1. I have connected 5V to common anode of 3 segments but it is driven by 3.3V MCU.
2. Also tried removing by R11,R16, & R17 but its not working. What could be issue here?
3. As soon as I turn high(3.3V) on T1,T2,T3 all segments get together.

 

Are you saying all come on even if T1 T2 T3 are at logic HIGH. That's not surprising.

If T1 is at 3.3v and the base of the transistor is at 5 - 0.7 = 4.3v then there is 1v over your 2k2 base resistor which is about 500uA, easily enough to turn the transistor on and with a gain of 100 or more that's 50mA collector current.

Try a couple of small diodes (eg 1N4148) in series with the 2k2 resistor so the logic output needs to go below 2.1v before base current will flow.

 

Try a couple of small diodes (eg 1N4148) in series with the 2k2 resistor so the logic output needs to go below 2.1v before base current will flow.

Good idea.
Or an LED if you want more flashing lights.

 

Good idea.
Or an LED if you want more flashing lights.

That works too, though with <1mA flowing you might not get much light out. Also a red led might not drop enough and a green might be too much, they vary so much these days with no-name cheap ones.

 

Have a look at the schematic of this breakout -
https://www.sparkfun.com/products/12009

 

Have a look at the schematic of this breakout -
https://www.sparkfun.com/products/12009

That's another option but arguably overkill for this specific problem which is a simple level shift.

 

That's another option but arguably overkill for this specific problem which is a simple level shift.

2 resistors and a mosfet?

 

Compared to 2 diodes, in this instance, yes. In any case this isn't really a level-shifting issue, it's a biassing issue as its not interfacing between logic levels as such.

 

Use a second transistor to set the correct levels as shown below, but the display control logic (T1 T2 T3) must change to High=on Low=off.

 

That works too. It depend on whether the TS wants to change his logic/code. Personally I would, as negative logic always annoys me.

For completeness, using the original logic, here's the one extra component solution... using a 2.4v Zener diode (plus change the base resistor to 1k5 for a little extra drive to run the segments at 15mA, or leave at 2k2 for 10mA/segment).