Need SMPS suggestion

Doesn't the spectral power distribution curve shift to the right when you increase the filament power (say to automatically compensate for the glass getting black)?

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Hello,

the problem is this: 12V and 24V halogenic bulbs need to run on a pretty constant (0,3%) voltage.

So far we've been using toroid transformers, but the problem is that as the transformer warms up the voltage drops by about 0.5V (this is a 24V toroud with three 150W bulbs connected). These are lamps used in puppet animation, so if the studio turns off the (hot) transformers at night and turns them back on in the morning, there is a visible light jump in the picture.

The problem isn't really the fact that the things run hot, but that they take several hours to reach a stable temperature. Leaving the lights on during the night would not only be a waste of power and bulb lifetime, it could also create a light jump due to lamp degradation.

Another possible problem could be variations in mains voltage, but measurements have shown that that is not the case.

Regulated SMPS comes to mind as a solution. I'd like to have an individual supply for each lamp with the output V settable to 80, 90 or

100%. The actual voltage is of course not as much an issue as its constantness.

Given that lightbulbs are simple loads, I thought about getting rid of the output rectifier and caps (just a small rectifier for the feedback path). But this would probably create EMI emission issues with the lamps connected via 20 feet of unshielded cables. Or would a bit of series inductance at the output get rid of this?

Another solution could be driving the transformer not with hard switching, but with a (say) 100kHz sine wave. This would have to be some kind of resonant scheme, but I've never heard of it.

Or do I have to bite the bullet and go DC?

Suggestions?

robert

If your aim is to correct this - not to make money. That's what, 20A? Or 10W variance. It occurs to me that a small transformer, fed from a 10W amp, and a true RMS servo circuit or similar could do the job, connected in series.

It's difficult to understand just exactly what you're talking about? The power-supply is just a step-down tranformer off the line and you want

0.3% voltage regulation?

I've been involved with medical instruments that use light sources to measure color spectrum changes in blood and tissue. The solution there was to use a photodiode to measure a sample of the light output from the lamp and regulate the voltage supplied to the lamp. That takes care of line supply variations, temperature changes, and bulb degradation over time.

Jim

That's what post-pro is for. ;-)

Cheers! Rich

I have no doubt that this is true, but from the original post, it sounds like Robert is doing Claymation, and needs to either maintain the color balance from one frame to the next, even though it might be overnight, or invest in some post-pro color balancing equipment/software.

I had the privilege once of being invited to a pre-release screening of a movie that was "a work in progress", because they wanted to survey audience members to see if they needed to do anything to make it more successful. The movie was "Who Framed Roger Rabbit?".

It was freakin' awesome. The color balance was crap, the continuity was crap, and they hadn't even finished the animation yet - I got to see some of the armatures in scenes where the toons had to do physical stuff, and there was at least one green-screen, where it was still in green-screen, I could spot the wires the actor was dangling from, and I could see the green sticks with the little white 'X's on them to define the sight lines. This was the scene where Eddy Valentine is in free-fall in ToonTown, and Bugs Bunny and Daffy Duck come skydiving alongside of him, and pretty much say, "OOps! No parachute, eh, Eddy? Good Luck!"

But after post-pro, (and drawing in the toons) it was a very fine movie indeed. :-)

Break a Leg! Rich

Have you tried an external forced air cooling scheme on your transformer? If you want to change transformers you can look into the parameters like surface area, core geometry, finish and insulation to reduce heat. Better location within equipment. Or you could go for the exotic ones with low resistance magnetic wires.

Constant 0.3% voltage? Halogens are known to work at 90% of their voltage rating and if well enclosed 80 to 70% of that. At the power levels you are working at I would say you have no choice but to go for Switchers. SMPS are super efficient (even >90%) so if you are not using them now you are throwing away money . Traditional linear power supply efficiencies are way below that. Switchers are known to have large input (and sometimes output) currents so you'd need a soft start circuit set up to limit that ... to extend your lamps lifetime to even double their normal rating . The web is full of people selling that stuff. You seem to have in mind something like a dimmer ... well that too could work as a soft start circuit. Switch the lights on at lower voltage then increase it for max. illumination.

DC versus AC. DC is known to attract positive ions from one side of the filament to the other causing irregular evaporation in the filament. The effect though seems to be insignificant on the lamps duration or efficiency .

Lamp degradation by extended use. Impossible because the whole process is self sustaining. Evaporated tungsten filament when deposited on the glass bulb reacts with the halide in the gas and the resulting compound is deposited on the filament.

EMI emissions ... well maybe later or someone else will cover that .

"Go easy on the whisky"

theJackal

To get at least 0.3% constant voltage on your lamps an undervoltage tracking arrangement circuit may work . You can surely find many on the Internet. They are basically based on the principle that the negative end of a capacitor connected to the emitter of a transistor tracks a reduction in voltage of the load. This turns the transistors emitter negative turning it on. The action progressively turns on 2 other charged capacitors which drive current to the load to prevent any further drop in load voltages.

"Go easy on the whisky"

theJackal

Going DC would simplify the whole exercise substantially.

Use a timer to turn on the lights a few hours before you get there.

--
Regards,
  Bob Monsen

It seems that you really need to get your video monitoring improved, like a wavefrom monitor/ vectorscope, and a decsent monitor and suitable hardware so you can match the stuff you shot yesterday with today's stuff.

I doubt that 0.3% voltage change is going to alter the colour temperature significantlly,in TV terms anyway, but it might affect the luminance so you really need to do a split screen with yesterdays stuff and alter the camera iris to match the Y sig. Just tweak the lighting controller to match

martin

Yes but with 15 lamps, ideally each with its own supply, it becomes tedious.

robert

My client does. Yes.

Many shots don't get digitally post-processed at all (it's 35mm all the way), and light jumps are very hard to correct for because not only the temperature of the lamp but also the film response to different light levels changes, requiring a lot of tweaking.

A great movie. But pre-screenings of half-finished films, as interesting it is to the professional, sadly ruins the experience of the movie as a piece of art. Re-watching it doesn't help much.

robert

Is it?

The subject line says "SMPS".

robert

Yes. But these lights are commercial 24V halogenic heads connected via 2-wire cable, and I'd like to keep it that way. Plus the entire lamps get incredibly hot (250F or more) and I don't know just how calibrated a PD would stay under those conditions.

robert

That's another option. The fact that the transformer gets hot (and the resulting reduction in output voltage) in itself is not really the problem, the problem is that it takes so long.

Forced air cooling would significantly reduce the heat up and thus the time to reach stable temperature, as well as the magnitude of the error.

robert

Not an option. Professional light meters (for photo and film use) measure 1/10 of a stop. The light jumps I'm talking about are half that and they are a huge effect. Of course it is possible to measure light more accurately, but the instruments get more involved and this is no option for practical use in a film studio.

Absolutely not an option. You wouldn't believe what f*ck-up potentials are already in a tightly-scheduled commercial shootings. They need to be eliminated by simple, robust, no- interaction solutions. These are not hobby projects.

robert

Too much work. Postproduction time is scarce and expensive.

You have no idea.

This is full blown 35mm theatrical stuff. So is the camera.

robert