Questions: 1 - How can I to "tune" the output clamp, or it can be necessary to tune output RC clamp ? Or I have other problems?

Attached PIXL data showed the PIVS is 136V for LNK4149 design, the Vrr of output diode is 300 V. The 300 V output diode should be fine without RC snubber circuit unless there is very high voltage overshoot due to high leakage inductances from transformer and layout. The transformer leakage inductance can be reduced with split primary winding. The leakage inductance due to layout can be reduced by minimizing the primary and secondary power switching path as shown in recommended layout in the LNKPH datasheet.

2 - What is the relation temperature of LNK420 and output diode clamp ?

There is no relationshio between LNK420 temperature and output diode snubber circuit. It might be necessary to increase the core size to reduce the primary and secondary turns.  23 secondary turns might be too high for 75 W output design. Also it is necessary to have spilt primary winding to minimize the leakage inductance.

 3 - In design 1, why it dont start without otput clamp ?

Maybe  there is very high voltage overshoot without diode clamp, and the output diode might be broke down due to high reverse voltage stress.

4 - see attached pixls files. coud you please help in 75W design ?

Answered in Q #2.

Dear PI-Skywalwer, too many thanks for your atention. Some more considerations:

1- I dont measured leakage inductance, but, I reasembled the transformer in design 2, and now It´s much better. Please verify that there is some problem in PI XLS designer software because in "Transformer primary design parameters" we have 47 turns of 1 x 24AWG. This is 2 layers to be splited in one layer of 23 turns plus one layer 24 turns.  BUT, in "Electrical diagram" we have 23 turns of 2 x 24AWG plus 24 turns of 2 x 24 AWG. This is 2 layer + 2 layer. you can see this on 75W and 50W design. My first prototip of 75W desing was made 2x24 + 2x24AWG. Now I have reassembled the transformer with 1x24 + 1x24 and is much better: 

Now, the temperature of LNK420 is 55C ( the same temperature. I think I will increse heat sink), D block 73C ( firs was 89 ), D clamp 64 ( first was 82), and Z clamp 75 ( first was 92).

2 - Any Idea to use the same core NEE30 ( custom) to obtain 75W ina a safe way ?  

Dear PI - Skywaler, 

About the 23 secondary turns.  I changed VO, form 38,4 to 37,2 ( a more acurate value), and changed KP from 0,55 to 0,65. The secondary turns are now=18 turns.  If I change kp to 0,67, the CMA is higuer than 600Cmil/A and NS=17.  Could you please tel-me some thing about this ?  can we use higuer vaues of KP, for example 0,7 or 0,8 ?  I think there is no problem in use CMA higuer than 600Cmil/amp, correct ?

again, many thanks. 

Higher Kp will reduce the inductance and reduce the number of winding, and there is no issue with Kp = 0.7 or 0.8 for normal operation.

Low Kp value is recommended to meet EN61000-3-2 class C harmonic current limit. 

Sometimes, specially in high power application, it is necessary to ajust suggested transformer construction in the PIXL software. This is the reason why there is editing winding information menu in the Transformer Construction page in the PIXL.

NEE30 is safe to use for 75 W application, but the efficiency can be improved with larger core.

Now I´m changing KP, to a higuer value, 0,6 or 0,7 to see what hapens, I see perfectaly in Pixls that this reduce de secundary turns and indutance. You say that 23 turns are a big number for this power, than I´m doing some tests with less turns.  The harmonics contens are higer, but I´m not afraid with this at this moment. I prefer to use Nee30 for padronizations porposes. Thank you very much.