Building a high voltage Film Cap, but its failing across the board

[crashtriple3]

new to this project and have limited knowledge, but the failure is consistent thru all specimens so I am hoping someone could shed some light.

Film cap design is rated to 1650v

Failure-
Cap initially failed around 800v consistently, the curing time was adjusted and a ramping cycle was introduced to ramp up temp and down temp and peak curing temp is now held for 9 hours and max voltage reached has increased to two groups, many failures in the 1500 range and then about the same amount reach the
1800v range before failure.
Failure occurs in the top layers and shows damage inward about 1/4" inch (overall diameter is approx 3.1")

Adjusting the curing process gives me significantly better results, does this sound like a curing issue?

Also due to be noted that the failure always seems to occur 3/4" clockwise from the stitching and then runs parallel to it from end to end.

Failure looks like bubbly marks or acne on your skin, then when the layers are pealed back you can see the depth of the failure.

It is not a dielectric failure.

Just purchased a new high speed film winding machine, and film tension is set to around the 650 range,

Could this be a process issue? Is an adjustable tension setting needed when winding? I am wondering if my top layers are not being wound tight enough and various air gaps are being introduced.

Or could this be a curing issue? Noticed the substantial different in my results when my curing times were adjust from a straight single temperature curing time to an adjustable ramping setting.

Any input or thoughts would be appreciated, very new to the making of film caps.

 

[vk6kro]

How do you know it has failed? Does it go short circuited?

What capacitance are you looking for?

Do you need to use the capacitors, or is making them the point of the exercise?

You can certainly buy capacitors with that voltage rating. Microwave ovens have one that is about 1 uF which will stand at least 3000 volts DC.

 

[crashtriple3]

Because it goes snap crackle and pop and yes it shorts out and the layers of the cap bubble and blister. when I cut open the cap and start to peal back the layers you can see the holes that have burned thru and when you trace them down I find that the damage goes about 1/4" depth.

strictly building a cap, I'm an ME so my knowledge of this experiment is vague I was brought on to inspect the build process and machine set up of these caps. I am just wondering if someone in the capacitor manufacture field could help just in the sense if anyone has experienced failures of this nature, mainly to rule out a common mistake that make have occurred in the design or manufacturing process.

 

[sophiecentaur]

How uniformly smooth is the metal foil? Are there any high spots / wrinkles? They could be reducing the gap in places. Is there so much tension in the winder that the dielectric is being squashed? (worth asking)
If you work out the film thickness and the dielectric strength, how many volts would you expect it to handle and how far away from it are you failing? i.e how near the theoretical limit are you?

 

[crashtriple3]

I did find what looked to be a "dent" that slightly showed on the surface and as I pulled back some layers it still showed.

I was un-able to get an answer from the EE's about the theoretical value which is a valid point, so I decided to calculate it myself. Do you an equation I can use? I'll check what Google has for me.

 

[crashtriple3]

Also i have been doing some reading and this is a complete shot in the dark but could a thiner layer of insulation on the exterior of the cap cause this problem? Basically the cap is wound and the 10-15 layers of thin cover wrap is put on. There is no other dielectric encloseure the cap is placed into prior to testing. Again a shot in the dark...any thoughts?

 

[Carl Pugh]

You can never work insulation to it's rated voltage.
The thinner the insulation the high the volts/mil you can use. Make multiple low voltage capaciors and put them in series for a high voltage capacitor.
Check capacitors for corona (partial discharge). The circuit used in the military document MIL T27*** works great. You can probably get MIL T27*** from the WWW.
A cheapie corona detector is an AM radio tuned to where there are no stations and minimum static. Slowly raise voltage on capacitor, when static appears on radio, reduce voltage on capacitor and then raise voltage again. If static appears at the same applied voltage all the time, it's corona.
If you listen carefully, you may be able to hear when the corona starts.
Capacitors for use on AC are much more difficult to design than capacitors used for DC.
If oil filled capacitors can be used, they are preferred for high voltage.

 

[crashtriple3]

there are design limitations I have and the size/shape is set in stone. So I am told this cap should be well capable of the voltage levels we are testing too that is the information I have to go off. My goal is to determine where/why this failure is occurring...production, curing, testing...etc

But basically what your saying is the exterior layer of insulation of the cap has no relevance on the caps voltage handling what so ever? I am not talking about the insulation between layers I am referring to the exterior insulation of the wound cap from atmosphere. Can the thickness (or lack of) of this layer hinder its performance and create this type of failure I am noticing?

Also during the initial loading of the cap shouldn't the voltage be applied gradually I.e. ramped up to 50% soak for few hours then ramp up to 75% soak for few hours and then ramp up to peak/failure voltage. We need to hit 1650 vdc for 10 sec but the testing has basically turned into raw peak handling voltage which ideally shoudl be over 2000vdc

 

[sophiecentaur]

I did find what looked to be a "dent" that slightly showed on the surface and as I pulled back some layers it still showed.

I was un-able to get an answer from the EE's about the theoretical value which is a valid point, so I decided to calculate it myself. Do you an equation I can use? I'll check what Google has for me.

Dielectric breakdown strength will probably be specified in Volts per meter. If you know your thickness (in m, remember) then that will tell you the breakdown voltage to expect. But the small radius corners will increase the field strength at the edges by a bit - is that where the breakdown appears?

I did wonder whether you had made the dielectric strip width comfortably greater than the metal strip width.
I saw a suggestion to use capacitors in series, to increase breakdown voltage. This can only work reliably if you include a (very) high resistor chain to maintain the middle voltage at half way. You can achieve the same capacity value as for two in series by just using a double thickness of dielectric tape. Or, by doubling the length of the roll, get back to the required Capacity.

 

[crashtriple3]

film thickness: 4.8mm
width: 125mm
edge margin: 2.5mm
metal:aluminum, HE
edge resistance: 1-4 ohms/square
body resistance: 5-10 ohms/square
material: polypropylene, ht treophan
k=2.2 +/- .1

these are the film material properties, the cap ends up being 3.2 max diameter and 5.025 max length with approx 3216 turns...handling voltage is supposed to be 1650vdc but failures occur around 1400-1600

 

[sophiecentaur]

If you want to know how it will take high voltages then you need to know its dielectric strength. Without it. you can't. All you can do is to try increasing the thickness until it stops failing. If it means your Capacitor is as big as a bean can then you need another material. Why did you pick that particular material?

 

[jim hardy]

film thickness: 4.8mm
...
3.2 max diameter and 5.025 max length with approx 3216 turns...handling voltage is supposed to be 1650vdc but failures occur around 1400-1600

can that be right? A roll of 3216 turns of 4.8mm film in 3 inch(?) diameter?
3216 X 4.8mm is 15 meters.
If mils it's 15 inches.

What do i misunderstand?


this guy says ~500 volts/mil for his polypropylene film
http://www.boedeker.com/polyp_p.htm

which would be 2400 volts for 4.8 mils
1800 volts should be limit for 3.6 mil material

i have no idea if his is typical.

4.8mm is 0.189 inch - seems too much for your package size

at two thicknesses per turn,
3.2 inch diameter divided by (3216 X 2) = 4.97E-4 inch, not quite a half mil.

but a half mil film is way overstressesd at 500 volts, i think, if that guy's polypropylene is typical..
Can you lay a caliper on your film?
And double check that turns counter? Perhaps it's geared 100::1 .

What'd i miss?

 

[crashtriple3]

actually you didn't miss anything, more like what did i miss


i meant 4.8um

 

[jim hardy]

i meant 4.8um

4.8 um X 39.37 = 188 microinches, 0.188 mils?

Okay. Here's a link to a polypropylene supplier's information brochure

http://www.google.com/url?sa=t&rct=...kN3iDQ&usg=AFQjCNFxcR-2G8rDmEZi4nMoxs8fD-vFaw

Prototype Capacitors
The high-K film was metallized and wound into prototype capacitors with different capacitances. The metallization
is a thin layer of aluminum with high surface resistance for high dielectric breakdown strength.
10 μF capacitors made with 5 μm thick film exhibit high dielectric breakdown strength and high energy density.
Most of the 10 μF capacitors can pass 1,500 V breakdown test, and more than half of them can survive 2,000 V
charge-discharge test.

Does that mean nearly half of them made with 5μm fail 2,000 V ?

Take Karl Pugh's advive and set a cheap AM radio near a capacitor as voltage is increased.
If it snaps crackles and pops shortly before the cap develops acne, i'd suggest your engineer call the film supplier's engineer .