pick your brain regarding the theoretical maximum DC potential that one of these caps can withstand vs their nameplate AC voltage rating.
I have a spare 150 kVAR, 7960 VAC (60 hz) rated (measured C is 6.45 uFd) that I have routinely charged up to around 20 kV (about 2.5X its AC rating) from the rectified output of a 14.4 kV pig - while toying with the idea of using it for a storage cap for a possible AC to DC resonant conversion of my big coil. I was also hard short discharging it before the internal bleeder resistors had sufficient time to bleed off any significant amount of the charge – say within 5 seconds of initial charge up) with no seeming ill-effects to the said capacitor. I realize that there is a necessary and significant degrading of a capacitor’s voltage rating when dealing with it’s AC rating vs its DC standoff. IIRC, thanks to the negative effect of increased AC frequency to a capacitor’s XC, the higher the AC frequency, the more significant this degrading must be to keep the capacitor within safe voltage limits when applying AC to it.

So is there a good ballpark multiplication factor for the AC voltage rating of a 50 or 60 hz AC rated cap to obtain a reasonable maximum safe DC charge voltage?

I seem to recall seeing some of the smaller, consumer grade PFC or “motor run” capacitors that had a dual AC and DC voltage rating that was 440 VAC/1000VDC or 660 VAC/1500VDC. Is a 2.27X multiplication conversion factor a reasonable assumption with 50/60 hz AC vs. DC rating with these larger and robust utility style caps or could they reasonably be expected to have an even more aggressive multiplication factor?

Also, one more itty-bitty question. ;^) I noticed that the nameplate on my newly acquired 13,800 VAC @ 0.25 uFd rated GE protective caps did not have a stated biL rating. Am I to assume based on their nameplate voltage rating that they should be good for at least a 110 kV or possibly 125 kV biL, since that seems to the typical biL rating for other electrical  utility components that operate at these voltage levels?

Thanks and hoping that all have a wonderful Memorial Day,
David Rieben

Sent from Mail for Windows 10

From: Bert Hickman
Sent: Monday, May 22, 2017 7:30 PM
To: Tesla Coil Mailing List
Subject: Re: [TCML] **External Email** Re: Utility PFC Caps


It can be either.

Some (but not all!) utility PFC capacitors contain internal fuses. These
disconnect failing capacitor rolls, reducing the overall capacitance of
the unit while still keeping the rest of the capacitor rolls in service.
After the fuse blows, the reduction in measured terminal capacitance is
sudden and is typically in the range of 10 - 100% depending on the
internal electrical configuration of series and parallel strings.
Externally fused PFC capacitors usually fail catastrophically, blowing
their associated fuse and, infrequently, expelling insulators or
rupturing their cases.

Lower voltage PFC capacitors use self-healing metalized film technology.
These caps show a more gradual reduction in capacitance over time
(usually 5% or less) unless they're subjected to prolonged over-voltage

High-voltage PFC capacitors use multiple capacitor rolls connected as
one or more series strings. If a cap in a string fails (i.e., shorts
out), the capacitance of the affected string increases, and the voltage
stress on the remaining rolls in that string increases. The higher
voltage stress may cause other caps in that string to subsequently fail,
resulting in a cascade failure of the string. Any HV cap that has
significantly higher than nominal capacitance should be replaced in the

During preventive maintenance, utilities will look for any signs of case
bulging or leakage. Capacitor current or capacitance of suspect units
may also be measured. Capacitors with significantly higher or lower
reactive current or capacitance may be flagged for replacement.

Any surplus HV cap that has a measured capacitance that is significantly
_lower or higher_ versus the nominal faceplate value (allowing for
expected tolerance) or that shows bulging or leakage should be avoided
by the HV experimenter/surplus scrounger. The capacitor has already
partially failed and will only further degrade over time.


Terry Oxandale wrote:
Sorry, can't remember. Too many birthdays since then.

-----Original Message-----
From: Tesla [mailto:tesla-bounces@xxxxxxxxxx] On Behalf Of Ed
Sent: Saturday, May 20, 2017 7:05 PM
To: Tesla Coil Mailing List <tesla@xxxxxxxxxx>
Subject: Re: [TCML] **External Email** Re: Utility PFC Caps

Which way did the capacitance 'migrate'  Low or high?


On 5/20/2017 4:36 PM, Terry Oxandale wrote:
We would remove these caps from the bank once they migrated outside a specific tolerance, typically 2-3 per year, and hopefully before an internal rupture occurred (bulging).

Sent from my iPhone

On May 20, 2017, at 10:15 AM, David Rieben <drieben@xxxxxxx> wrote:

Hi Bert, (and Terry, if you're still following this thread),

Yes, I was rather pleasantly surprise to find that the more updated dielectric system of these types of caps renders their latent dissipation factor suitably low for the rapid cycle pulse discharge duty of being employed as the main tank cap of a large RSG/pole pig driven Tesla coil. I was also quite surprised to locate this type of cap in a relatively low capacitance, thus rendering their C within a suitable range for Tesla coil usage  in combo with an adequate ceiling voltage rating (13.8 kVAC). Locating more than one of these identical 0.25 uFd rated (single phase) units meant the possibility of seriesing just two to obtain 0.125 uFd @ a whopping 27.6 kVAC rating! Their GE model # is 9L18CCL101 (whatever that means) and the only drawback is the single bushing (w/casing ground). However, this minor caveat can easily be addressed by simply 'floating' their seriesed outer casings and using the two bushings as the inputs. Not that I really needed these, but considering the
en fickle nature of availability of affordable capacitors of this caliber, I like to keep a good supply of suitable replacements in case the 0.1 uFd @ 75 kVFD Maxwell pulse cap that I am currently using in my big coil lets out its 'magic smoke' ;-))

I have two of these ~33 lb. beasts on their way to me now. They are used with no guarantee, but having no visible 'bulging' of the outer casing and having been simply removed from decommissioned switch gear in a fully operational environment, I figure the odds are pretty high that these beauties are still quite functional and well within their nameplate specs. The real kicker is that I will have <$150 in both of them, including S&H :-) Gotta love that!

Happy sparking,

Sent from my iPhone

On May 18, 2017, at 11:22 AM, Bert Hickman <bert@xxxxxxxxxxxxxxxxxxxxx> wrote:

Hi David,

Your plan sounds excellent. As Terry mentioned, you really don't want to tear into these caps. I've also been there, done that, while autopsying a failed/ruptured 175 pound GE pulse cap. NEVER again - what a MESS... :^/

Even worse, unlike castor oil or mineral oil, many of the low-flammability dielectric fluids that GE (and others) used were really nasty, foul-smelling solvents. When the cap failed, the leaking fluid literally "ate" the backing off the indoor-outdoor carpeting below.

Your two caps in series should work great just as they are, while increasing the bang size a bit. A little tweak on tuning and ballasting and they work great. And, they should be virtually bullet-proof.

Play safely and make big sparks,


David Rieben wrote:
Hi Bert, (and Terry O),

Tesla mailing list
This email and any attachments are for the sole use of the intended recipient(s) and may contain confidential information. If you receive this email in error, please notify the sender, delete the original and all copies of the email and destroy any other hard copies of it.
Tesla mailing list

Tesla mailing list
Tesla mailing list

Tesla mailing list

Tesla mailing list

Tesla mailing list