Leaded snap-in capacitors / Electromagnetic Projectile Accelerators / Forums

Forums

4hv.org :: Forums :: Electromagnetic Projectile Accelerators

« Previous topic | Next topic »

Leaded snap-in capacitors

1 2 next

Move Thread

LAN_403

grythumn

Fri Jul 22 2011, 02:44AM

Registered Member #4021 Joined: Thu Jul 21 2011, 10:09PM
Location: Hanover, MD
Posts: 11

Am still in the "gather parts and decide what I want to do" stage, but ran across these on ebay:

CDE doesn't have datasheets for these that I see, but they have datasheets for the current screw-mounted equivalent. The $/J ratio seems excellent. Practically, they won't work with a bus-bar design, but PCBs (albeit generally with 1 ounce cu) are pretty cheap nowadays. A dual sided board with solid power planes and central power connectors ought to have an average resistance of about half a milliohm.

Anyone worked with a bank of these, or with a capacitor bank with a PCB backplane?

-Bob

klugesmith

Fri Jul 22 2011, 04:59AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Welcome to 4hv.

Were you looking at the CDE PF series datasheet at

? The snap-in series at

seem very similar, except it stops at 360 V. Maybe an industrial customer special-ordered some 500 V (PF series) parts with snap-in terminals.

I see no problem with connecting many of these together with circuit board stock, as long as you put reasonable engineering limits on your load. Don't plan to short it with a screwdriver, that would be abuse even of a capacitor by itself. I agree that your caps are probably intended for photoflash applications, with discharge times on the order of a millisecond, and that their maximum ESR is probably between 100 and 200 milliohms.
It would be a pretty inefficient design if the load impedance (per cap) isn't much greater than that.
And a waste of copper if the interconnect R per capacitor is much, much less than that.
You probably know the rule of thumb, 1-oz Cu foil has a sheet resistance of 1/2 milliohm per square. Could you use help to figure its impulse current damage threshold, in Amps-squared-seconds per inch of width squared? Or amps * sqrt(seconds) per inch?

One might even design the foil pattern with fusible links for protection, in the unlikely event of one capacitor failing with an internal short!

I have no firsthand experience with a bank of many aluminum can electrolytics. But analyze PCB-based power distribution networks as part of my job. These days a single IC can take 10s of amps at VCC < 1 volt, and want that power with sub-milliohm impedance from DC to 10s of MHz. At such low impedance levels, high frequency performance is limited at the PCB level by the inductance of via arrays, and spreading inductance in closely spaced power planes.

Forty

Fri Jul 22 2011, 03:26PM

Registered Member #3888 Joined: Sun May 15 2011, 09:50PM
Location: Erie, PA
Posts: 649

I have a bank made out of 6 snap in caps. I used some strips of copper to make busbars on it. mounted the caps on a thick strip of plastic so the bank couldn't twist or bend. drilled holes in the copper for the cap leads and soldered them in. rather permanent.
If you were going to use all those capacitors in a single bank, then you'd have difficulty either way you did it. You'd need a lot of copper clad board, with careful design of the traces, but it would look really nice when finished. or you could slap some drilled copper bars across them.

grythumn

Mon Jul 25 2011, 04:40PM

Registered Member #4021 Joined: Thu Jul 21 2011, 10:09PM
Location: Hanover, MD
Posts: 11

Thanks for the replies!

Forty, I had thought about drilled busbars, and I'm glad to hear it is viable. I think for this first bank, I'll stick with the PCB if the preliminary calculations show that it is feasible; I'm much better at assembling PCBs than metal working. :) The idea floating in my head is a dual sided PCB, with the +500V on one side, ground on the other, plated through-holes, and not so much traces as a solid sheet of copper foil with appropriate sized holes to prevent arcing around the PTHs. Multilayer boards often dedicate a few layers to just ground and power, I'm just omitting the signal layers. :) I've been reading a bit, and most of the board houses say that the copper foil on the outermost layers is closer to an ounce and a half after plating, which should help lower the resistance of the PCB.

Klugesmith, I haven't actually laid out a PCB, even a simple one like this, before, so I'd appreciate help with the thermal calculations, and the fusible link idea is great. I'm going to try and sketch out a basic PCB this week in eagle and try some tools to simulate the response of a bank with WAGs on some of the capacitor parameters based upon the PF datasheet. I've mostly worked with digital stuff thus far, micros and FPGAs, where you either have a dev board or the project is low enough power and frequency that you can run it on a protoboard with a regulator and a few bypass caps. :)

-Bob

cduma

Mon Jul 25 2011, 08:27PM

Registered Member #1822 Joined: Fri Nov 21 2008, 08:04PM
Location:
Posts: 300

I think you would be much better off just using a couple of large caps.

Dr. ISOTOP

Mon Jul 25 2011, 10:07PM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

Those caps are a pretty good price per joule, but if you look hard enough you'll find screw-mount caps for nearly as good of a deal.

grythumn

Mon Jul 25 2011, 10:46PM

Registered Member #4021 Joined: Thu Jul 21 2011, 10:09PM
Location: Hanover, MD
Posts: 11

cduma: It would simplify things, yes, but aside from the probably mythical Mouser capacitors, every other option I've seen is 3-20 times the cost. 25 caps of these caps, for $63.50 shipped, gives a nominal bank capacity of 3.75 kJ. Granted that doesn't include interconnect costs or increased risk of a few bad caps.

3 500v 10000uf caps would give the same bank capacity for... $314 or so, assuming a discount on shipping, from the same vendor.

Price calculations mostly from digging around on ebay

Am I missing a good source or need to wait for a good deal on ebay?

EDIT: Thanks, bwang. I'll keep an eye out for deals on big screwmount caps while firming up plans on the actual coilgun design and drafting a PCB as a fallback position.

-Bob

Forty

Tue Jul 26 2011, 02:00AM

Registered Member #3888 Joined: Sun May 15 2011, 09:50PM
Location: Erie, PA
Posts: 649

the junkbox has a nicely priced bank.
i'm still hoping those 'mythical' capacitors i found are real lol. they still haven't responded to my email i sent about them.

grythumn

Tue Jul 26 2011, 06:43AM

Registered Member #4021 Joined: Thu Jul 21 2011, 10:09PM
Location: Hanover, MD
Posts: 11

Forty wrote ...

the junkbox has a nicely priced bank.
i'm still hoping those 'mythical' capacitors i found are real lol. they still haven't responded to my email i sent about them.

I'm sticking with 'probably mythical' but wish you luck. :) Digikey wants $200+ for the same cap. :) As for the junkbox, aside from a few low uf HV caps from a busted inverter and monitor, I don't have any caps of significant capacity. Lots of sub-25V stuff, not worth messing with.

-Bob

Forty

Tue Jul 26 2011, 07:32AM

Registered Member #3888 Joined: Sun May 15 2011, 09:50PM
Location: Erie, PA
Posts: 649

this junkbox. 891J for 35 bucks

1 2 next

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