Jump to content

Victron Multiplus-II Grid-Feed: What am I missing?


JO+e

Recommended Posts

Hi all,

I have a little issue that I haven't quite managed to figure out. It isn't really my intention to feed to the grid, however the fact that it doesn't seem to be working when I've enabled it (as far as my limited understand goes anyway) is bugging me, so I'm hoping somebody can point out what I may be doing wrong.

I have a Victron Multiplus-II 48/3000/35-32 with a system pretty much like this: Victron Grid-tied setup from scratch

I have ESS set up and working (Venus OS on a Raspberry Pi), but I originally noticed that the MPTT was being throttled as soon as the battery was (almost) fully charged. This was confirmed by adding additional load and observing the solar yield reported on the Venus UI (which increases in response to additional load).

The ESS page notes as follows:

Quote

When an ESS system is able to produce more power than it can use and store, it can sell the surplus to the grid; and when it has insufficient energy or power, it automatically buys it from from the grid.

But even with "Feed-in excess solar charger power" enabled in the ESS settings I rarely see any grid feed. It almost seems like the grid set-point overrides grid-feed. I've fiddled with a lot of settings (Venus only) and went through the ESS manuals briefly some time ago but haven't managed to find anything that works. I'm definitely a novice with the Victron stuff, so hopefully somebody can give me some pointers, or tell me what other information is required (at the moment I'm not sure what to provide to assist).

Running Venus firmware v2.33 on the Pi (think I've tried 2 or 3 different versions since the install)

Multiplus reports firmware version 453.

Any help/input will be appreciated.

 

Link to comment
Share on other sites

18 minutes ago, JO+e said:

Any help/input will be appreciated.

What version of Venus are you running? Try the latest candidate release of v2.40. There is a change in there that might help you.

Let me expand on how it works. The venus device will tell the solar chargers to charge the battery to 0.4V higher than the Multi is configured for. The Multi sees this overvoltage and slurps it off (you might slurp off coffee from a very full cup, that's how I visualize it) and feeds it into the grid.

It doesn't work if it does not see this 0.4V overvoltage, and in a few rare cases I've seen that the Multi's calibration plus voltage losses over the cable from the solar charger can be enough to account for the whole 0.4V. If the solar charger sees 0.4V above the setpoint it will throttle, and if the Multi does not see that 0.4V it will not feed in. So it is possible for these two systems to end up in a type of equilibrium that limits the power being fed back in.

This is why SVS (shared voltage sense) was introduced way back in v2.20 (or was it before that? I cannot remember exactly). But SVS occasionally has other issues, so usually you want to leave that off (you can turn it on to see if it helps).

But, in v2.40 there is a change where the Multi and the solar chargers sync voltage, pretty similar to SVS but not involving the battery (which is often very slow, eg BYD only sends a voltage every 2 seconds). And this might also help to make it work better.

Link to comment
Share on other sites

Thanks @plonkster!

Running v2.33 but will give v2.40 a bash tomorrow night (unfortunately the updates wipe my manual tinkering to get the Pylontech and RS485 HAT to work, so not going to tackle that now) and report back.

I've also just enabled SVS and will see how that fares tomorrow - it really helps knowing how this all works together, so your brief but insightful explanation is truly appreciated.

Thanks for the super quick assistance!

Link to comment
Share on other sites

  • 1 month later...

My apologies for the extremely late feedback, but it took me some time to get around to doing some proper investigation.

So it would seem there were possibly two issues originally. SVS and then later Venus fw 2.40 (SVS disabled) seem to have fixed the one issue where grid feed was not working as expected, but what I soon realised was at some point the MPPT charge controller would start throttling when the batteries reach 100% charge.

Subsequently I updated the firmware on the MPPT (v1.46) and Inverter (467), installed an additional battery (Pylontech US3000), set up the ESS assistant again and still the same result. What I have noticed is that there appears to be something wonky with ESS and the MPPT. The bulk LED on the charge controller does not flicker as indicated in the documentation I saw (which is supposed to indicate external control), but does show ESS under state in VictronConnect and shows external control in Venus. The second oddity is that Venus (and the inverter) will show that the battery is in the absorption charge phase when everything works fine (i.e. battery at or near 100%, grid feed working), but at some point the amber absorption led indicator on the MPPT charge controller comes on and that's when the limiting occurs (solar yield limits to ~1900W). I don't understand why these indications aren't in sync between the MPPT and Inverter, or if this behaviour is expected, but it seems rather strange to me, and results in unexpected behaviour. I did however read that the MPPT switches to constant voltage mode when in absorption, so I'm not sure if this is part of the cause. I would have expected that the charge controller LEDs would continue to show external control?

Here's an illustration of what I'm seeing:

1524346688_MPPTLimit.thumb.jpg.2d3e0a1932327d0d1a43b6eba948c50f.jpg

 

Each time I've gone to check on the system when the yield drops to ~1900W I've found the absorption LED lit on the charge controller. Unfortunately VRM only lists external control for the MPPT state so I can't show the correlation in the image (the behaviour after the highlighted time period coincides with the MPPT leaving absorption, clouds, and partial shading from about 3PM so don't worry too much about that).

Any thoughts? Is this behaviour expected, or does it seem like something may not be working as expected?

Any and all input would be greatly appreciated!

Link to comment
Share on other sites

Let me explain how it works, then you can maybe figure out what is going on.

When you feed excess PV into the grid, the solar chargers are instructed to charge the battery to 0.4V higher than the normal charge voltage. I don't want to go into too much detail about the "charge voltage", it depends on many things, but generally it is whatever the BMS requested or what the Multi was configured for.

After this, the Multi will feed as much power into the grid as is required to eliminate this "over voltage" and get the voltage down, while the solar chargers will work as hard as they need to to raise the voltage by 0.4V.

As long as your solar chargers and the Multi is calibrated fairly closely, and in the fast majority of installations they are, this works very well. The additional proviso is that your cables must be adequately sized. If you have an 0.4V voltage drop on the cable at power level X, then the maximum power the system can feed in is X. End of story.

Once the calibration differences combined with any voltage drops on cables come to 0.4V, feed-in is limited.

One way to get around it, which is unfortunately not easy, is to increase the offset. For this you need to root the venus device and set the value internally. This is also where things get even more tricky, cause the Pylontech batteries switch off at 54V, so you don't have much space to work with. The charge voltage for a pylontech battery is hardcoded at 52V, which technically leaves you with a whole 2V or room, though I would not increase the offset by more than 1V.

So here is how you do it. After rooting the venus device, ssh to it and run this command:

dbus -y com.victronenergy.system /Debug/BatteryOperationalLimits/SolarVoltageOffset SetValue 0.6

This will add an additional 0.6V to the 0.4V causing the solar chargers to aim for 53V and the inverter to aim for 52V.

If that improves the feed-in, then you have confirmed the problem. If not, then something else is wrong.

Remember that this setting is volatile... it is lost on reboot. Other things are needed to make it permanent. I'm telling you about it for testing and educational purposes 🙂

Edited by plonkster
Link to comment
Share on other sites

Thanks @plonkster.

I literally just increased the offset as suggested, checked Venus and a few seconds later saw how the battery jumped from 93% (53.something V indicated in Venus  - it was however in absorption, so something with the display is iffy) to 100% and the yield throttled from 2600W to 1900W =/ I checked and the amber absorption LED on the MPPT (150/100) is lit, so ?

I really don't think the cable voltage drops are the issue, but I can try measure these. 

I'm assuming there is a LOT of debug info available if you know what you're doing (sadly I don't), but if I know where to look I might be able to make more sense of it, so if you have any other pointers I'd be very grateful.

 

Edit: Thanks to some cloud the battery discharged a little and it's back up to 2.6 - 2.8 kW. Amber LED off. When it was limiting though, I noted that it was pulling around 200 W from grid with the setpoint at 0 W. And as I'm typing this it's back to 1.9 kW (battery voltage indicated as 53.4 V in VRM).

Edited by JO+e
Updated info
Link to comment
Share on other sites

1 hour ago, JO+e said:

absorption

Ignore the so-called charge state. When using a Lithium battery with a BMS it means nothing, the Multi goes into that state when the battery reaches the requested charge voltage, that is all.

Once you are logged into the venus device, you can also run "dbus-spy". That gives you a text tool to browse the various services. Then navigate to com.victronenergy.vebus.* and from there to /Hub/ChargeVoltage. That's the voltage the solar chargers are aiming for.

Now, if the batteries are closer to 52V than to 53V, then feed-in is working correctly. On the other hand, if the batteries are at 53V, that means feed-in is restricted for some reason (temperature perhaps).

I know the Multi derates to 60% if it gets hot... and 60% of 3000W is 1800W. Maybe?

Link to comment
Share on other sites

21 minutes ago, plonkster said:

Ignore the so-called charge state. When using a Lithium battery with a BMS it means nothing, the Multi goes into that state when the battery reaches the requested charge voltage, that is all.

Sorry, the reason I'm fixated on absorption is that there is a correlation between the behaviour (limiting to 1.8/1.9 kW) and the amber LED on the charge controller, indicating absorption, so as far as I can tell something on the charge controller is causing the behaviour.

/Hub/ChargeVoltage indicates 52.4 V. The MPPT charge voltage is however indicated as 53 V (I'm guessing this is the 0.6 V offset?).

I also considered temperature at some point, but figured that would be a more linear degradation (I guess I was wrong), and also without the correlation to the charge state that I've noted. Is there anywhere I could confirm the temperature or derating? I've never seen a temperature alarm, but I'm also not sure if this would be expected.

 

Link to comment
Share on other sites

24 minutes ago, JO+e said:

Sorry, the reason I'm fixated on absorption is that there is a correlation between the behaviour (limiting to 1.8/1.9 kW) and the amber LED on the charge controller, indicating absorption, so as far as I can tell something on the charge controller is causing the behaviour.

This just means it is at the charge voltage. It correlates with the fact that the battery is full, but it does not cause any of the behaviour.

The MPPT is backing off because if it doesn't it would go over the voltage it has been told to stick with. It can only ramp up the power once the voltage drops, which will only happen once the Multi feeds the excess power into the grid. The Multi is supposed to feed power into the grid attempting to deflate the battery to 52V, while the MPPTs will attempt to keep it pumped up to 53V.

Your Multi is only "deflating" at 1800W (ish) instead of the expected 2400W, therefore the MPPTs have no choice but to back off. Because the voltage is "up there" the Inverter's internal state machinery determines that it has reached absorption. This state does not in itself do anything. The only thing it does is it times how long it has been in this state, because in a legacy setup that time determines when it will go to float... but with a lithium battery with a BMS, there is also no float.

So really... just ignore the "absorption". It just means "the battery is (almost?) full" 🙂

Keep in mind that when running grid-tied the Multi is always limited to 80% to avoid overheating.

Is the Multi running hot perhaps?

Edited by plonkster
Link to comment
Share on other sites

Such a good explanation, even a simpleton like me can follow, thank you!

How do I go about confirming whether the Multi is running hot? It's physically hot to the touch, but is there any way to retrieve temperature readings from it - I'd prefer empirical data to my anecdotal observation (MQTT, console, etc?).

It isn't exactly installed in the most ventilated or coolest spot in the house - it's in the passage cupboard where the geyser used to be installed, so quite a lot of space, but like I said, very little ventilation. For perspective, the passage ambient temp is currently sitting at 32 C, so it's a lot warmer than that in the cupboard itself.  I could quite easily move one of the temperature sensors I have into the cupboard, but again, that's just going to give ambient so I'm not sure how useful that will be.

Or do you maybe suggest running a test with additional cooling to confirm whether it's derating as a result of temperature?

Link to comment
Share on other sites

1 hour ago, JO+e said:

Or do you maybe suggest running a test with additional cooling to confirm whether it's derating as a result of temperature?

Electronic engineers uses these "air in a can" things, its an aerosol can with compressed air for cooling things down. Then you'd spray it directly on the heat sinks inside (while it is running) and see if it improves. Of course engineers also do the reverse... if they are looking for a bug, they use a heat gun to make things hotter. One method I've used is plain air form an air compressor.

Of course you can also use an IR temperature meter to measure how hot it is inside, and so on.

It does sound to me like it might be a temperature issue.

Edited by plonkster
Link to comment
Share on other sites

So rather than try additional cooling I opted to try something else that previously didn't work i.e. set the ESS mode to "Keep batteries charged" (previously with the grid feed not working it would limit the PV power to whatever load was present without feeding anything to the grid) and observe the behaviour in Venus. Although today was around 2-3 C cooler than yesterday I didn't note any throttling of the MPPT. The overall yield for the day was roughly the same (1% less) as yesterday (with the throttling), but I think it was a bit more cloudy today as well. I will keep it set like this for the next few days and see if I notice any throttling, but so far it doesn't seem to be related to temperature.

Just for my benefit, I'm not really sure of what the Multi looks like inside, but I'm assuming the temperature you're referring to (and associated heatsinks) would be for the FETs?

Lastly, I did note that the reported battery voltage was a lot more stable in this configuration, staying roughly at 52.3 V the whole time in contrast to jumping above 53 V as observed yesterday.

Link to comment
Share on other sites

10 hours ago, JO+e said:

Just for my benefit, I'm not really sure of what the Multi looks like inside, but I'm assuming the temperature you're referring to (and associated heatsinks) would be for the FETs?

Yup. I was referring to the FETs.

10 hours ago, JO+e said:

Lastly, I did note that the reported battery voltage was a lot more stable in this configuration, staying roughly at 52.3 V the whole time in contrast to jumping above 53 V as observed yesterday.

Indeed, cause it is feeding the excess into the grid and therefore lowering the voltage.

I don't really understand why KeepBatteriesCharged would cause this difference. It shouldn't... I suspect it's mere coincidence and that the cooler day probably has more to do with it.

Link to comment
Share on other sites

45 minutes ago, plonkster said:

Yup. I was referring to the FETs.

Cool, thank you. Always eager to learn a bit more about how things work.

47 minutes ago, plonkster said:

 I suspect it's mere coincidence and that the cooler day probably has more to do with it.

Quite possible, yeah. As mentioned, I'll keep it in this configuration for a while and see how it behaves. Additionally, I've moved one of my Z-Wave temperature sensors into the cupboard to log the ambient temperature and see whether there's any correlation (crude as it may be) to throttling. 

Is there any way to monitor how much power the inverter is supplying (from DC) in VRM or otherwise? I noticed yesterday that the PV Charger (Venus) would report more than 2.4 kW at times, and it's a bit difficult seeing what the inverter is actually doing (AC loads + Critical loads - Grid =? PV Charger - I'm aware there will be some losses though)

I'll report back once I have any further detail from the testing. I do however suspect that there might be an issue with the configuration of my system as several of the charger settings were not according to the parameter values detailed in the Victron reference info (I didn't install/commission/configure the system initially), but let's first see about confirming or ruling out the temperature component.

Thanks again for all the assistance, and your patience with me 😉

Link to comment
Share on other sites

  • 2 weeks later...

Mr @plonkster, following some further testing I think your assessment is correct as it's doing the same thing in the current configuration.

Yesterday:

MPPTLimit_16_01_2019.thumb.jpg.577a09bd6088afe1f9c899089c640abf.jpg

Today:

MPPTLimit_17_01_2019.thumb.jpg.c88da9c7e52602c9a89f7fde75b01f20.jpg

 

Ambient got to around 37 C yesterday (no additional cooling) and 35 C today (with a fan blowing directly towards the inverter, but with the cover on).

So now I'm not quite sure where to from here (I don't want to void the warranty, but was thinking of removing the cover and testing with the fan again - maybe measuring the heatsink temperature with a DS18B20), but it's got me wondering:

  1. Is this expected? Should the inverter really be overheating/derating at 35C ambient, or is it possible there might be something wrong with the inverter? My personal feeling is that this shouldn't be the case, but I'm also hardly an expert.
  2. From our earlier discourse I'm assuming an alarm/notification isn't to be expected under this condition?

If there might be something wrong with the inverter I'd like to confirm it before contacting the retailer.

Any further guidance would be greatly appreciated.

 

Link to comment
Share on other sites

49 minutes ago, JO+e said:

If there might be something wrong with the inverter I'd like to confirm it before contacting the retailer.

There are no known issues with the 3kva inverters. The manufacturing defect was limited to one batch of 5kvas. So I doubt there is anything wrong with it. It appears to derate beautifully to 60%, which is a known characteristic.

I do agree with you though that 35°C seems low to derate that much. It only derates that heavily around 60°C, if I recall. But we must not confuse ambient with the internal temperature of the inverter.

You will not void the warranty by taking off the front cover. The VE.Bus socket is in there and that's they only way you can even access it for programming, so it would be absurd. Your idea of measuring the heatsink temperature seems like a good one to me.

Link to comment
Share on other sites

Cool, thank you!

Hopefully some further poking will help resolve my concern. I wasn't even aware of the defect that affected the 5kVAs, so it was more of a general question regarding anything that could potentially lead to overheating. I know it has an internal fan that sounded a little iffy to me since I got it, but it didn't sound really broken (and does still spin).

Sorry for the stupid question about the cover. As mentioned previously, I didn't perform the installation and thought it was just the black panel at the bottom that comes off for all the wiring/VE.Bus etc. I just have a routed ethernet cable I fiddle with occasionally :)

Let's see if I can get around to getting a sonoff set up with a temp probe over the weekend and see what it's getting up to :)

Link to comment
Share on other sites

  • 2 years later...

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.

×
×
  • Create New...