Jump to content

PV Failure


Recommended Posts

I wonder if someone could diagnose what and why the following happens to my system from time to time ?:

I have a 5kW Mecer inverter, 4 X 200 Ah AGM batteries and 6 X 340 W PV panels generating between 1700 to 2000 watts on good sunny days. I have an ordinary geyser for which I had made a 1,5 kW element which allows me to use solar during the mid morning to heat the geyser. On occasions the Inverter/PV Panels suddenly shut off the solar input but within seconds the inverter recognises solar power coming in again and continues as before?  I would be interested if anyone has experienced this event on their systems and to learn what may be the problem?

Please find a snip of a photo of the timeline of this event.

Thank you

PV Failure 19_03_21.JPG

Link to comment
Share on other sites

I omitted to say that the model of the inverter is a SOL-I-AX-5P and the Max Solar voltage is 145 VDC

I wondered whether the input voltage from the panels could be the problem when it approaches the max voltage - in the above it is 125 volts and then Zero??

Thank you

Link to comment
Share on other sites

Not an easy fix! Are there any conditions that are the same every time this happens? e.g. time of day, geyser element on etc.

You need to identify if there something that is contributing to the fault and better still is if you can replicate the fault. 

Link to comment
Share on other sites

37 minutes ago, Richard Mackay said:

Not an easy fix! Are there any conditions that are the same every time this happens? e.g. time of day, geyser element on etc.

You need to identify if there something that is contributing to the fault and better still is if you can replicate the fault. 

We cant replicate the fault as such but it has only happened when we are using the geyser element, which means we are using the full power of the panels, for a long period of time. It takes around 3 hrs to fully heat the geyser and it is during this long constant use when it happens. It does not happen if we only heat the geyser for an hr or so ie when clouds interfere with the production from the panels. We would then have to use utility to top up the geyser.

 

Link to comment
Share on other sites

2 hours ago, Richard Mackay said:

It sounds like an overload or heat related issue. Does the inverter have a log of the faults that have occurred? If does it register this fault?

No Watch Power has an event log but does not record that kind of event, only the events that effect the inverter like AC failure and recovery for load shedding and low battery etc.

 

Link to comment
Share on other sites

5 hours ago, Richard Mackay said:

I suspect that heat is causing the inverter to shut down

Agreed. 125 V should be OK, but the SCCs seem to run hotter with high voltage like that.

The fact that the shutdown only happens when the inverter is working hard for a long time suggests a thermal issue.

Consider flipping the fan direction; it helps with cooling. There should be a maximum temperature recorded somewhere. If you use fully patched firmware (e.g. 72.20e if it's not a 64 V model), you can see the temperature on the LC Display.

Link to comment
Share on other sites

  • 4 weeks later...

Thank you guys for your comments, I have managed to record two subsequent incidents where the inverter has dropped the incoming PV power. As mentioned it only happens on a clear day when the inverter is working hard. The attachments reveal the parameters at the time of the events and it would appear to me that the PV voltage may be the cause of the inverter "letting go"? In all cases that I have recorded the incoming voltage is about 123-125 volts and is very brief as is evident in the attachments, it soon (within a minute) goes back to normal, and I have not experienced it happening twice on a day.

If it is voltage causing the problem I'm interested to know if this could be termed a fault with the inverter, to be referred to the manufacturer/supplier, as the rated volts is  145 VDC and this event happens long before that??

Thank you

IMG_0896.jpg

IMG_0922.jpg

Inverter Drops PV.jpg

Link to comment
Share on other sites

12 hours ago, Bushtracker said:

In all cases that I have recorded the incoming voltage is about 123-125 volts and is very brief as is evident in the attachments, it soon (within a minute) goes back to normal,

So this has been described as a PV "dip". 124 V is probably just the open circuit voltage of your panels. OV cells usually have a Vmp of about 0.5 V (less a bit at working temperature), and about 0.6 V at open circuit. So when your panels are drawing no current, you'd expect about 98 * 6/5 = 118 V, not far off the 124 that you see. So is the higher panel voltage cause or effect?

Despite the advertised 60-115 V MPPT range, the 145V max SCCs start derating their output at around 130 V. At 140 V, they should still be delivering up to 67% of rated power. So I don't think that 124 V Voc could be the cause of the dips. The dips have been noted in King inverters, and some firmware updates were delivered that was supposed to have fixed it. I don't know if they really fixed the problem or not. But Mecer don't offer Axpert Kings with their branding, as far as I know. [ Edit: wrong! See next post by Calvin. ] The SCC in a King is presumably very similar to that in an Axpert MKS / Mecer with 145 V max SCC.

There was some discussion about temperature affecting the output. Do you get temperature recorded in the logs? It will be the maximum of 4 measurements, but if the SCC temperature is high enough to start affecting output, then it will be the highest measurement. You mention that it only occurs when the inverter is working hard, the above dips are soon before noon... so temperature could be a factor. @Calvin has suggested that the SCCs run hotter when wired 3S of 72-cell panels, than if wired 2S. But I have had someone from Papua New Guinea (hot climate) say that it's better to run them 3S, and he never sees overheating.

Edited by Coulomb
Link to comment
Share on other sites

49 minutes ago, Coulomb said:

But Mecer don't offer Axpert Kings with their branding, as far as I know.

Consider your knowledge extended 😀. See https://www.mecerpc.co.za/Mecer-SOL-I-AX-5KP-p-223718.php

 

49 minutes ago, Coulomb said:

@Calvin has suggested that the SCCs run hotter when wired 3S of 72-cell panels, than if wired 2S

This was certainly the case when I had 2 of my inverters wired 3S and one 2S.  The 3S inverters regularly (when SCC production approached 4000W) ran the SCC temp to 87, at which point it would de-rate the MPPT a bit to control at a temperature of 86-87C.  The 2S inverter never got close to even 75C.  When I changed the two 3S inverters to 2S the problem went away entirely.

Some points to note:

  • This was on main firmware 71.86
  • I had poor ventilation in my inverter room at the time. Ambient could get up to 40C.
  • The fans in my inverters were still pulling down.  I have subsequently reversed them.

The drop in PV was only a few hundred watts, not the complete drop-out reported above.  I had plenty of those in earlier firmware, but it was resolved  somewhere around 71.92.

 

Edited by Calvin
Link to comment
Share on other sites

Thank you all for your comments and observations. I was under the impression that this event was something seriously wrong with the inverter but it is evident that this "dip" is not uncommon? Should I worry?

Any advice on how to update my firmware would be much appreciated. Present display is UI 70 40

@Calvinsuggested that temperature may be a contributing factor to the event? Unfortunately I don't see temperature recorded on the system but this is the end of a mild summer for us and the last week has not been hot. Typical temps at the moment are between 25 and 27 degrees C.

I am due to install additional panels to my system in a few days and wonder whether this will exacerbate the system

The 6 panels are connected in 2 strings of 3 in series and wish to add 6 more!

@ThatGuyasked me whether this drop out occurs when the batteries are in Float - I have not observed this so when it happens again ill try and check

Thank you

56 minutes ago, ThatGuy said:

 

 

Link to comment
Share on other sites

Further to my reply earlier I think the reason for the Inverter dropping the panel output is due to excess operating voltage under load. The Open circuit voltage is rated at 145 volts but the operating voltage is 60-115 volts, for my Mecer Inverter. In my case, when the inverter lets go of the PV input, it is registering 123- 124 volts!! 

I will ask my installer to re-configure the panels and have three strings of 2 instead of 2 X 3, and see if that fixes the problem?

Thank you all for your time

 

 

 

Link to comment
Share on other sites

On 2021/04/12 at 3:38 PM, ThatGuy said:

It seems to be a few second shutdown that happens when the inverter switches from float back to absorb. My assumption is that it only happens under load, because if there's a brief drop in solar input (cloud or something) then the load is enough to remove the surface charge from the battery which drops the voltage low enough to cause the inverter to think that it needs to switch back to absorb/bulk charging.

Brilliant! Now that you put it like that, I can't see why I didn't think of this before. Perhaps I did and dismissed it for some reason. For reasons that are beyond me, when an Axpert goes from absorb stage to float stage, it does so via "zero charge" stage, which means that the charge current is zero for about 15 seconds. This would be a "dip".

Someone mentioned today (and I think others too) that their dips only tend to happen in the afternoon. While that means it might be internal temperature related (it doesn't have to be a hot day, just hot inside the inverter case), it could also be because it's in float mode, which often happens after about 11am, depending on many factors. Some batteries might be higher internal resistance, with the battery voltage sagging under load, or some batterys may possibly even be disconnecting because the settings are at the Pylontech (and other) recommended voltages, which are too high for Axperts.

So either the BMS has disconnected, or the internal resistance of the battery (which is higher for most batteries when fully charged) causes the battery voltage to drop steeply, especially if a load happens to come on (contributing to the intermittent nature of the problem), then the battery voltage may dip low enough to go back to the bulk charge stage. On 48 V models with factory firmware, there are two paths from float to zero charge stage, and hence back to bulk charging. One is 4.0 V under the float voltage setting for 10 seconds, no exceptions. [ Edit: That 4.0 V changes to 1.2 V with the fully patched firmware (LFP flavour only). That sounds pretty hard to achieve, unless the BMS has disconnected the battery and the loads are extremely light, or the loads are quite high and the battery internal resistance is high. Perhaps those with under-sized batteries might be more at risk here. The other path out of float stage is via what I call the "leaky bucket". Here, you only need to be 1.0 V below the float voltage, but for 30 minutes. But this time, if you go back over the (float voltage - 1.0V) threshold, you don't start the 30 minute timer again, you count it backwards at half the rate it goes forward. You can think of this as water dripping into a bucket with a hole in it, such that when the water is coming in, the overall water loss is 1/2 of what it is without the water dripping in. So the bucket can still empty, but it might take a bit longer than 30 minutes, depending on how often the battery stays above the threshold. But, assuming that I read the firmware correctly, and that your firmware does the same as the one I'm familiar with, then you shouldn't get two dips in a row that are less than 30 minutes apart, unless you happen to have a very high load.

For this theory to be correct, the dips should always be of short duration, around 10-30 seconds, depending on how often your monitoring software samples the PV power. Any dips of say 5 minutes duration, assuming monitoring every minute or faster, would not fit this theory.

Perhaps there are multiple factors at work; perhaps some are having the SCC stop due to an internal temperature alarm, others have dips per this theory, and perhaps others are suffering some other issue, e.g. PV over-voltage, or just intermittent hardware (e.g. cracks in printed circuit tracks, components intermittently failing, etc). Another possibility is buggy firmware, either  DSP firmware or solar charge controller firmware (if separate; some models combine these functions into one firmware).

Edited by Coulomb
Added that the 4.0 V drop to re-start bulk charging changes to 1.2 V with the LFP flavour of fully patched firmware.
Link to comment
Share on other sites

7 hours ago, Bushtracker said:

Unfortunately I don't see temperature recorded on the system but this is the end of a mild summer for us and the last week has not been hot.

As mentioned above, ambient temperature doesn't have to be particularly high for the internal temperature to reach a level where the solar charge controller may throttle back maximum current, possibly to zero. But thinking about this, I would expect more of a gentle slope towards zero power, not necessarily reaching zero every time, and not plummeting quickly to zero and back again just as quickly as we seem to see.

Link to comment
Share on other sites

47 minutes ago, Bushtracker said:

I will ask my installer to re-configure the panels and have three strings of 2 instead of 2 X 3, and see if that fixes the problem?

Be aware that this is probably not a 30 minute operation, so it could be costly. Extra wires and fuses have to be installed, possibly more conduit, possibly a second combining cabinet, and more.

So perhaps hold off while we see how well @ThatGuy's theory works out.

Link to comment
Share on other sites

49 minutes ago, Coulomb said:

For reasons that are beyond me, when an Axpert goes from absorb stage to float stage, it does so via "zero charge" stage

Hi @Coulomb,

I have an unusual setup, which may shed some light on this.

In order to get around things like premature float, my controller sets both absorb and float to 52.7V each morning and resets both to 50.5V when the batteries are charged.  

What I can see from my log files (2 second intervals) is that nothing happens when the inverter mode changes (absorb <-> float) - no SCC blips or reset, nothing.  However, when the charge voltages are changed down the inverter tries to get the voltage down to the new level.  It does this by essentially reducing the PV input so that the loads are supported by the batteries, often (in low load conditions) to zero.  This looks like a reset but in fact does not seem unreasonable.  I also see that in high load conditions the PV input drops but not to 0.

So, perhaps the zero-charge stage is not related directly to the change from absorb to float, but simply to the fact that there is a new (lower) target voltage that can only be achieved by taking some power from the battery for a few seconds.  In my case (8 x US3000) the change to 50.5V normally happens in about 2 minutes and leave the SOC at about 99.5%.

 

Link to comment
Share on other sites

5 hours ago, ThatGuy said:

I thought the charging-mode threshold was around 2.0V, not 4.0V...

Oops! I forgot to mention that with fully patched LFP flavour firmware, that 4.0 V becomes 1.2 V. I forget who is running what model inverter and what firmware.

5 hours ago, ThatGuy said:

Why would the inverter switch to absorb on my 1C lithium packs that can source over 340A for 0.1s before the BMS disconnects - it seems to me that I should never get the 4V drop that would be required for charging state change. I've never seen it drop more than max 2V...

Right. But 1.2 V could happen, so if you were running fully patched firmware, you could get back to bulk charging when needed.

Quote

What kind of measurement does the inverter do during those 10s? Does it only take 2 measurements (IE one when the voltage first drops 4V and then again on a timer after 10s), or is some kind of averaging filter applied to the measurement over 10s?

Nothing so elegant. It makes about 500 measurements in that 10 seconds, and if even one of them has the battery voltage above the threshold, the counter resets to zero and it needs to see another 500 measurements after the last one that was above the threshold. That's what I mean about "no exceptions".

Link to comment
Share on other sites

I happened to notice a dip myself today. It was an unusually sharp decrease in sunlight, while the car was charged in pulling 6 A (so around 1.5 kW load, quite modest).

857363599_PVdip.png.76c45a71d8bfc50554cd54442bee67a1.png

When the battery is nearly full (it was at about 98.5%), the battery voltage does jerk around a fair bit. Especially if the charge current happens to change sign, from charging to discharging. I'm thinking that the 1.2 V drop threshold for going from float back to bulk (via a dip) is perhaps too low. The battery readings involved are also not load compensated. We had the threshold at 1.0 V for a while, and increased it to 1.2 V. Maybe it should be 1.5 - 2.0 V.

My battery voltage readings have excessive noise on them; they're not from the inverter. I haven't bothered to add the extra capacitors to smooth them out more.

Of course, Bushtracker has an AGM battery, and won't have LFP patched firmware, so this doesn't apply to him. But it's an indication of how easy it sometimes is for the conditions to be right for a dip.

Edited by Coulomb
Add more verbage immediately after the image.
Link to comment
Share on other sites

On 2021/04/12 at 4:22 PM, ThatGuy said:

Remember that as Coulomb said above - the inverter only starts limiting output through SCC at 130V. So although it would be "even more better" to go 2S, operating as you currently are at 95-100V definitely won't hurt anything. The 123-124V you're seeing is NOT operating voltage, since the SCC has shut down at this point and is basically open-circuit, and even this voltage is within limits. 

I'm sorry I am not au fait with the workings of the SCC, Float and Bulk charging  and need further explanation as what is going on during and prior to the "dropping" of the PV power input? Is this an SCC issue controlling the operating voltage?

My WatchPower is set to display a new reading in 30 sec so when the drop occurs the PV input power drops to zero. When one is standing nearby and hears the inverter suddenly slow down, one can literally observe the PV input power as zero displayed at the inverter and watch as it regains the power input again within seconds.   So it is not always recorded as Zero as that depends on the actual time of the event - that is the reason for the voltage reading above zero in the photos below.

I now have another 5 incidents over the last few days of the PV input power being dropped! I shall not approach the electrician until  there is a concrete explanation of the reason for this event. The manufacturers have not given me a clear answer as yet either. 

Thank you all, much appreciated for your comments .

IMG_0927A.jpg

IMG_0896A.jpg

IMG_0922A.jpg

IMG_0928A.jpg

IMG_0932A.jpg

Link to comment
Share on other sites

40 minutes ago, Bushtracker said:

I'm sorry I am not au fait with the workings of the SCC, Float and Bulk charging  and need further explanation as what is going on during and prior to the "dropping" of the PV power input? Is this an SCC issue controlling the operating voltage?

We're not sure what's going on either, and are throwing around theories to be shot down by evidence, or not, as the case may be.

Your dips seem to be short duration; by the next set of data (30 seconds afterwards), it's all back to normal. So I'd guess what is happening is that the inverter is in the float stage, and your PV input is just barely able to cover your geyser load on a good day. Remember that there are other loads, there is a ~7% loss through the inverter, perhaps 10% from PV to load, and some 70 W (wild guess for a 7200 W model) of self consumption. So a sudden load, such as a fridge switching on, will cause a dip in battery voltage, as the battery goes from perhaps a slight charge to a ≈1000W discharge for several seconds. My guess is that it's enough to trigger the switch from float to bulk charge.

To switch from float stage to bulk stage, all the inverter-charger has to do is to change its target battery voltage from around 54 V (float) to absorb/bulk (around 56.4 V typically). But inexplicably, Axperts don't do it that way. They first go to zero current for 15 seconds or so, then stay at 1A charging for another 15 seconds or so, then ramp the charge current back up over another 10 seconds or so. Perhaps they do this because it makes sense on first start-up, and use the same sequence for float-to-bulk, figuring that this would be a quite uncommon occurrence.

This "startup sequence" is the dip that you experience. Although completely unnecessary, it's "normal" for Voltronic Power inverter-chargers to do this, and nothing to worry about.

It's still a nuisance, so I'm still mulling about how these infrequent events could be avoided, or at least made less common. 

Edited by Coulomb
Link to comment
Share on other sites

On 2021/04/12 at 5:38 PM, Calvin said:

What I can see from my log files (2 second intervals) is that nothing happens when the inverter mode changes (absorb <-> float) - no SCC blips or reset, nothing. 

Correction:  this should have read "when the inverter mode changes (absorb -> float)".  On the change from float to absorb the SCC faithfully does a reset as stated by @Coulomb.

 

Link to comment
Share on other sites

6 hours ago, Coulomb said:

My guess is that it's enough to trigger the switch from float to bulk charge

I've seen it happen in the other direction though. The battery is full and needs to go from bulk to float. What I saw and recorded today is that the load decreases, panel voltage and battery voltage rises and the dip starts. It definitely happens when there is a regular switching of a load and what I suspect is that the SCC can't handle the sudden change in power either up or down and overshoots then needing to dump power which triggers a reset of the algorithm which starts again from zero. If this happens often enough the SCC freezes at 0 and needs to be reset. I attach data recorded by ICC every 5s of a dip and freeze that happened today. Unit is a Axpert king. The screen shot is an excerpt for those that don't want to download the excel file.

image.png.14bfc585f912d5b0d66ebf3efcc03ad6.png

Data from dip 14.4.2021.xlsx

Edited by Shadders
Link to comment
Share on other sites

13 hours ago, ThatGuy said:

The BMS is connected with SOL battery setting (Soltaro), and commands a 55.4V charge voltage limit (Float = 55.4V, Bulk = 55.4V). So mine should still require the 4V drop to switch charging modes... I still don't quite understand how I'm getting this value then.

Well, 4 V below 55.4 V is 51.4 V, which is easier to achieve for 10 seconds than a "real" float voltage setting. With Watchpower saving data only every 30 seconds, you might miss such dips.

Quote

Could problems like this could be caused at other times during the day (IE when sun is NOT so weak) due to the slow response of the SCC? Could a rapid drop of the PV voltage (not necessarily all the way to 60V, but maybe just a 20-30V drop) cause the SCC control algorithm to undershoot the target voltage that it's tracking so that the controller is targeting a voltage that's too low?

Oh, definitely. The SCC power tracking is quite slow.
 

Quote

It might be that the SCC algorithm just isn't tuned properly to handle rapid drops in voltage - EG they've implemented a PID algorithm to track MPP, but due to China reasons they've neglected to set a value for D, or haven't tuned the I value properly...

I spent a fair bit of time looking for the PI algorithm in the SCC firmware. It's all HCS08 code (8-bit, lots of stack manipulation 🤮), and finally concluded that there isn't any PI control code. They just use ramps. There may be two ramp rates (slow and medium). No PI, and definitely no D (Derivative) control for fast response. That stuff takes time to tune, and may be difficult to get right with different sized panels. So they just keep it simple. Unfortunately, that results in plenty of over- and under-shoots of battery current and voltage.

Quote

Can confirm your "leaky bucket" analogy at least!

Voltage drops 1.0V below 55.4V @ 15:05

PV drops as inverter changes charging modes from Float to Bulk @ 15H35 - Perfectly 30minutes later!

Heh, nice to see my code reading confirmed. Thanks!

With your high float voltage, yours might be more susceptible to leaky bucket dips. "Unlucky dips"? That could be a new slogan 🤔. At least, they will be a minimum of 30 minutes apart, usually much further apart.

Edited by Coulomb
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...