Xantrex XW6048 - fault F49 - DC over voltage

Re: Xantrex XW6048 - fault F49 - DC over voltage

Hi Bill, I have checked/disassembled the batteries and cable connections very early on. Would I be wise in turning off the power and checking each battery under an invert load? If there was a bad cell, wouldn't it be there all the time?

The problem only appears when the batteries are in a float state and Sell is on???

Thanks, Nick

Re: Xantrex XW6048 - fault F49 - DC over voltage

RCinFLA wrote: »

A clip on d.c. amp meter and measuring individual batteries would help diagnose the issue.

I doubt the solar charge controller could drag the battery that high even if not in sell mode. The XW should suck up all the solar controller excess, but if you can catch it in the fault, with the clip on ampmeter check solar controller line and main inverter lines to see who is sourcing the current and how much current.

You could have an intermittant high resistance on one of the battery cells. You did not say how long the batteries were subjected to 70 vdc. They will be damaged if subjected too long to this level, unless one of the cell is high resistance.

How much current is going into the battery at the 70 volt fault point? If it is low current, say less then two or three amps, then one cell is high resistance. I assume you have access to the individual batteries so check how the 70 vdc is distributed across the batteries. Make sure you check at battery post terminals, not the cable terminals. If one is showing high voltage, it is the bad battery. If voltage is equally distributed and current is higher then a few amps then I would suspect the XW.

I would be hesitant to put any replacement batteries on the system until the problem source is identified.

Hi RC, don't have much of the sophisticated equipment you specifiy above - in any case, the 70V is not on the batteries for very long (millisecond). Moreso, the 70V is very transient - when this problem begins to occur again, it just occurred to me that I could trap the battery voltage at the moment of the fault. Through 4 cycles of the problem, I could use my Fluke 115 to trap the high limit of each battery.

That gives me something to go on.................

Re: Xantrex XW6048 - fault F49 - DC over voltage

In reading through the posts about this problem again, it occurred to me that I may be misunderstanding how the Sell function works. When sell is on, is the current being sold inverted from the batteries or is it inverted from the source arrays via the Charge Controllers????

It would seem that since the sell function is directly proportionate to the array output, that the batteries are not the source???

Need a little clarification here??

Nick

Re: Xantrex XW6048 - fault F49 - DC over voltage

Solar Guppy wrote: »

Are your running 58V or higher sell voltage?, which will have the battery's more than 100% charged and they are not buffering enough and the chargers/inverters are not regulating in some transient condition, I ask as other have come to this forum saying that was recommend setting and its wrong.

For the XW and Sell mode, 100ah of battery's for every 1kwatt of PV your selling to grid is a minimum requirement for proper operation ... strange thing can happen with to small a battery bank, such as nuisance trips and at higher powers mppt tracking stalling on the charge controllers.


With such a small battery bank I would recommend you set the sell voltage to 50.8V, which is as close as one can get to be battery's fully charged ( 12.7V * 4 ) yet get some buffering by the battery's. Also, make sure the XW-6048's charger is turned off, its not needed as you have 5kw of PV to maintain a fully charged your AGM battery's.

This should offer some improvement and validate your need a bigger battery bank.

Its also important your using the Battery Temperature Sensors so the voltages are compenstated for. They come with the XW's, but you didn't mention if they were installed to the XW-6048 or XW-Mppt's

Grid Support Voltage has been set to 57.0 and Max Sell Amps to 27.0 since day one. On each of the three failures and while talking with Xantrex tech, both these and other setting have been lowered with no effect on the then ongoing problem.

There is one BTS installed on only one of the controllers, but the batteries have never even felt warm to me. I do have the 2nd BTS here on my desk - the installion guys suggested only one was needed?? Should I install it and where?

Have turned off the XW-6048's charger - will report on the next event if any. Also will report on the batteries and cables under load as Bill suggested.

Many thanks............

Re: Xantrex XW6048 - fault F49 - DC over voltage

Have done all as requested. I will also test the batteries under a load and report, and if the problem happens again, I will test each battery for a high voltage level.

Many thanks for the fine input,
Nick

Re: Xantrex XW6048 - fault F49 - DC over voltage

As promised, I'm getting back to everyone on the battery test - still have the test to perform under the conditions of the fault. I may or may not (if the problem never happens again) be able to report on that part of the test.

Used this storm here in California (read that as a power outage) to check my batteries under an inverter load of 300-2500W for 2+ hours. During the outage and after the power came back on, the individual battery voltage levels were all within .01VDC with the first/lowest reading (1 hour) at 2500W being 12.12 and the second reading (2 hours) at 400W was 12.17. When the power came back on at just under 2.5 hours, the readings were 12.30.

Seems like the batteries are operating correctly.

And BTW, all readings were taken from the cable ends and not directly from the battery terminals/posts.


Nick

Re: Xantrex XW6048 - fault F49 - DC over voltage

RCinFLA wrote: »

I think you have something that may be a combination of items unique to your setup and grid environment. If you have not seen the post I wrote a week or so ago take a look at the link below and the diagram at bottom. This may give you some understanding of how the inverter interacts with the grid for the XW series.

http://forum.solar-electric.com/showthread.php?t=5903

You really know this stuff.........

One of the challenges to the inverter/grid parallel syncronous design of the XW is the inverter is building a grid waveform match based on assumption it is dealing with a sinewave. The grid has numerous possible pertubations from a dip when a heavy load like central air conditioner compressor starts up to a sharp high voltage spike when a lightning bolt strikes the power lines in the area. The XW must put up with, and handle these pertibations.

Power here is pretty clean and our lights never flicker or dim. We have no air conditioner and the only motor is the fridge.

A lightning strike to grid will likely fall outside acceptable parameters setup within the XW and cause it to release from grid for a small period of time (AC1 relay opens). A heavy load startup surge, like air conditioner compressor only causes a short small dip in grid voltage as seen by XW. It will stay grid engaged and adjust to follow this dip but there is some time lag in the tracking process.

This adjustment is continously happening while the inverter is on and grid locked (AC1 relay engaged, inverter active). The time it takes (10-30msecs) to recognize and make the adjustments results in some random 'pushing' and 'pulling' from battery as the 'gear ratio' of the inverter PWM is modified to track the new grid voltage. The result of D.C. current fluxuations going in to, or out of, the battery is kept track of by the XW monitoring the battery voltage and any charge overage due to a sudden AC voltage increases is made up by some discharging of battery pushing back to grid or load (and vice-versa for a sudden dip in grid voltage).

Understanding this, the only different between this fluxuation in power flow due to grid fluxuations and the 'sell mode' is sell mode is intentionally pushing to grid for longer period of time.

When you say you turned off the charger you are turning off the inverter and going to bypass mode where only AC1 relay is engaged but no interaction with battery is happening because inverter is off.

Isn't there still interaction with the batteries from the charge controllers? And, if as you say, the inverter is in Bypass mode, it is still selling to the grid, so the published mode SellMode???

Now for some possible theories for your issue.

If your local grid has a lot of pertubations you could be subject to more overages to instantaneous charge current to battery, even though you are selling. The fact you have small A-hr batteries may aggrevate this. Do you see flicker in lights often? Does XW momentarily disconnect from grid often?

Aside from the fault 49 issue, and occasional power outages there are no other anomalies.

On the battery side, battery line inductance can create transients in DC voltage seen by XW. Are your battery lines long, at least 2/0 gauge? Are the positive and negative lines kept together as much as possible to reduce line inductance?

Battery cables are 8', half of which is in a 1.5" pvc conduit.

The A.C. side reaction time of the XW is limited by the 60 Hz waveform. The D.C. voltage measurement could be quite fast. It likely has to be slowed down by XW to better match the reaction time of the AC side. Maybe for your setup/situation the D.C. measurement in XW is still too fast.

One other thing I just remembered. The grid-tie regulations require the XW to be intelligent enough to quickly release from grid when grid goes down (or open). One thing that may only happen in sell mode is a periodic monitoring of grid impedance. One common method for doing this is intentionally, slightly off-phasing the grid pushed voltage to detect the extra current that should happen when grid is good. Not sure if this is a factor but it is something different about sell mode.

See responses in bold above.

One thing that the above doesn't really address is why the unit starts operating correctly again, only after disconnecting from the grid by opening the main breaker at the meter panel momentarily??? I believe that the XW inverter has a software or hardware glitch.


Many thanks, Nick

Re: Xantrex XW6048 - fault F49 - DC over voltage

At the moment, I think I have this under control, at least to the extent that if it occurs again, I may be able to bring it back on line easily and very quickly as described in my last post. Should that not cure the problem the next time, I will be sure to report my findings.

And I will report back six months or so from now as to whether or not this problem has resurfaced.

In any case, Xantrex should be interested in this thread, there does appear to be a serious flaw in the way the XW is performing under these circumstances, whatever they really represent.

Thanks again, Nick

Re: Xantrex XW6048 - fault F49 - DC over voltage

Helo Dan and thanks for the relevant info. I have recently made contact with Xantrex again, wanting to pass on this thread to the product manager whose name was given to me in August. She is no longer the PM and I have yet to receive an email address and name for the current PM. The below is what I am planning to send to the PM.
_______________________________________________________________
Hi there, As you may have gathered from the previous case number (00132874), my 6048 has been experiencing Fault F49 (and once it starts, there has been no known fix or workaround) off an on for a year now, with no logical explanation. I have gone to great lengths outside of Xantrex support to find the cause of this problem and may have stumbled upon some relevant information.

As you know, your technical support staff are trained to analyze this kind of problem, in part, by determining what the battery pack looks like, ie, how large or small it is or isn't. Every time I have called for help, the stock response has been, "you need a larger battery pack" and of course, I have only spoken to Nick Klein. And while you folks may have found that in your experience, that solves the problem, in my experience (and I have been programming, repairing, trouble shooting and building computers since the 60's), the issue is related to an underlying glitch of some sort - either hardware, software or a combination of the two.

So, I posed this problem to a solar forum, whose collective wisdom appears to be very useful - in particular, especially in this case.

The basic problem is that, there is no common denominator that relates to the onset of the issue - it can happen on a warm day, on a cold day or on an overcast/foggy/rainy day and once it starts the only cure for the infinite loop has been to turn SELL off, which of course defeats the purpose of a Grid Tie. At no time have any of the batteries ever tested out of spec or even remotely felt warm (BTS is connected).

In the early stages of the incidents, beginning a year ago, there was no evident fix that would bring the system back on line and somehow, weeks/months later, the system miraculously would begin working again when I arbitrarily set it back to SELL. It was only when I tried to bring it back on line, after weeks of F49, F49 in an infinite loop whenever SELL was turned on, that I discovered that the resolution was to turn off the AC power source (Grid connection) at the main meter panel, similar to a brief power outage, that I determined that this may be a work around. That was the last time that I experienced the problem, which was in late September and at which time it had been out for five weeks, until that fateful day when the power went out briefly and I subsequently and coincidently decided to turn SELL on once again. It's been on ever since without any disruption.

The problem at that point was that there was no reason that it wouldn't happen again and there was no logic for it to have occurred in the first place, at least none that didn't point to hardware and/or software. If it was really an improperly sized battery bank, it would seem logical the problem would occur frequently and/or right off the bat when the system was first turned up.

I have attached a PDF file that details the back and forth conversation regarding this problem that occurred on the Solar Forum - this thread was started on the heels of my possible discovery of the work around. In a nutshell, here is what was suggested by those participating, all of which seems very logical to me, but obviously does not address the underlying problem.

Reset the battery charge level from the recommended 57.0VDC to 51.8VDC.
Turn off the 6048 AC charger option and rely solely on the charge controllers to float the batteries.

To date the problem has not recurred (and during this time the power was out for a day and the batteries performed perfectly), but the real question becomes, is the above the real long term fix for this problem - only time will tell for sure?

No matter, in view of the fact that the above changes were made subsequent to the "turn off the power" workaround, there remains a some very significant questions - in other words, we really don't know if the problem is just not happening, at the moment, as has been the case since day one, or, have the above changes contributed to the avoidance of the circumstances surrounding the event onset? Regardless, the following questions need answers:

Why does the 6048 not correct the F49 situation once whatever the underlying cause(s) disappear(s) - in other words, why the infinite loop and why not reset the conditions that caused the problem and to begin the test for the problem from the beginning?

What does disconnecting the AC/grid do to the internal workings of the 6048 that would allow it to once again begin selling to the grid successfully? My experience with this kind of workaround suggests a memory variable, one that contains residual/incorrect data and with which the 6048 determines an F49 condition exists, is being reset due to the loss of AC power (total supposition on my part about the inner workings of the 6048 ). However, by itself, this would account for how the workaround is successful.

The Xantrex recommended fix (more batteries) for this problem may have solved the issue for some case in the past, but that kind of fix may be masking the incorrect workings of the 6048. In other words, I believe that a larger battery pack (if that really has fixed this issue in the past) is merely hiding a problem in the system and is effectively doing your customers a disservice by causing them to redesign their systems and to find more space and money for more batteries that really aren't required.

Many thanks,

Nick

BTW, I expect to discuss this personally with your software/hardware engineer at some point after the above is given serious consideration and I am prepared to escalate the fix for this problem all the way to the top.