Need urgent help for a buck-boost converter drive circuit.

[arif1996]

Hi...

I am using a standerd buck boost configuration. A inductor, diode(60EPF)
capacitor and a MOSFET switch(IXFN80N50). The buck boost is designed s
that it will handle 1.8KW power. All the rating of the component value ha
come through calculations and ensures that the converter will operate i
continuous conduction mode.

During designing, I considered that the variation of the input voltag
should be between 50-250 volt and output voltage should stay at 120 v. Bu
I don’t have any control on the output and I also don’t need that for m
purpose. I will be trying to track the maximum power of a wind turbine
All the algorithm will be done to a PIC 18F452 and it will generate th
required PWM signal and thus switch the buck boost. Now PIC can only giv
us 0-5 volt. So,I used a driver stage(with TC 4420) to drive the MOSFEE
which will give me 0-15 volt proportional to PIC signal. Driver and PI
has isolated ground with an optocoupler. Siganl from the driver i
connected to the gate of MOSFET and ground of the driver connected to th
source i.e. high side driving configuration.

The problem is :

Lets say, switching frequency is 100KHz and duty cycle is 50% from the PI
, so the driver stage ouput is 50% duty cycle with 0-15 volt voltage leve
and frequency is 100KHz to drive the MOSFET. Its ok. I have seen it.Now i
I increase the input voltage level, say…10…20…30…gate signal ok…switchin
the MOSFET…getting output voltage.

But if I increase further…say…40…50…60…gate signal distorts i.e. i
doesn’t show me 100Khz and so as the 50% duty cycle…gate signal jus
vanish after a certain input voltage...getting output voltage but tha
doesn’t make much sense.

I am stacked on this point and trying to search some literature but ther
is not much information about any practical solution. Waiting for any kin
of suggestion.

Today i have bought IR2117 and IR2110 to use for a driver circuit. The
have isolated ground and are useful for both high and low side driving.
have calculated my input driver circuit current to charge up the MOSFET G
cap and i have found that if a driver circuit can give me 6 amps then i
will be ok. But in the data sheet of IR 2117 i didnt see any informatio
about how much current it can give? It says Io+ 200/400mas..i am confuse
in this point too.

Will it be able to drive such high power level(around 2KW) MOSFET? I hav
seen that they are suitable but i need to be sure from someone or need t
understand what does it mean?

A lot of questions but hope some former will guide me through the righ
way. I am stacked on this issue for along time.

Thanks
Arif

 

[Ken Smith]

Hi...

I am using a standerd buck boost configuration. A inductor, diode(60EPF),
capacitor and a MOSFET switch(IXFN80N50). The buck boost is designed so
that it will handle 1.8KW power. All the rating of the component value has
come through calculations and ensures that the converter will operate in
continuous conduction mode.

I notice that you say "standard buck boost" and then "a MOSFET". What is
your circuit really?

Is the circuit that causes trouble like this?

PIC -----> OPTO ----> driver -----> MOSFET

Where along the chain does the 100KHz stop being seen?

 

[arif1996]

Buck-Boost:
Input DC----->switch(u need a MOSFET at this point)--->inductor(i
parallel)--->diode(in reverse bias)---->capacitor(i
parallel)--->inductor(for filtering purpose)--->capacitor(in parallel fo
filtering purpose)---Output DC--->

To drive the MOSFET:
PIC(generating 100KHz PWM wave) -----> OPTO (isolation)----
driver(increase the voltage level in proportion to PIC signal) -----
MOSFET.

 

[Ken Smith]

Buck-Boost:
Input DC----->switch(u need a MOSFET at this point)--->inductor(in
parallel)--->diode(in reverse bias)---->capacitor(in
parallel)--->inductor(for filtering purpose)--->capacitor(in parallel for
filtering purpose)---Output DC--->

Is it this:

Switching device
----
---------O O---+---[L]---+---
! !
--- ---
^ ---
! !
GND GND

If yes then it is only a Buck converter.

 

[arif1996]

Lets drop the discussion about whether it is buck or buck-boost...i wil
solve the problem by myself..its hard to send a lot of mail without an
output...

value
has is

Buck-Boost:
Input DC----->switch(u need a MOSFET at this point)--->inductor(in
parallel)--->diode(in reverse bias)---->capacitor(in
parallel)--->inductor(for filtering purpose)--->capacitor(in paralle for
filtering purpose)---Output DC--->

Is it this:

Switching device
----
---------O O---+---[L]---+---
! !
--- ---
^ ---
! !
GND GND

If yes then it is only a Buck converter.

 

[Genome]

Hi...

I am using a standerd buck boost configuration. A inductor, diode(60EPF),
capacitor and a MOSFET switch(IXFN80N50). The buck boost is designed so
that it will handle 1.8KW power. All the rating of the component value has
come through calculations and ensures that the converter will operate in
continuous conduction mode.

During designing, I considered that the variation of the input voltage
should be between 50-250 volt and output voltage should stay at 120 v. But
I don't have any control on the output and I also don't need that for my
purpose. I will be trying to track the maximum power of a wind turbine.
All the algorithm will be done to a PIC 18F452 and it will generate the
required PWM signal and thus switch the buck boost. Now PIC can only give
us 0-5 volt. So,I used a driver stage(with TC 4420) to drive the MOSFEET
which will give me 0-15 volt proportional to PIC signal. Driver and PIC
has isolated ground with an optocoupler. Siganl from the driver is
connected to the gate of MOSFET and ground of the driver connected to the
source i.e. high side driving configuration.

The problem is :

Lets say, switching frequency is 100KHz and duty cycle is 50% from the PIC
, so the driver stage ouput is 50% duty cycle with 0-15 volt voltage level
and frequency is 100KHz to drive the MOSFET. Its ok. I have seen it.Now if
I increase the input voltage level, say.10.20.30.gate signal ok.switching
the MOSFET.getting output voltage.

But if I increase further.say.40.50.60.gate signal distorts i.e. it
doesn't show me 100Khz and so as the 50% duty cycle.gate signal just
vanish after a certain input voltage...getting output voltage but that
doesn't make much sense.

I am stacked on this point and trying to search some literature but there
is not much information about any practical solution. Waiting for any kind
of suggestion.

Today i have bought IR2117 and IR2110 to use for a driver circuit. They
have isolated ground and are useful for both high and low side driving. I
have calculated my input driver circuit current to charge up the MOSFET GS
cap and i have found that if a driver circuit can give me 6 amps then it
will be ok. But in the data sheet of IR 2117 i didnt see any information
about how much current it can give? It says Io+ 200/400mas..i am confused
in this point too.

Will it be able to drive such high power level(around 2KW) MOSFET? I have
seen that they are suitable but i need to be sure from someone or need to
understand what does it mean?

A lot of questions but hope some former will guide me through the right
way. I am stacked on this issue for along time.

Thanks
Arif

I think the type of supply you are describing has been known in the past to
old English people as a non-isolated flyback converter, Americans call it a
buck-boost. The output is negative with respect to the input and, in your
case, you have added an additional stage of LC filtering.

Your mosfet is 'high side' referenced so you need some form of level
shift/isolated scheme to drive it. You might be being caught out by the fact
that the source goes negative with respect to your system ground when you
switch the mosfet off.

If you are using some sort of bootstrap scheme to supply power to your
driver then you end up adding the 'flyback' voltage to your original bias
voltage so as you increase the input voltage the output voltage also
increases until you are exceeding the voltage rating of something somewhere
and it starts misbehaving.....

Note that the IR level shifted drivers only work with a positive level
shift, the source of the driven mosfet must always be positive with respect
to the ICs ground (there might be a -5V max allowed). Otherwise the IC gets
blown up.

Since you don't seem to mind that the output is negative with respect to the
input then I would suggest you look at a CUK converter. Here the mosfet is
ground referenced and your input/output ripple currents will be much less.

DNA

 

[arif1996]

I think the type of supply you are describing has been known in the pas to
old English people as a non-isolated flyback converter, Americans call i a
buck-boost. The output is negative with respect to the input and, i your
case, you have added an additional stage of LC filtering.

Your mosfet is 'high side' referenced so you need some form of level
shift/isolated scheme to drive it. You might be being caught out by th fact
that the source goes negative with respect to your system ground whe you
switch the mosfet off.

If you are using some sort of bootstrap scheme to supply power to your
driver then you end up adding the 'flyback' voltage to your origina bias
voltage so as you increase the input voltage the output voltage also
increases until you are exceeding the voltage rating of somethin somewhere
and it starts misbehaving.....

Note that the IR level shifted drivers only work with a positive level
shift, the source of the driven mosfet must always be positive wit respect
to the ICs ground (there might be a -5V max allowed). Otherwise the I gets
blown up.

Since you don't seem to mind that the output is negative with respect t the
input then I would suggest you look at a CUK converter. Here the mosfet is
ground referenced and your input/output ripple currents will be much less.

DNA

you are right..i am using a high side driving..but i tried using isolated
ground from the signal and driver using an optocoupler..but thing seeems
to work for a low voltage input to dcdc but not for a high voltage and the
problem is there..

I can go through a cuk approach but the problem is the time..i dont have
much time to design a cuk again, solder it and test it...i also go through
the driver stage by keeping the modification using a transformer..but again
the question is the time...

In the recent days i am trying to search for such problem..i want to give
a try for IR2117 IC..the problem is...in the data sheet they have a
recemmonded connection to drive the high side but i had the problem with
the COM in..data sheet said that it should be logic ground(i.e. signal
ground)..then how about the Vcc ground..where i should place it..Vcc that
i am giving to the circuit as 15 volt and coming from an another dcdc
converter which takes 5 volt input and give a 15 volt output with a
isolated ground i.e. ground different from the PIC signal..what should i
do in this case...if you have a connection diagram for IR2117 then it
would be helpful for me...

lets put it this way...i am giving supply to IR2117 as 15 volt which has
different ground from the signal...signal is coming from a PIC...output
Voh and Vs should go to the MOSFET gate and source..i connect a
oscillloscope to Voh and Vs and it should come up with a signal which will
maintain the same signal from the PIC only with a increase voltage
level...plz let me know what i am doing wrong here...

Thanks a lot for your help. Thats why new people come to forum..that
someone who has experience will help them and will try to explain the
wrong steps...

Thanks
Arif

 

[Genome]

you are right..i am using a high side driving..but i tried using isolated
ground from the signal and driver using an optocoupler..but thing seeems
to work for a low voltage input to dcdc but not for a high voltage and the
problem is there..

I can go through a cuk approach but the problem is the time..i dont have
much time to design a cuk again, solder it and test it...i also go through
the driver stage by keeping the modification using a transformer..but
again
the question is the time...

In the recent days i am trying to search for such problem..i want to give
a try for IR2117 IC..the problem is...in the data sheet they have a
recemmonded connection to drive the high side but i had the problem with
the COM in..data sheet said that it should be logic ground(i.e. signal
ground)..then how about the Vcc ground..where i should place it..Vcc that
i am giving to the circuit as 15 volt and coming from an another dcdc
converter which takes 5 volt input and give a 15 volt output with a
isolated ground i.e. ground different from the PIC signal..what should i
do in this case...if you have a connection diagram for IR2117 then it
would be helpful for me...

lets put it this way...i am giving supply to IR2117 as 15 volt which has
different ground from the signal...signal is coming from a PIC...output
Voh and Vs should go to the MOSFET gate and source..i connect a
oscillloscope to Voh and Vs and it should come up with a signal which will
maintain the same signal from the PIC only with a increase voltage
level...plz let me know what i am doing wrong here...

Thanks a lot for your help. Thats why new people come to forum..that
someone who has experience will help them and will try to explain the
wrong steps...

Thanks
Arif

Your X-Remote-Host: is 142.162.49.157 which, with the use of an IP address
locator places you in Mount Perl, Canada. With a bit of Googling about we
get to....

http://www.mun.ca/

Which matches your e-mail address. I was guessing because you sound like a
student hitting the deadline on some sort of project. I am also a
bastard..... Oohhhh, Faculty of Engineering and Applied Science....

http://www.engr.mun.ca/home/

OK, thinking about it I could have got there quicker by just corrupting your
e-mail address a bit.....

Anyway. You posted via, User-Agent: Electronics-Related.com usenet HTML
frontend. So that was via the WWW, a web page. Unfortunately the site you
visited not only stores and displays your message to other people who use
that particular website...... it also posted it to UseNet.

UseNet is 'not a forum'............. it's a community of Ill Will.

Bugger, I hit the wrong button and replied to your e-mail address the first
two times. That means there are messages in your in-box at the university
that give the game away. I don't know about the privacy policy in your
department but someone might have a sniff and see what you have been up
to......

Ho Hum

Stop Panicking, it does not matter. Damn, this might appear in the 'forum'
you posted to as well as UseNet. That means people might get the wrong
impression about the professional standards of your university.

Tell you what, I don't care, someone needs a good kicking for placing you in
this position.

If you have been landed with this thing and this is as far as you have got I
can only assume that you have inherited some lecturers pet project that he
gives to all of his students and they have messed it up already because he
is not much use at teaching them in the first place and have not a clue
anyway.

1.8kW wind turbines with PIC controlled load regulating things.

Mong mong mong mong.

Based on your description..... there might be two things. First one is,

How you are connecting your oscilloscope to make the measurement? OK OK, I'm
sorry, but it's not unknown for something to go wrong if you forget about
how the bits are grounded. Use a differential probe.

Second one is.... what sort of opto-coupler are you using? I don't know
enough about using such things in that particular location but you are going
to need an extremely high common mode rejection (shielded) device. The
source of your mosfet will be screaming up and down at huge amounts of volts
per nano-second and that will mess up the transmitted signal.

Seriously, it sounds like you've been given a piece of shit and you should
deserve better.

I have forwarded this to your Associate Dean.

Like I said, I'm a bastard.

DNA