19 replies to this topic

[newtoolsmith]

Hi friends,

I decided to start this topic because I noticed some problems designing a high current power supply.

For electrolysis we need a high amperage but low voltage. If no special transformer is available, you can add new secondary windings to a microwave oven transformer (MOT).

I did this and wound two parallel secondary windings onto the MOT core. The wires are 16mm? PVC insulated copper cables. The addition gives 32mm? - enough aerea for 200A. The windings consist of five turns only and give 4,3VAC each.

The main problem is the rectifier. Across a diode you will measure aproximately 0,7V. Each bridge rectifier has two diodes in the current path so you will have a voltage drop of 1,4V over the bridge.

If you carry a current of 200A over the bridge rectifier there will be 1,4V*200A=280W lost power!

The second disadvantage is that 4,3VAC will give 6VDC when rectified and smoothed down with capacitors.

THIS IS WRONG!!

There are 1,4V to be substracted and 4,6V will be the resulting voltage we can use. Theese 4,6V @ 200A give a power of 920W.

When comparising the uesd and the lost power, we get a efficiency of approx 75% ( 920W/(920W+280W)=76,6% ). The transformer itself does his very best to dissipate some power too.

I found that 4,6V is quite low - my actual experimental cell runs at 4,9V and I would like the voltage to be a bit higher.

SECOND WAY:

I think I will use windings in series and reach 8,6VAC@100A. After rectifying and smoothing I will have 12,2VAC with 1,4V loss. This gives 10,8V.

The next step is to run two cells in series. Each cell will then be operated at 5,4V. This is what I wanted because you need this value of voltage to prepare perchlorates.

The rectifying efficiency increases to about 88%.

Another advantage is the lower current. This gives the possibility to use standard rectifiers.



There is one problem left: I dont understand the sign on top of the MOT.



First the MOT is called to be a 500W transformer, then the primary power consumption is said to be 990VA and last but not least the secondary Power output is said to be 1773 VA. The heat winding for the magnetron is not interesting since it delivers under 40VA.

That is impossible but I think the 1773VA are realistic (it was a 700W microwave oven and the efficiency of a microwave oven is too low to be worth any discussion).

Any ideas?

MfG
newtoolsmith

Edited by newtoolsmith, 03 October 2006 - 09:46 AM.

[BrightStar]

As you are winding your own secondary (easy to do with a toroidal transformer by the way), why not use the following configuration:



This would decrease your diode losses as the voltage drop will only be 0.7v. With bifilar winding, it would be just as quick to make.

I question the need for smoothing the supply. Rectified AC should be fine for your cell.

[newtoolsmith]

@Brightstar:

I thought about using the M2 bridge too. It might be the best choice.

As you said I already have two windings and can use them as a series with midpoint.

This morning I searched my electronic devices and found some really big thyristors with included freewheeling diode. I will check their current capacity.

MfG
newtoolsmith

[pyrotrev]

For maximum efficiency use Schottky rectifiers, these have a much lower forward voltage drop than standard types. Even so at 200A they'll need significant heatsinking. As mentioned the centre-tapped configuration gives lower rectifier losses than a bridge at low voltages (though at the expense of more wire on the transformer), you can get dual diode packages to make this easier e.g. the ST microelectronics STPS24045TV (45V, 240A ).

[newtoolsmith]

Now I know how to do it:

will use the two secondary windings in series and use the midpoint as grount potential.

Bridge rectifiers will be used because they can manage to carry 35A and they are much cheaper than schottkx diodes.

Only two diodes of each bridge will be used. Theese are the two diodes who are orianted towards the cathode of the bridge. The voltage drop will be around 0,55V.

Three rectifiers used parallel will be enough because the 16mm? winding can give more than 100A even if its used only on every second half wave.

MfG
newtoolsmith

[lavenatti]

You may want to think about adding balancing resistors after the diodes if you are using them in parallel. Some very low resistance (<1 ohm), high wattage resistors will force the current to travel equally through the diodes. Without the balancing, it is possible that one diode will "run away" and carry more and more current until failure occurs, usually with lots of unpleasant smoke.
A bit of nichrome wire is usually the cheapest way to go for this type of application. The commercial resistors may be a bit pricey.

[pyrotrev]

Now I know how to do it:

will use the two secondary windings in series and use the midpoint as grount potential.

Bridge rectifiers will be used because they can manage to carry 35A and they are much cheaper than schottkx diodes.

Only two diodes of each bridge will be used. Theese are the two diodes who are orianted towards the cathode of the bridge. The voltage drop will be around 0,55V.

Three rectifiers used parallel will be enough because the 16mm? winding can give more than 100A even if its used only on every second half wave.

MfG
newtoolsmith

You'll be very lucky to get 0.55V drop at 35A, most silicon rectifiers will be 1V+ at that sort of current, download the datasheet and check the graphs!. And current hogging will likely be a problem as Lavenatti points out, the balancing resistors will add further voltage drop. Not a problem but do think about heatsinking, otherwise you'll get smoke long before you ever make any pyro!

[newtoolsmith]

@lavenatti:

Right said! Diodes even tend to decrease their resistance when getting hot and this will increase the current difference between the diodes.

Wont the resistance of the cables be enough? I dont really want to add more resistors.

@pyrotrev:

good joke!!

Someone said somewhere: "Der optimale Arbeitspunkt stellt sich bei minimaler Rauchentwicklung ein."

In english: "The working conditions are optimized when it smokes least."

Heatsinking is absolutely necessare as we are talking about at least 50W of heat. A fan will be used to get cool air in. A heatsink is available - I will have to cut it because it is much too big for this purpose.

The 0,55V are taken out of a datasheet. I noticed that the many bridge rectifiers dont differ much in their values. Notice that only one diode will be used in every path and not two in series.

MfG
newtoolsmith

[refireworks]

When about 25amps at 5Volt is enough then you can consider a low cost solution by reusing the power supply of an old/outdated PC.

I have good experiences with it. Be however aware that these switch power supplies sometimes need to be loaded to work properly. So when using this solution first checkout on the internet the characteristics of the supply.

[wjames]

out of interest.....

I used to buy OLD arc welders - and i mean proper, OLD+OIL FILLED arc welders.


Wouldnt that be an easy plug and play soloution ?

Older versions with out 3000vdc starters were 10vDC off load, and about 7v on load - all the way upto and including 1000 amps.

Just a idea.

[Arthur Brown]

Looked at an electric welder from B&Q but 70v open cct volts really isnt on! A PSU ex a PC can be up to 500w and usually the beefiest rail is 5v which suits chlorate production.
There are lots of other things to consider before getting a PSU and trying to feed a cell with it.

[Arthur Brown]

Can anyone give me specific instructions to turn a PC PSU on! Whrn powered up from the mains it just sits there standing by, There must be a way to turn it on!

[goldmagnet]

Hi there

For those who can't be bothered building one and want a professional solution a radio ham PSU works great, I have a Palstar 1~15 volt varible 60 amp unit.

They cost about £180 new, but there are plenty of ham junk sales on and you can pick them up there.

Edited by goldmagnet, 06 December 2009 - 10:50 AM.

[CCH Concepts]

has anyone considered just making a power supply. you can buy pwm ic's that require minimal passive components, attached this to a mosfet array and you have a basic power supply. you can then get more complicated and add surge protection, thermal protection and so on. but you could make it for quite cheap and as long as its correctly fused worse case scenario is you would need to replace the power stage.

this ic has suggested curcuits in the data sheet.


pwm controller

200 watt transformer

[Alexander]

A lot of power supplies can't regulate all the way to zero current and are damaged by running with no load. This page has a method that looks sound. Shorting green to ground should do it, somewhere I have a note of the pin number on the motherboard connector.

http://www.wikihow.c...ab-Power-Supply

There is a £13.67 500W (30A 5v line) EZcool power supply which has a very bad reputation and an 800W (51A 5v line) Storm PSU for £26.42. I'm looking at scan computers prices.

The electrolytic cells I made (long time ago now) were very non linear in current draw. These weren't chlorate or perchlorate cells but I'm concerned that a fixed voltage regulated power supply will either not run the cell or be in overload. It may be possible to modify the circuit of an ATX supply. If the coils can be rewound this could provide a huge amount of current for the money. Modifying a commercial PSU seems no more hazardous than building a supply from scratch.

The easier DIY designs have some drawbacks, like a live output if the transistor latches in the buck converter. I'm looking into safer designs like a flyback converter. Schottky diodes have the drawback of low voltage ratings. The biggest advantage of going this route is that you can control how the power is regulated.

Rewinding a microwave oven transformer seems like a good plan, compllete microwave ovens in the 700-800W range are available from astonishingly low prices £35-40. These would have transformers inside rated over 1kw. I haven't been able to find the transformers cheaper then the oven so far. They aren't magnetically limited and could waste a lot of power if the secondary voltage is wrong. At high current 50A+ a single diode could have a drop of 2V.

I'm still scratching around in a lot of directions. This thread has sparked a lot of ideas.