Replacing Halogen Bulbs with LED Bulbs

This is happening to me too; slightly different system though so take my results with a grain of salt, test for yourself. 

Sensor Racks, ETCNet2 & sACN both control via Cobalt-on-Congo and Legacy Unison. The  http://www.tcpi.com/spec-sheets/ELITE/Designer-Elite/Designer-Elite-DIM-PAR-Series_print.pdf is what we liked* in a PAR38 size, and my local tech support guy allows as how it's one of the least horrible options he's seen. (He estimates his clients getting this kind of request a couple of times a month, these days.) The fade got down to about 6% before dropping out: if they pull the trigger on this purchase I'll try to create a custom curve for them from the desk. No perceptible flicker, and we tested without the existing diffuser lenses in place. However: we did see them briefly stick on at the bottom of a 5-second fade, couldn't duplicate it, but it does indicate caution. The whole circuit behaved as one, no popcorning within the bank of 12, but I wouldn't want to rely on the idea the whole roomful wouldn't do the same thing unpredictably and probably not in unison. In concert. Simultaneously. 

*By "liked," honestly, what I mean is "was surprised to not completely hate."  Building a fade that is more immediately perceptible to the audience--steeper at the top, I think--will be a challenge. I'll fling an email at the pros for more input, but my only other concern right now is the potential for camera flicker in casual video shots. If we light the audience specifically for IMAG, I'm sure we'll put tungstens on them, but I'm not sure without that there might not be a problem.  Hope it helps.

Thank you all. Very helpful. We are going to isolate one circuit, replace those lamps with the TCP bulbs, work on the dimming curves and see if it will give us acceptable results.

Please repost your results!  There are a lot of interested parties!  Thanks!

Do a slow out, say 20 count, as well as 5 or so. Same counts coming up.

Video the results and post.  

Video is problematic for this sort of thing. This is especially true with most people's default video device: their phone. The automatic exposure adjustments will try to compensate for the changing light levels as you dim and the resulting video will be very misleading at best.

Even a good camera that allows for manual control may still fail to show things that your eye would pick up, or may show things that are imperceptible to the eye. 

It turns out that finding a way to objectively assess dimming performance is a pretty difficult task to accomplish.

Video is problematic for this sort of thing. This is especially true with most people's default video device: their phone. The automatic exposure adjustments will try to compensate for the changing light levels as you dim and the resulting video will be very misleading at best.

Even a good camera that allows for manual control may still fail to show things that your eye would pick up, or may show things that are imperceptible to the eye. 

It turns out that finding a way to objectively assess dimming performance is a pretty difficult task to accomplish.