John,

It is complicated to try to utilize variable pitch blade technology in these small size wind turbines. They use it on the big ones, but I only know of 1 small wind turbine design currently sold that uses centrifugal weights to vary pitch mechanically on a smaller wind turbine. This wind turbine is fairly expensive.

http://www.magnet4sale.com/Talon5-Variable-Pitch-5KW.html

It would be nice to have a computer controlled variable pitch small wind turbine, but it would probably be very cost prohibitive.

Edward

 

Hello again,

I ordered 5 new MOSFETs for this wind turbine controller and also a soldering kit. I was going to practice some on a scrap circuit board (desoldering and soldering) before working on the actual board. I've done a little soldering, but not a lot.

I've been busy with other things and the winds have been calm lately, so this project got put on the sideline for a while. Today, we had some strong winds and I did some testing on the wind turbine controller.

I discovered that the controller limits power output of the wind turbine to 40A. Right at 40A, the controller begins pulsing the MOSFETs to the 4kW dump load and that loads down the wind turbine and keeps the turbine RPM slower. I was able to determine that the controller is limiting turbine speed to around 550 RPM. That is too low for this particular wind turbine. This wind turbine is safe to around 800 RPM. I'm loosing a lot of power with this artificically low limit.

With another controller, I've seen peaks of 65A from this same wind turbine, but the reason that I am not using that controller now is because it has relays and I didn't think the design was as reliable. This controller that I am testing now is solid state and has MOSFETs. It was designed for a different wind turbine which needed to have turbine speed limited to a lower RPM.

I thought perhaps the frequency of the 3 phase AC output of the wind turbine was being monitored. I thought maybe at a certain frequency that relates to a certain wind turbine RPM, the circuit board would open the MOSFET gates to dump power to the 4kW load resistors. This in turn would load down the turbine and slow it down. However, I do not see any wires coming off of the AC side of the 3 phase bridge rectifier. The only wires going to the circuit board are coming off the DC side of the bridge rectifier. As a result, I don't think my theory of the circuit monitoring incoming 3 phase AC frequency would be valid.

All that to say, I am wondering if there was a way for me to change something on the circuit board to allow it to reach higher power output (say 60A)?

Would that current limit of 40A be programmed into the IC on the board, or is it something that I could change by changing a component?

I noticed there was a 16.0000 Mhz oscillator crystal on the board, and wondered if that might have something to do with the RPM limit?

I've got pictures of the circuit board earlier in this thread, but I'll post one here highlighting the oscillator crystal.

Thanks for any feedback or ideas you might have!

Edward

 

However, I do not see any wires coming off of the AC side of the 3 phase bridge rectifier. The only wires going to the circuit board are coming off the DC side of the bridge rectifier. As a result, I don't think my theory of the circuit monitoring incoming 3 phase AC frequency would be valid.

Don't give up that easily. There would still be a ripple on the DC side, plenty to determine rpm from. So we can't rule that out.

All that to say, I am wondering if there was a way for me to change something on the circuit board to allow it to reach higher power output (say 60A)? Would that current limit of 40A be programmed into the IC on the board, or is it something that I could change by changing a component?

Tough to say. If it's not measuring rpm (and I suspect it is), then it could be measuring current with a hall effect sensor. Can you see anything that might be wrapped around the main current bus, that could be giving a signal back to the board? Another option is measuring voltage across a shunt. If that were the case, you could just replace the shunt resistor with a smaller ohms one, so that more current would flow at any voltage measured across the shunt. That would fool it.

Couldn't this thing be triggering simply on voltage, which is proportional to rpm? So if you load it down it doesn't need to send amps to the resistors?

 

Thanks for the good insights Wayne.

There is no hall effect sensor that I can see on the board. There is no current shunt either. There is a digital ammeter on this controller, and it's getting it's signal from the circuit board. No shunt. Looks like the circuit calculates the current somehow. Today, I wired in an analog DC ammeter with internal shunt to just see how accurate the digital ammeter really is. Turns out that the analog and digital readout match almost exactly.

There is a signal wire with "C+" marked on the board that goes to the DC side of the bridge rectifier. I wonder if that is where the circuit is measuring a ripple that is used to monitor frequency & RPM.

Then again, when I observe how the controller works, very precisely at 40.0A, the controller kicks into braking mode to try to slow down the turbine. Makes me wonder if somehow the circuit checks to see what the calculated current is (which seems to be accurate with acutal current output). When it hits 40.0A, then perhaps the IC is programmed to trigger the gates on the MOSFETs to slow the turbine.

I don't think it's determining RPM based on voltage, because the voltage is clamped to the 48V battery bank voltage. I use the wind power to heat hot water when the battery bank is full. Even if I am pulling off more than 40A from the battery bank and the battery bank voltage is not too high, the controller still goes into braking mode and sends power to the 4kW load resistors. So, I can see that they have it programmed to slow down the wind turbine at that 40.0A output (~ 550 RPM) in all situations. I would like to see if I can figure out how to change that limit to 60.0A (~ 650RPM). That last 100 RPM makes a HUGE difference in power output and gives me an extra 1000W of power. I'd hate to lose that. Especially since the factory says the turbine is safe to 800 RPM, so I'd still have plenty of safety margin by limited speed to around 650 RPM which puts out around 60A (3000W).

 

Good detective work. I agree it sounds like it knows what the current is. Someone that knows how that might be working will have to chime in as to whether there is some hope of manipulating it.

I know in my work I'd be able to deduce the load by how much the voltage drops between the ripple peaks. A big load will make a filter cap essentially useless, draining it fully on each cycle, whereas the cap will hold voltage at the peak if there is little or no load. Also, once you've characterized an alternator by charting out power-vs-load-at-rpm curves, you can estimate current based on the rpm and DC-side voltage. You need a computer to do that, but it appears there is a controller on your board. If I'm right, then something that alters what the controller sees for the output DC voltage would allow it to keep raising current until the voltage gets to the programmed level, making it think it's at the 40A limit.

I wonder if you couldn't LET it channel current to the resistors, but somehow replace those resistors with more batteries.

 

Thanks again Wayne. I'll see if anyone else can think of a way that I might be able to modify this existing circuit board to alter the power limit.

True, that the power from the 4kW resistors could be used to do something more useful. Interesting idea about using it to charge batteries. Right now, the energy sent to those 4kW heating resistors is not totally wasted as it produces heat that is put into my house. My system is installed in my utility room inside the house, so the heat is beneficial in the winter. In the summer, it's a different story! I'd be adding heat which the A/C would have to work harder to remove.

If I could figure out how to increase the power limit on the circuit (from 40A to 60A), then the MOSFETs would be dumping much less power to the 4kW resistors.

 

The more I think about this, the more I suspect that the 28pin IC on this circuit board must monitor the current output of the wind turbine. I'd suspect that the IC is programmed to begin electromagnetic braking at 40.0A. This 40.0A current is precisely when I see electromagnetic braking beginning. If that's the case, then the IC would need to be reprogrammed, which is something that I would not be able to tackle. At first, though perhaps I could just change the oscillator crystal to a different frequency, but I am just guessing as I am not sure what the purpose of the oscillator crystal is in the first place. I admit that I am not an electronics expert.

If some of you experienced electronics experts would look at the pictures of the cirucit in this thread and let me know if you think it's the programming of the IC that would need to be changed, then I would really appreciate it!

Thanks,

Edward

 

From looking at the board, it would appear that the uC controls the current. It would be impossible to modify it without destroying the purpose of the board.

So I guess there are two things different between this model and the next one up. 2 Mosfets, and a different program on the uC!

Perhaps you can call the manufacturer and order a replacement uC for the higher power model? Probably not, but maybe worth a shot.

 

If it is measuring/sensing current, changing the crystal probably won't do anything. That function is probably based on an A/D conversion. If you could find the reference voltage for that conversion, then you might be able to change it.

If it is frequency, then I would increase the crystal frequency a little (say to 20 MHz) and see what happens. That chip looks like it has had its markings ground off, so you can't find its frequency specifications. If you go a little higher in freq., it will probably still work. But if you go a lot higher, it may not continue to work reliably.

John

 

If it is measuring/sensing current, changing the crystal probably won't do anything. That function is probably based on an A/D conversion. If you could find the reference voltage for that conversion, then you might be able to change it.

That part I can't really tell. It appears to be a multilayer (4 or more) board, as there aren't many lines leading to the uC, mostly the high current traces on the outer layers. So I have no idea what pins are connected to other spots.

It is possible that the uC is measuring the drop across the green resistors on the bottom side of the board, but that is just a guess. I do not see anything else that would be a "sense resistor" other than the traces themselves.

 

I tried the numbers that I could make out on those green tubular devices on the back side and came up blank (literally -- no hits). They could be current sense resistors. Also, note the relatively small gauge wire from the main rectifiers. I wonder if those are providing some sort of current sense (e.g., a voltage drop across a rectifier) or frequency to the control board?

So, is it frequency or current? (Thinking out loud)

Do you have a crystal either higher or lower than 16 MHz, but close? I would be tempted to try it and see what happens. Cut-off changes = frequency. Cut-off doesn't change = current. Since the current cut-off seems quite sharp, I would make any frequency change small if going higher to reduce the chance that other factors that might affect the result. On the low end, there probably is no such a limit.

But then, I am an experimentalist.

John

 

Hmmm... Tinkering with the clock is good idea. Instead of replacing it could you somehow change it enough to detect? Load it, chill it, something like that.

OP, wouldn't this thing start dumping even if there were little or no load? Doesn't that mean it triggers on rpm?

 

I guess the exact answer would be if it stops when exactly 40.00A is on the display. That would mean it is measuring a current sense resistor.

If it is stopping "around 40A", it could be measuring frequency, and for all we know, it is measuring both RPM and current. A LOT of code and functionality can be packed into even an 8 pin microcontroller with A/D ports.

Assuming it is capable of acting on Current, Voltage at 2 points, and RPM isn't out of the realm of possibilities.

 

Thanks guys for kicking this around. Yesterday when we had high winds, I was paying close attention to the controller. It was dumping power to the 4kW resistors at 40.0A consistently. It's probably not coincidental that the 40.0A dump point is so consistent. I am getting the impression that this is probably limited by current. Perhaps they also incorporate RPM.

I just thought of another clue in all this. Previously, I used a different 2kW wind turbine with this same controller. It had a larger generating unit that was designed to spin at lower RPM. It's been a while since I was running that wind turbine, but I'm almost certain that the electromagnetic braking was kicking in at the same current of 40.0A.

Voltage is another parameter that causes this controller to dump power. When battery bank voltage reaches a preset limit (batteries are fully charged), then the MOSFETs also dump power to the 4kW heaters/resistors to prevent the batteries from being overcharged.

So, I'm pretty sure that current is what is being monitored. I know that voltage is monitored for sure. Not sure about RPM though. I'm suspecting NO on the RPM, IF (big IF) my memory serves me right and the other wind turbine (which ran at a different RPM/frequency) was dumping at the same 40.0A current limit. I'm just not 100% sure that it was the same 40.0A current limit. Back then, I didn't pay as close attention to the way the controller was behaving.

I contacted the company where I got the controller from and inquired about buying just the 3kW circuit board. They did not reply. I suspect that they would not do this. They did say that they would sell me a complete 3kW controller for $438. More than I'd care to spend. I would consider it if I knew for 100% sure that it would dump based on power output (current) and not RPM. If it dumps based on RPM, then I'll still be stuck with the problem of power being dumped too soon before my turbine reaches it's optimal operating range.

Wish I could modify this existing controller but might not be feasible. It's one thing where it's hard to safely experiment, because once the wind is roaring and the wind turbine is running at peak power and RPM, then that is not the time to be making any mistakes! I've had a "run-away" wind turbine before that was overspeeding out of control, and it was a scary situation!

Thanks for the great help.

Edward

PS. jpanhalt, I don't have any other crystals to experiment with, but I'm more and more thinking that this circuit does not monitor frequency (RPM). I could be wrong, though. So much of this is way over my head in terms of electronics. I'm learning a lot in the process.

 

I agree with your assessment, particularly in light of what you found with the slower turbine. Any way you can get a look inside and pictures of the higher current unit?

John

 

One thing you might do is probe the pins of the processor when the turbine is running at various states of power production. If there's a current sensor, it must be delivering a variable input to the processor, and you should be able to find that. Having located it, you should be able to trace the circuit that feeds it, back to whatever the sense element is. On the other hand, if there's no analog input that measures current, all you'd find would be a pulse input which measures rotational speed.

With large currents it's quite easy to set up a "shunt" which measures voltage drop, just by using a length of cable of known resistance per unit length. You'd be able to tell if something like this is going on by looking for small wires attached at different positions to a larger wire.

 

The is what appears to be an LM358 (U4?) on the board. That could be the input buffer for an A/D done in the MCU. Might be interesting to trace its connections.

John

 

Thanks for the good ideas.

John, I don't think I'll be able to get a picture of the 3kW controller. The guy I'm dealing with is in the US, but the controller would be special ordered from a factory in China. I asked for the factory's information, so that I could contact them myself, but I was not given the info. When I inquired about ordering just the circuit board or asked some detailed questions, he just ignored and didn't reply to those e-mail questions. The US person only offered to order me an entire 3kW controller at a high price. I think it boils down to the fact that he is only motivated if it will be profitable enough for him to help.

John P, To probe the chip, would I use a multi-meter set to voltage? Where would the positive probe go? What about the negative probe?

I can imagine I could use the positive probe to check the pins on the chip, but then I'm not sure where the negative probe would go.

Thanks again,

Edward