Sorry for the delay. Back at the bench now. I tried with 2.4k on the series resistor for the LED. Still no luck.
View attachment 319996

No luck when the anode from D1 is connected to the Ref of D2, when I remove this the LED turns on at 27.2V as shown below:
View attachment 319997
Tried to make it as clear as possible in the picture!

Its difficult to see the TL431 connections with your wiring style.
However, the TL431 on the left looks to be missing a ground connection (center pin is anode).

I can only suggest re-checking all connections. I've bread boarded the circuit and it works, but I'm not very happy with it.
In post #17, I suggested using an LM339 instead.

 

Sorry about the sketchy wiring job, I was hoping it would make it more clear in a picture if everything was spread out, but clearly that was not the case.
I'm avoiding using the LM339 due to needing this component to work for up to 50V.
However I got the circuit to work on the bench perfectly with some changes:


I'm very happy with how this configuration is performing. Will be trying a modified version for my 48V window version soon if I get a power supply to test that high!

 

Sorry about the sketchy wiring job, I was hoping it would make it more clear in a picture if everything was spread out, but clearly that was not the case.
I'm avoiding using the LM339 due to needing this component to work for up to 50V.
However I got the circuit to work on the bench perfectly with some changes:
View attachment 320092

I'm very happy with how this configuration is performing. Will be trying a modified version for my 48V window version soon if I get a power supply to test that high!

Glad you got it working.
Does your schematic accurately show the bread boarded circuit?
I don't see how what you've done makes a difference.(?)