Help-How to boost +7.4v to +75v DC? (for electroluminescent wire driver)

[Genomerics]

I think the 'cunning plan' may be to sinusoidally PWM the H bridge to end up with a variable sine voltage across the EL wire rather than square waves... that's if it really does matter. Which, kind of raises questions.

If not and raw square waves are applied and the load 'does not care' then should the 'load' have capacitance then the bridge might suffer and even with a sine PWM it might be wise to have some inductance in series with the load. You might end up with a kind of Class D amplifier filter arrangement.

Genome.

 

[AleXYZ]

My original plan with the H-Bridge was to run the phases in quadrature, to make a modest 3-step emulation of a sine wave. Doing this also alleviates any concern about having to program or design for dead-band delays. However as the rail voltage needs have been going up, I now think that square wave may be necessary to get optimal RMS voltage on the output.

Today I took out one of my better (but still cheap) El-wire drivers and sampled its voltages when loaded with 6' of 2.3mm green el-wire. Here are the results:

Input: 5.00v DC @ 0.10 A
Output: 115v AC, 1.8 kHz, almost perfect sine wave

On the scope, it looks to be about 220v peak-to-peak. Sigh. This means I now need a minimum 220v rail, or dual rails at 110v each!


(El-wire on my bench experiments has been glowing...dimly... at voltages as low as 50v p-p, but one of the goals of this project was to make it really bright.)

 

[mtwieg]

My original plan with the H-Bridge was to run the phases in quadrature, to make a modest 3-step emulation of a sine wave. Doing this also alleviates any concern about having to program or design for dead-band delays. However as the rail voltage needs have been going up, I now think that square wave may be necessary to get optimal RMS voltage on the output.

Since the load is a capacitance, you'll definitely need some inductance in series to smooth out the current. Otherwise your current won't be sinusoidal at all. And if you're going put a series inductor in there, then hell, you might as well drive it at the resonant frequency, with no sinusoidal pwm modulation. Then that begs the question of how to size the inductance, since your capacitance will change depending on how much wire you drive. I supposed you could actively tune the drive frequency to match up with the changing resonant frequency, but that would probably have to be done manually (unless you incorporate some sort of current sensing feedback).

The upside of driving your load resonantly is that you probably won't need as much rail voltage, since the resonant circuit will step up the voltage somewhat.

Today I took out one of my better (but still cheap) El-wire drivers and sampled its voltages when loaded with 6' of 2.3mm green el-wire. Here are the results:

Input: 5.00v DC @ 0.10 A
Output: 115v AC, 1.8 kHz, almost perfect sine wave

On the scope, it looks to be about 220v peak-to-peak. Sigh. This means I now need a minimum 220v rail, or dual rails at 110v each!

How are normal drivers operated? Are they specced to certain lengths of wire, or certain frequencies?

 

[AleXYZ]

Thought I'd provide an update on this project. I was out for a week on vacation which gave me time to sit back and re-think my approach to the problem(s).

After a lot of semi-successful experimentation with boost topology, I decided that the efficiency wasn't good enough for what I needed. I am now going to implement a push-pull transformer inverter for several reasons:
(1) Higher efficiency
(2) Fewer parts
(3) Higher potential output voltage

I did some interesting experiments with multiple connected strings of El-wire of different colors and lengths at the same time. Optimal brightness seemed to occur at about 1800 Hz. To have output comparable with commercial El-wire drivers, I found that I needed a RMS voltage around 120-130 VAC. This means I need DC rails for the H-bridge as high as 200V.

Since this is to be a portable device (i.e. it should be wearable on one's body and not impede movement) it needs to have as small a footprint and weight as possible. Finding the right kind of push-pull transformer was difficult. After extensive time spent on the Digikey catalog I finally found they sell a few CCFL inverter transformers. I have some of these on order and will post results next week. They are much lower profile than power transformers and can handle up to 10 watts. Since I only spec'd 5 watts maximum output, I selected a 6 watt version by Bourns for this project.

CCFL transformers take input voltages ranging from 10v to 26v and have an output around 1300v. If I'm right, it should be no difficulty to use these transformers to make 150 - 200v rails. If I'm wrong, I'll go back to trying my hand at custom-winding a transformer.

For conversion from 200v DC to 120v AC, I have chosen the IR2210 H-Bridge driver with two 400v N-Channel MOSFETs. Actually there will be two sets of these to make a full bridge. One of the reasons I went with the IR series was that the power for the gate drives is derived from the low voltage (battery) source and not off of the high voltage rail. It's really pretty neat how they do this for the high side driver. I really want ALL of the power being inverted to high voltage to be used as efficiently as possible for the El-wire.

Thanks again to Genome in particular for pointing me in the direction of IR's gate drivers. IR's parts are expensive but are extremely well suited to the task.