Data Dropouts with data in GivTCP

For as long as I can remember since having the GivEnergy equipment installed and being able to monitor it all via the Home Assistant add-on GivTCP, I’ve had data dropouts from my garage + shed solar array but never from my House Solar Array but had never got around to investigating the problem until recently.
These dropouts do not show on the GivEnergy Portal as that information is only updated every 5 minutes.
So I'm wondering if anyone else has suffered with a similar issue.


The 2nd image above clearly show the dropouts on the Garage & Shed Solar

Graham_Harrison Yes, although not to that extent, but read on.

String 2 (its always string 2) will drop out/down occasionally for a couple of minutes. Used to happen fairly regularly - ie the time interval between dropouts was consistent and measured in minutes (45 mins or so between dropouts IIRC).

Its not the PV panels or wiring as the strings were swapped over and the issue remained on string 2.

This happens when there's no cloud, both strings are on the same roof, no shade affects one without affecting the other.

I positioned a fan under the inverter and that removed 95% of dropouts so IMHO its heat-related but not measurable by any of the temperature sensors exposed to the user via GivTCP. Those sensors show nothing abnormal when the dropouts happen.

This is on a G1 3.6kW hybrid and another one which I have access to (not mine, not the same location/install) shows the same on string 2. I think its a design fault.

I haven't bothered reporting it to GE because frankly what's the point? The portal (somewhat conveniently) never shows the issue (probably) due to the sampling time interval.....

Vestas thnaks for the input.
I know that my string 2 solar array definitely does NOT have an issue or fault as the myenergi eddi monitors the same array so I'm almost 100% confident that its something to do with the EM115's / data cabling.
My system has 3 x EM115
EM115 ID1 - Monitors Grid Import / Export
EM115 ID2 - Monitors House Solar
EM115 ID3 - Monitors Garage + Shed Solar
I initially thought it was the EM115 ID3 at fault but the other day it was temporarily swapped out with a replacement but unfortunately that made no difference.
This week I'm hoping to do the same with the ID2 then ID3 to see if that resolves it

Graham_Harrison Maybe I didn't make it clear - its not a "data" dropout, its a PV power dropout in my case and like you I know its not the PV array/wiring at fault.

ie its a real decrease in power, not a data recording anomaly/error.

Doesn't last long enough to be recorded on the portal usually and on the rare occasions it does then its a blip which would be written off by GE as a cloud passing over.

I purge the string PV power entities every night so I can only look at today’s graph, but I don’t see any such dropouts on either of my two Gen 1 hybrid inverters:

The inverter fans didn’t turn on at all today.

I do see grid voltage related shutdown spikes on the inverter power output when its really sunny, might be worth checking for that?

geoffreycoan Try turning off the fans please - preferably on a clear, sunny day. See if you notice any then.

I don't see the issue on days like today. Its only really apparent on clear sunny days - and nothing is near any limits in terms of voltage or current.

Edit - my initial thought was that the PV cabling was at fault, not enough conductor clamped properly or some such leading to overheating. Doesn't seem to be the case & was checked last year when they had to fix the roof leak.

Vestas I’ve extended the history retention so I can see what happens in more detail, and will do some tests with/without the fans being on

looking at Graham’s graph above he sees power dropouts right from first thing in the morning, wouldn’t think that the inverters were getting that hot by then?

If you pull up inverter power for the same period, do you see corresponding drops (although not to 0W obviously, as string 1 should prop it up)
That would atleast narrow it down to an actual drop off of PV and not just soemthing in givtcp (like the data smoother) getting in a twist and simply returning 0.

geoffreycoan Like I said I figured the likely cause was dodgy cabling at high amperage overheating so I pointed a fan at the cable intakes (left hand side) and that reduced the frequency. However that's the side of the inverter which gets hottest anyway so who knows what got cooled 🙂

You're on the same f/w so it'd be interesting to see what happens, although your setup is quite different from mine.

Graham_Harrison hasn't said what the cable length is from the meter in question or how its routed. There's plenty of interference sources which could take out over-length runs....

hoggy Not sure whether you're replying to me or not but if its me then this is an actual drop-off of PV power.

Thanks guys for all your input, I'm not really sure who to reply to so I'll add some more info for everyone then please shout up if I've missed answering anything.
My feed that is suffering the dropouts is made up of 3 systems
1 x East facing array with 1 Solis inverter (1320w)
1 x East facing array with 1 Samil Power Solar River (1750w)
1 x West facing array with 1 Samil Power Solar River (1750w)
The ID3 EM115 with CT clamp that measures the combined output of these 3 systems is located inside my garage consumer unit.
I'm in the North East of the UK and all my inverters are mounted inside my garage and can never get any direct sunlight onto them so they hardly ever get hot .
The data cable is cat5e (I think) which runs down from the garage to my main consumer unit in my house which is a run of approx. 15 metres and is bundled / clipped alongside several other cables along the outside of the house (so interference is very possible)

I've just looked at the Voltage & Current for the same time line and there are NO dropouts.
I'm not on my main PC right now but when I'm back on it in the morning I'll upload some more screen grabs

I see similar drop-outs, however for battery temperature, not string power.

Our inverter is connected via Cat 6A (shielded) ethernet cable to our router. When we used wifi, the drop-outs were much worse.

[wow, it took me 10+ tries to post the image!]

Graham_Harrison I'm in the North East of the UK and all my inverters are mounted inside my garage and can never get any direct sunlight onto them so they hardly ever get hot .

You inverters get surprisingly hot just with the process of converting AC to DC and vice versa so don't rule temperature out due to the inverter location.

You can see your givenergy invertor temperature (note spelling) in givtcp. I have fans underneath mine to blow cold air up the back of the inverters when they get above 40 degrees - mine are at 37.8 degrees right now for example

geoffreycoan You inverters get surprisingly hot just with the process of converting AC to DC and vice versa so don't rule temperature out due to the inverter location.

OK will take that onboard.

Something I haven't referenced as off up to now, I've only really mentioned my solar inverters is my AC inverter, my AC 3 inverter is located in my house utility room wall mounted and like my garage inverters never gets any direct sunlight onto it.
The AC inverters internet connection is via the stock GivEnergy wi-fi dongle running the latest (I think) firmware.

hoggy If you pull up inverter power for the same period

Graham_Harrison Its very clear its string 2 that is shutting off power and then that’s dropping PV power.

Nothing similar on mine yesterday but I only peaked at 1kWh.

Can you pull up grid voltage and invertor temperature for yesterday?

geoffreycoan Can you pull up grid voltage and invertor temperature for yesterday?

Do you specifically want "yesterday"?. The graphs I've been posting are from the 20th which was pretty much a cloud free day and that's very unusual in my part of the country.

Yesterday

I've just had a rummage through the data on the official GivEnergy Portal which the hope that sometimes that catches the data drops, and it does!!
19th & 20th were both pretty much cloud free days.


Both suffer dropouts to 0 at a similar time

Graham_Harrison Do you specifically want "yesterday"?. The graphs I've been posting are from the 20th which was pretty much a cloud free day and that's very unusual in my part of the country.

No I meant for the same time period as the other analysis, to see if any of those (grid voltage & inverter temperature) were contributory factors. In yesterday’s graph does show the temp spiking at 48 degrees which is warm but not excessive.

Graham_Harrison This is what it looks like on mine with a fan blowing -

and zoomed in -

PV1 power is the purple line, PV2 power is the other.

Edit - there's only one dropout on that day but a week before there were multiple dropouts like that as my wife unplugged the fan. Unfortunately that data has been discarded by HA as its outwith retention times.

Vestas the big difference with your situation though is it doesn't drop to 0 (ZERO) like mine always does

Graham_Harrison It does without a fan.

Right OK, I think we've all overlooked something a bit more simple - your on AC3 inverter and we are on hybrids? Obviously the Hybrids measure directly what the PV strings is doing.
In your case, all the AC3 has to go on is the flow reading from the EM115 meter, hence its a derived value (this is why on ac coupled systems, the portals PV string voltage will match the local grid voltage, and not an actual PV panel voltage as it has no way of knowing what that is & therefore simply multiplies the ct current by grid voltage to give you your PV wattage output).
If there are clashes on the comms between the AC3 and the EM115 meter, then although the output from the actual PV system may be static, the inverter will assume 0V x 0A = 0W and you'll get your dropout.

I'd start looking at the rs485 wiring between inverter and your em115 meter - is it using true twisted pairs or has the installer just picked two colours...that kidn of thing. Equally, giv kit doesn't seem to have provision for screens/drains on the cabling so if it's a long run adjacent noisy equipment it may be causing some issues.

https://community.givenergy.cloud/d/1858-incorrect-em115-wiring

SteveCook thanks for that link.
My system is wired correctly with a single wire from a twisted pair to the "+" and a single wire to the "-".
Am I right in my understanding that the wires on the CT clamps should be twisted? Mine weren't so I twisted them together and put some tape around every few inches to stop them unraveling again

Well I've made a few more adjusting but sadly no improvement

1. I set up a large fan blowing air directly at the AC inverter (& dongle)
2. I moved 1 of my (wired not mesh) wi-fi access points right next to the inverter (I have a very large wired / wireless network consisting of 5 ubiquiti access points)
   

[unknown] it seems unlikely its wifi/network related as if it were you’d see quite a few sensors going to zero or unavailable and you’re only seeing it on PV string 2. Assume you don’t see any errors in the givtcp logfile and at around the times it happens the inverter time continues being updated as normal (which I use to indicate successful polling)

Suggest it points to the wiring on the PV CT clamp, a poor connection on the EM115 connection at the inverter end, or the DC inverter on that string having problems. Doesn’t look like inverter temperature is a factor.

Can we see grid voltage to rule that out; although yesterday’s was well within tolerance

geoffreycoan Assume you don’t see any errors in the givtcp logfile

I've never actually looked, will do that now, cheers

geoffreycoan Suggest it points to the wiring on the PV CT clamp

That has been checked, double checked then I've even crimped on new ferrules to be absolutely certain.
Prior to that I swapped out the EM115 and its corresponding CT clamp to 1 I borrowed for testing purposes.

I'm not sure what you meant by "Can we see grid voltage to rule that out; although yesterday’s was well within tolerance"

Graham_Harrison I'm not sure what you meant by "Can we see grid voltage to rule that out; although yesterday’s was well within tolerance"

could you share the history graph of grid voltage for one of the days where you've seen these PV string drops to zero? You shared the graph from yesterday where the voltage looked to be within the national tolerance levels (230V -6% +10%)

When its a sunny day and I am exporting loads my inverters push the grid voltage up above limit and the inverter momentarily shuts itself down. Am wondering whether the same is happening with your PV inverters?

geoffreycoan could you share the history graph of grid voltage for one of the days where you've seen these PV string drops to zero

The dropouts occur EVERY day

It's a pain but I'd get some decent cat5 or 6 and temp run it from inverter to the em115 and see if they go away. This is maybe easier said than done depending on where each ends are!
Or just try switching pairs on the current setup to see if it's any better - you cable might have a nick in it somewhere kind of thing.

hoggy It's a pain but I'd get some decent cat5 or 6 and temp run it from inverter to the em115 and see if they go away

Strangley enough hoggy I was just pricing up a 30m length on Amazon last night with the intention of doing exactly what your suggesting

First thing this morning

I've put my investigation on hold for now as hopefully within the next week I will be having a 2nd AC3 inverter installed + EMS and the current data communication wiring will require a slight change

Graham_Harrison ... and you will have a whole new load of fun about how well the EMS works with GivTCP and the Givenergy portal ....