Problem reaching 4700K with Solux lamp [solved]

[torger]

The mystery continues, just got a reference tested lamp from my dealer. Unfortunately it was a 50W lamp and my DC power supply can only get it up to 12.4 volts then (I bought 35W lamps), and it became 3700K at that voltage.

Getting paranoid, I made visual comparison with a Grafilite Mode A3 fluorescent which others than me has measured to ~5500-5600K (I get the same result), and clearly the light from that is bluer.

I also did the white balance picker thing, with high CRI sources color temp can be guess quite well by the camera and indeed it shows almost exactly the same values as I measure (a little varmer as the paper is not 100% white).

The other picture attached with the brighter lamp is the 35W run at 16.4 volts, and then I get 4700-4800K, but voltage is way too high for that.

It doesn't seem likely that the measurements are wrong, considering both methods showing the same thing, and other lamp measurements matching other people results.

So I need to try with a different power source and see if I get some other result, there still could be some issue with my DC power supply. I have measured it, but only the voltage, maybe it does something funky when delivering power...

[WayneLarmon]

there still could be some issue with my DC power supply

Don't Solux bulbs normally run on 12 volts AC?  I don't have an official Solux fixture.  What I did was get a 60 watt 120 volt (USA) to 12 volt transformer that was in the outdoor lighting section of my local hardware store.  And a MR-16 fixture that was designed for outdoor lighting.  I wired everything together and taped it to a mike stand.

I have an AC variac (transformer with a movable tap, so the output can be varied from below normal voltage to above normal voltage.  Was used for troubleshooting electronics--to see if the device only flaked out on high or low line voltage.)  I was going to disassemble my fixture so I could get to the wires, plug it into the variac, and use my multimeter to monitor the voltage on the light fixture (after the 12 volt transformer).  I haven't done this yet.   (My multimeter is old and was acting flaky.  I wanted to get a new multimeter.)

With AC, it is tricky.  There is peak voltage, peak-to-peak voltage, and RMS.  I'm not sure which corresponds to DC voltage.  And I'm not sure which my meter measures.  (I was going to plug the fixture in without the variac and see what it measured on the lamp terminals.)

Does running on DC instead of AC make any difference?

Wayne

[GWGill]

With AC, it is tricky.  There is peak voltage, peak-to-peak voltage, and RMS.  I'm not sure which corresponds to DC voltage. 

For power, you need RMS, but it isn't difficult to convert if you know that it is a sine wave, and that's what is typically assumed with simpler meters.
It's much more difficult to measure true RMS when the waveform is not a sine wave - you need either a tricky thermal system, or some clever electronics or digital sampling and maths.

[torger]

Halogen lamps should work on both AC and DC. I always thought the the typical was DC and the Solux voltage numbers were for that, but it's surely worth investigating... electricity is not my best area of knowledge so I don't really know what differences to expect.

In this report they talk about a DC power supply though, so I think it should work. May be my power supply that is simply bad.
http://www.babelcolor.com/download/Light_under_control_2005-11-08.pdf

[JRSmit]

Halogen lamps should work on both AC and DC. I always thought the the typical was DC and the Solux voltage numbers were for that, but it's surely worth investigating... electricity is not my best area of knowledge so I don't really know what differences to expect.

In this report they talk about a DC power supply though, so I think it should work. May be my power supply that is simply bad.
http://www.babelcolor.com/download/Light_under_control_2005-11-08.pdf

typical halogen power supply delier 11.5volt on its terminals. Add to that the voltage drop over the cable a d you easily get close to 11 volts. You can check on the solux datasheet the drop in kelvin.

[torger]

Got some better measurement gear. The power supply meter does seem to show the correct result. The culprit are seems to be the lab cables I got with the power supply that is bad, I lose 5 volts(!) over them, leaving the lamp under-powered. I haven't yet been able to get proper cables, but it does indicate as soon as I get them the lamps will be on spec, like they've always been.

So it seems to be one of those user errors after all, I guess they need to add one more item to their list of user mistakes...

I'll report back when I've got to test with better quality cables. I think DC instead of AC should be okay, and hopefully it is because it's a lot easier (and cheaper) to get variable DC power supply than a variable AC power supply in a suitable range, at least around here.

[torger]

Did a quick test without having proper cables, but making an ugly-connect of the socket directly, and then the lamp holds the promise. The 45watt power supply I got get problems overdriving the 35watt lamp though, due to that it requires more than 3 amperes. I get it up to 12.4 volts though which makes one of my lamps just reach 5000K CCT which is all I need.

Actually reading on the description of the lab cables that says "max 2 amps", and actually divding 35watt / 11.5volts = 3.04 amps had been a good thing to do. Than I'd known that the cables were not adequate and the power supply (indeed 45 watts, but max 3 amps) a bit too weak. Duh...

Although it seems like my current supply makes it with the smallest possible margin, I may sell this and get a proper one just to punish myself :-).

[JRSmit]

Bewaren that when still cold the lamp draws a lot more ampère. Regulated power supplies need to be oversized by factor of 3 to delier the startup current.

[degrub]

Or use what we in the US call a "ballast"

[torger]

In this case we want the light warm and stable, so delivering extra current early on should not be needed, right? Or are you referring to short circuit protection?

Anyway, bought a bigger power source now and some decent cables, going to test after work today.

[torger]

So at last, with a proper power source, got a 0-30V 0-5A DC lab variable power supply, and proper cabling I got it to work as promised. All my fault, no shadow on Solux.

With one of my Solux 4700K 35W lamps I got 5000K CCT at 13.2volts, 3.25 ampere, and 5500K at 15.8volts 3.6 ampere.

Both the D50 and D55 simulation looks fine, although running it at 15.8 volts is really pushing it of course. For some quick camera profiling shots should be okay.

I've attached another alternative to simulate daylight, using a standard 3000K halogen spotlight and put a film photography 80B light balancing cooling filter in front. Note that the scale reaches 1.4 to fit the bump making it look a little flatter than it should compared to the others. My conclusion is that this filter method does not work well enough, using a Solux is better.

[Iliah]

In this case we want the light warm and stable, so delivering extra current early on should not be needed, right? Or are you referring to short circuit protection?

Anyway, bought a bigger power source now and some decent cables, going to test after work today.

If cables got warm, the resistance goes up. Cables need to be 3x-5x current rated.

[torger]

I'm using 10A cables now, seems to work fine, the old one were only 2A designed for measurement...

[AlterEgo]

I'm using 10A cables now, seems to work fine, the old one were only 2A designed for measurement...

so you hook the cables directly to the bulb contacts then ?

[Iliah]

I'm using 10A cables now, seems to work fine, the old one were only 2A designed for measurement...

Yes, 10A cables is quite a guarantee. Especially if they are copper.

[GWGill]

Yes, 10A cables is quite a guarantee. Especially if they are copper.

The rating of a cable depends on the parameter used to determine it. Typically this will be a temperature limit - i.e. what current can you pass through the cable before it gets so hot that the insulation is in danger of melting ?

For the purposes of this particular exercise, the important parameter is the voltage drop though the cables, which translates to the required maximum cable resistance.

[Iliah]

The rating of a cable depends on the parameter used to determine it. Typically this will be a temperature limit - i.e. what current can you pass through the cable before it gets so hot that the insulation is in danger of melting ?

For the purposes of this particular exercise, the important parameter is the voltage drop though the cables, which translates to the required maximum cable resistance.

There are many factors, including cable length, that affect voltage drop. The rule of thumb is to have 3x to 5x Amps rating against what a generic brand manufacturer states on a cord (that is, if it is stated 10 Amps, one can expect 2-3 Amps without a serious voltage drop), and that is for reasonably short cables. 10 Amps power transmission is a copper wire of approx. 1mm diameter (18  AWG); but in household (short cords) they can use 0.4mm - which, in turn, is rated for 0.36 Amps only for power transmission. One can use a calculator like http://www.rapidtables.com/calc/wire/voltage-drop-calculator.htm