Power Supply Design

From your post I get the feeling it's a one off. The solution below may not be the cheapest design for multiple supplies but for a one off the junk box could provide the parts.

Take 3 transformers f.e. 110V/24V with f.e. a 10VA rating . Serialize the primaries and parallel the secondaries. Watch out for correct phase. Take a fourth transformer 110/24V with at least 3x10VA rating. Connect the lot back-to-back, i.e. connect the 24V winding of the no.4 transformer to the paralled 24V windings of the triple. Mains input is on no. 4, a 330V output on the serialized triple. Give safety a thought, use split bobbin transformers.

Rgds, Jan.

Hmmm? have we got this right ??

--
Regards ......... Rheilly Phoull

[...]

Yes.

If you may think "back-to-back" does not work, please tell me why you think it does not work.

Rgds, Jan.

Go hunting for an old valve (tube) radio/amp/testgear whatever and remove the power tranformer. If your really lucky you might find something with a complete regulated power supply.

"J.Koning"

** Because :

1. You do not know AC from DC.

2. You have ignored crucial voltage ratio, regulation and saturation issues.

Ie - you have never tried it.

BTW

Are all Google Groupies septic tanks ?

........ Phil

The cheapest solution is undoubtedly a bridge rec + reservoir cap. I'm puzzled by the other replies!

If you're on 110v, get a 240 transformer, feed 110 to the 110 tap and take not-quite-240 from the 240 primary tap. As has been mentioned, transformer regulation means you wont get 240, but you should still manage 300v after rectification.

Note these 2 supplies are not isolated, so are not safe for all projects.

NT

I ignored the rectification issue. But I consider that as self evident.

Not with four trannies, with two back-to-back, yes. Works fine.

Once you are in, yes.

Nrgds, Jan.

Hi, Brian -

It can be done with 4 diodes and 2 capacitors (and and electrical outlet) where the output capacitor is the one you want to charge. Look up voltage doubler. It will double the peak line voltage (to about 330VDC).

You will need to use adequate capacitors and diodes. Also, you will need to protect the circuit from inrush with a resistor.

Take note that there is no isolation from the mains. If you are charging an isolated capacitor and then removing it from the charger, this should not be a problem (other than possibly electrocuting yourself).

Good luck.

John

You need two diodes, your basic 1 amp 1N4007's will do just fine.

And you need a capacitor of about 1/20th the microfarads of your large capacitor.

hook them up like this:

small capacitor minus lead to the hot lead of the AC line, (use a polarized line cord, the hot wire is the one that goes to the NARROW prong.)

The plus lead of the capacitor to the cathode (bar) end of diode #1. Also hook the anode (triangle) end of diode #2 to this point.

Other end of diode #1 (the triangleend) to the neutral AC line wire (the wide prong).

The cathode of diode #2 goes to the PLUS side of the capacitor you wish to charge.

other side of the big cap goes to AC line neutral (as does the cathode end of diode #1.

--
Now BE VERY CAREFUL!

You have a deadly contraption there.

The camera supplies have one advantage that most AC powered capacitors do not have. They don't mind being essentially short circuited for the long time that the capacitor has low voltage at the early stage of charging. This sort of supply uses the step up transformer as an energy storage device. The transformer is loaded with a certain amount of energy in the form of magnetic field, then that energy is dumped into the output. The lower the output voltage, the longer it takes for the energy to dump. The recharge part of the cycle waits, patiently till this part of the cycle is done before putting more energy into the transformer. This is called a flyback supply.

Most AC powered supplies will try to get the voltage up to normal very quickly by putting out as large a current as they can manage, even if this high current causes some parts like the transformer and rectifier diodes to overheat.

So if you are going to try to scale up camera supplies, I think you should be looking for another current limited approach like another flyback type supply.

And winding transformers is not so bad. It is kind of neat, that with just a little care and patience (and the right materials), you can produce an excellent device, as good as a commercial unit. Try that with an integrated circuit, or even a diode.

In article , wrote: [....]

If you split all the connections between the windings, you can get 110VAC isolated. Then if you use a doubler like this:

--- -------+--->!-------+----------- +Out ) ( ! ! ) ( ! --- ) ( ! --- ) ( ! ! --- --------------------+ ! ! ! --- ! --- ! ! -----!

If you drew it (see below) it would help some people. View in fixed font

-- --

360

But V out DC after the filter is going to be too high. If you use your idea with three transformers instead of 4, you get ~240 AC and ~336 DC

Ed

A simple voltage tripler, As it sounds like you need little current, for safety, I'd run it through 2 similar back-back transformers, say

110-12-12-110. Or use a 110-110 isolation transformer, but that will probably be more costly.

If you are runing off 230vac, just a diode and cap will get you 300v. A simple doubler will get you plenty more, all at the expense of lousy regulation.

If running off 12vdc is the requirement, maybe hack an el-cheapo low-power inverter through a transformer and rectifier?

What's the application? You will more likely get an optimum solution with more input data.

Barry Lennox

"J.Koning"

** You IGNORED the final DC voltage issue.

Dickhead.

** Not with "off the shelf" 10 VA e-core types it dont work fine. ** I and neither a Google Groups user nor an American.

BUT you assumed the OP was.

And that the USA uses 110 volt power.

.......... Phil

If you mean to the multiplier, the transformer R will do that. If you mean to the transformer, all should be well, unless its a large toroidal. So I'm not sure it needs anything.

NT

Lots of suggestions, but a lot depends on how much capacitance you are charging and how quickly you want to charge it. Does it need to be isolated from the power line?

How about some more details?

Mike

YAwn.......

Ok alot of people have asked for more details so here they go.

I have access to both US mains power (115 VAC RMS @ 60HZ) and a 13.8 VDC benchtop supply. The caps are 3,700uf at 350v rated 400 surge (currently there is a bank of 6 of them in parallel). I was hoping to charge them up to between 300-330v. I have been using a disposable camera board to charge up only a single one to 300v and that takes about 5-10 minutes. (The caps are electrolytic type and I know they are not suited at all for a pulse discharge but thats what they will be used for as they came to me free).

(The caps are electrolytic type and I know they

That depends on how large the discharge pulse is and the type of caps. I did a test a few years ago and discharged a small bank of 3-1500uf

400v computer grade caps into a coil. The resulting current pulse was 850A peak. I triggered it with a microprocessor that counted the discharges and finally shut it off after 11,000 charge/discharge cycles. I then measured the discharge pulse and there was no measurable decrease in peak current.

First and no offense indended, but

22,000uf charged to 300v is extremely dangerous so please be very very careful! Also, Caps used in pulse discharge applications have been known to explode, so enclose the entire capacitor bank in a sturdy box to contain shrapnel. Especially when you're trying to get the power supply working. If you decide to power it directly from the powerline enclose the entire finished unit in a nonconduting box of some kind.

You will have a tradeoff betwee charge time, convienience, safety features and cost. You can apporximate the charge time assuming a constant current by using T=CV/I whre T is time (seconds), C=capacitance (Farads), V is final charge voltage(volts), and I is the charge current(Amps). If you do not use a constant current the charge time can be estimated by T=6*RC, where R is the current limiting resistance in series with a constant voltage supply.

For example,

.0222F charged at a constant 0.1A rate to 330V will take a bit longer than .0222*330/0.1 or 72.6secs.

.0222F charged from a 330V power supply through a 1K current limiting resistor will take about 6*.0222*1000 or 133secs.

Whatever you decide on, keep the following in mind.

The capacitor bank will be a dead short at the begining of charge and that some form of current limiting is a must! It can be a current limiting transformer, a simple resistor, or linear current limiter. I've used a triac driving a voltage doubler even tho I've been told that that's kind of a no no. It worked well for me at low current.

A switch mode power supply is simple in concept, but very difficult to design, so I would stay away from that approach. You would most likely spend a bunch of money in blown up parts trying to get it to work unless you can find details on building one that meets your needs.

Also be sure to include some way to make sure that the caps can never be over charged, well at least try to keep them from overcharging, or you will be needing that shrapnel containing box for sure. The line voltage can get pretty high(130v), so watchout for fluctuations there.

Don't forget to add a bleeder resistor to discharge the capacitors when powered off.

And of course a proper fuse or circuit breaker for fire protection.

An easy way to do it would be a small variac powering a voltage doubler and a high wattage resistor between the voltage doubler and the capacitor bank. Just slowly bring the voltage up manually while watching the voltage on a meter. Maybe add an ammeter so you can crank it up as fast as you can while keeping the current to a safe level. Note that most variacs are NOT isolated from the power line, they are autotransformers. Cheap and simple, but not very convenient and prone to getting over anxious and blowing something up.

One other thing, be sure to disconnect(relay?) the charge circuit just before discharging or you will be drawing current from the charge power supply. If you are using an scr to discharge, the current won't drop below the minimum current required to commutate the scr.

I built a 5000J cobalt magnetizer with a .072F capacitor bank charged to 375V and finally ended up using a 500VA isolation transformer and a

24v 4A transformer with secondaries in series feeding a heavy duty voltage doubler for the base supply. The base supply is connected to the capacitor bank through a high voltage 1.6A linear current limiter and a power relay. I have 2 seperate circuits monitoring the charge voltage and current. One shuts down the current regulator when the caps reach 375V, this regulates the charge voltage to compensate for powerline voltage fluctuations and to allow user adjustment of the final charge voltage, and the other circuit drops the power relay and shuts down the current limiter if the capacitor voltage goes over 380V or the current exceeds 1.75A. This arangement safely charges up in about 20secs.

Good luck and be careful!

Mike