Clipping

[kitestrings]

One advantage of working from home more is that I’ve been able to view, and tweak settings for the turbine.  The Classic controllers each have two relay outputs, one logic and one that is PWM.  We use one of the PWM outputs to divert to a preheat tank for dhw.  This works well, and allows us harvest more available power, especially from the solar (PV).

The other PWM output we have connected to a resistive load bank in the crawl-space.  This provides a homemade “clipper”, as some refer to it.  It gives you another electrical diversion for when winds are high, or highly variable, and the batteries are getting closer to their set point.  The load bank is wired on the 3-phase side of the rectifiers (there are two), and the Classic sends pulses based on a user selected V input to the controller via a SS relay.  It works well, and particularly when we’re away from home gives me added peace of mind.  If the sun emerges in the afternoon from the clouds in the morning say, and it is windy…eventually you need a place for the power to go, or a means to shut it down.  In our case, the clipper starts to engage slightly below the setpoint for the furling actuator/shutdown.  This is all somewhat independent of the mechanical furling, right.

So from the start, we first charge the batteries; three level regulation as normal.  When the batteries are not using all the power, we divert opportunity load to water heating.  This is on the DC side.  When both are satisfied, or a combination where the generation outpaces load, the clipper starts to be introduced – this can take about 4 kW of load (more actually, but we don’t run that fast).  If the voltage goes a tad higher, or lingers too long, the furling actuator shuts down the turbine.  Here's what it all looks like:


Yesterday, it was very windy so I was experimenting.  It is interesting to see that you can lower the V set-point and effectively cap the voltage, and hence rpm… to a point.  What I hadn’t thought about too much is this –

When the controller is pulsing to the load bank there is this happy marriage, where the batteries are taking load, and the load bank is also getting tickled more and more to accept the excess.  The settings give you both a threshold (Vin high), but also a pulse width setting.  To illustrate: if the setting is say 105V, and the width is 3V, the relay will start to pulse at 102V, with increasing frequency above that and goes full -on- at 105V.  What the controller “sees” at that point is what I believe I’ve missed.  When the relay is fully on, the voltage out of the rectifier collapses.  The controller sees -0- volts, and disconnects (goes to ‘Resting’).  Usually the furling actuator engages before this, but when I’d intentionally lowered the voltage setting the audible ‘click’ of the controller dropping out was hard to miss.

This may be more apparent to some of you, but then I went for years thinking the word “spatula” had an “r” on the end ;>].

[bigrockcandymountain]

That's pretty neat to hear about.  My 3 phase load bank is pretty high resistance, so with the relay fully on, it still passes voltage to keep the controller charging.  Mine are just 1500w 120v water heating elements wired delta.  It sounds like yours is better suited to controlling the voltage. 

[DamonHD]

This "emergent behaviour" from complex interactions is fascinating.

Even for my *tiny* off-grid system it's taken me about a decade to get most of this stuff (nearly) right, though there are still for example parts that do the right thing for the wrong reason...

Rgds

Damon

[JW]

That's not Delta its WYE

[JW]

That's pretty neat to hear about.  My 3 phase load bank is pretty high resistance, so with the relay fully on, it still passes voltage to keep the controller charging.  Mine are just 1500w 120v water heating elements wired delta.  It sounds like yours is better suited to controlling the voltage.

oops my mistake WYE or star when compared to Delta has less current which would enhance battery bank capacitor characteristics.

[kitestrings]

I realize I did not explain part of this well at all.  It is simple in concept.  I described that,

“When the relay is fully on, the voltage out of the rectifier collapses.  The controller sees -0- volts, and disconnects (goes to ‘Resting’)”.

This isn’t quite what I’d intended.  There is a sharing of turbine output, as I describe earlier, but when the relay completely changes states (full on) the voltage can drop below the cut-in voltage of the controller.  The controller then disconnects and goes to “Resting” (Resting is state the controller is in before reaching cut-in, or in low-light conditions when in PV mode).  The output in Resting is -0- watts.

As Damon questions, how can it take so long to flush out some of these nuances… and they seem so obvious after the fact?  This is all true, but real-world is not always in slow-motion, where you can break things down play by play.  In the situation I’ve described there are several interrelated things at work.  It may start with a big swell of wind, and there is some charge controller and load diversion in play.  When the rotor winds up though, it is also likely to be furling at or near the peak rpm observed.  We engage a variable load, and electro-braking if you will, but the ability of it to provide counter torque is determined by rotor velocity; and the velocity is rapidly changing.  At its apex we introduce quite a bit of load, the rotor may also be stalling hard as the turbine fully furls, and this I believe is when the Vin to the controller drops rather suddenly.

It can also be challenging visually/audibly to follow in real-time.  In our case there are current limits programmed into the controllers that can get deployed (amber LEDs), the pulsing of the PMW output is a pulsing amber light, the trigger of furling actuator is a red LED, and as I mentioned there is a poignant relay ‘click’ when the controllers go from MPPT to Resting (or vice-versa).

Lastly, despite my resistance to admit it, there are things that I forget, or get by me, that might not have been missed by a younger me ;>].  I hope this helps. ~ks

[kitestrings]

JW, I think you've got this now, but yes, ours is parallel pairs of resistors in wye; bigrockcandymtn has his is delta.

[JW]

http://aaronscher.com/Circuit_a_Day/Impedance_matching/series_parallel/series_parallel.html

[bigrockcandymountain]

I realize I did not explain part of this well at all.  It is simple in concept.  I described that,

“When the relay is fully on, the voltage out of the rectifier collapses.  The controller sees -0- volts, and disconnects (goes to ‘Resting’)”.

This isn’t quite what I’d intended.  There is a sharing of turbine output, as I describe earlier, but when the relay completely changes states (full on) the voltage can drop below the cut-in voltage of the controller.  The controller then disconnects and goes to “Resting” (Resting is state the controller is in before reaching cut-in, or in low-light conditions when in PV mode).  The output in Resting is -0- watts.

I kind of guessed you meant watts output instead of volts.  Mine the happy spot is about 90 or 100v dc.  Thats where it will run in a good wind with the load bank fully on.  The 3 phase is seeing a voltage of 64vac to 71vac, which makes sense to me. 

Is your 6 ohms for the load bank individual resistors, or measured across the wye?

[kitestrings]

Right, so that would be about 90-100VDC into the controller, right?

We're running a bit higher than that, I think I have it set ~105-115VDC IIRC.  Let me think... ours are 6.8 ohm wire-wound resistors, pairs in parallel, so R = 3.4 ohms, as in:


In theory something like this, but we don't go anywhere near the upper range of this chart:


Someone will comment of the wire choice, connectors, location, but this is what I built.

[bigrockcandymountain]

Yes correct, 90-100v in to the controller. 

Wow, that looks great.  I don't see any obvious wire choice problems. 

[kitestrings]

A follow up on "emergent behaviour", as Damon described...

We have two Midnite Solar (MS) Classic controllers, each capable of comfortably taking about 2.5 kW from our turbine.  I've been trying to figure out what the h%*! has been going on.  Something changed recently that seemed to be causing the first controller to want to unload - revert to 'Resting' - just as it should be current limiting, and passing excess load to the second one.

I'd forgotten this, but when we first set this up we had to work through a similar error that was being triggered by something called "OCP".  OCP, or over-current protection is something that is set up to help protect the FETs.  Long story, but OCP it seems does not work well with wind mode in this configuration.  We ended up spinning our alternator up, in dynamo-fashion (using our car).


I was then able to film, and share this with MS Tech's years ago.  It took a lot of finesse, but thanks to Ryan (who dabbled with wind, & sometimes posts here as 'Half-crazy') we worked through a means of turning this feature off.  It involves changing two settings in the (Modbus?) programming.  I'm inserting them below in hopes in may assist anyone else that should encounter this event.  It may never surface with one controller; I don't know:




What I'd forgotten, is that when you do a reboot, or firmware update, one of those settings reverts to the default setting, so you have go back in and check them.  A few weeks ago, I had to repower one of the controllers because one menu/screens appeared to have frozen (this may have been related to the DST time change).  I'd totally forgotten to re-check this setting.

Once again I videoed the event last week, so as to slow it down to frame-by-frame pace.  I woke up to this "Ah-huh, well-duh!" moment over the weekend, and we are happily back in the saddle now.

When working properly, it is quite slick.  There are 16-programmable steps in each power curve.  The current limit is set to match the last step of the Classic A.  The second step of Classic B (the first is set lower, but -0- amps, to wake up the unit) starts at the ending voltage for A, so anything in excess of the current limit shifts to Classic B.

My memory it would seem is not improving with age.  I'm adding a(nother) laminated reminder card to the wall by the controllers for future reference.

[kitestrings]

Very happy to have this sorted out now.  The winter winds are back on a regular basis now, and this is when we depend on the turbine.

I've tweak the power curve on this one more time.  Generally, I just lowered the voltage where power starts to transfer to the second controller.  With the current limit set to 40A, it allows enough margin that I can keep the upper limit to about 2.5 kW on this controller.  Here they are through a range of winds (Classic 'A', cut-in is set at 59V, or ~92 rpm & 6 mph):






Classic "B", awakes at 78V, and takes modest output above that point.  It receives any output in excess of the 40A current limit of Classic "A".  The amber light is the current limit indicator.  This occurs at approximately 92V (~140 rpm).  Classic "B" looks like this:

[kitestrings]

A couple more pics - begging Mod's forgiveness - Classic "B" at 98V:


Above, or a bit above 100V, the clipper load starts to engage, helping to limit peak Vin and rpm.  At 115V, the power furling is triggered and shuts things down until things calm down.

The daily logs then look something like this.  On this day I started the turbine at 7 or 8 AM in the morning, and these were the numbers we logged from then until midnight:

[SparWeb]

Someday (could be soon) I just might need your "classic" tales.

[SparWeb]

Yup that day may have come...

If I had a choice between a Classic 150 and an Outback Flexmax 80, to be a MPPT controller for a wind turbine like mine, which would you recommend?

I'm looking at an upgrade to my 10-ft by swapping the generator with a higher output unit and regulating it with a MPPT controller.  I don't believe I would attempt a multi-stage system like yours.  The one MPPT controller (whichever I choose) goes in series between the power rectifier and the battery.  It also makes sense (I think) to keep my existing Tristar as a diversion controller completely independent of what I'm doing with the MPPT controller. 

This is being made possible by a kijiji deal on either an Outback or a Classic (my choice) for just 450 bucks.  Canadian.

[SparWeb]

The choice is moot: The Classic sold, but I could go get the Outback.

Furiously reading the install manual to figure out if this is the right thing.

From the User's Manual:

The FLEXmax Charge Controller is not usable for direct regulation of wind turbine input and OutBack cannot warranty its use in these applications. In wind turbine applications, the FLEXmax is recommended as a diversion controller.

Hmm

[SparWeb]

Here's an emphatic "no":
https://forum.solar-electric.com/discussion/16708/can-i-connect-a-wind-turbine-to-outback-flexmax-80-charge-controller

Oh well, forget it.  I'll wait for the next deal on a Classic before I do this change.

[kitestrings]

Sorry it didn’t work out. I was going to ask if the OB would work with wind -

We’ve been real happy with the MS, and in particular with their features and support. The only other consideration is the model (150, 200, 250). The 150 has the highest amperage capability, but if your turbine may might go above the voltage limit, that might move your choice.

I’ll keep my eyes open for anything in this lane. Best, ~ks

[SparWeb]

Thinking about using the Classic again.
Would I need a clipper? 

Note that Midnite sells a "Clipper" of their own, but I see you've rolled your own.  I already have a diversion load so could I continue to use it as a diversion? 
I may want to build a bigger diversion load (>2kW) for future expansion, so if I had a Classic in my system, would I want to have a clipper configuration, or keep using a separate diversion controller?
Would that even be necessary, if I had a Classic?

I'm not considering anything so large that I'd have to stack two Classics together.

The driving curiosity here is to try using MPPT to regulate wind in such a away that I can measure different power curves on the same WT, with and without.

[bigrockcandymountain]

Yes, in my opinion you need a clipper.  It limits the overvoltage by adding a load on the 3 phase lines before it gets to the charge controller.  When a strong, quick gust hits, it keeps the voltage under control until the mppt can catch up.  It only really works for a few seconds per gust.  Some clippers have a shorting switch, extra voltage monitoring relay, etc.  I think the Midnite clipper also includes the rectifier. 

Fyi, my "clipper" is spread all over the utility room.  It isn't contained in one neat box like midnite.

Mine is a homebuilt. Nothing fancy.  It has a 3 pole ssr that is triggered by the classic  voltage in.  When it goes above about 120vdc it starts diverting power.  I have a dedicated 3 phase dump load consisting of 3 off 1500w 120v water heater elements wired delta in the same big water tank as the main dump load.

I think it would be ok without it in most cases, but it makes me feel better knowing that it is in place.   

I always intended on putting in a latching voltage monitoring relay set to about 150vdc.  If it goes above that, it would short all 3 phases and shut everything down until i went and reset it manually.  That still hasn't been installed. 

I haven't played with the power curve since installing new blades.  I probably should, but realistically won't get there until winter.  You'll have tons of fun tuning the curve and logging results. 

After all your hard work building a data logger and matching your blades to the power curve of your generator, it might feel a bit too easy with the classic.  That's the only drawback i can think of.

[SparWeb]

Thanks, that helps.
I'm planning several changes at the same time to work with this, so you might know what's in my system now, but I'm talking about changing it all.

For fun.

Because I can.

I really should post separately about this...

If i'm working with about 35VDC / 100 RPM then cut-in for a 48V bank happens at about 140-150 RPM, and when turning at 400 RPM then the open-circuit voltage is about 140 VDC.  With the clipper working on the AC side of the rectifier then it sees 40VAC RMS at cut-in and about 100 VAC at 400 RPM.  Above that, the furling tail should be working, but in gusts it could go higher.

In those conditions there's about 20 Amps to be shed if the clipper is working at full potential.  100VAC / 20A = 5 Ohm to give me a starting point for the resistors.  I think I can re-use the 2 Ohm resistors in my current diversion load if they're in star.  Need to check.  Might not have any margin for gusty days, though.

[kitestrings]

Hi Guys,

So, I read your question slightly differently.  If you are asking do you need a MS Clipper, I'd say no.  Both bidrockcandy & I have done it with homemade load banks.  I think you do need one or the other though, as he's suggested.  Ours (and I believe his) operates on the Aux2 output, which is the PWM relay.  We use the "PV V on Hi" option, so you can select the voltage and pulse width.  We switch with a SS 3-phase (75A IIRC) relay just before the rectifier to the resistors.

Looking at your numbers though, I think you will need to have your cut-in quite a bit higher.  With the Classic you need some headroom above the battery voltage for the thing to work with well.  You'll probably want to be up around 58-60V, so maybe 165-170 rpm if my math is right.  At 140VDC then, I think you will see 100 VAC (line) and ~58 VAC (ph), which would be about 29A with 2 ohm resistors in wye, or about 5 kW total; 800+ watts/ea.  Check my math because I may have misunderstood something.  I suspect you may be starting well below this level though.  The Classic is pretty user-friendly so you can fine tune accordingly.

[kitestrings]

I'll just add one other thought here...

brcmtn mentioned, "...When a strong, quick gust hits, it keeps the voltage under control until the mppt can catch up."

It's not so much that it doesn't respond fast enough IMO.  The Classic is first and foremost a charge regulator for the batteries.  Once the batteries have had their fill, so to speak, the Classic reduces the charge levels to follow whatever charge mode you're in, most likely float at the end of sustained blow.  If this reduced charge requirement is less than the available resource, the turbine is unloaded more and more, the voltage climbs, and this can happen very fast.

I think it is true that often the 'clipper' is often just burning off a little bit of the peak power points during gusty conditions.  If you're into you're well into a sustained wind event though, this loading can be the only means of staying within reasonable rpms, unless you shut down.  We shut down with the furling operator if limits are sustained.  Passive furling in theory protects things, but if you're unloaded there can be a lot hard cycling go on.

[bigrockcandymountain]

Right you are kitestrings.  I forgot that aspect because my wind controller is always in bulk, so it never unloads the turbine at all.  The main dump load resistor burns off the excess power after it gets to the batteries.

 It's not the correct way to set it up.  I have it set up that way because my "clipper" dump load is a little soft for keeping the turbine under control all on it's own.

Even so, I have recorded some "higher than they should be" peak voltages on the classic input.  Here is what I blame it on...

My rotor is pretty heavy and spinning pretty fast on a windy day.  It is a pretty big gyroscope and as such has a lot of gyroscopic stability, meaning it takes a larger force to yaw than the same rotor that isn't spinning.

I think you get a hard gust that changes direction and hits the rotor face on, and before it can yaw away from the wind and furl, it spins up and for a few seconds makes more power than you can deal with.  I also have a fairly large rotor and small charging ability, so that contributes as well.

That's just my theory.  Of course on a turbine direct connected to batteries, you would never notice this.  It's only looking at the logs on the classic that i noticed it. 

[SparWeb]

Thank you very much, both of you.

One takeaway lesson that I'm deriving from this is that even if I replicate the systems you have on the ground, I will still see different operation of the turbine.  I believe my folding-tail furling is much more responsive than what you are using.  So far through 10 years of developments I seem to have hit upon sensitive tail furling.  Not a reason to rest on my laurels, but that I've been watching this for years with a digital tacho and rarely does it get into the high 400's.  That's with the 10-ft rotor and 3-HP conversion combination I've been flying up until recently.  I'm working on upgrades so that will make many quantities into variables.  They're all dependent.  I could eat these words after I'm done making a bigger one.

I can only expect about 2kW from the 10-ft rotor, and only in a burst before furling takes over.  Never personally seen more than 1.2kW.  I'd still prefer to not under-rate the diversion load or the clipper, so I won't use anything less than a trio of 1kW resistors.  Probably more so that I can make a bigger prop in the future.

The SSR should be rated several times higher than the expected voltage or current going thru it.  I want no less than 250-300 Volts and 50 Ams.

How does the Classic determine RPM?  How much distortion can it tolerate?
That's probably a question for Midnite to answer...

Here's the open-circuit wave on my o'scope:

[SparWeb]

I'll just add one other thought here...
... Once the batteries have had their fill, so to speak, the Classic reduces the charge levels to follow whatever charge mode you're in, most likely float at the end of sustained blow.  If this reduced charge requirement is less than the available resource, the turbine is unloaded more and more, the voltage climbs, and this can happen very fast.

Thinking some more about this.
What I observe with a WT hard-wired to the battery bank and a diversion load to regulate the voltage is that the batteries are being continuously charged AND discharged during a sustained windy day.  They do get warm.  This is OK and maybe good - considering where the sustained wind comes from. On a cold winter day I've never had any second thoughts about this.  In the summer, heat is a problem and I do give the WT a break in the summer for this reason.

The clipper takes away the excess when the battery is full, and doesn't pass it through the battery bank at all.  So that phenomenon would stop with a clipper.  Mixed feelings about that.  My battery bank is outdoors, and Calgary does get cold.

Tuning the MPPT curve won't stop the Classic from being a battery charge management system; it's designed to be that.  So maybe there's a place for a diversion controller and dump load in the system.  Still speculating, so I could be getting to the point where I'm ready to call Midnite and get their advice.

...if you're unloaded there can be a lot hard cycling go on....

We may be misunderstanding each other, but unloaded at any time for any reason is a red flag.  My claim about a furling system that works is only valid if the wind turbine being furled is electrically loaded.  I accept that the loading of the WT is reduced by the action of the MPPT, but "zero' load is not where I want my WT to be, ever.  I wasn't expecting you to mention an "unloaded" rotor, so it's making me wonder just how drastic the Classic will back off the duty cycle.

[kitestrings]

I think your description is spot on brc.  There are conditions, as you've described, where the turbine just can't move fast enough before the next gust, perhaps shifting direction, tries to wind it up again.  Having the dump load available on a continuous basis like you do is not a bad thing at all.  We don't currently have that capability.  We preheat dhw, and when that is satisfied, we lose that power sink.  I'd often thought it would be good to have an equivalent fan/coil load when the water heater load comes off.

I do like to be able to periodically allow the batteries to charge, absorb and float without the influence of the diversion load, but I find it is generally easier to do this with the PV on good solar days.

I think we would all agree keeping the turbine loaded at all times is best.  In our case we would need a lot more discretionary load for that to happen in all conditions.  When the sun breaks out we have about 3.2 kW of solar added to the mix.  Our normal operating mode goes something like this:

The charge controllers are attempting to first bulk, absorb and then float the batteries.  Once they are fully charged, or whenever we bump against the the upper charge limits set, the voltage starts to climb if there is no place else for power to go.  Initially the load diversion comes into play.  Through one of the Aux 2 relays the Classic diverts, or burns off, excess available power into the water heater using "Waste Not".  Waste not, also known as "Louis Louis" (like the song, for 'use it or lose it') allows you to set a threshold relative to your charge setpoints.
 
When the batteries are full and the water heater is satisfied, the other Aux 2 relay starts to engage our homemade clipper via the 'dump relay' in the rectifier box.  This one, unlike the diversion load, is on the 3-phase side of the rectifier.  It's mission is just to keep the voltage input into the CC's below safe limits (250VDC, 180VAC, and I don't like to be anywhere near this high).  Slightly above the clipper settings (IIRC ~90V, 5-volt width), we have the furling actuator furl the turbine and shut it down (at ~117V I believe).  The strategy is just to save the wear and tear once the batteries are full and we start into these higher-paced conditions.

Regarding some of your questions Spar:

How does the Classic determine rpm?

It doesn't, only indirectly by monitoring voltage in to the Classic (and then attempting to follow your power curve).

"So maybe there's a place for a diversion controller and dump load in the system" (?)

Yes, absolutely I'd say.

"...so it's making me wonder just how drastic the Classic will back off the duty cycle."

I think you will find unless there's a place for the load to go, the controller will remove all but what is required to float the batteries, hence the need for the clipper (and/or load diversion).

[SparWeb]

It's great learning about the ways your systems operate, in such fine detail.  Many thanks as always, all of you.
Still pondering the possibilities - all very interesting and enticing.

[bigrockcandymountain]

I do like to be able to periodically allow the batteries to charge, absorb and float without the influence of the diversion load, but I find it is generally easier to do this with the PV on good solar days.

I'm curious why you say this? The reason i ask is that when I charge with the generator, even though the absorb and float voltages are set lower, the batteries seem more fully charged. (Higher resting voltage etc) Is this what you are seeing by turning off the diversion load and charging?

I think my batteries are somewhat happy with my charging program.  They are 7 years old right now and still working decently.  They are a bit undersized because we were a family of 2 when we bought them and now there are 5 of us.  They use what i would consider "lots" of distilled water.  About 12 gallons a year.

[kitestrings]

Yes, it really is just to avoid unnecessary out-gassing and water consumption.  There are conditions where the local loads, with diversion load, will never allow you to move out of the bulk cycle.  So, you are basically 'boiling' away water all day.  For the most part I don't sweat this too much, FLAs are pretty robust and frequent overcharging, within reason, is probably better than undercharging (or outright neglect).  Just every so often I forego a bit of energy harvest (preheating water) in favor of the long-term health of the batteries, or so goes my logic. 

If I periodically turn off the diversion, with favorable conditions, then the day is usually long enough that we can get out of the constant current mode.  In the absorption stage we're getting at that last ~10-15% in a very controlled fashion without heavy outgassing.

With "Waste not" you can also moderate the delta to your setpoints.  I usually have a lower setting when it is just the PV (~-1.0V), which can be many days on end in the summer; but increase it (~1.5V or higher) on days where we have wind too.  All things equal though, the resting voltage will land a touch higher without diversion in play.