PFC Flyback led driver (LT3799) questions.

[treez]

Hello,
Please could you answer the following questions about the LT3799 PFC’d flyback LED driver?

LT3799 datasheet:
http://cds.linear.com/docs/en/datasheet/3799fa.pdf

Why is the current sense reference voltage just 100mV?
Other chips, such as LM3450, and all others, have at least half a volt for current sense reference voltage due to noise issues, why is the sense voltage so low for the LT3799?
Can the LT3799 be used with an opto-isolator as in the attached pdf schematic? (In this schematic, I have connected the internal error amplifer thats inside the LT3799, as a voltage follower. LTspice simulation of LT3799 based LED driver is attached, with the optocoupler feedback.

Why does the LT3799 have to use an internal multiplier block and why should it have to sense the input voltage? After all, the NCL30000 PFC flyback controller does not have either of these..
NCL30000 datasheet:-
http://www.onsemi.com/pub_link/Collateral/NCL30000-D.PDF

(the NCL30000 simply uses constant on time, which is all that is needed to give Power factor correction, when a Critical conduction mode flyback is used with virtually no primary bus capacitance.)

 

[treez]

Also, how does the datasheet LT3799 circuit do overload protection?.....there is no feedback from the output....so how?...I thought it was illegal to have an offline power source without overload protection......as you know, you cant rely on the bias winding falling below the operating voltage of the controller during overload because the ringing of the leakage inductance means that you still get high enough voltage on the bias winding, even if the output is totally shorted....so isn't it illegal, the shown LT3799 circuit in the datasheet?

 

[treez]

Hello,
Another question on the same lines...........Regarding control chips for offline LED drivers….
The LM3447 claims to be able to accurately regulate current in LEDs with an offline flyback and no opto-coupler feedback needed.
So why then does ti.com also have the LM3450, which claims to do the same thing, but needs not only opto-coupler feedback, but also a separate additional secondary side switching current regulator?


LM3450
https://www.ti.com/lit/ds/symlink/lm3450.pdf

LM3447
https://www.ti.com/lit/ds/symlink/lm3447.pdf

In other words, why would anyone buy the LM3450 when the LM3447 does the same job with far less components?

 

[BradtheRad]

Why is the current sense reference voltage just 100mV?
Other chips, such as LM3450, and all others, have at least half a volt for current sense reference voltage due to noise issues, why is the sense voltage so low for the LT3799?

As a general principle, by creating a low voltage across the sense resistor, it can then have a low Watt rating.

Half a volt across a sense resistor is practical if Amperes are low.

Sorry, don't know the answer to your other questions.

 

[treez]

Also regarding PFC flyback LED drivers, the LM3447 has an “FF” pin and a “FLTx” pin. The FLTx pin is supposed to decrease power throughput when the average input voltage decreases as when phase cut dimming takes place.
However, the FF pin is supposed to increase primary current as the average input voltage decreases (in order to maintain constant power to the load when the Vin falls.)

As such, the FF pin and the FLTx pins oppose each other….won’t this cause problems in the control of the PFC flyback?

LM3447 datasheet:-
https://www.ti.com/lit/ds/symlink/lm3447.pdf