How to dim a light without changing its color temperature

[Steve Smith]

Well, it would not be very difficult for an engineer to develop a continuous-current regulator for a LED, the main problem is that its efficiency would be very lowi and it would heat much, much more than a PWM regulator.

Most PWM circuits run at a very high frequency, perhaps 50KHz. Putting a small bit of capacitance on the output would make it DC. Although I doubt that it would affect a meter, even without it.


Steve.

 

[Brett Rogers]

I scored an old Kyoritsu camera tester for free, by being in the right place at the right time recently. It incorporates a variable light source across several EV, not stepless, but accurate. Also tests shutter speeds, of course, but it needs the light source to do that. Point is that, if you are in the USA (as your details suggest), if you keep your eyes peeled, you ought to be able to find yourself an earlier unit in good repair at a price that will not break the bank in your country. You've said you want to check shutters and exposure meters, surely, the easiest way to do this is to acquire a tester that is designed to do just that, rather than re-invent the wheel?
Cheers,
Brett

 

[jlbruyelle]

Most PWM circuits run at a very high frequency, perhaps 50KHz. Putting a small bit of capacitance on the output would make it DC. Although I doubt that it would affect a meter, even without it.
Steve.

Sure, but this is an addition to a PWM regulator. I was replying to a post about avoiding PWM, which means that I was talking about linear regulators - which are not a good idea in this kind of application.

BTW SBC sensors have an extremely high bandwidth actually: I measured one up to 2 MHz and I could have improved it with a better amplifier. Although I'm sure most meters have a lowpass filter to avoid problems with flickering light, I'd be surprised if there was no exception.

 

[Chan Tran]

I bought this light bulb

EcoSmart 100-Watt Equivalent R20 CEC Dimmable LED Flood Light Bulb in Daylight 5000K (1-Bulb) A20R2050WT2001 - The Home Depot

Enhance your home with the EcoSmart R20 reflector LED light bulb. This 100-Watt equivalent LED light bulb uses on 12.5-Watt of energy to produce 1470 Lumens of Daylight (5000K) color temperature. Perfect

www.homedepot.com

to replace the modeling lamp in my Novatron flash head. I put together a dimmer circuit and dim it. To my surprise the color temperature doesn't change as I dim it (it stayed at 5200K) and also it's not pulsed either. I used a fast light sensor and an oscilloscope to verify this. Now I have a light that I can vary intensity without changing the color temp as well as it's continuous.
Now what is a good design for a mixing chamber? I intend to use 2 of these lamps.

 

[koraks]

To my surprise the color temperature doesn't change as I dim it

LEDs don't change color temperature.
And dimmable LED bulbs indeed shouldn't flicker. Although that's not an absolute; it depends on the driver circuitry. Some do flicker, usually at net frequency (60Hz where you live). Again, it all depends.

Now what is a good design for a mixing chamber?

How about a cardboard box, painted white on the inside, with a piece of white cloth or white plexi? Or maybe a small softbox.
Not sure what the physical requirements are. Is this supposed to go over the Novatron? Maybe post a photo or two.

 

[Chan Tran]

LEDs don't change color temperature.
And dimmable LED bulbs indeed shouldn't flicker. Although that's not an absolute; it depends on the driver circuitry. Some do flicker, usually at net frequency (60Hz where you live). Again, it all depends.




How about a cardboard box, painted white on the inside, with a piece of white cloth or white plexi? Or maybe a small softbox.
Not sure what the physical requirements are. Is this supposed to go over the Novatron? Maybe post a photo or two.

Oh no. I simply need to screw the bulb into the Novatron head that's all. There is no need to do anything. But out of curiousity I wonder if I can use the bulb for what I wanted to do when I started this thread over a year ago. I want a lighted diffused surface about 4 to 6 inch round or square. I want the luminance to be equal everywhere on the surface. I also want the light intensity to be variable. I use this to check camera exposure meter. I can adjust the brightness to a certain level using a spot meter to verify that then put the camera to it to test the camera meter. Thus I need to keep consistent color temp because camera meter respond differently with different color temp. I think meters were calibrated with 4700K light but 5200K is close enough I would say.
By the way the solid state relay I used to dim the light

https://www.digikey.com/en/products/detail/sensata-crydom/MCPC2425C/752698?utm_adgroup=General&utm_source=google&utm_medium=cpc&utm_campaign=PMax%20Shopping_Supplier_Sensata_8118_Co-op&utm_term=&utm_content=General&utm_id=go_cmp-20504611347_adg-_ad-__dev-c_ext-_prd-_sig-Cj0KCQiAhc-sBhCEARIsAOVwHuQq9cKv9NmEmEgfvR748IYIEn2B7EmDJ4GZDfAByrZvIFSyncBc0fUaAsWAEALw_wcB&gad_source=1&gclid=Cj0KCQiAhc-sBhCEARIsAOVwHuQq9cKv9NmEmEgfvR748IYIEn2B7EmDJ4GZDfAByrZvIFSyncBc0fUaAsWAEALw_wcB

its output I am quite sure is pulsed.

 

[koraks]

Ah, I see. Well, you could experiment a little with a box and two diffusion screens of some sort at a distance and the light source at a good distance from the first diffusor. While a double-diffusor setup comes at the cost of a considerable amount of lights, it tends to do a pretty good job at evening out any hot spots.

In terms of making the thing adjustable, I'm a bit hesitant when it comes to using a regular bulb on a dimmer, because the dimming behavior may not be as linear as you'd like it to be, and I'm not sure if the dimming range is sufficient. Again, it depends...just give it a try and see if it's good enough.

 

[Chan Tran]

Ah, I see. Well, you could experiment a little with a box and two diffusion screens of some sort at a distance and the light source at a good distance from the first diffusor. While a double-diffusor setup comes at the cost of a considerable amount of lights, it tends to do a pretty good job at evening out any hot spots.

In terms of making the thing adjustable, I'm a bit hesitant when it comes to using a regular bulb on a dimmer, because the dimming behavior may not be as linear as you'd like it to be, and I'm not sure if the dimming range is sufficient. Again, it depends...just give it a try and see if it's good enough.

I was thinking of using 2 bulbs for good intensity. I don't need higher than LV15. I would check the intensity with a spot meter as I dim it. I do not depend on the potentiometer position (or input voltage) to set it to certain level and thus linearity isn't a problem. I found the diming range is more than what I need. I need from LV15 down to LV9 or so.

 

[koraks]

I was thinking of using 2 bulbs for good intensity.

2 bulbs is one stop faster than a single bulb. A single bulb of twice the power achieves the same. Two bulbs of twice the power = +2 stops. Etc. Two bulbs should work fine, as should one bulb, or four or whatever takes your fancy!

I do not depend on the potentiometer position (or input voltage) to set it to certain level and thus linearity isn't a problem.

OK, good.
Btw, my concern about dimming linearity not so much applies to the dimmer as such, but mostly how the LED fixture responds to it. But ignore this; if you're happy with the real-world performance, then that's all that's relevant.

 

[Chan Tran]

2 bulbs is one stop faster than a single bulb. A single bulb of twice the power achieves the same. Two bulbs of twice the power = +2 stops. Etc. Two bulbs should work fine, as should one bulb, or four or whatever takes your fancy!



OK, good.
Btw, my concern about dimming linearity not so much applies to the dimmer as such, but mostly how the LED fixture responds to it. But ignore this; if you're happy with the real-world performance, then that's all that's relevant.

I understand what you say about linearity. For example if you supply the light bulb with 50% power the light intensity isn't half. So I would use a light meter to actually monitor the brightness and adjust until the meter indicates the level I need.

 

[koraks]

Yeah, that'll fix whatever linearity problem that might exist.

 

[MTGseattle]

Here's something that may or may not be appropriate; A vintage GE Footcandle selectometer

Light is behind a fairly large brass iris aperture that is controlled by a slide lever.

General Electric Foot-candle Selectometer (No Model Number) Works!! S5468 | eBay

In good working condition. It is in good condition for its age. Item is sold as described. Dan Batson, DGD Inc.

www.ebay.com

 

[ic-racer]

Since this thread started I found the plans for a nice light source on the internet. Since building it, I don't know how I got along without it.

 

[Bill Burk]

Here's something that may or may not be appropriate; A vintage GE Footcandle selectometer

Light is behind a fairly large brass iris aperture that is controlled by a slide lever.

https://www.ebay.com/itm/3133785667...sMprmvOeo/4IZgGXwXe1ouZ87D|tkp:Bk9SR6jopa-ZYw

That's been on my watch list and the seller made me an offer, but it's not going to work for testing light meters. It's more suited to find out "how many footcandles make for comfortable reading"

 

[MTGseattle]

I was wondering about that. It seems like at a close distance (inches), the light would be the same regardless of aperture. I think the 1/2 stop increment scrims make the most sense so far.

 

[MTGseattle]

@ic-racer has built a pretty cool device. I've read through the thread about it.

 

[Bill Burk]

This issue is really kicking my butt.

The light standard (the blue box out of frame on the left) produces 100 foot lamberts which I calculate to be 32 candles per square foot.

I simply cannot get any Weston Master II to read 32. Best is 25.

This difference is so consistent that I think the difference might be the law of inverse squares falloff. You see, I cannot hold the Weston cells directly against the opal glass of the standard. There’s a lip on the standard and the meter cell is also set back, for about 3/8 inch distance where I just can’t get closer.

But I can match the light of the standard with a variable brightness test box (by aiming a Spotmeter). And here I can try different geometries.

I created a small tube with a translucent face that I can insert into the Weston, flush against the bubble lens in front of the cell.

This raised the reading a sixth stop above 25. But I am looking for 32 when the light matches the standard.

I bought a cheap color temperature meter but it doesn’t recognize tungsten light. It’s so stupid.

Meanwhile I am trying to hit the maximum reading off the Weston when other meters tell me the light at the panel is 100, as I vary the color temperature with 80b filter or higher voltage.

Backing up a step…

For the first pass checking light to readings, I poked holes in pieces of foil to attenuate the light, and then fine tuned with the variable light.

A see-saw pattern emerged as I made different range foil screens which needed the fine-tune brightness turned up or down. Readings that should have been a third-stop apart were nearly identical.

So these cells are extremely sensitive to tungsten voltage variations.

My latest rig ties the cell to an enlarger rack.

So I am going to get in range and vary the light by raising the cell. At close range 3/8 inch rise drops the reading about a third stop. And I know it doesn’t change the color temperature.

So my next set will be with the test lamp at full voltage. Ranged by foil, trimmed by raising and lowering.

We’ll see if that gives me a 32.

 

[Bill Burk]

@ic-racer has built a pretty cool device. I've read through the thread about it.

I’ve met with Serhiy and it’s great he put the plans out.

I can’t use LED’s for my current project because Selenium cell meters bring their own problems to the table.

 

[Bill Burk]

I’m getting 32 at 32 with the diffuse face tube insert and the test lamp cranked near full.

The Chinese color temp sensor tells me it’s 2400K

This helps. But of course now I want to “calibrate” at 4700 K and it’s not going to be easy since I can’t push the needle up.

 

[MTGseattle]

Look at all of that interesting old stuff! You're having fun right Bill? That's what counts.

 

[Bill Burk]

Yes, it’s turning into fun. A bit more success today.

The main improvement is a brass holder I made for Sekonic slides to attenuate the light. The set of slides ASA 40 to 100 (at 1/60) provide 1/3 stop changes in light. Also setup a dedicated microscope stand to fine tune. It can raise the cell enough to reduce light by two stops. And I have a three stop filter (Sony speaker grill sandwich).

Taking advantage of the peak sensitivity…

This Weston Master II just calibrated (with bright tungsten light 2458 K) over the range 1.6 to 50 candles per square foot.

 

[ic-racer]

Fill me in, how are you calibrating it? I have a "Master II" but did not think it could be calibrated.

 

[Bill Burk]

Fill me in, how are you calibrating it? I have a "Master II" but did not think it could be calibrated.

Right. You don’t. You just clean and balance the meter and then discover whether or not the assembly is calibrated.

This meter is determined to be calibrated for these conditions:

With a tungsten source adjusted in brightness as needed to make the remaining steps best meet the goal…

Aim source at a diffuse screen, adjust screen luminance by a combination of physical attenuation and adjusting source to screen distance until the Sekonic L-758DR indicates the nominal sample value, taking spotmeter readings off the diffuse panel.

Note readings of milliamps, voltage, and Weston Meter dial.

Samples were checked in third-stop increments from 1.6 to 50 candles per square inch on the exposure meter scale.

 

[eli griggs]

I need a light source to test camera exposure meter as well as testing camera shutter speeds on auto.
I used to use a Beseler computerized 45 dichoic head for this. I just dial in equal filters to create neutral density.
I am thinking of a few ways.
1. Copying the dichroic head design using a mechanism that moves the filter in and out of the light path but instead of using filter I would use just an opaque gate to reduce the amount of light.
2. Inserting ND filter in the light path.

Dimming the light by reducing power would lower the color temperature.
Any ideas?

My Rozeann bought me a Lume Cube RGB Panel Pro for Christmas which is brighter than the slightly smaller RGB Go.

The newest light has adjustable colour temperature to 7500-7600 'daylight' and can be fitted to a honeycomb screen for more 'directed' light coverage.

The smaller RGB Go only goes to 5600 daylight and both drop way down into the warmer light range, household or golden hour type light.

Both have their power levels digitally controlled from 0 to 1 to 100% so the light temperature should stay the same across the range.

These are not large lights and ambient lights should be eliminated while testing, but both should give a fairly close target enough illumination for testing.

These lights can make thousands of colours with their LEDs, no gels needed.

By the way, the larger light can be controlled by your phone, once Lume Cube Bluetooth software is installed on it.

There are other larger panels available, but I'm just sharing what I have to hand.

These should be light enough to measure shutter speeds with a simple, home made diode/LED shutter tester with a mini jack plug installed and plugged into your laptop, loaded with running "Audacity" Software.

The Audacity Software runs a split second recording program and the light, in front of the camera lens opening, and it's audio jack, on the shutter tester, will perfectly read the light duration, coming through the shutter from opening to the closing of the curtains.

You then trim the recording to the duration of the "exposure", faster than you can say "exposure", and you have your shutter speed, which you can test again and again.

I hope this helps and, Good Luck.

 

[ic-racer]

Right. You don’t. You just clean and balance the meter and then discover whether or not the assembly is calibrated.

This meter is determined to be calibrated for these conditions:

With a tungsten source adjusted in brightness as needed to make the remaining steps best meet the goal…

Aim source at a diffuse screen, adjust screen luminance by a combination of physical attenuation and adjusting source to screen distance until the Sekonic L-758DR indicates the nominal sample value, taking spotmeter readings off the diffuse panel.

Note readings of milliamps, voltage, and Weston Meter dial.

Samples were checked in third-stop increments from 1.6 to 50 candles per square inch on the exposure meter scale.

Ok, thanks. I saw some wires and potentiometers and wondered if you were changing resistors or something.

Maybe I'm missing something, you are measuring the meters selenium cell's output directly in mA and V? Why not just observe the readings of the scale?