Sciencemadness Discussion Board - Capacitor source?

Capacitor source?

elementcollector1 - 25-9-2013 at 18:55

I've been looking around for a good source for a particular capacitor for high-voltage experiments: 330 volts, and 80 or so uF. I can't seem to find anything that's under $20, so does anyone have an idea where I might find these?

12AX7 - 25-9-2013 at 19:10

Photoflash..?

What kind? Polarity, ESR/ESL...

If you're looking for film, that sounds about right.

Tim

elementcollector1 - 25-9-2013 at 19:19

Yes, photoflash. It's not for anything photography-related.

BobD1001 - 25-9-2013 at 20:07

These may definitely fit your bill. They are surplus disposable camera flash circuits. Each one has a 330V 120uF capacitor. I will verify these specs as I have about 12 of them sitting in my electronics parts cabinets.

http://www.goldmine-elec-products.com/prodinfo.asp?number=G1...

Or if you looking to just buy one: http://www.goldmine-elec-products.com/prodinfo.asp?number=G1...

Apparently, there is also a very nifty way to make a quick and effective Geiger count out of the board itself!

Bob

-I just took a look at mine and they are rated 350v and 160uF

[Edited on 26-9-2013 by BobD1001]

[Edited on 26-9-2013 by BobD1001]

elementcollector1 - 25-9-2013 at 20:31

Quote: Originally posted by BobD1001

These may definitely fit your bill. They are surplus disposable camera flash circuits. Each one has a 330V 120uF capacitor. I will verify these specs as I have about 12 of them sitting in my electronics parts cabinets.

http://www.goldmine-elec-products.com/prodinfo.asp?number=G1...

Or if you looking to just buy one: http://www.goldmine-elec-products.com/prodinfo.asp?number=G1...

Apparently, there is also a very nifty way to make a quick and effective Geiger count out of the board itself!

Bob

-I just took a look at mine and they are rated 350v and 160uF

[Edited on 26-9-2013 by BobD1001]

[Edited on 26-9-2013 by BobD1001]

Perfect, thanks!

IrC - 25-9-2013 at 20:48

http://www.goldmine-elec-products.com/prodinfo.asp?number=G1...

If you need larger capacitance this one is 470 uF at 400 volts.

BobD1001 - 25-9-2013 at 21:21

Quote: Originally posted by IrC

http://www.goldmine-elec-products.com/prodinfo.asp?number=G1...

If you need larger capacitance this one is 470 uF at 400 volts.

I purchased 5 of these exact ones, and wired them all in parallel for a nice little railgun project! Man do they put out some SPARKS!

elementcollector1 - 25-9-2013 at 21:36

Quote: Originally posted by BobD1001

I purchased 5 of these exact ones, and wired them all in parallel for a nice little railgun project! Man do they put out some SPARKS!

Funnily enough, my project is a coilgun...

BobD1001 - 25-9-2013 at 22:58

In that case I would highly recommend the 470uF capacitors. They are excellent! The 160uF ones are fun for a short time, but you quickly get power hungry... before you know it you'll have 12 wired in parallel still wanting greater projectile velocity. The 470uF really do the trick, and are a good deal.

12AX7 - 26-9-2013 at 04:24

Yes, recommend indeed. Photoflash caps have huge ESR on purpose. Snap-in style caps (note how the pins look to be riveted on the base) generally have quite low ESR.

I once did some photoflash tests with a xenon tube and various capacitors. I measured the discharge rate with an inductive current sensor, by probing the voltage across the capacitor, and with a photodiode.

A "general purpose" 100uF 350V type performed about like any photoflash: yeah it makes light, but the discharge is slow (several miliseconds!). After just a few pulses, the capacitor had obviously warmed up a little.

A high-current film capacitor (10uF) showed 5 microseconds discharge, which is essentially limited by the speed of xenon gas itself (it takes about this much time for the gas to ionize).

A snap-in (220uF 400V) showed a discharge curve of only 10 microseconds -- 20 times the value of the film capacitor, yet only twice the discharge time (half of which was the tube ionizing).

To further verify my measurements, I also rigged an air spark flash (3nF at 10kV or so), which discharged in 100ns or so -- the wiring was sloppy, so the inductance was quite noticeable. (The xenon tests were too slow to see much inductive kick.)

Tim

elementcollector1 - 26-9-2013 at 11:20

Quote: Originally posted by BobD1001

In that case I would highly recommend the 470uF capacitors. They are excellent! The 160uF ones are fun for a short time, but you quickly get power hungry... before you know it you'll have 12 wired in parallel still wanting greater projectile velocity. The 470uF really do the trick, and are a good deal.

For a similar price ($6/3 larger capacitors v. $5/10 lesser ones), the larger ones would add up to 1410 uF while the lesser ones in the bulk package would go to 1600 uF (according to BobD's earlier finding that each was 160 uF). I've always heard that the more capacitors the better, so that if one breaks you have several more. Unless there's something I'm missing, and the large capacitors would lead to a more powerful shot?

BobD1001 - 26-9-2013 at 11:40

Quote: Originally posted by elementcollector1

From my calculations for the energy in each capacitor: E=.5(C)(V^2)

For the 470uF 400V capacitors: E=.5(470*10^-6)(400^2) = 37.6 Joules

For the 160uF 350V capacitors: E=.5(160*10^-6)(350^2) = 9.8 Joules

So the 470uF 400V capacitors each store approximately 4 times the energy of the 160uF 350V capacitors. (3.83 times the energy to be exact).

This is assuming that elementcollectors charging circuit can charge to 400V, otherwise they would not reach their full energy potential. I also prefer using the larger capacitors due to the fact that wiring a few of these in parallel looks much cleaner than wiring many of the smaller ones in parallel.

[Edited on 26-9-2013 by BobD1001]

elementcollector1 - 26-9-2013 at 11:45

So it's not about capacitance, but joule output. I see.
My charge circuit is the disposable camera's flash circuit. Not sure if that can do 400V, as it's rated for 330-350V.

MrHomeScientist - 26-9-2013 at 12:58

For an educational outreach program I do, the other co-host designed and built a coil gun. Here's a short video of it in action: http://www.youtube.com/watch?v=2G5_VEmBDf8 (his channel, but it's me pressing the button)
We're launching old aluminum hard drive platters, with a rope to keep them from flying all over the place.

He's the electrical engineer, but I do know some specs of the device. We use 6 huge (small soda can sized) capacitors, wired such that there are two banks of 3 capacitors in series, with the two groups of three wired in parallel. The system at full power is 1500 joules at 1.2 kV. In the video, we only used half of that so we didn't dent the roof

I think they were 400V capacitors, but I'm not sure about their capacitance. That may be calculable from the info I gave - like I said, I'm not the EE!

If you want any more details let me know, or just send him a YouTube message directly and "tell 'em I sent ya!"

BobD1001 - 26-9-2013 at 14:51

Quote: Originally posted by MrHomeScientist

For an educational outreach program I do, the other co-host designed and built a coil gun. Here's a short video of it in action: http://www.youtube.com/watch?v=2G5_VEmBDf8 (his channel, but it's me pressing the button)
We're launching old aluminum hard drive platters, with a rope to keep them from flying all over the place.

He's the electrical engineer, but I do know some specs of the device. We use 6 huge (small soda can sized) capacitors, wired such that there are two banks of 3 capacitors in series, with the two groups of three wired in parallel. The system at full power is 1500 joules at 1.2 kV. In the video, we only used half of that so we didn't dent the roof

Given the full output power of 1500J, we can determined that each capacitor value was:

1500J/6 Capacitors = 250J/capacitor

250J=.5C(400V^2)
250=.5C(160000)
.001563=.5C
C=.003125 F

So each capacitor had a value of approximately 3125uF given that you recall the correct values.

Here's one heck of a capacitor for the job! http://www.ebay.com/itm/EPCOS-B43310-A9159-M-15000uF-at-400V...

[Edited on 26-9-2013 by BobD1001]

elementcollector1 - 26-9-2013 at 16:19

BobD: I think I'll stick with the smaller ones, thanks. That one would not fit into my design specifications, on account of simply being too ridiculously large.

MrHomeScientist, that is an amazing device, but is your partner sure about wiring the capacitors in series? This could lead to some nasty accidents, from what I've heard.

I've found two more source sites - Digikey and Jameco. Try this capacitor for size:http://www.digikey.com/product-search/en?pv63=497&FV=fff40002%2Cfff80009%2C340085&k=capacitor+350V&mnonly=0&newproducts=0&Col umnSort=0&page=1&quantity=0&ptm=0&fid=0&pageSize=25
(not actually planning to use this, just wanted to see what the biggest one was.)

Digikey lists a 400V, 3300uF capacitor as being $34.18 at the lowest. 400V, 470uF are listed as $5.78 at least.
Jameco doesn't seem to list a capacitor at high enough capacitance...

I think I'll stick with the 400V/470uF ones then. Is there any way to calculate the best amount of copper winding per capacitor (assume 32 AWG) so that nothing is over/underpowered?

BobD1001 - 26-9-2013 at 17:25

I mostly posted that giant capacitor for kicks, however I would love to have one to make a nice coilgun or railgun with! Let us know how your coilgun project fares, and if you ever need a nice coil wound I have a lathe. Just provide the wire and I'd do it free of charge since you've been so helpful with great input twoards my periodic table collection

IrC - 26-9-2013 at 18:04

My squisher (or thrower depending upon mood) has 120 of the capacitors BobD1001 is showing you except they are 8,000 uF/450V. Which are arranged to give me 3,200 uF at 4,500 volts stacked three in series to yield just under 1,100 uF at 18 KV. I use two 10,000 volt 100ma Tektronix (1KVA ea) transformers, and a few microwave oven diodes feeding through a bank of 75 watt 10K resistors to charge my capacitor bank. A 5 KVA Variac runs the parallel transformers and I have never allowed the bank to get above around 11 KV. Easy way to run at much lower voltages for less dangerous experiments which is mostly where I have operated it. The capacitor bank itself is in a box, you never know and while it has never occurred no doubt shrapnel is a real possibility. Two nice shiny 2" trailer balls form the gap with a sewing machine needle spaced between but offset slightly as a trigger. Years ago I built a strobe trigger circuit for the needle. Now I use the board from Goldmine mentioned above (operated by push-button to create trigger pulse that would have triggered the Xenon tube) to power the needle with an indicator light and voltmeter showing state of capacitor bank. Gap assembly in a box with sound dampening because the damn thing is really loud. Every so often expect to buy new trailer balls and needles.

Over the years I have used many loads from 3" Cu tubing squisher coils to rail guns, exploding wire forming, among other ideas at the time. Anyway series is no problem if you do it right. Keep balance, and equalization. Do not put 100 uF in series with 470 uF for example. Have from 150K to 680K (value used voltage dependent) 2 watt resistors across each capacitor paralleled with around 3300 pF 3KV capacitor. Keeps transients from punching holes in the dielectric (at least not too often). Plus across each capacitor is an NE-2 in series with around 180K 1/2W or so depending upon voltage. This shows a cap still holds a dangerous voltage. Safety first. However never forget the neon goes out around 55 volts. Just imagine 1,650 volts and 1,066.6 uF, and neons are all dark. Lethal. Always look at voltmeter and never trust it either. Before I play with hooking up a load I make sure power is disconnected, and I short it dead out with a shortened single piece of car jumper cable removed from the pair and using two of the giant alligator clips.

I also power a relay with a bleeder on the output which is open with power on. If I kill the power the bleeder starts bleeding the bank down through N.C. contacts. If I shut it off and leave I want to know it will be a dead bank later. Still look at lights and voltmeter anyway (yet don't totally trust either ever) as I learned long ago a very heightened sense of self preservation. Things like I never wear rings, watches, nothing metal on my fingers, wrists, no metal at all. Safer that way. Mad science is fun but I only have one life to live, which I value above whacky fun.

elementcollector1 - 27-9-2013 at 10:27

BobD, thanks for the compliment! It means a lot to know I helped someone.
Would using a plastic or metal barrel be better? Sources say plastic, as the metal dampens the strength of the shot.
Also, I can't find where to put my charging button or where to put a status LED (to tell when the charge is ready). I'll see if I can get a picture of my circuit sometime later.
As to winding coils, does it really require such precision? I could wind the wire loosely, shove it upward so that it's wound tightly, and repeat the process, yes?
One thing I don't understand and really need help with is how to wind the coil. My typical strategy was to wind the wire forward down the core, and then back, but wouldn't this cancel out fields? Do I need to wind it forward, run a wire back, wind it forward again, and repeat?
I'd rather not die to science either, and never wear jewelry anyway. My clothes have metal zippers and attachments, though. This will likely not be a problem for a coilgun of the strength I'm building, and if things start pulling towards it I can always find some mu-metal and place it around the outside.

MrHomeScientist - 27-9-2013 at 11:14

EC1: We definitely keep safety in mind when working with it! When not in use, he shorts the capacitors with a length of wire across the terminals. We use big aluminum bus bars to attach the caps together. It can be pretty scary: they very quickly build up a large amount of charge just from gathering static electricity from the air. So you have to put the safety on immediately.

Our coil platform is a piece of wood on the bottom and a sheet of 1/4" (I think) lexan on top, sandwiching the coil. The coil is quite thick copper wire that we wound in a single plane, which sort of looks like this but with the central hole much smaller:

We wound it by drilling a hole in the middle of the wooden platform and lexan sheet, and putting a dowel rod through it which we then wound the coil around. We pressed the lexan on top to keep it from springing out of shape. It wasn't easy!

We also put rubber feet on the underside of the four corners of the platform - turns out this was a bad idea. Thanks to Newton, the downforce created by many launches has cracked the platform and lexan sheet completely in half, and now the coil itself is the only thing holding it together! Using particle board was not a great idea either, but it was what we had handy at the time. We're in the process of rebuilding it.

EDIT:
"Digikey lists a 400V, 3300uF capacitor as being $34.18 "
I believe these are exactly the capacitors we use.

[Edited on 9-27-2013 by MrHomeScientist]

IrC - 27-9-2013 at 11:51

"My typical strategy was to wind the wire forward down the core, and then back, but wouldn't this cancel out fields?"

Not if you do not change the direction of winding, i.e., CW or CCW looking from one end. Good idea to put a thin paper or plastic sheet wrap pre-cut to size between each layer, also coating each layer with something to keep turns from shifting in large magnetic fields. If your using wire as in a small low energy coil some kind of acrylic or epoxy is good. The idea being locking the turns from any movement. Different story of course with tubing coils but rigidity of turns is still important so think about how your going to do it. Very surprising the amount of force trying to move the winding. For an indicator LED you might consider a resistor from your HV terminal (no idea voltages so cannot guess at values) connected to a string of high voltage zeners (varistors work too and 130 AC is typical, will work fine with DC for this) in a string to one end of say a 10K to 50K pot, other end to ground. Center tap feeding base of NPN transistor through series say 1K, emitter grounded. Collector to LED through proper resistance (say 680 ohms (for 12 volts and standard LED) depending upon low voltage source and proper LED current) to a DC source say 8 to 12 volts. The pot allows you to set the point where the LED lights when the HV is at the chosen voltage. As to zener or varistor voltage total figure you want say zero to 5 volts on the pot center tap when the HV terminal is at the desired voltage. I would choose the resistor feeding the zener string to limit current to a maximum of one or two milliamps. If you have 400 volts then around 200K 2W sounds reasonable since you would have around 800 milliwatts heating the 200K resistor. Just a ballpark example. Handy though as you can trigger each shot as the LED lights up knowing your chosen voltage has been reached.

[Edited on 9-28-2013 by IrC]

elementcollector1 - 27-9-2013 at 12:02

MrHS, what did your coilgun cost? Well over $200, judging from the 6 capacitors you used.
IrC, this does make sense. What about vinyl electrical tape (or is that too thick)?

IrC - 27-9-2013 at 12:20

Quote: Originally posted by elementcollector1

MrHS, what did your coilgun cost? Well over $200, judging from the 6 capacitors you used.
IrC, this does make sense. What about vinyl electrical tape (or is that too thick)?

You posted while I was editing to add a crude indicator light idea. While I have done the tape thing winding transformers not a good idea unless you can firmly lock the turns with acrylic or something letting it dry before adding the tape layer. The turns would find the smooth tape allowed them to move around. Of course we are talking about a small low energy setup. The entire approach is poor at higher powers. Look at MrHomeScientist's tubing coil. Far better since you will soon be craving more power. What do you think it was breaking the base plate He was describing? There is really a great deal of force on the turns as power goes up. If you read my insane yet true description above, I mentioned a coil made of 3" Cu tubing. No I did not make the coil. Found it at a long gone surplus place in Phoenix. Had no idea who made it or what the hell it was for but in 1986 I grabbed it for ten bucks. Since the yard had around a hundred old Klystrons of the 2 MW and higher variety I guessed it was some kind of tank or loading coil. No idea really though. Anyway at around 11 KV on my bank (~65,000 Joules) I actually bent the thing. No I never got it pretty and even again. I don't know how the hell they made a coil of 3" heavy Cu tubing to start with let along trying to bend it back.

Just realized "MrHomeScientist's tubing coil" is wire not tubing. You guys should try 1/4" Cu tubing. You can buy shrink in long lengths, slide it on (meaning go larger in ID for the shrink than you think or a few feet along it will be very difficult). Use heat shrink air gun to shrink it for entire length. Then wind the coil similar to the pic of MrHomeScientist's wire coil.

Edit to change EC1 to MrHomeScientist, as to who built the coil in the pic. I have really got to read posts more carefully.

"My clothes have metal zippers and attachments, though. This will likely not be a problem for a coilgun of the strength I'm building"

That was funny but burned out my minds eye picturing a demo you were giving where your pants went flying. I imagine zippers are not magnetic and while high dI/dT generated magnetic fields would exert force even on brass zippers I think you are safe to assume no clothes will be flying. I can say even in my 65 kilo-joule shots no one was driven insane from seeing me pant-less.

[Edited on 9-28-2013 by IrC]

MadEngineer - 28-9-2013 at 16:11

Quote: Originally posted by elementcollector1

MrHS, what did your coilgun cost? Well over $200, judging from the 6 capacitors you used.?

Hi all, I'm the evil twin of MrHomeScientist and the one that built the coil gun he posted. He showed me the thread and I wanted to jump in with any info on my design. I'm an EE by degree but shamefully threw this device together with back of the napkin calculations.

Items in the coil:
6x 400V, 3300uF Caps $34.18
1x 1200V 5KA (peak) SCR $93
Copper bus bar (don't rember wait we paid)
2 DPDT relays
Misc wire, including the wire we used to wind the coil
Buttons and switches

I also implement a safety discharge through another 50W resistor on power failure.

Originally we used a variac, MOT and 50W resistor to charge the bank, though I bought one of these for $400 to replace all that messiness

http://www.amazing1.com/content/download/CHARGE_Instructions...

That lets me dial in a charge set point in a "set it and forget it" kind of way; which is great as this will be used during live performances.

Using the SCR (Silicon Controlled Rectifier) meant our trigger system was solid state, no loud spark gaps here. Though that limits how much voltage we can achieve. I also purposefully used wire and not tubing so as to regulate discharge current somewhat by keeping the coil impedance higher than if I used tubing. I didn't want to stress the SCR too much, another limitation of going solid state.

The result is a system that delivers about a 1 ms impulse at 2kA initial inrush, well within the 2x safety margin of the 5kA SCR.

Here's a better Photo of the device: Sorry, It's on Facebook
before we added the nice charger.

So without the nice charger, cost comes in around $450

MadEngineer - 28-9-2013 at 16:26

Quote: Originally posted by IrC

What do you think it was breaking the base plate He was describing?

The coil housing was a poor design, particle board and plexy compounded by a bad decision to put rubber feet on the 4 corners. The stress of launching caused the wood and plexy to crack down the middle. I'm debating on casting the coil in resin next.

IrC - 28-9-2013 at 18:03

Wasn't actually asking a question, rather pointing out to him the movement of the turns was causing the breakage due to the tremendous force applied by the magnetic field. Is why I went to tubing but your right about the low impedance being hell on the switching device. I wasted many monster hockey puck SCR's before going to the trailer ball triggered gap. Loud but cheaper in the long run. You guys are getting me interested in playing with mine again it's been stored in sections for years now. I also tried the casting but it also breaks apart soon. I used a bunch of JB Weld once and it broke apart faster than clear acrylic casting resin. Never had any luck with any of those ideas. Probably OK for under a hundred joules I suppose so worth trying.

elementcollector1 - 28-9-2013 at 22:07

Quote: Originally posted by MadEngineer

Welcome to ScienceMadness! It's not shameful at all to build such a cool thing.

Quote: Originally posted by MadEngineer

So without the nice charger, cost comes in around $450

Hmm. I don't think I'm willing to spend this much on my coil. Thanks for the more precise info, though - it really does help!

[Edited on 29-9-2013 by elementcollector1]

IrC - 29-9-2013 at 00:22

Search around for surplus and think of hacker sources. You can build it for a fraction of cost if you do. Medium size setups more fun really. Less headaches, less dangerous. Besides you can shoot it more often. Myself, I think the ideal size to build at minimum is a quarter squisher. Added bonus is decent rail gun power range. That means in the 5,500 joule range but that's just me.

http://205.243.100.155/frames/shrinkergallery.html

http://www.matsceng.ohio-state.edu/~daehn/metalforminghb/tab...

http://www.powerlabs.org/railgun.htm

elementcollector1 - 21-10-2013 at 14:47

As it turns out, there's an indicator light built in to the camera flash circuit I hacked to build the coilgun. It's some sort of tiny neon tube shaped like a LED, and glows red when ready. All I have to do is unsolder it, stick some magnet wire in there so I can place it where I want to, and resolder it.
Anyway, construction is still ongoing - right now I'm fussing over a sort of electronic 'trigger' that consists of two pieces of blank PCB with large dots of solder on each one, and a wire in the solder. While the first board is stable, the second is on a hinge that is connected to a small thing shaped roughly like a standard gun trigger, and works about the same - when the trigger is pulled back, the two boards touch due to the solder and short. These are soldered to the two tiny prongs that stick out on the side of the camera circuit - visible here on the middle right:
<img src="http://www.spudfiles.com/spud_wiki/images/e/ef/Camera_flash.jpg" width="600" />

Apart from that, the only things left to do are solder the charge button, wire the coil and buy some capacitors. I'm planning to use a solderless breadboard, so I can add more capacitors if necessary.

[<a href="u2u.php?action=send&username=bfesser">bfesser</a>: reduced image size(s)]

[Edited on 21.10.13 by bfesser]

bfesser - 21-10-2013 at 15:13

Quote: Originally posted by elementcollector1

I'm planning to use a solderless breadboard, so I can add more capacitors if necessary.

I wouldn't recommend using a solderless breadboard for prototyping high-voltage or high-current circuits. You may experience unexpected shorts and arcing between the rails.

Metacelsus - 21-10-2013 at 15:35

Very interesting. In fact, I am working on a coilgun right now. I found eight 330V 1200uF electrolytic capacitors while taking apart some TVs (for the flybacks). I have absolutely no idea why they were in there, but they're going to make an awesome (825 J) coilgun. I prototyped a 12V to 330V boost converter on a perfboard yesterday, but I have yet to test it out. I think I shouldn't have problems with shorts or arcing, given the relatively low voltage compared to the breakdown voltage of air. Pics will come soon, at least if I remember to post them.

watson.fawkes - 21-10-2013 at 16:00

Quote: Originally posted by elementcollector1

It's some sort of tiny neon tube shaped like a LED, and glows red when ready. All I have to do is unsolder it, stick some magnet wire in there so I can place it where I want to, and resolder it.

The classical NE-2 neon indicator lamp has a breakdown voltage of around 90 V, so you'll want something with better insulation than magnet wire. If you were buying it new, you'd specify AWM (appliance wiring material) with 300 V rated insulation. You won't need much current; 24 gauge should suffice.

elementcollector1 - 21-10-2013 at 16:20

Quote: Originally posted by watson.fawkes

Hmm. I have 12-gauge solid wire, as well as 12-gauge speaker wire. I also have some fairly thin wire that may well be 24-gauge, although I'll have to check. Thanks for the tip-off!