I was not aware of ripple causing electrode issues, but I never claimed to know much about electrolysis. That is why I said to evaluate the requirements. For just one device, reusing a microwave oven transformer with a different secondary makes sense. No clue where the production line came from.
In this forum sometimes threads wander way far from the original topic, which is a big part of why I sometimes skip sections. And there are some folks chained to every possible fear who love to dump that on us, and so they lose all credibility with me very fast. The cure for fear is knowing just what to do in the situation.
The package will be a bit heavy and with that high of a constant load current you will need heavy duty connections. And I do suggest watching it closely for the first few hours.

 

Looking closer at my drawing, I showed the second transformer center tap connected to a ground. That may cause some confusion. THAT ground is simply the negative end of the rectified circuit. Here's a clarified drawing:

 

The only I have down solid from posting is that the way I thought this would work is completely wrong and I shouldn’t do it the way I was going to try.

If you'd started with the question "I need 5volts at about 24amps, with low ripple, whats the best solution", I would have said straight off:

Buy a 5v, 30A SMPS for $30

or

Repurpose an old PC ATX supply

I would never have suggested the approach you're hell-bent on because its completely the wrong approach. I don't know where your ex-friend got those capacitors and that huge choke from, or how much ripple he really had, but I can say its completely the wrong approach where that sort of current is required. And replicating it without access to his resources is likely to be a very expensive exercise. Big transformer supplies like that date from the middle of the last century (I used to build them too) and pi-filters are not for high-current supplies, the magnetics just don't work.

I hear your anecdotal evidence regarding ripple. I did some reading of research papers on the subject. Apparently >13% ripple reduces conversion efficiency by 4%, but efficiency on the best commercial electrolysis cell is only 60% anyway. Your losses on that power supply would put conversion efficiency at around 30-35%. There is some research evidence that ripple above 30% (unsmoothed full-wave rectified is 48%) coupled with excessive voltage above the natural cell voltage causes impurities in the electrolyte to etch and corrode the electrodes. So better voltage regulation and cleaner electrolyte is at least as important as ripple (commercial units apparently circulate electrolyte through a filter to stop build up on the electrodes). But zero ripple is not only not easily acheivable on the route you're taking, but is unnecessary. 10 - 15% ripple would certainly appear to be acceptable. But that wouldn't change my recommendation.

 

. I thought I read that somewhere someone mentioned a production line.

Not in errorat all, That was me answering a question he asked about why his latest MOT had soldered output terminals, terminals that were broken.

 

I would never have suggested the approach you're hell-bent on because its completely the wrong approach

I usually drop out of threads after it ends up being about water electrolysis. It usually means they are trying to make hydrogen or the oxygen to run a motor, while it works to a certain point it isn't efficient or in the end worthwile. The output is so small that it isn't worth doing it. Welding gas suppliers don't go down this route when making it because there are much cheaper and faster ways of making hydrogen, chemically.

 

Looking closer at my drawing, I showed the second transformer center tap connected to a ground. That may cause some confusion. THAT ground is simply the negative end of the rectified circuit. Here's a clarified drawing:
View attachment 211827

THANKS, Tony. This is exactly the circuit that I have been trying to describe.Even the colors are right, although I never mentioned the colors..

 

They could also be electrolyzing ammonia with a peroxide additive. The problem is when the result contacts the metal electrodes it forms an unstable metallic salt.

 

If you'd started with the question "I need 5volts at about 24amps, with low ripple, whats the best solution", I would have said straight off:

Buy a 5v, 30A SMPS for $30

or

Repurpose an old PC ATX supply

I would never have suggested the approach you're hell-bent on because its completely the wrong approach. I don't know where your ex-friend got those capacitors and that huge choke from, or how much ripple he really had, but I can say its completely the wrong approach where that sort of current is required. And replicating it without access to his resources is likely to be a very expensive exercise. Big transformer supplies like that date from the middle of the last century (I used to build them too) and pi-filters are not for high-current supplies, the magnetics just don't work.

I hear your anecdotal evidence regarding ripple. I did some reading of research papers on the subject. Apparently >13% ripple reduces conversion efficiency by 4%, but efficiency on the best commercial electrolysis cell is only 60% anyway. Your losses on that power supply would put conversion efficiency at around 30-35%. There is some research evidence that ripple above 30% (unsmoothed full-wave rectified is 48%) coupled with excessive voltage above the natural cell voltage causes impurities in the electrolyte to etch and corrode the electrodes. So better voltage regulation and cleaner electrolyte is at least as important as ripple (commercial units apparently circulate electrolyte through a filter to stop build up on the electrodes). But zero ripple is not only not easily acheivable on the route you're taking, but is unnecessary. 10 - 15% ripple would certainly appear to be acceptable. But that wouldn't change my recommendation.

He used a Samsung transformer a model I currently have but damaged the primary. I have another almost identical model. The whole reason he did it this way was, it was cheap, easy, and like I’m listening to you guys I listened to him. I would consider him an expert in chemistry and he claimed it yielded the best results. The caps were the only expensive part. I’m not hell bent I was just trying to replicate. I find it interesting that I was able to find another person using the same method to create a psu for that purpose on YouTube. It is possible that he got that method from that YouTube video, but it is also likely he would have informed me of the channel as a resource. It is completely possible that I misunderstand ripple/voltage regulation as it pertains to electrolyte/electrode health and efficiency. There could have been some form of regulation I am unaware of.
ATX power supplies are 3.3, 5, and 12 volts. I would have to purchase or find a very high wattage unit to get a high amperage, right? Then I would have to set it up for variable voltage which still requires a voltage regulator on the 12 volt supply.

 

THANKS, Tony. This is exactly the circuit that I have been trying to describe.Even the colors are right, although I never mentioned the colors..

Tony and Mr.bill, am I right to assume that you are in the camp that believes what I’m trying to do is feasible? I base my assumption on providing the schematics and agreeing with each other. Do you think it is ridiculous to even try, like Irving is suggesting?

 

Btw Irving I don’t know how this picture didn’t upload, but here is his scope while running that setup. Should have been with the other pictures. Idk maybe it was and doesn’t change anything.

 

Tony and Mr.bill, am I right to assume that you are in the camp that believes what I’m trying to do is feasible? I base my assumption on providing the schematics and agreeing with each other. Do you think it is ridiculous to even try, like Irving is suggesting?

Certainly from a power supply point of view the project is feasible. A fair amount of work, but certainly it will deliver as requested. No comment about the chemistry aspect, only about the electrical part. I am not convinced that a $30 power supply with those specs would be very reliable.

 

Certainly from a power supply point of view the project is feasible. A fair amount of work, but certainly it will deliver as requested. No comment about the chemistry aspect, only about the electrical part. I am not convinced that a $30 power supply with those specs would be very reliable.

Irving was talking about 30 dollar power supply for delivering 5 volts. After he mentioned using an atx power supply and regulating the 12 volt rail to produce the 7 volts, I believe that’s how it would be done. I would think maxing out an atx power supply would be just as wasteful and bad as using this transformer, but I can’t say for sure. I saw the schematic, so should I research building a bridge rectifier. I’m sorry I keep going back and forth, but I would like to start on this. I can’t imagine it is so dangerously complicated and wasteful that I shouldn’t do it. Should I build the bridge rectifier, or buy one, if building is the way can you provide some insight.

If the atx power supply is better I’m fine with that too.

As for the filter, tony didn’t have that in the schematic, is that because it isn’t needed, or because he was taking it one step at a time?

 

hat schematic was only to explain the concept. The filter is certainly needed to provider smooth DC. One step at a time, for sure.
The flaw in the ATX power supply idea is that the high power section is the 5 volt part, or now, the 3.3 volt part It could be done to have one supply of 3.3 volts in series with another supply of 5 volts, which would deliver a total of 8.3 volts, just a bit more than the requested 7 volts. And I am not sure if we ever got an answer as to exactly how close it had to be.
At 30 amps there will be volyage drops and so it may be that starting out with more will deliver 7 after all the voltage dropping happens in the wires.

 

hat schematic was only to explain the concept. The filter is certainly needed to provider smooth DC. One step at a time, for sure.
The flaw in the ATX power supply idea is that the high power section is the 5 volt part, or now, the 3.3 volt part It could be done to have one supply of 3.3 volts in series with another supply of 5 volts, which would deliver a total of 8.3 volts, just a bit more than the requested 7 volts. And I am not sure if we ever got an answer as to exactly how close it had to be.
At 30 amps there will be volyage drops and so it may be that starting out with more will deliver 7 after all the voltage dropping happens in the wires.

It seems to me that it would be easier and more accurate to wind a transformer.

 

I have been reading about this stuff as much as I can while still going to school and work. To all warning against this, is it possible to pull to much power from an outlet without popping the breaker? Is this something you were warning me about without saying it?

 

The power available at a normal US 120 volt 15 amp outlet is 15amps X 120volts= 1800watts. That is a maximum design value. The power that is desired from the power supply package is 30amps X 7 volts=210watts. So there is an over 8 times safety margin as far as load capability. So the total power is not close to being a problem. The reason for my caution is that 30 amps is a fair amount of current and any inadequate connection point will tend to overheat.

The comments about using an ATX power supply were just to explain what could be done, not by any means to recommend that option. Changing the output voltages of those supplies requires knowing just which parts to adjust and understanding how that circuit functions. THAT part would be a challenge for most electronic engineers.

 

Long Post Warning, but I'm going to do this once, then I'm out of here. With the exception of a couple of posters (and they know who they are) there's a lot of misinfo on here based on little knowledge/experience. So here's my summary.

I'm not an expert but I do like to have facts and I like to read base material rather than rely on partial info. So the first thing I do is some research.

The electrochemical cell voltage of an electrolysis reaction varies from about 1.2v to 2.2v, never higher. Therefore, allowing for various losses at high current, apparently mainly in the interface between electrode and electrolyte due to gassing, and some cable losses, 5v is more than adequate. If you need to go much higher at the power supply output then you're losing 50W or more in the cabling which is crazy (2m of 10AWG = 6mOhm, at 25A this is ~4W) Conversion rate is roughly proportional to current but losses increase as the square of the current. The law of diminishing returns applies.

Commercial units use carefully regulated low voltage per cell, moderate current and stack cells in series, so total voltage on the stack is 100v or more to get close as possible to theoretical conversion rates of 60% as previously stated.

I may not have all the facts at 100% but I now have a better feel for it. What is needed is an efficient, regulated, low voltage, moderate current, variable supply with <5% ripple.


That picture is close to useless. It coud be looking at a 7v rail, theres no way to tell. lets give it the beneft of the doubt and say its ac coupled so only looking at the ac component. Again that could be 1v/division or 1mV/division, without that info it tells us nothing. (why are all these photos so out of focus you can't get useful info out of them?)

It seems to me that it would be easier and more accurate to wind a transformer.

Which, sadly, just shows how little you know or understand - thats actually the least accurate and definitely the hardest way to go... but don't let me stop you trying.

Irving was talking about 30 dollar power supply for delivering 5 volts.

The 5v supply was just by way of example, you can have other voltages by adjustment. Those $30 supplies are pretty reliable, but for even higher reliability go for a higher wattage supply. I've used 400W versions in a CNC machine that runs virtually 24 x 7. Or go for a branded supply, such as Meanwell, 5v @ 40A is ~$40.

After he mentioned using an atx power supply and regulating the 12 volt rail to produce the 7 volts, I believe that’s how it would be done. I would think maxing out an atx power supply would be just as wasteful and bad as using this transformer, but I can’t say for sure.

ATX supplies are SMPS - they are 80% efficient or more, compared to the 50% you'll be lucky to get with alternator rectifiers. Once again, you have a lot to learn.

Their voltage regulation is better than 0.5%. SMPS is the way all power supplies have gone, you rarely see a linear supply these days for anything bigger than an amp or so. We are not maxing out the supply, its barely using 50% capacity of a typical 500W box; again, if it worries you go for a bigger ATX supply.

The flaw in the ATX power supply idea is that the high power section is the 5 volt part, or now, the 3.3 volt part It could be done to have one supply of 3.3 volts in series with another supply of 5 volts, which would deliver a total of 8.3 volts, just a bit more than the requested 7 volts.

There is no flaw, this is common practice. I've built several high current supplies for battery chargers for LiPo packs this way.

An ATX supply of 500w designation or more will easily give 5v @ 40A which is more than enough. Bigger wattage supplies will give proportionately more. There's a wealth of videos on YouTube on how to make ATX supplies produce a variable output on the 5v rail from near 0 to about 8v. Personally that's the way I'd go on this. Its not a challenge to do this.

@Hutch2793 I'm not saying it can't be done your way. But sourcing very large electrolytic capacitors with low ESR to handle the ripple current in the capacitor (you can't just use any old capacitor or they'll overheat and can explode, you need a high ripple current rating) will be expensive. And that choke is going to be really hard to source, I've searched and the best I can find is over $130. Your best bet will be to wind your own on those discarded cores, but you'll need to experiment with the number of turns - more turns = lower ripple but more losses/heat. It just seems like a lot of effort and cost to provide a mediocre solution. As a learning exercise maybe it has some value, if you bother to go and learn how to run the math; I've tried to give you a flavour but I get the feeling you don't really want that, you want to be told do this, do that. If your interest is more in the electrolysis then go the other route.

You said you wanted to be a professional engineer. Kudos to you. But start by doing the research and getting the facts so you can make sense of what others say.

 

I don't see where winding a transformer is that hard. I have wound a couple and while it can be tedious and require careful attention I would not call it hard. The one limit is the turns ratio, so the output voltage will vary in whatever steps the turns per volt is. So there may be some experimenting required. The voltage change with load will depend on the windings resistance, so wire size does matter.
One thing not mentioned yet is that the transformer supply is more forgiving of line voltage transients and spikes than most switchers, and it will not generate interference that radiates to disturb other systems. The magnetic fields may not be as small as some would like, but it is not likely that one will sit next to that supply for long times.
LARGE filter capacitors will be expensive, that is a down side, and they might even be a challenge to find.
Winding the filter choke will be an experiment because the core material magnetic properties will be unknown, but that is OK. The DC resistance of the choke will need to be considered and it will matter. Copper wire resistance tables will allow you to do all of the figuring for that part, though.

The project will be an educational one and the experience will be useful in an engineering career, eventually.

 

I see the answers to the following questions have been answered, but it seems reiteration is necessary.

As for the filter, tony didn’t have that in the schematic

No - I didn't. I was focusing only on rectification methods.

is that because it isn’t needed, or because he was taking it one step at a time

Whatever filtration you desire - that's up to you. Whether you listen to experienced advice or choose to go a different way - that's again, up to you.

As for seeing things on YouTube - I saw a guy with a small DC motor connected to a pigtail of copper wire with an additional pigtail supposedly making the motor run. He demonstrated that the motor ran vigorously when connected to a 9V battery and how it ran slow - but ran - on just the pigtail. What you DON'T see is the hidden hearing aid battery inside the motor. The pigtail simply completed the circuit. The demonstration went on to show that it wasn't being done via hidden magnetic coupling beneath a table because he lifted it into the air and walked around showing the device running. It was a cleaver hoax, but it was just that - a hoax. Also saw a guy use a motor to spin a generator then switch the generator to supplying power to the motor - and it ran. Then he added things like a drill press, a grinder and a band saw, and all supposedly ran on just the power being produced by the generator. But physics proves that can't be legitimately done. Beneath the table were secret connections the viewer could not see. Why do people manufacture these hoaxes? Because of advertising. They get paid when their video's are watched. They make money from advertisers by tricking you into seeing how they did it. Shame on them. And I've even watched some of these videos thinking "I'm going to do that too and get free energy." Well, actually I've never been that dumb. But I have watched some of them out of curiosity to see if I can figure out how they did their magic trick. Yes, "Trick".

You can try and replicate that video if you feel inclined to do so. But when asking skill and expertise (not referring to myself, I'm a noob too) LISTEN to what's being said. To keep asking the same question over and over in the hopes to get the answer you want isn't likely going to happen. I must admit some frustration with the lack of attention you've been paying to what's being said. For me to say "Use two diodes instead of four or six" and have you ask "So are you saying I should use just two diodes instead of four or six right?" is frustrating.

Bottom line, build it as you see fit. Then discover the results were either like the guy on YouTube or perhaps the experience of those of us on AAC were right all along. The choice is yours and yours alone.

Peace.

 

Thanks to Tony for the explanation about youtoob. I had wondered why there was really dumb stuff showing. Now I know "it IS all about the money" and no thoughts about honesty. That explains a whole lot. and it provides a good insight into the validity of what I have seen there.

One more consideration about the power supply using a section of a microwave oven transformer, which is heat. Mostly those ovens are used for only short periods of time. For constant operation there might be a heat issue, which is one more reason to pay attention to the supply for the first few hours of operation. But if it is OK after 3 hours then it should be OK to keep it on continually.