Low PV Voltage Readings

pmcall

I have a 6kW grid-tie/battery back-up system consisting of:

60 Kyocera 120 watt panels ( 3 strings, 108 volts each string )
3 Trace C-40 configure as Charge Controllers ( one for each PV string )
2 Trace SW4048 stacked with GTI
8 Trojan L-16H 6 volts batteries (in series)


During the day the PV input voltage readings at each of the C-40s is the same as the battery input readings at each of the C-40s. When measure the PV voltage at thePV disconnect box I get somewhere between 80 and 95 volts.

Would appreciate any pointers as to what to how to go about troublshooting. Would bad fuses in the PV disconnect box be a cause for this? Bad C-40s?

nigtomdaw

Re: Low PV Voltage Readings

Wow thats 7.2 kw of power, I doubt that 3 C40.s have gone bad , looking forward 2 an answer. I think more info will be required and heres me thinking 4kw is the holy grail !

BB.

Re: Low PV Voltage Readings

From a quick cruse of the net... The C-40 is a PWM type charge controller. When passing maximum current, the PV input will probably be within a volt or so of the battery terminals (think PWM--pulse width modulation--imagine a switch simply turning on and off quickly between the input and output--100% charging current, switch closed. 50% charging current, switch is only closed 50% of the time. 10% charging current, switch is only closed 1/10 of the time)...

For an array this large, it would make sense if you use a MPPT type controller (Outback MX-60, etc.)... Those controllers find the Maximum Power Point (Tracking) and the input voltage will most likely be much higher than the battery bus when all is working well (think like an AC transformer--but is really a DC to DC converter circuit which behaves like a transformer).

I assume that a string of 20 panels is really 4 strings series x 5 panels in parallel. Assuming your panels are:

Module: Kyocera KC120-1
Max Power Voltage - Vmp 16.9 Vdc
Open Circuit Voltage - Voc 21.5 Vdc
Voltage Temp Coeff - Vtoc -0.0824 V/°C
STC Rating - Pmp 120 Wstcdc
Max Power Current - Imp 7.1 Adc
PTC Rating 105.7 Wptc

4x Voc (21.5vdc) = 86 volts at 77F cell temperature (voltage will drop on warm days)

5x Imp (7.1 amps) = 35.5 amps of current per controller...

So... If your battery voltage (state of charge) is low, the controllers should be dumping maximum current into your battery bank (PV voltage = Battery bank voltage). Once the batteries have done most of their charging, the controller(s) should begin to throttle back on their current--and you will see the RMS voltage ("average" voltage) of the array start to rise.

If the array voltage never rises (and you see current of 20-35 amps through the controllers), then either your batteries are never getting charged (too much load) or your controllers are set to the wrong charging voltage (set too high?) or not working correctly.

The last possibility is that (if you are in a hot climate and get temperatures at, for example, 95F) you do not have enough panel in series to charge your battery bank.

Using the Xantrex array calculator (for easy numbers), at 95F ambient, one would expect Vmp to be roughly 52.8 volts (assuming 4 panels in series--guessing from your Voc of 80-95 volts).

To fully charge your battery bank you should probably want the array to support 62 VDC (for equalization) or 58 VDC (normal charge), plus 1-2 volts for controller drop (60-64 VDC total from the array)...

So, on hot days (and cool batteries), it is possible that your array current will drop as it is not able to supply full current to the batteries while charging (and leaving your batteries--possibly--chronically undercharged even though, from a "wattage" perspective, you more than enough power to keep your batteries charged--again, assuming reasonable loads).

You can fine-tune the numbers (actual battery temperature range and your actual ambient and/or cell temperatures) as this may be a "close" call as to whether or not this is a "reliable/capable" system for your location.

But, from the size of this array--using an MPPT type controller would give you more power overall and always have enough voltage to fully charge your battery bank. (for example, instead of 4 panels in series x 5 parallel strings, you could use 5 panels in series with 4 parallel strings (give or take numbers of parallel strings--depending on the size of the MPPT controllers).

-Bill

pmcall

Re: Low PV Voltage Readings

Thanks for the response BB,

I am kind of a newbie to this (and a software guy to boot), my system was installed by a company that went belly-up a few years back. Attached is a picture of the schematics of my system, which I hope answers some of your questions ( and maybe educates me a little as well).

Other:
Until earlier today the float/bulk/equalize were set to the SW4048 defaults. Earlier today I set to 54/60/62 ( after reading and reading and reading the last couple of days).

One other thing to throw in. One of the SW4048's has always behaved flaky. If I don't have the MAX Sell AMPS set to something really low ( like 3 AMPS), when it starts selling back, the sell amps temporarily zooms to 40, 50 AMPs. Then the SW4048 loses sync with the grid, finds it 5 minutes later and resyncs, and then repeats the cycle of zooming.

BB.

Re: Low PV Voltage Readings

Well PMCall,

I took a stab at it... And according to your drawings, I was wrong--you have a 5 series x 4 parallel strings. And, if this is so, this should work perfectly fine with your batteries/charge controllers (high enough voltage to charge the battery bank in all conditions).

But, the Voc voltage does not make sense (assuming that I have everything correct)... Assuming the Kyocera KC120-1 panels and you are around San Diego--so 70-75°F days are probably pretty normal, assume 100°F cell temperature:

100°F-77°F stc / 1.8 °F/°C = 12.8 °C cell temperature rise over standard test conditions

21.5 Voc + -0.0824 V/°C * 12.8 °C = 22.55 volts (at 100°F cell temperature)

22.55 volts * 4 series panels = 90.2 volts @ 100°F
22.55 volts * 5 series panels = 112.75 volts @ 100°F
Your Measurement of Voc = 80 and 95 VDC @ ~70-80°F???

Your plans say 5 panels in series, your measurements seem to imply only 4 panels in series... So now what. Is one or more panels mis-wired (in all arrays)? Or is your volt meter out of calibration? Was the system not wired to the drawing? Is it not the KC120-1 panels that I guessed? All of these are very possible situations here.

One thing that would help--do you have a typical maximum current reading you see from each of your C-40 controllers? You will either see around 28.4 or 35.5 amps maximum from each properly working controller on a nice clear day with clean panels (temperature should not matter--assuming your batteries need charging and/or you are selling power).

Also, what is your typical battery voltage range and temperatures (charging, resting, charged). Do you measure specific gravity on occasion too?

And, with the SW4048's selling power--you are correct. They are acting like your battery charger by load dumping (aka "dump controller") your excess power to the utility mains (that not needed by the battery for charging). The C-40's are really only needed to prevent the batteries from overcharging (such as when inverters are off-line or you are generating more power than they are programed to sell).

You have the system that all of us Grid Tie only guys want... Grid Tie plus full off-grid battery backed operation. Cool. :cool:

Lastly, you said that you wanted to diagnose your system. What is your question/problem/issue?

If it is the C-40 PV Bus ~= Battery Bus? If that is your question--I would believe that would be standard operation for your system. You would have to replace the C-40's with a MPPT charge controller (like the Outback MX-60 or possibly the new Xantrex MX-60) to see the PV bus voltage substantually different than the battery bus voltage (and MPPT might give you 10-20+% more power too). Or, you would have to program the C-40's to start cutting back on battery bus voltage before the SW4048 started cutting back (or simply turn of the SW4048's and see how the C-40's regulate to float, etc.).

Or is it the SW4048 misbehaving? Since you have too of them, and one behaves differently from the other with the same settings--then you should probably call Xantrex and see if they can help (it may be a faulty unit and it would be best to get it repaired under warranty--if possible).

Or is it some other issue that caused you to start looking at PV bus voltages.

Hope I am helping... I should probably take more time and shorten this post...:roll:

-Bill

PS: I am pretty sure that SW4048 have their own AC to DC battery chargers internally too... So how those are set may cause interesting things too (like charging from your AC mains at times instead of from your solar panels).

niel

Re: Low PV Voltage Readings

pmcall,
i believe they speced the wrong controller for that configuration. imho all 3 controllers are blown due to too much voltage. this type of controller is not able to step the higher voltage down to the battery voltage you need. this means a 12v pv for a 12v battery and a 24v pv for a 24v battery. see the specs here: http://www.xantrex.com/web/id/72/p/251/pt/25/product.asp
5 pvs in series will produce an open circuit voltage over 100v and an mppt type with voltage stepdown features of the better controllers is needed. this high voltage exceeds the max v limit speced for the c40. there are a few to choose from like the mx60 or the new xw60, but not limited to these. these will need other accessories to accomodate the multiple controllers too and if it is not in the drawings(didn't check) you should also have a battery temperature sensor to allow the charge voltage to be altered for temperature changes. i believe before you venture and buy more stuff that an evalution by everybody here may be in order. for instance with 60 pvs and 2 good downconverting mppt controllers you may be able to have 2 banks of pvs containing 30 pvs each and in series/parallel arrangement of 5sx6p. if that isn't possible and you don't want the wiring arrangement changed you may wish to run 3 of the better downconverting mppt controllers at added cost of an extra controller.

nigtomdaw

Re: Low PV Voltage Readings

Somethings are not right on the PV front here.

A.60 x 120 watt panels are 7200 watts or 7.2kw PMCall states that he has a 6kw grid tied battery based system . 1200 watts is a lot to loose or seemily in this case gain ?. So further investigation are needed PM on exactly how many panels and types you have b4 the clever guys can help you out.

B. Neil according to spec the C40 can handle a Max PV Open array voltage of 125 volts DC * (yes I was surprised also) It also can handle a max peak of 85amps
Your link specs the C35 not the C40

As this is a grid tied system primarily of which I have no practical knowledge I can only assume with a massive 105 amps pouring into the relatively small battery bank with no draw uless a grid outage that the C40s will spend most of there time in float mode and the grid tied side doing its thing. If your batteries are fully charged then where is the problem, if they are a little low you will be able to adjust each C40 (pots inside finger friendly) to alter the bulk and absorb settings .

As far as grid tie goes its over to someone else !

nigtomdaw

Re: Low PV Voltage Readings

Looking over this thread again if BB is right and PMcalls voltage reading supports his calculations then each string has 4 PV panels in series then parralled 4 times gives 16 panels in each string. 3 strings gives 48 panels not 60.

48 x 120 watt equates to 5760 watts 0r 5.76 Kw alot closer to PM,s 6KW original description.

Anyway only theory wait and see

Solar Guppy

Re: Low PV Voltage Readings

The Schematic does show 5 KC-120's in series in each string ...

First thing is shut everything down, pull all the fuses in the combinor boxes and measure the Voltage from the arrays. You should see about 100V from each string ( 12 measurements needed as you have 12, strings of 5 panels )

In typical mild sunny weather. If you have lower voltages, I'd bet you have KC-120's from the 2000-2001 date code.

Read here for more info on the KC-120 issues

http://www.solar-guppy.com/forum/viewtopic.php?t=233&highlight=kc120

I suspect you will endup with a all new array on Kyocera's Tab ...

The remainder of your system could benifit from newer equipment and at a minimum a better understanding of the programming of your SW's. C40's and SW's will be losing about 30-40% of your potentional, I'd recommend a pair of Xantrex XW-MPPT controllers and a Xantrax XW-6048, the 6048 is a 6 KW, 240 vac ( dual phase, with gridtie ) ... if you sell the C40's and SW's, I'd bet you could upgrade for about 2K. Anyways, start with the array and work forward on resolving your issues

niel

Re: Low PV Voltage Readings

nigtomdaw wrote: »

Somethings are not right on the PV front here.

A.60 x 120 watt panels are 7200 watts or 7.2kw PMCall states that he has a 6kw grid tied battery based system . 1200 watts is a lot to loose or seemily in this case gain ?. So further investigation are needed PM on exactly how many panels and types you have b4 the clever guys can help you out.

B. Neil according to spec the C40 can handle a Max PV Open array voltage of 125 volts DC * (yes I was surprised also) It also can handle a max peak of 85amps
Your link specs the C35 not the C40

as to the 7200w stc that would be correct for 60 kc120s if that's what he actually has.
yes i did notice afterwards that i read the wrong controller therefore it didn't blow out the c40s, but i didn't feel it that important to change my post at that point as i still say they had designed for the wrong controller. a system like that is for a better downconverting mppt controller. if you go with a c40 design it would waste far less pv power when matching the inputs and outputs more closely.

nigtomdaw

Re: Low PV Voltage Readings

Neil I think our posts will more than likley turn out to be redundant. As I think if the installer completed as the blueprint shown then SG post, the missing volts, and the failing KC120s will probably be on the money

When you look at the KWs of PVs and the relatively small battery capacity I would think that the system designers paid more attention to the grid tie, and 3 basic C40s cost less that one of the 2 MX60 that would be required and even with the advantage of better harvest from a MPPT unit I dont think it would ever be an issue with 6kw pumping 125 amps into that battery bank when the grid failed. However if it indeed there was 60 x 120 w panels giving a calculated 7,2 kw array then the system could push out 150 amps which could cause problems We also dont know when the system was comissioned and if the MX60 was born then ?

We will have to wait and see ?

crewzer

Re: Low PV Voltage Readings

I have a 6kW grid-tie/battery back-up system consisting of:

60 Kyocera 120 watt panels ( 3 strings, 108 volts each string )
3 Trace C-40 configure as Charge Controllers ( one for each PV string )
2 Trace SW4048 stacked with GTI
8 Trojan L-16H 6 volts batteries (in series)

During the day the PV input voltage readings at each of the C-40s is the same as the battery input readings at each of the C-40s. When measure the PV voltage at thePV disconnect box I get somewhere between 80 and 95 volts.

Would appreciate any pointers as to what to how to go about troublshooting. Would bad fuses in the PV disconnect box be a cause for this? Bad C-40s?

This may not be as wacky as it might appear. The C40’s specs indicate they can handle an array consisting of five KC-120’s wired in series (84.5 Vmp STC, 107.7 Voc STC). In bulk charge mode, the battery voltage will essentially set the array voltage. Array current of 7.1 A Imp x 4 = 28.4 A shouldn’t be a problem for each C40.

Translation: Configured this way, the 7,200 W STC array will never operate at 7,200 W. 5,000 W is the likely maximum operational power (59.2 V x 28.4 A x 3). It just appears to be an unusual but conservative design.

The target absorb- and EQ voltages for a “48 V” Trojan battery bank are 59.2 and 62 V. Five Kc-120’s wires in series @ 16.9 Vmp STC = 84.5 V. Allowing for hot weather and voltage drop in the wiring, this could work OK.

It would be possible to measure “…somewhere between 80 and 95 volts…” at the inputs to the controlelrs when they're in absorb mode and operating in PWM. The DVM would be measuring the duty-cycle weighted average of the battery voltage (59.2 V?) and the array’s open circuit voltage (~100 V?).

It won’t take long for your array to fully charge the battery bank in the morning. Accordingly, your system probably spends a lot of time with the controllers operating in either absorb mode or float mode. Unusually high PV array voltages would be a natural consequence of these operational modes.

Do you live in a hot climate?

Regards,
Jim / crewzer

mike95490

Re: Low PV Voltage Readings

Solar Guppy wrote: »

In typical mild sunny weather. If you have lower voltages, I'd bet you have KC-120's from the 2000-2001 date code.

Read here for more info on the KC-120 issues

http://www.solar-guppy.com/forum/viewtopic.php?t=233&highlight=kc120

I suspect you will end up with a all new array on Kyocera's Tab ...

The man with the brains - the KC-120's & low output, never clicked in my skull

nigtomdaw

Re: Low PV Voltage Readings

Jim thats just added possible solution 3 to the mix,

1. Incorrect number of panels
2. SG KC120 Failure History
3. Creezer not as wacky as it seems

Anyone want to bet ?

pmcall

Re: Low PV Voltage Readings

Wow, thanks for all the suggestions. I have tried to answer most questions throughout this response.

Verified that I have KC-120's, but date is 8/2002 (not 2000/2001). System was installed in 10/2002.

My question about voltage at the C-40s came about because I had a guy come out to look at my system earlier this year when I noticed that my sellback had decreased dramatically.

He took a bunch of measurements, and then consulted with Xantrex and then returned with instructions to reboot the system because the PV voltage input was "low" at the C-40s. So he rebooted the system, and the voltage readings were again in the 80-95 range.

Been having low sellback, so I checked the voltage at the C-40s, and it was low. I repeated the reboot instructions ( basically disconnect all power to the system for half an hour or more, and then bring it back up).

When I turned off the PV disconnect during the bring up sequence, I would intially measure voltage 80-90, but it would then decrease in a matter of seconds to the battery voltage.

Battery bank use to be 2 strings of 8, but I blew some batteries, and the prices are now outrageous ($417 each for Trojan L16-H) so will wait until I can't have a single 48v string before I buy again.

We just got through 2-1/2 weeks of no grid power (due to the fires) with the one string of 8, so I think I can make it a little longer. Everything in our house is electric (hot water, stove, secondary water pump, etc), and the entire electric panel is "behind" the inverters. Also lived for 2+ weeks off grid after the 2003 fires.

We live in the mountains SE of downtown San Diego. In the summer it is almost always above 90, sometimes 100-110. In the winter it gets down to the mid 30s (occasional flurries), in between it is typically 65-90.

I am going to measure the voltage at the PV combiners as suggested by Solar Guppy to see what is going on there.

Will report back once I get those measurements.

niel

Re: Low PV Voltage Readings

"My question about voltage at the C-40s came about because I had a guy come out to look at my system earlier this year when I noticed that my sellback had decreased dramatically."

wild guess here, but are the pvs dirty with soot? it would not just be from the fires because this goes back farther in time, but fires could've added to it too. when did you clean then last or did you ever? I'd be surprised if this is the problem, but hey, i have to ask.

Solar Guppy

Re: Low PV Voltage Readings

80-90V is to low ... somethings wrong with the array if the disconnects are open ... you should see 100V no load each string. Looks like you will need to open up all the JBOX's and meaure the Voltage at each panel .. you should see at least 20V ... I bet you will find panels putting out 10V ... now one other know issues could be the bypass diodes, if you find a KC120 with low voltage, remove the bypass diodes ( they just connect at the screw terminal block ) ... if its the Diode, really easy fix, you can get replacements at Digikey for pennys each.

I wouldn't be surpised if you have date codes mixed in and/or 2002 panels have issues as well ... pretty sure it was 2003/04 when Kyocera realized they had probelms and changed their process

Also, 10/2002 the MX60 was just comming to market and in that time frame C40's/SW's were the ONLY option for offgrid/gridtie ... No FX's at that time

mike95490

Re: Low PV Voltage Readings

pmcall wrote: »

Verified that I have KC-120's, but date is 8/2002 (not 2000/2001). System was installed in 10/2002.

Are those panel sticker dates ? Getting panels mfg in Aug, installed in Oct, is pretty speedy shipping & handling.

pmcall

Re: Low PV Voltage Readings

Reviving the thread.

So I finally crawled under the KC-120s and tooks ome measurements. I disconnected the panels from the inverters first. If I took the readings right, looks like about 25 of the 60 panels went bad.

Note that on a couple of the panels with low readings; I took the bypass diodes off and re-read, result was 0 volts (instead of 8-9 I was getting). Was a pain to get the diodes back on, so I didn't do this for all 25 panels.

Looks like I will be talkgin with Kyocera.

Solar Guppy

Re: Low PV Voltage Readings

Join the club, for contact info/help check out this post on my forum

http://www.solar-guppy.com/forum/viewtopic.php?t=233