Ultra-efficiency through adiabatic design

[Orthofield]

Rats, once again I forgot to attach the file.

orthofield

[Orthofield]

Hi Smudge,

I've been reading in control thermodynamics, which is the general area of which adiabatic charging would be a small portion, and I find this article discussing conditions for minimum entropy production in the Coffee Cup Hotel to be revealing.
The coffee cup hotel is a ridiculous method of heating your coffee by moving the cup from one room to the next, each room a bit warmer than the previous one. If one moves the cup to a room, waits until it is warm (equilibrates with the temperature of the room) then moves it into the next room, the production of entropy is reduced. In an infinite number of rooms, the process of heating the coffee is completely reversible.
This approach is similar to that where the capacitor is charged with a voltage only slightly higher, that is, low current charging.
The authors in other articles point out that entropy reduction and power generation (efficiency and power) can both be maximized for only certain systems, and there is often a trade off due to the limits of time the process can undergo. Another name for control thermodynamics is finite time thermodynamics.
This research area bodes well in terms of energy conservation. Just from the example of the coffee cup motel, one can imagine energy saving devices. For instance, as an amateur cook, I find the idea of 'ramp heating' in analogy to ramp charging to be appealing. Assuming the heat capacity of the food on your electric burner is the same through the cooking process-- a big assumption-- the heat could be delivered much more efficiently to the food if the temperature of the food was constantly taken, this information fed back, and the burner heated to just above that temperature. This is analogy to the coffee cup reaching equilibration, or the capacitor being charged just above its voltage. This 'ramp heating' for stove, or home, seems fully possible, and in some cases could save much energy over the usual process. At least for a slow cooker-- I'm not sure about faster applications.

orthofield

[ion]

Hi Orthofield

Regarding the ramp to setpoint idea, this is commonplace in modern temperature and process control instrumentation and is available as a standard feature on most of the newer low cost entries.

There are many nuances regarding temperature control in the food cooking industry, sometimes a ramp to setpoint is not what you want e.g. in making french fries, you will get a soggy product.

[Orthofield]

Hi ION,

Glad that this invention is already being used. I don't think it's being applied as widely as it could be, since the electric heat in my house just runs until the temperature is reached, rather than equilibrating by stages :-)

orthofield

[Orthofield]

Hi All,

Here is an interesting article about the charging of a capacitor. Any circuit that simply switches voltage across a capacitor to charge it is wasting a lot of energy.

"The problem is particularly interesting to treat in finite-time thermodynamics, because
we are faced with very stringent conditions along the time axis. We have
only a very fixed time to our disposal to (dis)charge the tiny capacitors. Indeed:
computations are performed at a well defined and fixed pace, governed by the clock
frequency; pictures are written at a well defined pace, governed by the frame rate.
We can conclude that both in a computer and in a display, we have to charge and
discharge a huge number of small capacitors in a well dened but very short time,
with a minimum of power dissipation. The challenge is the following: how to change
the voltage source Vc(t) in order to do better (i.e. to dissipate less) than the classical
1/2CV 2 ?"

orthofield

[T-1000]

Hi All,

Here is an interesting article about the charging of a capacitor. Any circuit that simply switches voltage across a capacitor to charge it is wasting a lot of energy.

There is third method about which no one are talking about - the charge of capacitors with very short pulses of over voltage with capacitor charged voltage control circuit which switch off charging when capacitor gets fully charged. This is mainly used within BEMF capacitor charging circuits including Tesla coil...

P.S. In http://www.overunityresearch.com/index.php?topic=2358.msg38917#msg38917 I was doing some tests with one of those circuits.

Cheers!

[Orthofield]

Hi T-1000,

I don't completely understand your circuit, but it does look like it has the possibility of something like resonant charging, since there are some inductors there. If there is charging direct from capacitor to capacitor there will be losses, but yours look more like a CLC circuit, with shut off after the 1/4 cycle-- resulting in the appearance of a spike. A spike doesn't preclude the resonant charging as long as the spike shape is controlled by the inductor.

orthofield

[Smudge]

I think this is what T-1000 was referring to.  You can charge a capacitor in a time far less than the CR time constant by applying overvoltage and the shorter the time the less energy loss you get in the R.  Huge overvoltage will give you almost zero energy loss.

Smudge

Esit.  Except that calculations show this not to be true, you don't get almost zero energy

[T-1000]

In akula's LED circuit I was charging electrolytic capacitors with 40V+ voltage BEMF spikes and that allowed me to have constant load from array of LEDs attached for 21-25V range on capacitor. Also the beauty of BEMF with LEDs is, it can power way higher rated ones than your circuit is consuming for making these spikes.

Here is photo of 8 1W LEDs array in series fully lit up from 11.99V and approx. 150mA input:

[Aking.21]

My system for charging a capacitor is simple. If I need 100 volts I put several caps in series in massive over-voltage! SO I'll put 2,000 volts worth in series together.
Then I take advantage of the charging curve. Funny thing is, some of the caps go negative.

[wattsup]

My system for charging a capacitor is simple. If I need 100 volts I put several caps in series in massive over-voltage! SO I'll put 2,000 volts worth in series together.
Then I take advantage of the charging curve. Funny thing is, some of the caps go negative.

@Aking.21

All that is doing is spreading the checkmate out over more capacitors.

I had prepared a post for @Magluvin at OU on the Energy Amplification thread, but will post it here. Maybe refer him here as well. I think this could be a major advancement but it will require some practice before it can be perfected.

The adiabatic design is just another checkmate condition, except that at least they are realizing that in order to advance in a certain line of experimentation, you need to start developing what I had coined as an "OU exclusion list". I had been eluding to this on an off for years, make a list of things that don't work and stay away from those conditions for 30-60 days but it's good to see this in a formal theory, although it is still a checkmate theory as it is proposed.

The thing is we live in a preset world with preset requirements. Our bulbs work off of 120/240vac mains and produce good luminosity. If we take a 12vdc bulb we require less volts but much more amps so we step the voltage down and increase the amps, rectify and we get this great luminosity as well. The problem is in all instances we consume the energy at the end of the game when the transformation is already complete. This is where we all know we are going wrong since when the stress of the load is applied the source starts backing up and losses occur due to the standard metrics we already know and accept as the main villains.

You see this is a problem that many in OU have never been able to understand, not even the best EEers here have ever tackled this question (as far as I have seen thus far) with some level of logic. It is a great challenge. Yes Tesla did high voltage but what did he use it for? If you cannot use it, you ar eleft to produce impressive lightning or sparks otherwise what's the use?

Think of the problem like a hot dog making machine or a hot dog extruder. The extruder is sending out the hot dog paste at 5 feet per second but the hot dogs need to be 6 inches long to fit inside the box. So what do they do? Yup, they put an adjustable cutter at the exit of the extrusion timed at 10 cuts per second and every second, they can fill one box with 10 hot dogs. Good business. This simple solution is not used in OU/EE, why I don't know.

So let's say you have a very efficient device, it consumes very little power but dammit it outputs 50kv when it reaches its operating frequency of 60Hz.

We will usually take that 50kv, dump it integrally into a HV capacitor (or just spark it Tesla Toroid style) then work to discharge fast enough to cut it up in usable pieces. This we do every day and each time, when that voltage reaches 50Kv (or any other highest voltage) it creates a backwards pressure on the output and that's when you start creating heat, flyback, back emf or whatever you can call it. Add that to the perils of switching 50kv and you realize this is not a good solution. Not a good solution because the reality will always be that if a primary is driven to a point where the secondary produces more output, the secondary automatically will become the primary and it will want to kill the dipole. This we cannot get away from as we drive our coils today.

So let's say you need 200 volts for the output. If you could scope the 50kv rise, every second you will see 60 rise and peaks of 50kv each. But if you could divert each rise 250 times per 1Hz, or, at 15khz per 60Hz, the output should not exceed 200 volts regardless of how high the actual output can rise in volts and the amps should increase proportionally as well.

So a simple hot dog extruder does exactly the job. Of course if the cutter breaks, then you will be in trouble but the EE solution could be a 250v safety zener of proper amps rating to dump any excess or at worst a spark gap if there is a full run away condition.

The experience and ability in EE required to do this goes way above my head man. Anyone would need to be very well versed to even consider it but the payoff could be so great. Even starting with a simple lower output tv flyback transformer would be a good start. The 15kHz mosfet base signal could come from any audio source so imagine an audio driver to control the HV reduced output frequency. The output load would receive 60hz with 15kHz hash. 

Actually we do this in reverse when we pulse our coil. So eventually if the two could be combined together, one efficient design primary and one efficient design secondary recovery. By not letting the secondary HV rise, you are promoting a freer primary performance, were the primary amperage would not be lost in the secondary rise. hahaha

Hmmmmm. I think I  just described a TK device. Naaaaaaaaaaaaa. Hmmmmmmmm.
Also @GK had been saying that SM used his stun gun circuit. Why not. Just don't let the voltage rise to the stun level, keep unloading it to a load as it rises.

So where it would definitely be good to look and practice is taking the HV from any regular flyback transformation and cutting it up and sending it directly to a load before it can rise to its peak voltage. This will keep the primary at it leanest operating parameters since it will never see a full rise on the secondary. So circuits where you can place a mosfet on the HV line, mosfet rated at 250 or 300 volts throughput run at 200 volts should keep it cool, or if you ever found a good diode with a 200 volt forward voltage cracking level would do it as well but it shorter bursts (mosfet is better).

You need to produce high voltage but not let it rise to its maximum high voltage. The layering this produces will be consumed by the load right away. The trick is to always start the HV output mosfet pulsing before you start the primary pulsing.

wattsup

PS: As usual sorry for long post.

[T-1000]

The algorithm with over voltage is straight forward:
Charge capacitor->watch its voltage over feedback loop->switch off oscillation circuit when the voltage on capacitor reaches required level->switch it on again when voltage on capacitor drops below minimum level

This is what was done with my correction in akula's BEMF charging circuit...

Cheers!

[lost_bro]

The algorithm with over voltage is straight forward:
Charge capacitor->watch its voltage over feedback loop->switch off oscillation circuit when the voltage on capacitor reaches required level->switch it on again when voltage on capacitor drops below minimum level

This is what was done with my correction in akula's BEMF charging circuit...

Cheers!

Hello T-1000

So here's the one chip solution:

LT3751   High Voltage Capacitor Charger Controller with Regulation,  does exactly what your talking about and a bit more.
I'll be integrating it into my solar powered fence charger project........

take care, peace
lost_bro

[Orthofield]

Hi Smudge,

Yes, although this overvoltage scheme sounds good, it generates a lot of heat in the wires, and is thus nonadiabatic.

What I'm getting at in this whole list is that a lot of ways people do things waste heat energy. Many ways that are fast and impressive do just that. The useful range thermodynamically is somewhere between super fast and cool looking, and thus heat wasting, and so slow as to be useful for the system at hand. In the case of a capacitor it is well known-- and should not be up for dispute-- that the losses in charging a capacitor are reduced by the number of steps used to charge it. The highest number of steps is a ramp. A modified ramp charge as shown in Desoete is universally acknowledged to be the most efficient way of charging a capacitor.

Practically speaking the issue with this sort of charging is that if the Mos or other switch is at anything other than 0 or infinite resistance, there are losses. That's why quarter resonant charging, which resembles the ramp, is often used instead. Both the resonant and ramp charging are used in experimental adiabatic computer chips which consume almost no power relative to a standard chip. In this case, the inductor's resistance becomes the big issue, but this can be minimized.

The attached paper discusses some of these issues, first laying out the typical scheme used in adiabatic logic, and showing how an approximation of it can be achieved for charging a supercapacitor by using a converter (where all the switches are either on or off, thus conserving power) to simulate the loss free, without the practical losses from using multiple switches, specialized wave generators, that may arise when using the adiabatic approach. This to me seems like the best scheme for using solar cells to charge a supercapacitor.

Hi Wattsup,

Great letter!

Wattsup, there is no checkmate in terms of energy, if things are actually done efficiently. There is no inherent large heat loss in an electronic system. It's not difficult to get near 100% efficiency in cap or inductor charging circuits with proper input waveforms. In that case, the heat loss has been vanquished. T-1000 may have referenced this when he talks about a feedback loop.
Then the issue after that is to add on your overunity principle you want to try.

This plays right into your point about overunity systems that don't play nice with the power grid. You say:

So let's say you have a very efficient device, it consumes very little power but dammit it outputs 50kv when it reaches its operating frequency of 60Hz.

Then you say, "we usually.. " and then you lose me. Why would you move away from the regime where high efficiency exists, towards one where it is much more difficult? Nothing I've ever thought about using, including parametric amplification, cannot be done at 60 Hz.

Now, in the 60 Hz realm, you can apply all these 'slow' adiabatic processes which save energy. BEMF behaves itself, can be tapped with normal diodes, transformers operate efficiently, etc etc.

What happens when you achieve your amazing OU system that operates at a Mhz and runs at 10 kV?
You're going to waste as much energy as you gained, converting that into something that can be used. And this may be what you mean by a "checkmate". But the checkmate is designed around a certain design psychology rather than anything inherent in Nature.

My belief is that Nature, or Spirit, or what have you, WANTS us to have overunity, and makes it slow and easy, just as much as it makes it hard and fast..the sexual overtone of that sentence was deliberate, and not really meant for humor. Female EEs-- and Taoists! -- might see things differently than many of the guys on these lists.


orthofield