Need advice, another questiong regarding LED bulb and enlarger.

I manage to get 2 A27 Led bulbs, 12w=75w 200º coverage. One is 2700K and the other 4000K. My dough is which one to use in my enlarger?
I have a meopta opemus III with a 75w Philips photocrescenta bulb, like this one here: http://www.lighting-gallery.net/gallery/displayimage.php?album=2555&pos=3&pid=78184
One of the comments specifies that the temp color is 2950K for that kind of bulb.

The color temperature will affect the contrast grade when using variable contrast paper. With the same contrast filter, the 4000K bulb will effectively give you a higher grade then the grade of the filter. The 2700K a slightly lower grade. How much the contrast will shift I can't tell from experience. With the 2700K bulb you may lose a high grade (4.5..5). With the 4000K you will likely lose some on the low side. So if you hardly ever use the highest grades, go for the 2700K, otherwise the 4000K. The 4000K is probably a bit brighter (more lumen) so that might be another plus for the 4000K.

Or try to find a 3000K bulb. Philips used to make 3000K bulbs which I prefer over the 2700K bulbs for household purposes.

I see, thanks!

1+ to spijker's explanation.

The halogen lights common in enlargers (~100W) typically have a colour temp of 3100 to 3300 °K.
I've used a Sylvania 4000K-LED and the result appeared around 1/3 grad step harder compared to new halogen bulb (same filtration, Adox MCC and Ilford MG IV); I did not, however, do any detailed investigations or measurements to this effect.
To my experience, the 230V enlarger bulbs used in my Meopta condensor enlarger are a bit warmer (lower °K), again the 75W is a bit warmer than the 150W.

With some LED bulbs, I've experienced a significant delay until the light output stabilized.

I'd appreciate, if others could share their experiences with off-spec bulbs in enlargers, too.

I know it would cut down some output, but could an LED enlarger bulb be corrected to 3200 K using lighting gels in the filter drawer?

Say 1/4 or 1/2 CTO to get the 4000k correct?

I've done a lot of work converting tungsten fresnels to daylight using HID bulbs and ballasts, and I can usually dial the temp to perfect daylight with CTO and minus green gels and an hour or so shooting raw images of a color chart with a DSLR.

As I understand it, the the spectral distribution of light from an LED is different than from a halogen or incandescent bulb. I'm not sure, but it may also be non-continuous.

So the distribution of contrast grades with colour heads or filters and variable contrast paper may also be quite different than with the original bulb.

Indeed, so much of this stuff was created with an expectation of tungsten light, with its CRI (Color Rendering Index) of a perfect 100.

I fight with this a lot as a digital video guy. Digital sensors are generally tuned to expect 5200k daylight. The bizarre color from fluorescents, LEDs, HIDs, even HMI lighting - a lot of gels required to get everything even. Shooting stills, hell - a couple Speedo packs and it's F16 at ISO 100, nice clean daylight. Get into motion work with digital sensors, and you need a pretty big budget to just assume your color will be clean and hopefully shoot under ISO 1000.

I'd be curious to see if an LED source could be gelled to work reliably with standard contrast filters. Since the contrast grades are essentially color correction gels, it **seems** like it would work. But there are bizarre spikes of green or yellow in what is being sold as "daylight" sources these days. I have some pretty strnage gel combos on set.

Thank you for the replies. I had not time to install and try the bulbs in the paper I use. Mainly I use ilford multigrade IV rc pearl and have 2 boxes of old ilfobrom grade 3.

I buy all my LEDs from Superbright, they have data sheets for their single bulb and components and the spectral distrubution is not continuous, and the tech guys there say you can expect some deviation from their data sheets, i.e. the bulb might not peak exactly at 640nm but close enough for government work.