How does IR2110's HIGH-SIDE Internal Circuity work ?

[goldsmith]

Hi Gary
If you use a mosfet that has an internally diode , it can handle your aim as well . i.e if your mosfet has not any internally diode , you can connect it externally !
I'm a bit confused that what you want to do exactly . because you talked about SG3526 . it is a complete driver ! are you trying an SMPS ?
Best Wishes
Goldsmith

 

[TripACT]

Hi goldsmith,

please see attached schematics.
the SG3525A feeds the opto-coupler and then the driver/buck.
the buck design schematic isnt updated so never mind that (i will place another hexfet).

yes this is a generic SMPS- buck converter

btw - why do we need the diode between source and drain.

i've read the two papers you wrote. great explanation but i dindt understand how any of the diode in parallel the mosfets will conduct when they are in reveresed bias.

 

[goldsmith]

Humm , SG3525 is a good choice for your aim because it can create enough dead time as well .
What is your problem now ?
A question , why you connected those out puts together with two diodes ?

 

[TripACT]

another quesion please: the bootstrap drive process should happen only once on the start-up of the circuit. correct? in steady-state the capacitor will always be full of charge ?

 

[goldsmith]

Of course . it is correct .
And another thing that can be interesting for you is , a transformer driver . ( but it's quality isn't as high as a mosfet driver .)

 

[FvM]

okay. so the "wrong" buck configuration was correct only for steady-sate period of time. not for the initial time. i understand.

so anyways, how do i connect the LIN ? is it ok to leave it grounded in the IR2110 ?

the only modification i need to do is connect the other hexfet to the LOW SIDE output of the driver.

i wonder how the buck configuration will be then. can you draw a sketch ? according to my buck schematics ?

The two switch topology suggested by goldsmith is a synchronous buck configuration. It requires to drive the low side switch with a complementary signal, with sufficient dead time between low and high side drive signals. In so far LIN has to be connected.

I personally dislike the idea of a driver circuit dictating the output configuration. It's O.K. if you want a synchronous buck for performance reasons, otherwise i would rather think of a different driver option.

 

[TripACT]

Humm , SG3525 is a good choice for your aim because it can create enough dead time as well .
What is your problem now ?
A question , why you connected those out puts together with two diodes ?

i tie the two outputs (output A and output B) because each output produce only 1/2 cycle. Output B is 180 degree phase shifted from Output A.
when you use these two outputs independently, you get two seperated outputs with 180 degree phase difference and only 1/2 cycle. Output A will result first 1/2 cycle and output B will result the latter 1/2 cycle. They both symmetrical. when you tie them up - you get a whole cycle.

why diodes?
because the SG3525A has a totem-pole configuration on the outputs. if you tie the outputs together without diodes, the outputs will interfere each on another because of the totem-pole configuration.

- - - Updated - - -

Of course . it is correct .
And another thing that can be interesting for you is , a transformer driver . ( but it's quality isn't as high as a mosfet driver .)

okay. so the "wrong" buck configuration was correct only for steady-sate period of time. not for the initial time. i understand.

so anyways, how do i connect the LIN ? is it ok to leave it grounded in the IR2110 ?

the only modification i need to do is connect the other hexfet to the LOW SIDE output of the driver.

i wonder how the buck configuration will be then. can you draw a sketch ? according to my buck schematics ?

thank you alot !

Gary

- - - Updated - - -

i have another question...i'll explode soon

you explained about LO and HO switching relationship according to LIN and HIN inputs.
does the bootstrap capacitor fully charge on the first cycle ? i.e when LO=1/HO=0 initially. and then when HO=0/LO=0 the capacitor supply drive voltage to the higher mosfet.
are there any more and more cycles of LO=1/HO=0? if not. its like a one-time switch for startup circuit. how will that be possible in reality ?

 

[goldsmith]

tie the two outputs (output A and output B) because each output produce only 1/2 cycle. Output B is 180 degree phase shifted from Output A.
when you use these two outputs independently, you get two seperated outputs with 180 degree phase difference and only 1/2 cycle. Output A will result first 1/2 cycle and output B will result the latter 1/2 cycle. They both symmetrical. when you tie them up - you get a whole cycle.

why diodes?
because the SG3525A has a totem-pole configuration on the outputs. if you tie the outputs together without diodes, the outputs will interfere each on another because of the totem-pole configuration.

Dear Gary
Thank you for that explanation , i didn't know about this action . but it still is not clear to me that how with this method we can increase the D.C . if you explain it i'll be grateful .
Let back to your problem :

okay. so the "wrong" buck configuration was correct only for steady-sate period of time. not for the initial time. i understand.

so anyways, how do i connect the LIN ? is it ok to leave it grounded in the IR2110 ?

the only modification i need to do is connect the other hexfet to the LOW SIDE output of the driver.

i wonder how the buck configuration will be then. can you draw a sketch ? according to my buck schematics ?

thank you alot !

Gary

Follow things that FvM told instead of Lin . he is quite right .
And sketch of what ? a buck with two switch ?

i have another question...i'll explode soon

you explained about LO and HO switching relationship according to LIN and HIN inputs.
does the bootstrap capacitor fully charge on the first cycle ? i.e when LO=1/HO=0 initially. and then when HO=0/LO=0 the capacitor supply drive voltage to the higher mosfet.
are there any more and more cycles of LO=1/HO=0? if not. its like a one-time switch for startup circuit. how will that be possible in reality ?

Don't confuse about boot strap method . at first cycle the circuit will recover itself . in practice some first cycles are not important . but if there are important for you , a simple pulse transformer can be good . but i don't think that transient time is important for you . isn't it ?
Good luck
Goldsmith

 

[TripACT]

dear goldsmith,

it is really nice that you - mighty goldsmith can learn something from someone like me. i'll take it as a compliment :grin:

about the PWM - i may added some confusion there - the two outputs waveforms are not exactly shifted but horizontally flipped from each other. they are still 1/2 cycle each. combining them together you gain full range D.C.
hope it is clear now.
please see the representative block diagram attached here, to understand the output section better:




about the sketch, please see my PDFs i just attached earlier. one of them shows the drive connection to the buck using asynchronous buck configuration.
how can i modify it to use two switches (synchronous buck configuration you suggested).

about the recovery time of the bootstrap capacitor. yes i understand you. correct.
what signal goes to LIN of the driver? again - see the two PDFs i attached earlier. one PDF shows the PWM (the SG3525A) and the other one is the second part schematic - the optocoupler, driver and buck.

Gary

 

[goldsmith]

Oh thank you according to the attached diagram now i completely understood that how with that way we can increase D.C .
Ok , tomorrow i'll write some papers about buck converters for you . i hope thus you can get the idea to complete your project as well .
Best Wishes and thank you for your clarification . ;-)
Goldsmith

 

[TripACT]

no problem dear goldsmith !

looking forward for your tomorrow's papers.

have a good night my friend

Gary

 

[TripACT]

btw - can you recommend any reliable app note to read about how power mosfet work in principle ?

EDIT: i'm looking for something simple and easy. not too scientific.

thank you

 

[goldsmith]

Dear Gary
Hi
I wrote somethings for you , but it isn't complete yet . because today i wan not in my home and i was pretty busy . but i promise you that i'll send it to you soon .
BTW : do you like SMPS , really ? if yes , can you give me your email ? if yes , if you stay just some minutes i can send you many references .
Best Wishes and sorry because of this delay .
Goldsmith

 

[TripACT]

hi,

well its not that i love it, but i have a project to hand and just need to explore the most of it in order to make sure i make everything properly.

its okay - any help is accepted happily while you can.

Also, i need to hand the papers tomorrow so no time for more explorations :|

Gary