Constant Voltage Power Supplies / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Constant Voltage Power Supplies

1 2 next

Move Thread

LAN_403

Dave Marshall

Mon Sept 15 2008, 01:29PM

Registered Member #16 Joined: Thu Feb 02 2006, 02:22PM
Location: New Wilmington, PA
Posts: 554

I've long had a number of ideas gleaned from my previous coilgun projects that I have wanted to roll into one huge project, but so far I've been held back by one key component I can't find a reasonable answer to.

I've come to the conclusion that capacitors simply aren't a viable power source for very high efficiency systems, and they complicate system design significantly. I want a power supply capable of delivering somewhere between 150 and 250v, and current on the order of 250 Amps with a nice steady power output.

So far I've considered a couple different options:

Battery Bank:
A bank of car batteries in series seems like it could meet the requirements set forth above, but I'd need between 12 and 20 batteries. At ~$50 each, and usually weighing more than 30lbs, this presents a pretty significant financial and logistical challenge. Since I don't need alot of capacity, a very small battery with a lower capacity would be quite sufficient, but I've yet to find anything like a 15AH SLA that can do more than ~50A, so now I have to consider a parallel/series arrangement. Sounds like a pain in the ass.

Generator:
Smaller and more convenient than a battery bank, but the cost issue still exists. To get a generator capable of sourcing the ~30kw I would want on the low end, I'm looking at several thousand dollars. It was suggested that I parallel a number of car alternators attached to a car engine, but I'm not mechanically inclined enough to be confident in this option, and the monetary outlay would still be substantial for the parts.

Mains:
If I had the balls to tap straight into the incoming 220v line from the pole, it would nicely meet my requirements, but that sort of thing makes me very nervous. Anyone have experience with doing this?

The largest breakers in my fusebox are rated at 60A, and are for the 220v three phase for the dryer. Given the very very low duty cycle for this application, getting on the meter side of those breakers seems like a viable option, but again...holycrapdeath.

So, I guess my options are some low capacity but very high current batteries (Anyone have any suggestions?) or the mains. Thoughts?

-Dave

aonomus

Mon Sept 15 2008, 02:31PM

Registered Member #1497 Joined: Thu May 22 2008, 05:24AM
Location: Toronto, Ontario, Canada
Posts: 801

What about a stack of alternators connected to an engine? Compulsator?

It just occured to me, couldn't you use a DC welder as a current source?

As for AC mains, it would be only sane if a breaker, and isolating transformer. As for giving the 250A spike would it be at all possible to find or make a gigantic (read 1-2ft in diam, probably 50-200lb in weight?) torroid as a reactor so that whenever you draw AC you don't cause the magnetic trip in your breakers to go off.

Don't forget that you'd have to rectify the AC mains to DC 170V (or whatever UK line voltage is peak to peak) and you'd likely blow whatever rectifier you had from the spike, probably heatsink and parallel a ton of bridge rectifiers.

Don't forget that the simple SCR switching is avalanche based so you'd never be able to turn off your coil without shutting off AC mains. You might have to go to MOSFET/IGBT which would let you do hard switching (although from what I've heard MOSFETs and IGBTs are much less forgiving for spike current).

If you are nutty enough to try AC mains -> DC -> Coil, you could always have several IGBTs and be able to make a multistage coilgun cheaper than buying a set of caps, scrs, and parts for each stage.

Either way I'd like to see whatever other people think

Steve Conner

Mon Sept 15 2008, 03:00PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Just find a lab with three-phase 120/208V service (assuming you're in the USA) and use a six-pulse bridge rectifier with huge stud diodes. A six-pulse rectifier gives a reasonably ripple-free output without any smoothing capacitors at all, and the diodes should be as strong as any huge SCRs or IGBTs you might use in your magnetic cannon of doom.

Single phase is out. It needs a big capacitor bank to smooth it, and you said you didn't like capacitors.

I once built a test rig to torture TVSs for an intrinsically safe application. It basically connected the TVS to the mains for one half cycle through a big SCR. I was getting 130A peak from a wall socket, and I had an isolating transformer in there to give a bit of reactance for current limiting. The higher mains voltage here helps.

...

Mon Sept 15 2008, 03:36PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

I would recommend using the battery method, I have my doubts that trying to suck uS pulses of 200+a out of your mains is not going to work well due to the inductance of everything. When you consider that the load you are feeding is a like 100 turns coil, and that there is hundreds of not thousands of feet of wire between you and the next transformer (which is a big inductor in itself...) I think the mains is going to look a lot more like a constant current power supply at the time scales you are looking at.

The batteries on the other hand have almost 0 inductance, but considerable ESR. If you sized the bank to have a fair bit of extra voltage (you would get the most power draw making a 500v bank and a coil that had the same ESR to get 250v across the coils resistance and 250v across the batteries resistance) such that the batteries voltage could sag significantly you could probably get away with much smaller batteries. I know that a 7a/hr sla will put out on the order of 200a when presented with a short across the terminals, you might be able to find a ESR value for the batteries you are looking at to see how that would work.

Dave Marshall

Mon Sept 15 2008, 03:39PM

Registered Member #16 Joined: Thu Feb 02 2006, 02:22PM
Location: New Wilmington, PA
Posts: 554

aonomus, I ruled out the alternator idea in my first post. It just sounds like an enormous outlay of time and effort for a system that likely won't be any smaller or simpler than a large battery bank, and then I have to worry about maintenance on the ICE thats driving it as well. The associated parts probably wouldn't be a whole lot cheaper than batteries anyway.

The point of this particular system will be to move away from the inefficient and overly complex capacitor/SCR topology. I'll be using transistors to switch the system, so constant current/voltage isn't a problem.

Rectification won't be a major problem, as I already have more than a dozen 200v 400A diodes. A few of those carefully employed should be just the thing.

I have 3 phase 220v service to my home, and can get ready access to it if necessary. My home's wiring is about 20 years old, and there isn't a master breaker in the box. I'll have to check and see if its outside the house near the meter. The 6 pulse bridge rectifier looks like it could be just the thing here. I've heard from several people that they've drawn 100+ Amps from a wall socket in pulsed applications, but that idea makes me nervous for prolonged use. Running a nice 4 gauge line to the 220v line coming from the meter and putting a set of 200A breakers on it sounds like a much more reasonable solution.

I could be satisfied with as little as 150A, especially if that 150A was from a rectified 220v line, which would yield >300vdc.

Dave

Electroholic

Mon Sept 15 2008, 09:40PM

Registered Member #191 Joined: Fri Feb 17 2006, 02:01AM
Location: Esbjerg Denmark
Posts: 720

If cost is not a problem, I would just use a capacitor/li-ion hybrid bank. or even lipo.
Good Li-ion cells can handle 20C peak, thats 40A to 52A for a 18650. And at 3.6V per cell, you don't need much to reach 200V. 50S 5P will give you somewhere around 180V 200A.

rp181

Mon Sept 15 2008, 10:57PM

Registered Member #1062 Joined: Tue Oct 16 2007, 02:01AM
Location:
Posts: 1529

i wouldn't risk drawing pulses with the old wiring, if your house burned down, insurence wouldn't pay.

Steve Conner

Tue Sept 16 2008, 10:51AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Is that as in 220V from phase to phase, 127V from phase to neutral? Sounds ideal

When you rectify that you get 300-something volts DC, the positive is 150V above neutral and the negative is 150V below. To get the six-pulse characteristic you need to hang your load between the positive and negative lines.

If you're worred about burning the house down, just use some big-ass fuses suitable for the wire gauge. Fuses will take higher peak current than breakers, but it's not peak current that starts fires, it's RMS.

I'd be more worried about complaints from the neighbours when you make the lights flicker in the whole street

rp181

Tue Sept 16 2008, 12:04PM

Registered Member #1062 Joined: Tue Oct 16 2007, 02:01AM
Location:
Posts: 1529

you could be the first ameture to build a compulsator.... wouldn't that be awesome?

OZZY

Wed Sept 24 2008, 01:22PM

Registered Member #511 Joined: Sat Feb 10 2007, 11:36AM
Location: Somerset UK
Posts: 55

Hi Dave

It sounds like you need a "stiff" power supply capable of delivering high peak power. This is the same problem faced by the guys who fit high power stereos in cars. They use a stiffening capacitor to bypass the load.

I would use the three phase 220V supply full wave rectified with some big ass diodes, and feed that into a large bank of capacitors. Remember that the capacitors are not there to smooth out voltage ripple from the mains, they are there to bypass the peak current required by the load.

The cap bank must have a lower impedance than the mains supply line for it to do any good. The cap bank should isolate the load from the inductance of the supply lines. The rectifier will only need to pass the average current which may only be half the peak current.

Ozzy

1 2 next

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