Circuit design help - power supply - Non FIRST

[Josh Siegel]

Hi guys (and gals),
First off, sorry that this post isn't exactly FIRST related. I posted this here, though, because people on CD know what they're talking about and are always willing to help. Plus, it may help our robot (among other things) if my theory works out.

Anyway, I need to design a power supply with dual frequency generators that can do the following, with a minimum of parts and cost:

1. Take in 12VDC at up to 30 amps (so this is a high wattage application).
2. From this, split the voltage into 6V and 6V (running in series) for use in two VARIABLE frequency sine wave generators (both use the same frequency – from about 5Hz-50Hz is what I’m looking for, though 0-150 would be great too) that are phase shifted from one another by a few degrees/radians/seconds
3. From the generators, I will rectify the outputs so that they are all positive
4. Then I will recombine the signal

What I’ve seen leads me to a few things…

First, I could use an op-amp or a wein bridge as my sine wave generator, but these do not handle high voltage or amperage well. If I could somehow get these to work without extremely expensive components (300 watt resistors and such), these would be great because the frequency is variable by a resistor, and I could share one POT between the two circuits (in parallel) so that the frequencies are the same in both circuits. The delay could simply be a variable 555 timer circuit, if my logic is correct – more time delay initially (“single fire,” if you will) means a bigger phase shift. The only thing I am unsure about is how to make sure I get full amplitude out of the circuit (6V in, the voltage oscillating from +/- 6V peak to peak rectified into +6V).

Second, I could somehow build a power inverter (DC to AC converter) and rectify the output, but I wouldn’t use a transformer and would somehow make the frequency adjustable, with a POT or similar.

Third, I could make a MAX038 or similar circuit, but my electronics skills are limited and the MAX038 has very low current/voltage limits.

The hardest part of this is making things to handle high-wattage. I hate to sound as simple as I am with electronics, but there has to be some way to use low-wattage resistors and other components in high-wattage circuits, either using the low-wattage circuit as a driver for relays or something else... Not that this is necessarily the best option; I'm just thinking out loud.

Since I'm using the output as a power supply and not to produce sound waves or anything, accuracy of the sine is not really critical as long as it is predictable. The key in this circuit would be efficiency and accuracy of the frequency, so that it makes the powered devices function as predicted.

Sorry if I'm being wordy or convoluted, I'll gladly explain as much as I can about this if it helps design the supply. I have some jpeg files of the desired output at every step, if those would help - they're on my desktop, so I can post them tomorrow.

Thanks a lot!
Josh

[sciguy125]

What exactly are you trying to do? Maybe there's a simpler way to do it. You also say that you're going to "recombine" the signals. What do you mean by that? You also mention power on the input, but how much power do you need on the output?

[Dave Flowerday]

Just a few days ago I ran into a site on the net talking about how to build a microprocessor-controlled DC power supply. It's a good article that discusses the principles behind the design, and I think you might be able to apply some of the same concepts to what you're trying to do.

http://www.tuxgraphics.org/electronics/200506/article379.shtml

Not really an answer to your problem but perhaps it will help a bit.

[Alan Anderson]

The hardest part of this is making things to handle high-wattage. I hate to sound as simple as I am with electronics, but there has to be some way to use low-wattage resistors and other components in high-wattage circuits, either using the low-wattage circuit as a driver for relays or something else... Not that this is necessarily the best option; I'm just thinking out loud.
Maybe a 400 watt audio amplifier would work. Generate and control your signals at low (manageable) voltage, and send them to the amp's input.

What kind of device are you going to feed this oddly specified output to?

[KenWittlief]

The most important question is whether you really need the output to be half wave rectified AC at specific frequencies, or are you going to use the output as DC? (filtered, ripple free)?

If you want 6V DC out then you are describing a switching power supply. There are many parts you can buy off the shelf to put one together.

If you really need a variable frequency half wave rectified sine wave then you are looking at an oscillator and what is in essence an audio amplifier.

[petek]

If you do need amplified & rectified AC, and have the budget, the easy way would be with an oscillator, a phase-shift network and a pair of power op amps - Apex's PA03 (http://eportal.apexmicrotech.com/mainsite/products/pages/op_amps/pa03.asp) for example.

[Josh Siegel]

Thanks for all the info so far, everyone!
Sciguy - As I said, I can't talk too much about specifics, but this unit will power an electrolysis chamber and resonate the electrolyte for increased output. That's where the frequency portion comes in.
The dual sine-waves with phase shift come in, because when shifted a minimal amount, the upper limit of the voltage is effectively the same as it was prior to generating the waves. The lower limit, however, is shifted slightly and therefore keeps the minimum voltage above 0 (say 6V, but that's just a guess). By not dropping to 0, production does not stop but rather becomes endothermic while the "active" production on the electrodes pauses and the product is allowed to clear.
The file at http://www.spline-designs.com/FrequencyGenerator.xls shows what I'm trying to achieve, though now I think if I could set a voltage range, I would only need one generator (ie., one sine wave from +6V to +12V at a variable frequency), which would make this much easier.

Dave - Thanks for the link, every little bit that gets me and others thinking helps :).

Alan - Good idea with the amp output, my only concern is the efficiency lost in the additional device, but I'm probably not going to be extremely efficient to begin with. The output is going to an electrolysis chamber, as described above.

Ken - I don't need DC in the strictest sense, otherwise I'd go straight in from my power supply (a DC car battery). I effectively need "pulsed" DC with a minimum "low" voltage, a max "high" voltage and some form of curve (not square wave) in between. Triangle wave would work, though I'd much prefer sine waves if possible, because sine waves have more of the type of derivative I'm looking for (sharp increase, quick level, sharp decrease).
I did think about tearing apart both an audio amplifier and a sine-wave inverter for this project, but the inverter isn't variable frequency and has a transformer, while I wouldn't know how to tune the amp.

Pete - Thanks for the link, I'll take a look at the data sheet and see what I can figure out about it.

Thank you all for your time in reading this, and I really appreciate the help.

[KenWittlief]

Oh! an electrolysis chamber!

Ive heard about that.

you sit in the chamber and it takes all the hair off your face, and it never grows back. Very interesting application!


:^)

[Alan Anderson]

I effectively need "pulsed" DC with a minimum "low" voltage, a max "high" voltage and some form of curve (not square wave) in between.
petek's power op amp suggestion seems perfect for this application. String together a simple oscillator, a full-wave rectifier, and an op amp with positive offset and appropriate gain, and voila: you've got what you're asking for.

[Kevin Watson]

Well, I would keep the power op-amp, but use a microcontroller to build a waveform table in memory. Once that's done, step through the table and send the output to a DAC, which creates your waveforms. Depending on your tolerance for distortion, you may need to filter the output to remove the Fourier step components. Here's a link to Google Direct Digital Synthesis (http://www.google.com/search?sourceid=navclient&ie=UTF-8&rls=GGLD,GGLD:2004-30,GGLD:en&q=Direct+Digital+Synthesis).

-Kevin

[KenWittlief]

From your description it sounds like you are experimenting with generating hydrogen from water.

Something to keep in mind: the overall efficiency of the system is the amount of energy you get out (the gas) divided by the energy you put in.

If you have to use amplifiers or other switching electronics to make the system work remember the power at the wall plug is the energy input, not whats across the electrodes.

The reason I bring this up: there are hundreds of websites created by people who think they have defied the laws of physics and come up with perpetual motion machines or motors that put out more power than they are driven with

and a common mistake they tend to make is pulsing the motors with low duty cycle DC pulses and measuring the output from the system as DC or AC. If you used pulsed power you have to get a true-RMS volt / amp / watt meter. Most meters (digital and analog) measure average voltage or current, not the actual RMS voltage. For a perfect sine wave, the average and the RMS values are the same. But when you start pulsing power a regular (average reading) meter will read less than the actual RMS going into the system

As a result, people measure 10W into their system, and 20W coming out, and think they have created an over-unity machine.

True RMS meters (like a Fluke 87) tend to be more expensive, so usually only people who really know what they are doing buy them.

[Josh Siegel]

I appreciate the warning, Ken -
You're right in that this is a hydrogen/oxygen electrolysis chamber, but I have taken into account the actual vs. apparent input. I'm not going for "over-unity" since I don't really believe that it's possible with current technologies, but resonance is a contributor to gas production (it is used by the US government in military electrolysis applications, not to help split water but to allow the electrodes to clear from gas bubbles and therefore maintain a larger surface area) and with voltage dropped below the threshold voltage, electrolysis becomes minimally endothermic. This circuit is just designed to help get closer to 100% efficiency and let me test out a few theories about stretching and relaxing hydrogen bonds, along with seeing if heat energy can be used to supplement the system such that our output may come close to our input and hydrogen may prove a viable storage option for energy transport on a small scale.

Now I'll take another look at the ideas presented, and see if I can set up a schematic in my mind before I try to build the supply.

[Al Skierkiewicz]

Josh,
What you need is a DC coupled power amp capable of offsetting the zero crossing of the sine wave to some voltage. This is possible with audio power amps but not at the power levels you have stated in your intitial specs. Your requirements will only give you a possible 4-5 volt peak to peak signal at 30 amps or somewhere close to 200 watts. A couple of function generators and a big stereo power amp would do the trick with a variable DC suppply to make the offset but would only make power up to about 50kHz.
What you are asking to do is dangerous in most situations particularly for someone without previous experience in high power amplifiers capable of the frequencies you have stated. I hate to restate the obvious, but you do realize the flamability of hydrogen? At the quantity that this process would appear to generate, you are going to realize a rather large volume of hydrogen gas.

[Josh Siegel]

Al, thanks for the words of warning regarding the amp's and their usage - I'm going to try to find some safe way of doing this, and certainly wont "play" with anything I feel unsafe with or that I haven't learned about. The relatively low frequencies should make this easier (0-150Hz, not kHz), but I understand the risks inherent in working with any high-wattage circuits.
Secondly, I do realize the flammability of hydrogen and I have flame arrestors/bubbler's and other precautions to ensure the safety around the generator. Given that it is not being stored and that hydrogen dissipates so quickly, the 1-1.5LPM that I've been getting isn't really a problem, unless presented with a flashback that the bubbler and arrestor will stop.

[Al Skierkiewicz]

1.5 Liters is still a lot of hydrogen if I read your post correctly. I don't want Bloomfield Hills to disappear off the map, I have friends out there. Remember, the hydrogen may not be dissipating but it may be gathering in high concentrations above your head. Remember the Hindenburg!
Most audio amps will go very low in frequency (some down to 8HZ) but only DC coupled amps will go to 0 Hz and you will only be able to offset the input waveform with a DC amplifier. You will need to be able to control the voltage input as the power amps will respond rather quickly to changes on their input giving you significantly more than 6 volts peak output.

[Alan Anderson]

Remember the Hindenburg!The lesson to be learned from the Hindenburg has little to do with hydrogen. Mostly, it's just a bad idea to paint a big bag of fabric with the equivalent of rocket fuel, especially when it's going to be put to use in an operation that tends to involve large discharges of static electricity.

But Al is right to warn about the potential for the hydrogen to collect in high places. If you're really generating more than a liter of gas per minute, I hope you have effective ventilation.

[Mike AA]

If you were closer I might be interested in buying your "excess" hydrogen for my fuel cell I'm looking at getting.

Hydrogen really isn't all that "bad" in terms of fear or explosions, atleast when bottled up. I guess if left to accumulate in a room it would. There have been tests that show when bottled hydrogen is about the same reaction as gasoline in a car if it were to get into an accident, possibly safeR.

Hadn't heard of using different frequencies to perform electrolisis faster, I guess I'll have to try that when I get my fuel cell.

-Mike

[petek]

Allow me to add to Al's words of caution about the flammability of hydrogen. I know a fellow who blew the doors and windows out of his mother's garage when he tried generating hydrogen by electrolysis. He was very fortunate that no one was nearby to be hit by the flying debris and that somehow the garage didn't catch fire.

If you go ahead with this, please make absolutely sure you have a safe (as in explosion proof) way to remove any excess hydrogen from the building - or better yet, do it outside.

[Danny Diaz]

There have been tests that show when bottled hydrogen is about the same reaction as gasoline in a car if it were to get into an accident, possibly safeR.This discussion brings back memories of the '96 Olympics in Atlanta, Georgia. The Georgia Tech Research Institute (GTRI) worked with the City of Atlanta and several companies in the Atlanta area to work on and demonstrate a hydrogen/electric hybrid bus to publicize Atlanta's efforts to help reduce air pollution (and also assess the performance of such a hybrid). I was unfortunate enough to come in just after the bus was developed and put into its short-term service, but I did get to ride the bus before it was shut down. Anyway, the designers and the City were extremely cautious about the bus and its fuel supply - so cautious the bus had hydrogen sensors running up and down the bus (outside and inside) in a fairly dense grid tweaked such that it could detect the smallest amount of hydrogen on the bus. It was tweaked so tight that the bus eventually got the moniker, "The Rooty Tooty Mobile", after the fact that passing gas on the bus would cause the sensors to go off and force the immediate evacuation of everyone on the bus. After one and a half days in service it was retired from active service as a shuttle bus and was used for research purposes from then on.

-Danny

[DonRotolo]

I'm not going for "over-unity" since I don't really believe that it's possible with current technologies,

...or according to modern theories of how the world works.

Josh, it appears that you want to generate a relatively large signal (tens of amps) in the frequency range of 5 to 150 Hz, with a lower limit of say 5 volts and an upper limit of say 12 or more volts. This is not AC, but pulsating DC, there is a difference.

If I have that right, then it's easy to do for fairly cheap.

First, use a low-power variable frequency source (555 timer, signal generator, Beastie Boys tape, etc) and feed that into a few-hundred-watt single-channel audio amplifier (like used for a car). Not terribly efficient, but not too bad either. Then, just connect a (5?) volt power supply between one of the amp outputs and "ground", and you have your 5 volt bias. Fiddle with the numbers (you'll probably want a 10 to 12 volt bias, really) and drop the electrodes into the water.

Note that you cannot power the bias supply from the same source as the amplifier, or you'll have a short to ground and other issues. (=smoke).

Good luck with that, let us know how it turns out, and yes, ventilate well, like outdoors.

Don

[Al Skierkiewicz]

...Then, just connect a (5?) volt power supply between one of the amp outputs and "ground", and you have your 5 volt bias. Fiddle with the numbers (you'll probably want a 10 to 12 volt bias, really) and drop the electrodes into the water.

Don,
This is not generally a good idea with modern audio amp design. The DC on the output that does not appear as part of the input signal will cause the amp to go unstable as the error correction circuitry tries to fight what it sees as an illegal output. This could have disasterous results for the power amp. The DC may also upset the damping/feedback control causing the amp to oscillate, sometimes enough to cause localized interference.

As to the notion that hydrogen is relatively safe, I think this can be attributed to the hydrogen cell research that has been publicized for the last few years. In this technology, the gas it locked up in a matrix that releases the gas in a very slow, low pressure stream useful for engines but slow enough to be non combustable under normal circumstances. I saw a demo where the container was shot with an incendary bullet and did light but did not explode. Hydrogen without precautions is explosive in a confined space.

[Andrew Blair]

Al, you're completely right. I would hazard a guess that several people in chemistry have done the experiment where you combine Mg with HCl, creating a few ml of H2. When you flip the containing vessel and light the H2 contained, it makes a popping noise. Remember: A few ml, in open air makes a loud pop! By performing electrolysis, both the O2 and the H2 are released, and if unseparated upon creation, is a reaction waiting to happen.

So, with a little electrolysis with enough current/voltage, in a confined space with both the O2 and the H2 mixing from uncontrolled hydrolysis, you have yourself a nice way to be seriously hurt. I will repeat. BAD!!! Unless of course full, absolute safety precautions are taken. I'm not being a jerk, but its important to be thourough. Sounds neat though, never knew pulsing DC would accelerate the process.

By the way, is hydrogen in a fuel cell matrix semi-liquid, or is it still far below that kind of pressure?

[sciguy125]

I was doing a little more thinking about this today.

You're essentially trying to work with a rectified sine wave, but keep it from dropping to 0V. Why not take the straight rectified signal, then add a DC portion to it? If your amplifier can do DC, then you can just add it before the amp. If not, then you'd have to add it to the output. Maybe put some other DC supply in series with the output of the amp?

[Gdeaver]

Hydrogen goes "BOOM" very easily and nicely. I would suggest extreme caution in this pursuit. You should have extreme fear of this research. You probably have not had the unfortunate experience to to see what a fuel air explosion does to a human being. I have. First, the smell of burnt flesh and hair stays in your nose for days. The sight of a human face blasted with shrapnel is a little unpleasant. The worst was the way he was flailing those stubs around that used to be hands. You get the picture?
If you still want to play, may I suggest the concept of micro batching. Your in the experimental stage. Why not scale down the experiment. A little "pop" is better than a room leveling "bang".

[Josh Siegel]

I'm well aware of the potential dangers, and never work with hydrogen indoors or around any sources of heat. I also ground myself before I work on the generator to prevent static discharge.

That said, I've been doing a lot of research on this topic over the past few days.

I know there are function generator kits out there (kit 6023 KT at MPJA, for example or the 3023 at Quasar) which use a "4 core" LM348 op amp to generate sine, square, and triangle waveforms. These use an op amp with an identical pin-out to the MXL1179 from Maxim IC, which can handle very high current and voltage - now, at first I thought that I could simply replace the IC, but then I thought about it and there's a) no way the traces would hold up to 30 amps, b) I'd still need high wattage resistors, and c) I don't know how to set up offset voltages.
That led me to think about the way PWM works, at least from my primitive knowledge of it. If there were some way for me to set a bias voltage on that function generator kit, I could change the potentiometer for the range I want and use a MOSFET to control voltage, couldn't I?

So what I'm thinking I'd like to do is this:
Function Generator (http://quasarelectronics.com/3023.htm) (powered by 9V or similar) - pseudo sine wave out to a rectifier (http://www.allelectronics.com/cgi-bin/category.cgi?category=search&item=FWB-66&type=store) - to a MOSFET (http://www.allelectronics.com/cgi-bin/category.cgi?category=search&item=IRF3205&type=store).

Am I thinking along the right lines? Is there any way to use the kit I linked with an offset voltage easily? Also, I'm not sure of the peak-to-peak voltage on the kit, and I assume my MOSFET selection should be based upon that and the offset (ie., 6V/30A MOSFET if it's 3V peak to peak and a 3V offset voltage).

The only problem I forsee with this circuit is the voltage in the secondary circuit (through the MOSFET) being susceptible to changes in the voltage of the control circuit. It also seems pretty inefficient.

[KenWittlief]

the efficiency problem you are facing is using any type of amplifier.

Amps work by taking a DC power supply output and holding back the part of the voltage that does not match the input. Subtracting the AC output from the DC power input leaves about half the input power dissipated by the amp (the Mosfet driver mostly) - so you are looking at 50% efficiency or less with this approach. The only way you can get around this is to use a squarewave output. This would keep the mosfet driver out of its linear range, so its either full on or full off. In either state its power dissipation is negligible. (this is how a Victor speed controller work).

If you want to approach 90-95% efficiency with a sine wave output you are going to need something like a motor generator set: a motor that drives an AC generator. The speed of the driving motor will determine the frequency of the AC output. The output can then be full-wave rectified.

Then all you need for your application is a DC offset (DC power supply in parallel with the AC rectifier). The diodes in the rectifier will keep the DC bias supply from back-feeding the AC generator.

The biggest risk you face is having something wrong with the DC bias. The O2 and H2 form at the opposite polarity electrodes. If your bias goes negative for any reason you will have a small amount of O2 in your H2 collector. Hydrogen and oxygen mixed together is subject to spontaneous combustion.

The only way you can be sure your electrodes are properly biased is with an oscilloscope.

[Gdeaver]

This is outside of my expertise but, wouldn't putting a large inductor after a FET give the desired results? Monitor the voltage after the inductor. When the voltage gets close to 12 volts turn the FET off. When it drops to 6 volts turn it on. The feedback could be done with op amps or micro controller. The output would not be a true sine wave but good enough. The frequency would be variable to the load. Kind of a switching power supply with sloppy voltage control. As far as safety, one should have the mindset that it is going to blow and design the apparatus to handle it. Murphy's laws dictates in will blow at the most inopportune time.

[KenWittlief]

Gdeaver has the right idea - you could make an resistor- capacitor or resistor inductor filter circuit that would turn a square wave into something approaching a sinewave.

For the current levels you want the inductor or capacitor would have to be very large. If you use an inductor you have to make sure it cant spike negative when the pulse is shut off. This is normally accomplished with clamping diodes.

[Dave.Norton]

If you want to approach 90-95% efficiency with a sine wave output you are going to need something like a motor generator set: a motor that drives an AC generator. The speed of the driving motor will determine the frequency of the AC output. The output can then be full-wave rectified.

If you want to see how to get a high efficiency power amp check out the sub-woofers from Sunfire. Carver has used a switch mode power supply on both rails, as the input waveform opens up the power supply on that rail tracks it. He then has a high current amplifier that only has 12 volts across it nominally, but can drive up to 300 volts peak to peak when required. It is fun to watch operating, but if you are ever looking at one be careful with it, 2 to 3 kW is not to be trifled with.

[Al Skierkiewicz]

Josh,
You can tell a lot about the function generator you linked to by looking at the power, the circuit design (quad opamp) and the picture. A single ended power supply on this opamp will render the output DC shifted to half the voltage of the power supply, in this case a 9 volt battery. Without a big capacitor on the output to block the DC (I can't see a big cap in the picture) the output will generate waveforms that center on half the power supply more or less so expect the output to be around +4 volts most of the time. In this application you are not really concerned about efficiency in the amplifier should you use one. What you are asking can be accomplished with a switching circuit in series with a variable DC power supply or a power amp where you can adjust the offset voltage of the output to be at some DC potential. With a power amp, the efficiency goes way down in this arrangement as the amplifier stage is drawing current all the time whether there is signal or not.

[KenWittlief]

for the DC offset power supply - is 580 Amps enough?

!

http://cgi.ebay.com/HP-6681A-DC-POWER-SUPPLY-0-8V-0-580A_W0QQitemZ7573639758QQcategoryZ58288QQrdZ1QQcm dZViewItem

:^)

[Josh Siegel]

580 amps? sure. 8 volts? uh uh. :p
Everyone, once again I'd like to say that your help throughout this design has been invaluable - You have helped me clear several things up in my mind, and I've decided that instead of using a sine wave (at least initially), I'll simplify my design to something that most of you seem to agree upon. I'm going to use a 555 timer to generate a square wave with an offset voltage and hook that up to a power mosfet, which should be fairly efficient and easy to build.
I'll go out and see if RS has the components in stock, or else I'll order from Digikey.