800vdc via pwm from main 240vac 50hz / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

800vdc via pwm from main 240vac 50hz

1 2 next

Move Thread

LAN_403

TechNerd

Wed Nov 18 2009, 10:21PM

Registered Member #2289 Joined: Thu Aug 13 2009, 02:49PM
Location: Jylland, Denmark
Posts: 19

Hi

I'm wondering if it's possible to create a 800v capacitor charger

using

240vac 50hz through a voltage doubler making about 380-480v
and then use a flyback at 100khz switching freq. to step it up to the 800-900v

I see a problem with the 50hz charging freq. of the voltage dubbler compared to the 100khz used in the flyback

please help

Proud Mary

Wed Nov 18 2009, 10:53PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Hi Techo,

I use doublers and triplers direct from 230V/50Hz mains electricity whenever possible.

For example, I will use a PCB-type mini isolating transformer with a doubler to create 600V for GM tubes.

In practice, I have found that a doubler using 100nF/1kV caps is good for 1mA before it sags,
far more than enough for GM tubes.

There is no need for the complexity of LOPT (slang: flyback) to generate 900V.

You need only select capacitors of suitable size, connecting them in series in the case of aluminium electrolytics, to produce your 900V.

May I suggest that you always place the needs of the application first, and then work out the voltage and current needed by that application?

In its raw form, no sensible answer can be given to your question, as you have not specified the current required by your application.

Try to think like this:

What are the power requirements of my application?

How can these be met within the project's budget?

Stella X

Patrick

Wed Nov 18 2009, 11:26PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

how much wattage do you need (ie. how long are you willing to wait for your capacitor to charge?)
how large is your final cap?

At 50/60Hz you will only have 120 pulses of power per second , so if you want a fast charging time you need large uF caps in your multiplier. Or you could full wave rectify the mains AC then switch at 50Khz on both your multiplier and flyback, but this is more complicated but then you can use small caps. It really depends on your purpose, $, and needs. (Skill too, so dont die.)

PS. I just want to second proud mary's thoughts too. you dont need a flyback really, if you have the skill/need you could just make a step-up inductor based smps,(Yet more complicated) since you wont have isolation with a multiplier alone. You dont need isolation right?

-Patrick

TechNerd

Thu Nov 19 2009, 11:22AM

Registered Member #2289 Joined: Thu Aug 13 2009, 02:49PM
Location: Jylland, Denmark
Posts: 19

Hi again

SPEC:
2 x 10mF - 400vdc caps in series (makes with my math 5mF - 800v around 1.6kJ)
charge time 1 or 2 minutes
is to be used in a coilgun

Technerd spec:

Electronics Technician
Surrounded by engineers at work so plenty to ask..

electronic components is not a problem to get
budget: well it's limited but almost anything is possible to get at work

hope it helps

Proud Mary

Thu Nov 19 2009, 01:01PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

TechNerd wrote ...

Hi again

SPEC:
2 x 10mF - 400vdc caps in series (makes with my math 5mF - 800v around 1.6kJ)
charge time 1 or 2 minutes
is to be used in a coilgun

My yellowing ivory slide rule - once the property of Otto Hahn - comes out dead on at 1.6 Joules.

(By the way, unless your electrolytics are a big as paint buckets, I'm sure you mean 5uF total, rather than 5mF!

Almost everyone has to do electrolytics-in-series to get over the practical 500V barrier, but there are practical aspects and limitations to this practice which show up the difference between a 'perfect' capacitance, and a real world iteration.

Real world capacitors have two very disagreeable properties - firstly, what is called ESR (Equivalent Series Resistance) and worse for you, I think, inductance.

I know absolutely zip, zilch, nada, about coil guns, but instinct tells me that there will be some definite relationship between
acceleration and the steepness of the wave front.

(A 'perfect' wave front for your purpose, I suspect, would be an impossible straight vertical line going from OFF to ON in zero seconds.)

Recalling that inductances in series are added, you'll see that seriesing two caps will double the inductance as well as halving the total effective capacitance. Any inductance, no matter how small, requires time to charge - what is called its time constant - which we sum up in the expression t = LR.

Now this small time interval imposed by inductance will slow down the devolpment of your wave front, so that instead of going up in the impossibly perfect straight line from OFF to ON, the wave front will now lean slightly to the right, slowly building rather than being a clean switch. (We are talking nanoseconds here).

So far as I can see from a quick twiddle with the slide rule, the best possible coil gun would have only one single turn in order to have the shortest possible rise-time of the wave front. This is most certainly possible, but to give a full account of it would make my reply into a whole chapter. But I'll sum it up for the interested as 'driving ultra-low impedance loads'

(Old forum regulars were probably wondering just how long it would take before I turned the subject to krytrons, thyratrons and pseudo-spark switches - so I'll leave it there!

)

Summary: two caps in series do not perform as well as one single one of the same effective value. I would advise de-rating the combined working voltage total by a good 20% to give the caps a reasonable life-time, and perhaps even more where they are being asked to look into what will seem to them to be not much above a dead short. The biggest risk to aluminium electrolytics is the inevitable negative-going voltage swing, which can be partly mitigated by the use of very heavy duty efficiency diodes. If you have access to PTFE caps of the same size, almost nothing commericially available can out-perform them in terms of dielectric relaxation time,
which comes to the front of the problem when you are trying to chisel off nanoseconds from your wave front rise time.

And what would I do looking out from the top of a pagoda? I would use dimethyl sulphoxide, (DMSO) the liquid dielectric for the girl who has everything. These experimental capacitors can out-perform just about anything else by a whole order of magnitude. A Rayleigh PFN made of these dumping into a single turn at the temperature of liquid helium would be my starting point for a really cool (

) coil gun.

PS: I have bought reagent-grade DMSO from a supplier in Poland, and found it to be of excellent purity at a good price.

TechNerd

Thu Nov 19 2009, 05:57PM

Registered Member #2289 Joined: Thu Aug 13 2009, 02:49PM
Location: Jylland, Denmark
Posts: 19

Proud Mary

By the way, unless your electrolytics are a big as paint buckets, I'm sure you mean 5uF total, rather than 5mF!

Ohh yes they are freakin' HUGE 9cm diameter * 25cm high.

and yes one capacitor might be better then the two of them in series.

all I'm interested in is an Idea of how to get a voltage supply of about 900v, creating a charge time of the caps around 2min...

got an 220vac to 2 * 22vac transformer - coupled backwards my math makes an 2200vac but with a very low current

Patrick

Thu Nov 19 2009, 06:59PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Sweeeeeeeeeeeeeeeeeeeeet caps, do be careful.

I second Proud mary's concerns though. I think a Boost converter is your best bet.

If your needing 1.6Kj Im thinking it would be way better to just use an inductor and catch diode, as Proud Mary says caps that large have unwanted side effects. (also it will whiegh a hell of alot too.)

-Patrick

Proud Mary

Thu Nov 19 2009, 09:21PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

TechNerd wrote ...

all I'm interested in is an Idea of how to get a voltage supply of about 900v, creating a charge time of the caps around 2min...

Then I don't quite understand why you can't use an ordinary voltage quadrupler circuit direct from the mains, which any schoolgirl could scribble down on the back of an envelope with an eyebrow pencil.

cjk2

Thu Nov 19 2009, 11:34PM

Registered Member #51 Joined: Thu Feb 09 2006, 04:17AM
Location:
Posts: 263

You will not be able to drive a 22v transformer in reverse with much more than 22v before it saturates. To say it in a different way: your idea of driving a transformer in reverse and applying more voltage than a given winding was designed for will not work.

Why not use a quadrupler as has been suggested before? I would recommend you drive it through an isolation transformer actually for safety.

wylie

Fri Nov 20 2009, 03:28AM

Registered Member #882 Joined: Sat Jul 07 2007, 04:32AM
Location:
Posts: 103

I think his original idea to use a "flyback" was really referring to any method of using a ferrite xformer for the final step-up. I'm just going to second everyone elses' comment that "its not really necessary." Unless you needed a low ripple SMPS type supply. (and coil guns dont)

I think Harry was spot on being more concerned with the ESL (the caps' inductance). You want the wave to be as fast as it can be when it dumps into the work coil, yes? Otherwise it's just going to burn energy as heat or pull on the projectile after it's already been accelerated, right?

So your storage caps and all the connections should have less L than your work coil. (EDIT: for maximum power transfer)

What about Steve's CCPS? If the power levels call for it i mean.

Feel free to ignore/correct any of this, just thinkin "outloud".

1 2 next

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