Small signal transfer function of offline LED driver?

[eem2am]

Hello,

I have designed an isolated, offline flyback LED driver (output current regulated) with no electrolytic caps. (Its 10W, 300mA LED current, 100KHz switching frequency)

The circuit you will see is , I virtually guarantee, the future of domestic mains lighting the WORLD over!!………………

The only way I can stabilise the LED current is to use the following compensation network.

WHOLE SCHEMATIC:
Xxxxxxxxxxxxxxxxxxxxxxx

(by the way, the actual circuit will use a quasi resonant flyback chip, but in essence, this is the circuit)


CLOSE-UP OF COMPENSATION NETWORK

XXXXXXXXXXXXXXXX

I got this compensation network’s component values by trial and error,
-please could you provide the small signal transfer function of the modulator, power stage, and error amplifier ?

(That is, the small signal, open loop transfer function of this power supply, so I can check the gain and phase margin for stability analysis)

zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz
Here are CLOSE-UPS of other parts of this LED driver……

PRIMARY SIDE

Xxxxxxxxxxxxxxxxxxxxx

FLYBACK TRANSFORMER

Xxxxxxxxxxxxxxxxxxxxxxx

OPTOCOUPLER FEEDBACK NETWORK (with current sense filter)

Xxxxxxxxxxxxxxxxxxxxxxx


Zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz
Here are some of the circuit waveforms:-

Primary DC bus

Xxxxxxxxxxxxxxxxxxxxxx

(-please admire the wide mains rectifier conduction angle)

LED CURRENT

XXXXXXXXXXXXXXXXXXXX

ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ
Now here is the really strange thing,

The current sense resistor voltage has 500mV of peak to peak noise on it at the switching frequency……….

Xxxxxxxxxxxxxxxxxxxxxx
CURRENT SENSE RESISTOR VOLTAGE

The above noise is there in spite of the 4th order RC filter

-I tried to filter away this noise with a 2nd order active filter as seen in the schematic but it does not help much, ………..
….since the FB pin of the opto-coupler driver still has 400mV of peak to peak noise on it at the switching frequency……….

Xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Opto-coupler driver FB pin voltage :

I cannot use a bigger secondary bulk capacitor because that would violate the whole point of this circuit…….
…it uses no electrolytics for bulk filtering , but instead film capacitors which are very bulky so I cannot use much of them.

(It is pointless using electrloytics in general LED lighting since the electrolytics will not last as long as the LEDs would last (10 years) )

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
Here is the LT Spice simulation file:

https://www.2shared.com/file/880936...____FLYBACK_MAINS_LED_CLAMP_OPTO__171009.html
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

So anyway, my questions are.

1. Please may you tell me how this circuit manages to work perfectly , in spite of the dreadful noise on the opto-coupler feedback pin?

2. Please may you provide the small signal transfer function of the modulator, power stage, and error amplifier/feedback network ?

(i.e. the open loop, small signal transfer function so that I can do stability analysis?)

Or please may you tell of the way I can work this out?

 

[eem2am]

Hello,

I have re-designed the circuit to use less components. –Also, even LESS bulk output capacitance!
….now just 4.7uF film capacitor , which, for a discontinuous flyback of 10W output is very good. - i pray that you can agree

Here is the new circuit………

WHOLE SCHEMATIC
XXXXXXXXXXXXXX

The idea of having the three feedforward capacitor paths is that having three such paths , rather than one, means that the caps can be of smaller value than if there was just one path.

This is necessary since the bigger cap values (4u7 plus) in X7R ceramic caps are very expensive .

CLOSE-UP OF TOP OF SECONDARY SIDE
XXXXXXXXXXXXXXXXXX

CLOSE-UP OF BOTTOM OF SECONDARY SIDE
XXXXXXXXXXXXXXXXXX

Here are the waveforms
LED current in green, DC Bus voltage in blue
Xxxxxxxxxxxxxxxxxxxxxxxxxx

However, I still have not managed to get rid of the switching frequency ripple on the Feedback (FB) pin of the opto-coupler driver.

The reference voltage of this driver is just 0.6V , so the peat-to-peak ripple here of 350mV seems excessive……

RIPPLE ON FEEDBACK PIN OF OPTO-COUPLER DRIVER
XXXXXXXXXXXXXXXXXXXXXXX

The simulation still works in spite of this large amount of ripple on the Feedback (FB) pin.
Will it be OK to just leave it in there?



Please may I ask these questions.?

Do you know how I could calculate the open-loop small signal transfer function of this current regulated converter?

Do you know if the large amount of ripple on the feedback (FB) pin of the opto-coupler driver is OK?

[By the way , here is the LTSpice simulation file]
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
https://www.2shared.com/file/887372...__FLYBACK_MAINS_LED_split_filter__171009.html