If you just want to power an LED, look up "joule thief". But PWM has many uses - including switching high currents. You need not be concerned about the differential between the energy stored and the energy used. Your circuit will pull out no more energy than it is designed to.

That's encouraging to know regarding energy consumption.

"Joule thief" is most interesting concept...I had a brief look on Google, will look into that one more.

My desired use (in a nutshell) for a capacitor bank, is to use it in driving a three phase DC brushless motor. Am waiting for the MCU IC's from the U.S. before I get to physically do anything.

Thanks for the reply

Your circuit will pull out no more energy than it is designed to.

Oh, I wouldn't say that, given that CraziestOzzy seems to want to only power an LED, implying that he intends to use only an LED connected to a capacitor. In such a case, the forward current goes up rapidly with voltage, once past the knee voltage of 1.5 to 3 Volts (depending upon the LED structure).

Some sort of regulation must be had, and PWM may well be the most efficient, but if it was me, I'd use a 1.5 V rechargable alkaline battery with no regulation.

I didn't say it would pull more energy than he intended it to, only that it will operate as designed. If I have a 1000V, 1000A power source and put a 1,000,000 ohm load on it, it will draw 1mA, regardless of how many gigawatts it can produce.

Since his question was specifically how to control the energy into his LED it's safe to assume that controlling the energy is his desired purpose.

And if not, well - LEDs are cheap.

lol....indeed they are

I guess I am (was) primarily concerned with the "hard stop" effected upon the controlling device that will sit between the source and the load if I am to use an "oscillation" type of control to trigger controlled release of energy from the capacitor bank. I am considering the use of PWM or a gate trigger on a yet to be chosen FET, to act in similar fashion to an inefficient (my view) linear voltage regulator.

An LED was an example of the lowest power consuming device I could think of for use in the analogy of using a capacitor bank in the face of a controlled load source and low power consuming device.

Perhaps I am to blame for any misunderstanding, so here is a quick mud map of my thoughts showing the Motor Control Unit (MCU) idea I have in mind...

The IGBT I have in mind is the IRG4PC50UD

The Thyristor I have as an option is the Thyristor-Thyristor module, TT162N

The PWM I have not decided on yet.

Both the IGBT and Dual Thyristor module are rated for serious work.

What I would like to know, is this concept of controlling a capacitor bank discharge for the purpose of driving a DC motor via a MCU IC at all feasable using a Thyristor module or set of IGBT's?

You can treat your capacitor bank as a DC source that declines in voltage over time (and load). Essentially the same type of controller used to run a brushless DC motor from batteries is all that you require. It will have to cope with extremes of voltage though.

Yeah, thanks for that. I have just received a set of MC33055 brushless motor controllers that have a voltage range of 10-30 DC. They are suited to my purpose of running a brushless DC motor.

The fun bit now is taking into account all the extra bits and pieces needed to make this work.

...lucky I got several IC's, not so much an issue if I fry the odd one or two

We drive (brushed) 28V motors using IR MOSFETs and a PWM circuit coded into an FPGA. But you can certainly get PWM motor controllers designed to adjust for a fluctuating input voltage.

http://www.maxwell.com/ultracapacitors/applications/industrial.asp

http://www.maxwell.com/pdf/uc/ISE.pdf

These systems have been out for a few years.

they manufacture a module

sized for city buses, that is charged in regenerative braking, and then augments the batteries in reaccelerating.

Thanks for that. I found most interesting the links you provided, especially this one LINK

If I understand your question correctly, a photo flash circuit does exactly what you are trying to do. The capacitor is charged from a power supply, and when the illumination requirements are met, the capacitor is disconnected from the flashlamp with a thyristor. This provides the correct illumination and also saves the unused stored energy to reduce recharge time and extends battery life.

Yes, very similar indeed. If I were to manually adjust the f-stop of the flash to something like 1/32 and set the distance from flash to subject to under 50cm, I would be able to trigger the flash between cycles very quickly (assuming unlimited power source) and would have to stop to prevent the xenon flash from overheating. I intend to drive the trigger faster compared to that of the flash scenario to operate a DC motor without discernible "cogging".

I will have to seriously look at this again. I have an old dismantled Bowens studio flash that is jammed full of caps. Will look at the PCB circuit and see if I can replicate similar operation without the bells and whistles.

You have got me revisiting an old idea and some old files on this one. Thanks for waking me up and giving me some ideas.

Brushless DC motors driven by a capacitor bank. Interesting concept. I believe you've already considered several complications in this approach. I hope you won't mind if I pose a few questions, comments and suggestions:

• How do you charge your capacitor? A capacitor is an energy storage device not an energy source. You will need to put power into this, first.
• The driver circuitry will draw some power all the time. So idle conditions from the all parts of the system will rob this capacitor of power.
• Since power losses seem to be critical, have you considered dynamic braking to recharge your capacitor? This can significantly complicate your circuitry though.
• Remember to test the current draw for your motor and driver circuit with both no mechanical load (free wheel) and a safety margin above the expected mechanical load. Ideally you will want to track these conditions through various accelerations.
• PWM can easily be your most efficient transfer of throttled current into the inductive load. Don't dismiss this option.
• LCD displays typically draw less current than an LED display will. Back-lighting concerns can remove this advantage though.

Probably several of my ideas and concerns are nothing but me thinking out of the box. But that can easily happen when the box is not known.

You have touched on some of the aspects I had to consider in the earlier stages of pondering.

The main one is the determination of load and the current draw induced by the DC motor. The second pointer is my chosen method of braking. Both points will have to be fine tuned (perhaps even reconsidered) once the motor is running.

The majority of my components run and are designed to be run on low power (even on standby) with batteries as a power source. Losses are may major concern, but will have to overcome one hurcle at a time.

Cheers for the pointers, always thinking out of the box

Crazy,

You maybe surprised to find that the highest density stored power for your application just may still be batteries. Keep this in mind as you research capacitor storage.

The ultra caps mentioned by someone here maybe just what your looking for with very large capacitance values for relatively small packaging. But if memory serves me well, they do have lower leakage resistance than other capacitor types. So long time storage of energy may not be practical.

Battery or capacitor though, I would recommend a high frequency boost/buck switching supply to regulate your circuitry. This will draw a minimal amount of bias current, thus maintaining efficiency. But this will also permit good operation over a wider range of supply voltages. This can also permit the back EMF from the motor to be steered to a capacitor for storage when using dynamic braking.

Sounds interesting, have fun.

cheers, the "boost/buck" scenario is the next step. Well advised and thanks for those tips. I will be running a combo of the two. Scored a 100+ NiCd batteries from the dump just recently and in the process of making charger that charges each individual cell separately. Have just finished drilling a few hundred holes into a double sided PCB and will post results in a week or two of the combo run of batteries and caps.

Supercaps....type maxwell