To Haxby: Thanks and more PC PSU questions

[domwild]

Haxby,

Thanks for your July 2008 answers re stacked PSUs and windmill MPPT. Sorry for not thanking you before.

Q1.: Stacked PSUs: Pls elaborate, you mentioned if one uses stacked PSUs, then it needs a small modification to do it safely.

Q2.: "PSU takes DC inputs from 90 to 350 V, check label". The label would mention the AC input, not DC, how do i stop the magic smoke escaping? Am amazed at the range of DC inputs you quote.

Q3.: May I ask for your critique of the following MPPT approximation similar to yours:

step=?
pulsewidth = ?
oldwatts=0
main:

read volts
read amps
calc watts
if watts >=oldwatts increase pulse width by step?
otherwise decrease pulse width by step?
oldwatts = watts

goto main

and let the Picaxe hunt to find a maximum wattage into battery from the windmill.

Q4.: What happens to the wattage produced when the Picaxe-driven (PWM) MOSFET has disconnected the batteries (load) from the PSUs, does it get lost? Do the small caps in the PSU store this wattage?

Q5.: The PSUs would be considered to be a load and would need to be disconnected from the mill until a certain RPM is reached otherwise the mill would never start. Is this correct?

So many questions, so little time!

An aside: As you are familiar with the Fisher & Paykel motor/generators, you may even be a Kiwi or Aussie, a fellow South-East Asian like myself?!?

Thanks,

 

[BeanieBots]

Please post a link to the original.

Q4. An MPPT controller with no load will keep ramping up the voltage until something breaks. This must be taken into account within your control loop.

Q5. The load MUST be removed when power output is below a certain limit or the mill will stall and not re-start. Similar issue to "Q4" except the other end.

 

[domwild]

Excuse my ignorance! Found my original article by searching this forum for "MPPT". The article is called "First-order low-pass filter". Now that I have found the article and for future reference, how do I give the forum the link?

Thanks.

 

[hax]

Hi Domwild, and thanks for the thanks.

Your answers to your questions: Note, do this at your own risk. DC is argueably more dangerous than AC.

1. Lets say you are stacking 3 power supplies. One of them is going to provide your negative line, one your positive line, and there will be one in between.
If you just connected them in series as you would a battery, you would see that they all short circuit out. That is because the negative of each supply is connected to chassis earth. And each one is shorting out the next.
To get around this, you need to open the power supplies and disconnect the point at which the negative rail connects to earth. But for safety, leave the first power supply (the one you get your negative from) alone. Otherwise all three powersupplies may float at dangerous voltages.

I won't go into too much detail here as it is not appropriate for the school kid audience, but you can google for multiple PSU power supply hacks to see how to do it safely.

2. Various power supplies work in different ways. But typically, the first thing you see in a computer power supply is a bridge rectifier. This converts AC to DC, and is generally the first component connected to the incoming mains line.

So for a 240V PSU, if you apply DC at around 300V instead of AC the power supply will happily work fine. I should mention that some power supplies have a whole lot of smarts monitoring the power line AC, and these may or may not work on DC.

3. I dont think your code will work unless you tune it to your particular mill. You can always increase the watts output no matter what the wind speed, untill the blade stalls. Then when the blade starts spinning again, the system will stall it again. My idea was not to measure watts out but to measure the wind speed with separate anemometer and to measure the mill RPM. The picaxe would use a lookup table and turn the load on or off depending on whether the RPM of the mill matched the available wind speed. This way the load would regulate itself and maximum watts would always be tracked. It would also work if you had a non linear TSR and a heap of other variables such as line resistance, battery level, battery impedance, mill impedance etc etc.

I think too many people are concentrating on generating the most watts without thinking about where the watts come from. The idea is not MPPT in the electrical sense, it is more about determining how much energy can be taken out of the wind at any particular point in time, and varying the load accordingly. All the circuit needs to know is the RPM of the mill and the current wind speed.

4. In my example when the mosfet turns off, it only turns off because the RPM of the mill is lower than the wind speed. The aim is to turn off the load so that the mill can increase in RPM. The energy is not lost, it is stored in the momentum of the blades, and can be collected again when the wind dies down. When the wind dies down, the mosfet stays on till the TSR is equal to the now lower wind speed.

5. I would keep the PSUs connected to the mill at all times. If the charge mosfet is off and the batteries are not charging, then the PSUs would not be too much of a load on the mill anyway.

Yes I am very familiar with the F&P motors. Many people are re-wiring these motors so that they can connect them directly to the batteries they are charging but I think that using the high voltage DC from these motors is a better way of doing it if you are competent and know the risks.

Using high voltage DC allows you to run the cables far further than your average 24 or 36V RE system. And you dont need to use big chunky diodes and switching gear either.

Another idea I had was to use a voltage doubler circuit on the low end. But all this would depend on what wind you want to capture, and this depends on your site.

I only wish I had a house in a good wind area. I would be up to all sorts of trouble, ehem I mean experiments. Unfortunately I am moving to an apartment so all my wind ideas will remain untested at least for the next 4 years or so.

Have a look at the LG motors coming out. They look more sturdy than the F&P's.

 

[domwild]

Haxby,

Many thanks for the reply.

Interesting comment re LG motors: The papers commented on a law suit by Fisher & Paykel vs. LG for some sort of a violation, possibly patent violation. However, in Australia we find the streets full of old F&P washers during kerbside pickup days but the LG washers are probably too new to be seen to be thrown away.

Your idea of a lookup table is similar to the one I saw in the precis to a Brazilian Masters thesis, where the measured watts are compared to watts expected from a lookup table and the error signal then drives the PWM. But yours is easier to implement.

The link for other contributors is given below (hope, hope):

http://www.picaxeforum.co.uk/showthread.php?t=9901&highlight=MPPT

Regards,