Drastically decreasing capacitor ESR

[jmaniglia]

Hey all,

I'm and Aerospace undergrad at Cal Poly working on my senior project, which has turned out to be more of a EE project than anything else.

Anyways, I'm building a capacitor bank for the firing of an electromagnetic rail gun. I'm attempting to make a PFN (pulse forming network) and capacitor ESR is KILLING me when it comes to buying a lot of cheap(ish) capacitors for my capacitor bank. Is there any way to practically eliminate ESR of a SINGLE capacitor. I am already connecting capacitors in series and parallel, and that can only decrease the ESR so much without buying a thousand of them. I would like to decrease ESR by orders of magnitude. Any thoughts??

One thought that I have had was connecting another resistor in parallel, that seems to increase current, but not the overall dampening of the system, as the PFN I'm working with is a slightly underdamped system, and the addition of capacitor ESR is causing it to dampen out to an unusable state.

PS: I am simulating these circuits using MATLAB and simulink, and it seems to be a very reliable method, so please do not hesitate on just throwing an idea out there with no mathematical backing. I WILL just throw it into MATLAB and see what pops out.

For any additional thoughts/questions about our railgun we are designing please email me at the email in my account here (jeffmaniglia@gmail.com), please don't post generic rail gun stuff in this thread, i'd like it to stay on topic.

 

[zomgineer]

Have you looked at the gigantic capacitors that they use in car audio systems? They have low ESR ('cause they're so big). They can be as big as 1 whole Farad or more. Maybe instead of using many small inductors and caps you could use only a few big inducutors and caps. Bigger coils with thicker gauged wire and massive caps will have less passive resistance.

Now I want to build a rail gun.

 

[jmaniglia]

I have looked into the large car audio systems. The problem with those is that their operating voltage, for inexpensive ones, is far too low. I'm looking at 1-2 kV for these babies! Most of the caps I'm looking at have a rating of 450-630 V

 

[zomgineer]

You can put caps in series to stack the voltage rating.

Edit: putting caps in series will also reduce the total capacitance.

 

[jmaniglia]

Exactly. Thank you for the ideas my friend, but that's why I've included decreasing the ESR with a single capacitor, since I have been successful on paper making a circuit with a low enough ESR, I've been unsuccessful in doing it within the budget...since to triple the voltage rating while keeping the capacitance you need 9 total capacitors...

 

[jmaniglia]

And sorry. to clarify, I'm attempting to get an ESR of LESS than 1 milli-ohm. I know, very lofty...

 

[waht]

Keep in mind that the ESR is an intrinsic property of a capacitor that arises from the physical construction and chemical make up needed to build a capacitor. What that means is that you are stuck with it.

In the general, the larger the capacitance, the smaller the ESR becomes. So you would have to hunt for a giant cap which would probably cost a fortune.

One way to reduce the effects of ESR is to hook up many capacitors in parallel. That way the ESR resistance of each capacitor will be in parallel which then divides the overall resistance. So 10 identical capacitors hooked up in parallel would reduce ESR by an order of magnitude

 

[jmaniglia]

Thank you what. I do understand that. And my system is one that already has around 20 to 30 capacitors in series and in parallel. series to decrease voltage (and sadly increase overall resistance) across a single capacitor, and parallel to rid myself of ESR. But I've run into a block. And I was wondering if anyone here knew of a way to add components (other than capacitors) to capacitors that would effectively decrease the ESR seen by the rest of the circuit.

I was experimenting with a resistor in parallel and it seems to increase the current through the capacitor and take away from current that the capacitor itself sees, BUT it does not seem to take away the dampening of the system. A semi-oscillatory system around a DC voltage (during capacitor discharging) was trashed with the addition of ESR into my circuit.

 

[zomgineer]

Let me introduce you to http://www.digikey.com/" . You can search for capacitors and you'll see that several types of caps come up. In particular, film capacitors tend to have high voltage ratings.

Here's a http://search.digikey.com/scripts/DkSearch/dksus.dll?Cat=131088&k=capacitor" that have some DC voltage ratings of 1000V and above (up to 5kV). There's 11 thousand there. You can select one or more options in each category to filter it down. I looked at some 1kV caps that had and ESF of 1mOhm and they tended to be around $60 or more.

If you want some kind of circuit topology that can reduce the ESF then the only thing I can think of is to put caps in parallel. Maybe you need to http://www.powerlabs.org/railgun.htm" . Read about his design. Engineers are really sympathetic towards students so he may help you if you Email him.

 

[Averagesupernova]

I had lots of suggestions until I read how high your operating voltage is. Keep in mind that putting caps in series can do nothing to your ESR except increase it. It may help you to increase your working voltage but that's about it.
-
A side note. There has been much discussion over the years about ESR in the car audio communtiy concerning the large capacitors placed in parallel with the power input to the amplifiers. It is pretty obvious, although not necessarily accepted in the car audio community, that a high ESR on a capacitor hooked on the power input for an amplifier is unable to supply current when the amplifier needs it. Jmaniglia, you are battling the same thing some car audio geeks are, only at a much higher voltage and you are smart enough to realize what you need to change.

 

[jmaniglia]

Well, now i will take that as there is not magic box to decrease ESR. NNNNNOOOOO! And i was hoping for an easy and cheap solution. =( Never works that way

 

[Carl Pugh]

You might recheck your calculations.
You are accelerating a physical body and should not require an extremely fast pulse?

You could use a water capacitor for fast pulses, however they have their own problems and you may not want to go there.

I have never built a rail gun or a water capacitor, so this advice is all theorical.

Good Luck, sounds like a fun project

 

[mheslep]

Given you are after a fast current pulse, it may be that series inductance is most of your problem (i.e. v=Lcap di/dt). If that's the case, then going with surface mount caps is probably the single biggest step you can take to reduce parasitic inductance, along with appropriate layout traces. It's also easy to parallel a large number of surface mount caps to bring down both resistance and inductance. Also, without checking, as I recall tantalum caps will give you the lowest ESR though they don't have high voltage ratings.

 

[mheslep]

.

PS: I am simulating these circuits using MATLAB and simulink, and it seems to be a very reliable method, so please do not hesitate on just throwing an idea out there with no mathematical backing. I WILL just throw it into MATLAB and see what pops out.
:

You'll want to add in the parasitics of your components (inductance, ESR, capacitance) to the model, i.e. every resistor has an inductance and every capacitor has a resistance. These parameters make a big difference in a fast current pulse system such as yours.