Micro-controller adjustabe Buck converter? (for MPPT)

[stevo]

Hi,

I am designing a MPPT (Maximum Power Point Tracking) device to be used
with a solar panel. In the design I need a high efficiency buck
converter whose output can be controlled by a micro-controller. The
general requirements are:

1. Vin min ~ 12V Vin Max ~ 22V
2. Vout mine ~ 13V Vout Max ~ 18 V
3. At least 0.5A of current
4. Adjustable output (in a way that can be controlled by a uC).

Anyone know of such a device? Do PWM buck converters do this? Most
adjustable buck converters that I have seed require 2 resistors to set
the output voltage. I suppose I could use a digital pot for this, but
that adds a lot of complexity.

Thanks
Stephen

 

[Tim Wescott]

Hi,

I am designing a MPPT (Maximum Power Point Tracking) device to be used
with a solar panel. In the design I need a high efficiency buck
converter whose output can be controlled by a micro-controller. The
general requirements are:

1. Vin min ~ 12V Vin Max ~ 22V
2. Vout mine ~ 13V Vout Max ~ 18 V
3. At least 0.5A of current
4. Adjustable output (in a way that can be controlled by a uC).

Anyone know of such a device? Do PWM buck converters do this? Most
adjustable buck converters that I have seed require 2 resistors to set
the output voltage. I suppose I could use a digital pot for this, but
that adds a lot of complexity.

Thanks
Stephen

If you're willing to learn the switching regulator theory, consider
getting a microcontroller with some ADC inputs and a PWM output, then
drive the FET straight from that. You'll be putting yourself into a
space where a software fault will let the smoke out of you drive
devices, but it'll be a minimum parts-count system, to be sure!

 

[John Larkin]

Hi,

I am designing a MPPT (Maximum Power Point Tracking) device to be used
with a solar panel. In the design I need a high efficiency buck
converter whose output can be controlled by a micro-controller. The
general requirements are:

1. Vin min ~ 12V Vin Max ~ 22V
2. Vout mine ~ 13V Vout Max ~ 18 V
3. At least 0.5A of current
4. Adjustable output (in a way that can be controlled by a uC).

Anyone know of such a device? Do PWM buck converters do this? Most
adjustable buck converters that I have seed require 2 resistors to set
the output voltage. I suppose I could use a digital pot for this, but
that adds a lot of complexity.

Thanks
Stephen

You could use a standard Simple Switcher sort of circuit, and sum a
DAC output into the feedback sense node of the switcher chip to push
its setpoint around.

John

 

[Klaus Vestergaard Kragelund]

Hi,

I am designing a MPPT (Maximum Power Point Tracking) device to be used
with a solar panel. In the design I need a high efficiency buck
converter whose output can be controlled by a micro-controller. The
general requirements are:

1. Vin min ~ 12V Vin Max ~ 22V
2. Vout mine ~ 13V Vout Max ~ 18 V
3. At least 0.5A of current
4. Adjustable output (in a way that can be controlled by a uC).

Anyone know of such a device? Do PWM buck converters do this? Most
adjustable buck converters that I have seed require 2 resistors to set
the output voltage. I suppose I could use a digital pot for this, but
that adds a lot of complexity.

I'm not answering your question, but a MPPT function - won't you be in need
of controlling the input current to the SMPS rarther than the output to
achieve the MPPT point?

Cheers

Klaus

 

[Spehro Pefhany]

Hi,

I am designing a MPPT (Maximum Power Point Tracking) device to be used
with a solar panel. In the design I need a high efficiency buck
converter whose output can be controlled by a micro-controller. The
general requirements are:

1. Vin min ~ 12V Vin Max ~ 22V
2. Vout mine ~ 13V Vout Max ~ 18 V
3. At least 0.5A of current
4. Adjustable output (in a way that can be controlled by a uC).

Anyone know of such a device? Do PWM buck converters do this? Most
adjustable buck converters that I have seed require 2 resistors to set
the output voltage. I suppose I could use a digital pot for this, but
that adds a lot of complexity.

How so?

Best regards,
Spehro Pefhany

 

[Joerg]

If you're willing to learn the switching regulator theory, consider
getting a microcontroller with some ADC inputs and a PWM output, then
drive the FET straight from that. You'll be putting yourself into a
space where a software fault will let the smoke out of you drive
devices, but it'll be a minimum parts-count system, to be sure!

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

If it's for mass production or consumer use that's the most cost efficient
way to go. Probably this could also be done with dual slope AD which
eliminates the internal ADC and shaves off a couple of nickels in cost. But
since dual slope conversion is slow Tim's cautionary words about smoke and
that kind of stuff would become even more important. Certainly a case where
I'd wear protective eye gear when testing.

Regards, Joerg.

 

[Tam/WB2TT]

If you're willing to learn the switching regulator theory, consider
getting a microcontroller with some ADC inputs and a PWM output, then
drive the FET straight from that. You'll be putting yourself into a
space where a software fault will let the smoke out of you drive
devices, but it'll be a minimum parts-count system, to be sure!

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

That's not just idle talk. We used an 8051 controlled PWM, and kept blowing
transistors and traces until the software guy got it right.

BTW, I don't see anything wrong with the digital pot. The advantage there is
that you can put fixed resistors around it to limit the range.

Tam

 

[Joerg]

Hi Tam,

If in doubt you can always have an external limiter which should not require
more than a transistor and a few discretes. That can be configured to pinch off
the switching device in case things run out of control.

Remember the rpm limiting governors on some older Volkswagens? These simply
interrupted the ignition when the engine was revved to high, creating that
stakkato sound. While not a healthy mode of operation this scheme was able
prevent the engine from blowing up.

The other thing that may help here is that the illumination and thus the source
characteristics on solars does not change too rapidly. The load is another topic
but if it is a larger battery it won't change in load characteristics that fast
either. Of course, one always has to anticipate a fault event such as a solar
panel losing contact or a battery cell to fail. That's where a hardware limiter
can come to the rescue.

Regards, Joerg.

 

[Genome]

| Hi,
|
| I am designing a MPPT (Maximum Power Point Tracking) device to be used
| with a solar panel. In the design I need a high efficiency buck
| converter whose output can be controlled by a micro-controller. The
| general requirements are:
|
| 1. Vin min ~ 12V Vin Max ~ 22V
| 2. Vout mine ~ 13V Vout Max ~ 18 V
| 3. At least 0.5A of current
| 4. Adjustable output (in a way that can be controlled by a uC).
|
| Anyone know of such a device? Do PWM buck converters do this? Most
| adjustable buck converters that I have seed require 2 resistors to set
| the output voltage. I suppose I could use a digital pot for this, but
| that adds a lot of complexity.
|
| Thanks
| Stephen

Since there are times when your required output voltage exceeds your
input voltage then a simple buck converter won't do the job for you.
It's output is of the same polarity but less than the input.

If you could accept an output of opposite polarity then buck-boost or
flyback would do the job for you. Unfortunately it suffers from high
input and output ripple currents which might hurt elsewhere, especially
the solar array.

A sepic converter would give you low ripple on the input and high ripple
on the output with the same output polarity either above or below the
input and non-isolated and as such is likely to be your easiest
solution.

Other isolated topologies would also meet your requirements and the best
would probably be an isolated Cuk converter with low input and output
ripple.

I'd still favour the sepic for simplicity.

Providing a programmable output is as simple as programming the supply
reference voltage by any means available.

Of course if you go for a full microprocessor control solution you'll be
able to solve the right half plane zero problem and light an LED when it
goes bang.

The pseudo code would be.

Begin
Set blown up LED bit high
End

You can actually avoid the pain of programming all that stuff by
hardwiring the LED with a suitable resistor to the microprocessor supply
so it lights up when power is first applied.

DNA

 

[Robert Reimiller]

2. Vout mine ~ 13V Vout Max ~ 18 V

Are you chargeing a lead acid battery? If so you probably want a range
more like 10.5 (pretty much discharged battery) to 16.5 (extreme cold
chargeing)

adjustable buck converters that I have seed require 2 resistors to set
the output voltage. I suppose I could use a digital pot for this, but
that adds a lot of complexity.

Not sure what the resolution is of digital pots are, I seem to
remember them being like 8 bits. The charger I did a couple of years
ago used a 12 bit DAC to set the converter output voltage. I used
three resistors, like so:

convertor-output-voltage --VVVVV-----+------VVVVV---- DAC-output
|
+--- convertor-feedback-input
|
<
<
<
|
Ground

 

[The other John Smith]

Hi,

I am designing a MPPT (Maximum Power Point Tracking) device to be used
with a solar panel. In the design I need a high efficiency buck
converter whose output can be controlled by a micro-controller. The
general requirements are:

1. Vin min ~ 12V Vin Max ~ 22V
2. Vout mine ~ 13V Vout Max ~ 18 V
3. At least 0.5A of current
4. Adjustable output (in a way that can be controlled by a uC).

Anyone know of such a device? Do PWM buck converters do this? Most
adjustable buck converters that I have seed require 2 resistors to set
the output voltage. I suppose I could use a digital pot for this, but
that adds a lot of complexity.

Thanks
Stephen

Hi, Stephen -

Let me say up front that my experience with solar cells is limited to the
SP-55 and SP-75 panels from Siemens (or Shell, maybe). These are 55 and 75
Watt panels for charging 12V battery systems.

I started a project to do just what you have in mind. It came to a
screeching halt when I looked at the current/voltage/temperature curves for
the panel. It turns out that I was just as well off by simply connecting the
panel to the battery directly when it needed a charge and disconnecting when
the battery was fully charged. The reason is that the panel is already very
close to producing maximum power output into a 12V battery after it reaches
operating temperature. And it will reach a higher temperature than you
expect. Remember that the panel is only about 10% efficient so the remainder
of the input energy is heat so that, for every square meter of panel, you
get about 900 Watts of heat (on a good day, at the equator). There might be
a slight advantage to using a constant power controller in the coldest part
of the winter, but maybe not.

All I'm saying is be sure to look at the temperature/output curves before
you put in a lot of work. Solar systems designed to charge 12V batteries are
already nearly at maximum output power.

John

 

[stevo]

I'm not answering your question, but a MPPT function - won't you be in need
of controlling the input current to the SMPS rarther than the output to
achieve the MPPT point?

Acutally, I thought the same thing. However, it appers that all the
literature on MPPTs points to controlling the voltage ouput of the
buck converter instead of the incoming current.

 

[stevo]

Are you chargeing a lead acid battery? If so you probably want a range
more like 10.5 (pretty much discharged battery) to 16.5 (extreme cold
chargeing)


Not sure what the resolution is of digital pots are, I seem to
remember them being like 8 bits. The charger I did a couple of years
ago used a 12 bit DAC to set the converter output voltage. I used
three resistors, like so:

convertor-output-voltage --VVVVV-----+------VVVVV---- DAC-output
|
+--- convertor-feedback-input
|
<
<
<
|
Ground

Yes, I will be charing 12V lead-acid batteries. But would I really
need to have an output voltage less than 12V? Your DAC solution looks
good. Thanks

 

[Klaus Vestergaard Kragelund]

Acutally, I thought the same thing. However, it appers that all the
literature on MPPTs points to controlling the voltage ouput of the
buck converter instead of the incoming current.

Well, then you haven't been reading all the literature. When I worked on the
MPPT for a power unit for the Rosetta satelite I saw plenty of papers where
they regulated on the input current. And so did we by the way

(We used a super buck converter to minimize input current ripple)

Cheers

Klaus

 

[Robert Reimiller]

Yes, I will be charing 12V lead-acid batteries. But would I really
need to have an output voltage less than 12V?

If you hook this to a battery that has been discharged to 11 volts and
your convertor can only go down to 12 volts, then there is no way you
will be able to run at the MPPT.

 

[Joerg]

Stevo,

Just in case you want to take a peek at uC switcher design anyway this app
note may be a good start:

http://www.microchip.com/download/appnote/pic16/00216a.pdf

Regards, Joerg.

 

[Martin Riddle]

Wait for the upcomming June or Aug issue of home power.....

Cheers

 

[Dave VanHorn]

I recently did a NIMH charger in an Atmel Tiny-26.
Efficiency wasn't a design goal, but it could have been.

The basic form of it is a PWM buck converter, with constant current output.
It could easily be reconfigured to constant voltage.
The 64 MHz PWM output was a big plus, letting me get to 500kHz switching to
keep the inductor small enough for my application.

 

[Tony]

Virtually any buck controller should be OK - your adaptive loop just replaces
the controller's outer voltage control loop (much better than adding another
loop outside the voltage control loop, as it's more tolerant of the load
impedance, eg battery or motor). I did this in the '80s, but without an MCU
(just counted down from the converter's clock, applied that as a periodic
disturbance to the loop, correlated the output power with the disturbance with
some simple analog circuitry, and used that to trim the mean setpoint. It worked
great, but never went into production due to "irreconcilable differences". My
guess is that you'll still need quite a bit of analog circuitry in there to find
the subtle trends in amongst the hash.

Wait for the upcomming June or Aug issue of home power.....

Cheers

Tony (remove the "_" to reply by email)

 

[Rich Grise]

8051!!?!?!??!!! [0]

I've done PWM with an 8051, and I've done PWM wiht a 68HC11, and
if I ever do PWM again it will be with Motorola, I think.

Speaking of PWM, I'm entertaining fantasies of a battery-operated
MIG welder again, where I'd need a buck regulator that could
regulate current up to about 65A with a source of 24V nom.
of Lead-Acid batteries and a welding arc for a load. Well, there's
no reason to lock into lead-acid, if any other style of battery
would make more sense for an app like this. Highest power density
is the operative word there, I think.

What would it cost to have a real engineer design something like
that?

Thanks,
Rich

If you're willing to learn the switching regulator theory, consider
getting a microcontroller with some ADC inputs and a PWM output, then
drive the FET straight from that. You'll be putting yourself into a
space where a software fault will let the smoke out of you drive
devices, but it'll be a minimum parts-count system, to be sure!

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

[0]
That's not just idle talk. We used an 8051 controlled PWM, and kept blowing
transistors and traces until the software guy got it right.

BTW, I don't see anything wrong with the digital pot. The advantage there is
that you can put fixed resistors around it to limit the range.

Tam