Eureka! I finally get it!

[Titantornado]

I know I asked you folks questions about it several times before, and at the time, I just couldn't get my head around just how it was working. (of course now, when I go back and read those posts, it make perfect sense)  It wasn't until I was reading the Xantrex C40 online manual, and saw the wiring schematic, that it finally hit me.

The best way I can describe it is to think of it like a spillway on a dam.  The lake is your battery, and if the lake starts to get too full, it dumps (diverts) it over the spillway.  Dang!  That's easy.  I don't know why I was having such a hard time picturing that.  I guess I kept thinking it (a controller) was switching back and forth between the batteries and the dump load.

Alright, so now I'm thinking controllers.  After reading many times here about the benefits of stepped diversion methods, what do you folks think about me getting three C40's so I'll have three steps of diversion.  Sure sounds like a better idea than my original plan of going with a single Flexcharge NCHC-24-100 controller.

Dang!  The elation of finally understanding something that has been alluding for quite some time.  :~D

[harrie]

Hi. Yes, I think its a good idea,of course depending on how much current your producing, maybe two 60s would be better? I have just installed one c-40, and is all I need being my input is less than 30 amps. I also added the romote digetal to the C-40, which is nice to keep track of total Amp Hours. Im still learning also, so others may have better answers. Lots of Fun tho. On My other mill, 9 coils 12 mags each rotor, 12 foot blades I have a 60 amp Morningstar controler that has performed very well also. I should have two 60s for that one, cause I have saw well over 100 amps at times, but havent smoked it yet and its been in operation for two years. I try and draw more amps from my 12 volt battery bank during the high winds.

Great Fun, Harrie

[SamoaPower]

"I guess I kept thinking it (a controller) was switching back and forth between the batteries and the dump load."

You were right both the first and the second time. It does act like a spillway (on average) and it DOES switch back and forth(rapidly).

Why do you want to use multiple units?

Why not just use the C-40 just like Xantrex says for diversion?

[Titantornado]

> You were right both the first and the second time. It

> does act like a spillway (on average) and it DOES switch

> back and forth(rapidly).

>

> Why do you want to use multiple units?

When wired as per the Xantrex diagrams for diversion control mode, (or for that matter, any mode) it doesn't switch back and forth, only on and off, rapidly, as you mention.  In fact, (and this is why it finally makes sense to me) the controller is on the "back side" of the batteries.  That is, the turbine is directly connected to the batteries, and the controller is attached to the batteries to chew up any extra power available.  My original thinking was the controller was installed between the turbine and the batteries, and somehow switching between batteries and dump load. (still not sure if that's what they are trying to do with the Flexcharge controller mentioned earlier)

Well, first, I'm under the impression that stepped diversion loads is good practice, and probably a bit more of a fail safe.  Using a single unit arrangement, should their be a failure of the controller or the dump element, you've lost all control of the turbine.  Heck, Amanda is building eight step dump controllers.  And second, I'm expecting (hoping) maximum current from the windmill in the 90 amp range.  Xantrex recommends derating to about 32 amps per controller, so this seems about right for a solution.

I'm always open to suggestions though.

[ghurd]

Why not 3 C35s?  Or 2 C35s and 1 C40?  Cheaper.

And 4 C35s 'could' cost less than 3 C40s.

I like the stepped diversion too.  My thought is first a smaller load, then a larger load, then a larger load.

ie: Step #1=10A, step #2=20A, step #3=25A.

If a C35 is overkill for 10A, why pay extra for a C40?

G-

[SamoaPower]

I guess it's a matter of perspective, Rod. Your description is correct, but from another viewpoint, you can say that a portion of the available current is being switched between the battery and the dump load to maintain the battery voltage at a certain level. Since the C Series controllers are PWM types, this is being done quickly with the amount of diverted current being proportional to the pulse width and the dump load resistence.

The primary purpose of any charge controller is to provide a charge algorithm to maintain the battery. As a point aside, in my opinion, NO available controller does this correctly for variable charge sources as we have. A secondary purpose for a diversion type controller is to provide a continous load for a turbine.

A stepped diversion controller is a rather crude, imperfect imitation of a PWM controller. Other than limited redundancy, as you point out, it offers no advantage over a PWM type. They usually also have single-point failure modes. If you lose a PWM controller or the dump load you don't necessarily lose all control of the turbine because it's still connected directly to the battery. You may end up cooking the battery if left that way however.

I think you will run into problems with the charge algorithm if you attempt to run multiple diversion C Series controllers. To my knowledge, there's no way to synchronize them. The absorption or acceptance charge phase depends on an internal timer which starts when the battery voltage reaches the bulk set point. At the end of an hour, it then switches to the float phase. You wouldn't want multiple controllers to be in different charge phases at the same time.

For higher current capability, the C Series will handle more than the nominal rated current if you fan-cool the heat sink, but not 90 amps. A better solution would be an external high power FET switch driven by the C Series controller. With multiple switches and dump loads you could gain some level of failure mode redundancy.

[Titantornado]

Ahh, I see your viewpoint now.

Cripes, just when I thought I had this all buttoned up, now I'm told that the controllers aren't happy working together.  I can't believe there isn't anything on the market to work with higher amperage requirements, without having to build something yourself.

(yea, I know it sounds silly to cry about building a controller while building a wind turbine, but that didn't involve electronics know-how)

[harrie]

Hi again Rod, The controlers are made to work together, Go to www.morningstarcorp.com, or www.xantrex.com, and pull up their manuels or the tristar 60, or the zantrex c40. the way I understand it, they can be set to different diversion loads that will be activated one at a time, depending on current encreases. I have talked to the Tech. people at morningstar, and that is what they suggested would take care of the problem I have with only the one 60 amp and sometimes over 100 amp from the mill.

Great fun, harrie

[SamoaPower]

Gee harrie, I don't know. I can only speak for the C-40 since I have two. The manual does not address the issue at all. I couldn't find any reference to it on their web site. It simply doesn't make good sense that multiple units could work in parallel, in diversion mode and maintain the correct charge algorithm. In series mode, yes, that works but is no good for wind.

I looked at the Morningstar site and likewise could not find anything indicating they could be run in parallel in diversion mode.

Take your choice Rod.

[Titantornado]

Yea, I went through the c40 manual and couldn't find anything stating they can be paralleled. I can see how the bulk charge timers would get all out of sync if you attempted to do so.  I guess the effect would be a much longer bulk charge time as each one would come into and out of play, extending the total time well over the one hour timeout.  How much longer? I got no idea.

Hmmmmmmmm, so once again I find myself at square one.  {   I think I'm about ready to make a plea to Amanda for an etched board and some components to put together.

[harrie]

Sorry guys, guess I should have read my manuels before I spoke, but I am sending Rod the tech support e-mail I received from Morningstar.

[Titantornado]

Yea, they sure do recommend it, don't they.  It's got me wondering what the charge cycle might look like, so I played around a bit with their charge profile diagram, and guessed at what it might look like.

So, first, this would assume all three units have reached their bulk setpoints.  The dashed lines would represent the bulk and float voltages of the first and second units.  Obviously, the voltage would be that of the third unit, and I would guess the amperage of units one and two would fall off sharply (and dumping to diversion load) as voltage would continue to climb, as represented in the graph above.  Because of the timer circuit, you can see how this would cause the absorption time to become extended.  How long?  Well, it would depend on how long it would be between unit one reaching it's bulk setpoint, and unit three reaching it's setpoint.

I know that's a bit hacky, but it looks like it should work, not?  Might require a bit more battery tending as far as watching your water levels.  Of course, this all assumes that the second and third controllers ever get their chance at operating, as depending on the bank size, they may rarely have to come on.  Also, I wonder how the charge cycle, (especially the timer) is affected by the constantly changing output of a wind turbine.  What happens if the bulk setpoint is reached, and the timer activated, then the wind slows and drops the voltage back down?  Does the timer(s) reset?

[Titantornado]

>  And 4 C35s 'could' cost less than 3 C40s

I couldn't find a deal that good, though, I would like to go that route with a fourth unit.  Right now, I can get three C40's for $360 shipped.  Best price found for the C35 has been $95 each plus shipping.  I think the C40 is significantly more robust anyhow.

[ghurd]

I checked quick when I posted. I think C35's were $99, C40s were $140 at the only place I looked.

"significantly more robust".  I think so too.  Honestly never used either one.

G-

[harrie]

Rod, I beleive they said do not parallel The units?

[SamoaPower]

Let's look at this in a little more detail Rod.

Say you have 30 amp dump loads on each C-40 and that the batteries will accept 10 amps at 100% state-of charge (SOC). Now, say we have conditions where the source availability is <40 amps. As the battery SOC increases, the terminal voltage increases to the set point of #1 where it starts diverting current to the dump load and the internal timer starts. (No, I don't think the timer gets reset when the voltage drops.) The timer times out and #1 switches to float phase and starts dumping more current to effect a decrease in battery voltage and current. So far #2 and #3 haven't done anything and the battery is at 100% (maybe).

Now, say our source current increases. #1 is diverting all it can, limited by the dump load resistence, so the battery voltage again climbs which is now into the overcharge region. At #2 set point, #2 starts diverting current and its timer starts. #1 still thinks it's in float phase so it continues to divert maximum current. It will stay in this condition until #2 times out and switches to float phase. This could represent a substantial overcharge.

I think you can extrapolate what happens with a further increase in source current and #3 kicks in. In effect, a good part of the whole charge algorithm has been defeated, partially negating the use of a controller at all. I've said essentially the same thing you did, but I don't think your modified curves represent what actually happens.

"Also, I wonder how the charge cycle, (especially the timer) is affected by the constantly changing output of a wind turbine.  What happens if the bulk setpoint is reached, and the timer activated, then the wind slows and drops the voltage back down?  Does the timer(s) reset?"

Exactly! This is why I say that NO available controller, to my knowledge, does it right. I think I know how to do it correctly, but it's number 37 on the 'to do' list. In the mean time, I monitor a lot.

[Titantornado]

> Now, say our source current increases. #1 is diverting

> all it can, limited by the dump load resistence, so the

> battery voltage again climbs which is now into the

> overcharge region. At #2 set point, #2 starts diverting

> current and its timer starts. #1 still thinks it's in

> float phase so it continues to divert maximum current.

> It will stay in this condition until #2 times out and

> switches to float phase. This could represent a

> substantial overcharge

Yea, that's a concern of mine too, and I thought about triggering #1 and #2 at slightly lower voltage than normal, (#1 @ 0.4v lower, #2 @ 0.2 lower, #3 normal voltage) but if your charging source never puts out enough to get to #3, the batteries won't get their appropriate charge.

Ugh!

[Titantornado]

Well, it does say not to parallel when used as a load controller, it says nothing about that when used as a charge controller. (not that it means you can/should)