Type III Compensation Loop

[bigkahoonaa]

What a headache.  I haven't had much success with the LTC3703 SMPS Buck controller chip.  I'm trying to get a unit that can regulate battery charge voltage to about 13.5 VDC, with wind mill input from about 14 to 100 VDC.  The LTC3703 can handle this large voltage range, but it needs Type III Compensation between feedback (BF) and compensation (COMP) pins:




I've tried the recommended method for calculating resistor and capacitor values from the chip's datasheet.  Anytime I do so, I can get nasty dips and spikes.  What gets me is that the datasheet recommends about a 10K resistor for R1, but every example in the datasheet has 10K for R2.  R1 is in the 100K range.




Am I missing something here, or should I use another method for calculating components in a Type III Compensation Loop?

Mau

[Flux]

Can't help you. You are a brilliant mathematician and clever with simulators but I seriously doubt that you understand what is needed to control a wind turbine.

If you are starting with a constant voltage regulator then you will never do much good.

Whatever method you use the end result must be to obtain a cube law power input curve to the turbine alternator ( power in follows speed cubed). Battery volts is fixed so you must track your charging current to let the input power follow that cube law.

If you can't do it that way then you need to do a mppt tracker. You move the set point a little and if the match improves you try a bit more, if it worsens you try the other way until you run at the peak power point for every wind speed. Nothing you can do with simulators here unless you can master all the variables of a wind turbine prop in the real gusty world with inertia and everything else.

Good luck

Flux

[joestue]

Build the compensation network from the ground up.

C3 and R3 should be tied to ground IMO

Get rid of C2 during testing, put it back in if the circuit needs it.

Just because the manufacture says to use a type III loop doesn't mean you need it.

Surface mounted components may dictate it's use, I don't know.

What is the bandwidth of the controller?

Back in the day when I was building buck regulators with 2n2222 transistors driving IRFZ46 mosfets and a KA7500 out of a PS, the bandwidth of any part of the system wasn't that much higher than the switching frequency, and I didn't have any instability with simpler loops, although the gain was far below what it should have been.

[Lumberjack]

As I understand it,the battery bank will pull the voltage down to its level. Usually once the voltage reaches a certain point the bank is fully charged and all further input is diverted to a different load.

Your best options for detection are frequency (prop speed) and the battery bank voltage for what it is worth.

[Opera House]

You have selected a mode of regulation where the input voltage can be anything and the output is a precise voltage.  I can't think of any case where that would be wanted.  I would be using a two input regulator with one input that does not allow the output voltage to go over a certain voltage and the second that turns on at a fixed input voltage, ramping up the current as the voltage increases.  I assume you want to dump as much current as possible into a battery with the regulator acting as an electronic transmission. Assume you are pairing this with a dump regulator too.

[definitionofis]

I completed my use of LTC3703 to use as a boost convert to boost 9v to 60v to charge my 48v battery. It has been a long struggle with destroyed parts.

I only tested it up to 110 watts, so far. It needs more testing when I get it attached to my wind turbine soon. I did a bunch of videos to document assembly and calibration. Here is one: http://www.youtube.com/watch?v=crSkmrQavXg