MPPT - simple, too simple ?

[Simon]

Attached is the PDf of my schematic. Basically the A section of ICi which is a LM393 is the heart of a basic SMPS but I have put a PIC in the feedback loop so that the voltage is controlled by the pic based on the current readings from R6 which is the sense resistor.

I have tried to implement a mosfet rectification in the SMPS output to make it more efficient but I'm unsure if i have protected against shoot through. The principle is that C4 will provide a delay after switch off of Q1 but it will instantly be cut off when the output of the comparator short C4. Is this safe or will spark fly ? at the end of the day it is solar powered so not the same sort of risk I'd get from say a battery supply.

I think I am yet to add soe circuitry to make sure it shuts down when dark and does not discharge the battery, but i have section B of the comparator to use up yet

[Neganur]

hmmmm I don't see any attachment

[RayJones]

Damn that's a simple circuit!
So simple it requires no parts whatsoever 

[Simon]

sorry yea forgot to attach  it is now attached

[Simon]

Damn that's a simple circuit!
So simple it requires no parts whatsoever 

yea i wish 

[Bored@Work]

From the datasheet:

The amount of current which the output device can
sink is limited by the drive available (which is independent of
V+) and the ? of this device. When the maximum current limit
is reached (approximately 16mA), the output transistor will
come out of saturation and the output voltage will rise very
rapidly.

(the datasheet doesn't even guarantee the 16 mA (typ.), but specifies 6 mA as the minimum current the device can sink)

Sounds like you need to current-limit the discharge of C4. Which in turn gives you an RC-circuit with a delay.

[blackdog]

Hi Simon,


It is not a good idea to put a processor in a feedback loop, it is always to slow to protect your Q1.
You can do it, but there is also a circuit necessary to quickly protect Q1 to limit the max current.
You can put a resistor in the source of Q1 and a PNP transistor to pull up the Gate of Q1.
This wil ofcourse lower the efficiency...
There are special "High Site" current sensor IC's for this function that can use small resistors 0.01 Ohm
It is all depending on de Current/power your source can deliver.

If Q1 is a IRF4905, it can handel 75A and more than 200Amp peak. No small transformer can do dat.
But, it is not easy to design powersupply's :-)

My 2 cents
Bram

[Simon]

Hi Simon,


It is not a good idea to put a processor in a feedback loop, it is always to slow to protect your Q1.
You can do it, but there is also a circuit necessary to quickly protect Q1 to limit the max current.
You can put a resistor in the source of Q1 and a PNP transistor to pull up the Gate of Q1.
This wil ofcourse lower the efficiency...
There are special "High Site" current sensor IC's for this function that can use small resistors 0.01 Ohm
It is all depending on de Current/power your source can deliver.

If Q1 is a IRF4905, it can handel 75A and more than 200Amp peak. No small transformer can do dat.
But, it is not easy to design powersupply's :-)

My 2 cents
Bram

The mosfet is controlled by the comparator, not the MCU, all the MCU does is set the threshold voltage of the comparator to control the output voltage. The MCU is not responsable for turning Q1 on or off and it does not need to react fast to for example turn the mosfet off mid cycle. All it does is decide the output voltage required and provide a suitable PWM output to the low pass filter that is basically a variable voltage reference for the comparator.

I did think of a high side current sensing chip but wanted simplicity, I'll use the builtin 0.6V ref of the PIC12F615 and acept a low resolution of ADC readings so lowering the resistance and voltage drop. I may put a high side sensor in later

[Simon]

From the datasheet:

The amount of current which the output device can
sink is limited by the drive available (which is independent of
V+) and the ? of this device. When the maximum current limit
is reached (approximately 16mA), the output transistor will
come out of saturation and the output voltage will rise very
rapidly.

(the datasheet doesn't even guarantee the 16 mA (typ.), but specifies 6 mA as the minimum current the device can sink)

Sounds like you need to current-limit the discharge of C4. Which in turn gives you an RC-circuit with a delay.

well looks like an output transistor will be required, C4 will be small in any case as the values of the resistors in series can be quite high due to the mosfets high gate impedence

[blackdog]

Hi Simon,

I understand your schematic, but... I try to explain what happens when short circuit or overload...
BOEM! boom and then you can go to the hairdresser 

Bram

[Simon]

Hi Simon,

I understand your schematic, but... I try to explain what happens when short circuit or overload...
BOEM! boom and then you can go to the hairdresser 

Bram

Well as I cut my own hair no worries 

Well I could put a fuse in should there worse really happen, or use a small resistor in series and a comparator to trip a forced condition that is safe. As for overload, well it is 5A of solar panel, so an overload that the panel can provide current for is not really an overload. Naturally if i was making it for more current I would be beefing a lot of things up and adding lots of protection

[Neilm]

Simon,

I suspect that you will have shoot through problems with this design. Your design from the FET control assumes that they will switch together. This does not happen in real life. What you will find happening is one FET turning on as the second is still part conducting - classic shoot through. There is nothing to ensure that both FETs will turn off and on at the same voltage.

There are several devices available that will take care of driving the FETs and ensuring that there is no shoot through.

Yours

Neil

[Simon]

what sort of devices am i looking for to act as drivers ?

[Zero999]

(the datasheet doesn't even guarantee the 16 mA (typ.), but specifies 6 mA as the minimum current the device can sink)

To be picky, the datasheet actually specifies the maximum current the device can sink is 6mA, when the output voltage is no greater than 1.5V so you'll almost certainly be able to get more than that, even at the worst case but the output voltage will rise. I agree it isn't wise to assume 16mA though.

[Neilm]

I would do a search for half bridge drivers on RS, Farnell and the like. Some of the results will be for N-type FET drives only, these will include the supplies for driving it.

Neil