Classic running warm, now dead

[Westbranch]

I see what you are doing , an interesting idea.  What uses 120V DC?

My panels are 140W 12v panels, not 24 so I will have 4S array configuration for 48V.
In winter time I expect to see upwards of 90V as we can see overnight temps dip to ~ - 40*C
The CC is in the basement so a bit of heat then is a good thing... when it is that cold outside

[offgridQLD]

I see what you are doing , an interesting idea.  What uses 120V DC?

His battery bank is 48v . His talking about his charge controller input from the PV array is 120v at times (3 series panels of around 40v at times)

I'm starting to have the same thoughts about reconfiguring one of my arrays back to 2s from (3s now) just to give the classic a easier life I I to have noticed it runs way hotter at 3s than 2s. Keep the input voltage closer to the battery voltage (yes more loss in the wires but I prefer that to loss in the charge controller = heat locked in the little aluminum box )

Kurt

[wrt]

Actually Kurt, in this circumstance I am not sure you would have more loss in the wire with 2s.

Usually having higher volts is considered advantageous because you have less amps to carry.

In this case you are reducing the panel count in the string, therefore amps remain the same while volts decrease. The remaining panels are distributed over more strings so you have a higher copper count to send the remaining amps and volts on. Probably not too big a loss at all.

The more strings part is the main reason I am reluctant to reconfigure unless it is clear that the classic needs it to live well in our climate. In my case I have to run 5 extra strings of cable + conduit + fuses +enclosure, bus bars, glands, MC4s, assorted racking and time. It all adds up quickly.

[vtmaps]

The more strings part is the main reason I am reluctant to reconfigure unless it is clear that the classic needs it to live well in our climate. In my case I have to run 5 extra strings of cable + conduit + fuses +enclosure, bus bars, glands, MC4s, assorted racking and time. It all adds up quickly.

It sounds like each of your strings needs a long cable to the combiner... usually the combiner is at the array and a single cable makes the home run to the controller.   If this is the case, the home run cable will have higher losses when you reconfigure to more (lower voltage) strings.  On the other hand, the increased cable losses MAY be more than compensated by the increased efficiency of the controller. 

Every case is different and the optimal solution depends on knowing the distance, the cable size and length, the panel specs and their configuration.   When you design a system, it is usual to use the full power rating of the panels to calculate the voltage drop and power loss in the cable.  The power loss in the cable goes as the square of the current, and the current is seldom at the rated specification.   In other words, with NOCT conditions the power loss in the cable will be much less than calculated at STC conditions.  On the other hand, reducing cable loss by making fewer (higher voltage) strings will cause the controller to be less efficient at all power levels.  Here is an example with some numbers calculated: 
http://midniteforum.com/index.php?topic=2008.msg18796#msg18796

wrt,
I'm not really sure how your 5.7 kw array is configured... why do you need so much more conduit, cable, racking, etc to reconfigure into more strings?

--vtMaps

[wrt]

My combiner is on the switchboard, each set of three panels has its own string back to its own fuse. At this point I have 10 separate strings, all are contained in conduit.

Running multiple combiners on the roof may seem like a better solution, and probably is in most cases where all panels are in the same area. However in my case the savings are not huge as my array is spread over the 15 meter width of the shed.

By the time multiple waterproof enclosures with the same amount of fuses in them are placed on the roof and wired back to the panels, Plus the heavier cable to run the higher amps from them back to the switchboard are accounted for I doubt there would be a huge costing difference.

As for the racking this again comes down to configuration, each set of 3 panels is on its own set of tilts to take advantage of winter sun, except for 12 which lay flat on the roof, these sets of tilts are equally distributed across the four different facets of the shed roof (the style of roof is called an American barn) this helps to spread generation across the day.

Because of this I only ever see a peak of around 4kwh out of the array, however it produces this for a long time.

I am probably making it sound more complicated than what it is but hopefully it clears things a little 

[Vic]

The string configuration here is essentially identical to wrt's -- strings of three real 24 V PVs STC Vmp of 106-ish V and Voc about 132 V.

Have run this off-grid system  for the last 9 + years.   First with the MX-60 CC,  and more recently,  using MN Classic 150s.

When in the later part of Absorb,  and in Float,  with light loads,  the Vin to the CC  often runs 120 Vdc  or more,   as PV loading is reduced,   the string voltage approaches Voc.

It was initially surprising how much less efficient the CC is  delivering 200 Watts output with 122 Vin,  vs delivering,  say,  1500 W  with 100 Volts input to the CC.   This efficiency reduction is certainly not linear -- there is some exponent involved.

Even the venerable,  efficient MX-60  would have its tongue hanging our,  frequently running its fan (200 W out,  122-ish Vin),  verses essentially never running its fan at 1500-ish W out and 100 Vin.

Did choose this string config,  because it laid out well -- strings of two wasted roof space,  racks of four PVs were mechanically weak,  and a wind concern,  and,  trying to rack three PVs and wire them as two per string was a real inefficient mess.   Also wanted the ability to try heating water directly from the DC PV,  rather than AC,  and  106 V was better for this than 71 V would be.

My "solution"  to the hot CC issue,  was to Air Condition (A/C)  the Cargo Container Power Room,  with a  6,000 BTU window A/C.   This works well,  and the MX stopped running its fan,   ever,   And the Classics are happier,  as well.   But in a hot climate with these high input voltages,  without A/C,  would bet that most any CC would be very thermally taxed.

The other main reason for A/C in the power room,  is to try to keep the batteries as cool as possible,  and this has also worked out well.

Just understand,  that   STC string voltages and losses in the CC,  increase by a surprising amount,  as the loading on the PVs decrease during Absorb and Float.   Working to keep the PVs loaded with heating/cooling and other opportunity loads  will help reduce the Vin,  and CC heating.

As we all know,  heat is not good for most electronics.

This is stating the obvious,  but just a reminder.   FWIW.    Good Luck wrt!   Vic

[zoneblue]

Echoing what Vic said, here even running strings of two 72 cell modules the classic does get hot in float. I have no idea why but it does. The batterys are charged, very light loads on the system, and yet the fan will be running. ITs counter intuitve, but vics explanation of the high input voltage is about as close of an explanation as ive heard.

[wrt]

So it is clear from the feedback received on this thread that 3s is a known problem for generating excessive heat within the controller.

As part of the solution to this problem and continual product development I understand it would not be viable to develop this out of the product, perhaps, as I had posted earlier, better education for users would be an economic solution?

Perhaps adapt the Midnite sizing tool to show a clear warning to avoid this configuration where ambient temp regularly exceeds 30c. Maybe a sticker on the display WARNING: GREATER THAN TWO PANELS IN SERIES MAY REDUCE PRODUCTS LIFESPAN IN HOT CLIMATES. Maybe warnings on the front page of installation manual. Bury it I the manual and the average installer will never see it 

It must be said here that I sought advice from two separate professionals when I reconfigured and expanded the system last. Both advised to change from the 2s that I was running to the 3s that I now have now. Both could not understand why I would have only 2s when 3s "was so much more efficient" and result in less cable runs.

Edit: As a curiosity, the sizing tool tells me I am fine to step 120v all the way down to 24v as long as I stick to 4 strings 

[RossW]

This is a disturbing thread.

I replaced 1000AH of AGM cells early last year with 16 x 300AH LFP (LiFePO4) cells. So far, everything has been very good.
I'm saving to put a second set of LFP....

I have six solar arrays, individually metered. They are geographically diverse - 4 on 2-axis trackers, 2 on single-axis, but all spread across about 50 metres (160'). Each array runs inside on its own feed cable and into the metering, protection and combining (via diode isolation). Arrays are 6 x 100W/12V (nominal, about 17Vmp), for a nominal voltage of 100V.

I have a Flexmax FM80, and a Midnite 150 (I can switch between one, the other, or a mix of both).

Summer temperatures are often 45C here, although my battery room is underground and ambient temperature rarely gets over about 30C.

As I've never seen, (or until now even heard of) failures going overvolt, I chose not to introduce another point of failure with high-voltage disconnect. I'm now worrying about the potential implications of that decision!

I have mitigated it SLIGHTLY with my battery cell monitoring system, which will alarm if any cell gets more than about 50mV above or below the string average, and/or if the total bank voltage gets above or below my preferred limits, but if I am away from home I'm now getting concerned!

(Graph looks a little furry, this is still a "work in progress" and I haven't calibrated the boards carefully, they're "near enough" but not "exactly right")


DC-rated Normally-closed breakers in the current/voltage range I want are hard to come by (and expensive).
I really don't want to blow the extra watts on an "always on" contactor - not to mention, the reliability issue.
I considered a sacraficial breaker across the input to the classic as a crowbar, but the breakers to the PV won't be tripped by it, because the PV won't make enough current to trip them! Blowing the 100A DC fuse on the input to the classic if I detect overvoltage is an expensive (and scary) step.
Replacing the breakers with solenoid-operated switches? A solenoid to turn it "on" (pulse only), and another to turn it off? Like a mechanically-latching relay? Sounds expensive, large and of questionable reliability. (Granted, it wouldn't get operated more than a few times a year, if that)

Anyone got any other thoughts??

[wrt]

There are a few point here that need to be looked at. As you stated before you have never seen or heard of overvolt failure. Midnite have assured us that this is an extremely rare way for this CC to fail (0.1% of FET failures result in this)

Your power room rarely gets above 30c. mine has seen at least 10 days this year at 35c and 4~5 at 38c

I think even though it is introducing another point of failure. Automatic hi and low volt disconnects are necessary to protect from the rare things in life (says the man with the bolted horse  )

Maybe if it is economical reduce the input volts as the others have been suggesting. Although I am waiting for Bob to read my post on the last page to find out whether these higher operating temps really matter.

It sounds like you are about to have around 15k worth of battery, a few hundred extra for control and latching contactors would be good insurance.

[zoneblue]

So it is clear from the feedback received on this thread that 3s is a known problem for generating excessive heat within the controller.

Well its forum wisdom that the greater the voltage dropped by the controller, the less efficient the controller operates. You probably know that even a couple percent differnce, when you are running 4kW ammounts to a lot of heat. This applies to any brand of mppt controller. Outback publishes effciency curves for the flexmax, and these are a good guide to horse trading pv cable losses against controller losses. If midnite could publish simialr curves for the classic that would a good help in this regard, however given the two products common heritiage we generally assume similar curves for the classic. ie approx 2% extra loss for each step up in string voltage.

[wrt]

Midnite do publish a power graph, which does suggest inefficiency as the input volts rise feeding a 48v bank. However it does not make the same point at lower voltages (the step down theory) the difference between 70v and 120v stepped down to 48v is around 11% loss.

When you look at the difference between 70v and 120v all the way down to 12v the loss is only around 5% you would think the efficiency losses would be massive from 120v to 12v.

Given the heat generated by stepping down from 120v to 48v you would think stepping from 120v to 12v would be unsurvivable in the coldest climate.

It needs to be considered that not everyone possesses forum wisdom untill they have a problem and start googling the problem or join the applicable forum (in this case Midnite solar) and in my case I was actually advised by my local Solar guy that seeing that I was mounting panels anyway that I should take the opportunity to change from 2 to 3s for the increased efficiency. Clearly they do not possess the forum wisdom either 

[Westbranch]

WRT, I like to think of it as the cautionary principle needs to be applied when heading for the upper limits of any electronic device....

that's when that guy Murphy steps in ....

http://www.murphys-laws.com/murphy/murphy-laws.html

[offgridQLD]

the difference between 70v and 120v stepped down to 48v is around 11% loss.

11%?  I take it your referring to 11% of the initial loss of say (2%)  so 11% of say 80w = about 11w more loss.

I hope its not 11% of 4000w = 440w...thats a heck of a  lot of heat to dissipate in a tiny enclosure.

I have 4000w of pv on one classic and 4200w on the 2nd classic . I would be keen to know how much energy (watts) in heat they both have to dissipate during worst case scenario conditions.

The 4200w array is the 3s range from 90 - 120v depending on loads conditions.

The 4000w array is 2s (different brand of panels to) ranges from 65 - 80v depending on loads,conditions.

Kurt.

[boB]

the difference between 70v and 120v stepped down to 48v is around 11% loss.

11%?  I take it your referring to 11% of the initial loss of say (2%)  so 11% of say 80w = about 11w more loss.

I hope its not 11% of 4000w = 440w...thats a heck of a  lot of heat to dissipate in a tiny enclosure.

I have 4000w of pv on one classic and 4200w on the 2nd classic . I would be keen to know how much energy (watts) in heat they both have to dissipate during worst case scenario conditions.

The 4200w array is the 3s range from 90 - 120v depending on loads conditions.

The 4000w array is 2s (different brand of panels to) ranges from 65 - 80v depending on loads,conditions.

Kurt.

Yeah, the Classic isn't going to dissipate 400+ watts.  ~maybe~ 100 watts at worst case and that would
be REALLY hot.

boB