question

mvader (Victron Energy) avatar image
mvader (Victron Energy) asked

Venus OS v3.00~36 available for beta testing

UPDATE 2023-04-28: new thread, v3.00~37

UPDATE 2023-04-22: v3.00~36 is now available; details below.

UPDATE 2023-04-07: v3.00~32 is now available; details below.


Good day!

A new version available, and a fresh new thread to put comments in.

Per just now, v3.00~29 is available for testing. Two minor updates compared to the previous version.

The status of v3.00 is the same. We're finishing up, ie. testing for stability, making sure all is as it should be; and making good progress with that.


What is new in v3.00?

  • Peak shaving for ESS
  • Disable the touch feature on a GX Touch: great for rental and other systems where the whish is to display system status while not allowing any other user interaction.
  • Improved ESS control speed - for EM540 meters only
  • RV-C canbus protocol for (mostly USA-built) motorhomes: lots of improvements
  • Many other small changes, see below.


Introduction to Venus OS beta testing

In case you don't know what this message is about, please start with reading this link, which explains the Venus OS beta program. Venus OS is the software running on all our GX devices, such as the Cerbo GX.


How to post an issue?

By posting an answer below. Do please keep all findings organised: one answer issue per issue. So first check the existing threads if your issue has already been seen. And if it does, add a comment saying "me too". And preferably some more details. And in case its not listed yet, add a new Answer.

Note that Answers and Comments are two different things on this portal.

Lastly, before posting issues, preferably first revert to the latest official release (v2.93), to double check if the issue you're seeing was present there as well. Regressions require a different treatment than other issues and bugs.

And include information about the results/differences in behaviour in your report.


Changes v3.00~32 -> v3.00~36

  • For the EM24, ET112, ET330 and ABB meters, the ESS control loop is reverted to be slower and (thus) more stable. The response times are now similar as they are in v2.93 / before working on improving the response speeds. The ESS response speed for the (faster metering) EM540 is still the same: much faster than the other meters. Its actually even a bit faster than it was before v3.00~35. As also discussed in the comments of v3.00~32, having to revert to the previous version for the slow meters is quite a disappointment. The reason is that while working stable and fast for some installations, its also unstable in others. Whether we'll still make improvements on it remains to be seen. In essence those meters are slow, and then to make a stable and fast control loop is difficult and bound to create issues in some system.
  • Fix connection getting lost every few minutes to MultiPlus 2x120V models. Bug was introduced during v3.00 development, and is not in v2.93. Thank you @derrick thomas for reporting
  • Some more fixes.

.

Changes v3.00~29 -> v3.00~32

General

  • Fix bug related to kWh calculations (= source data for graphs on VRM Dashboards): If the position of a PV-inverter was changed, or the AC input (grid/genset) was changed, this would not take effect until a reboot.
  • Fix bug that could cause the reported state for a system of six inverter/chargers or more to be stuck; when very quickly switching the system on and then off again.

Modbus-TCP

  • Add reserved registers to plug a hole in the solarcharger range between 3704 and 3707.
    • This gap was inadvertently created in Venus v2.80
    • This fix allows fetching the entire block from 3700 to 3730 in a single call.
    • This bug only affects MPPT-RS; the data in this block is related to Multiple trackers.

Venus OS large & developer:

  • Fix mDNS service advertisements for signalk-server; includes updating avahi-daemon from v0.7 to v0.8.
  • Add python3-can as optional package (see opkg in commandline instructions for how to install)


Changes v3.00~27 -> v3.00~29

DVCC:

  • Fix bug where data from the battery could get stuck/freeze until GX devices was rebooted. We've seen this happen in a handful of support cases; and it related to certain timing in the canbus messages. Once properly tested, we might release an interim official version v2.94 that includes this fix soon.


NMEA2000-out:

  • Fix PGN Fluid Level 127505 - fluid level instances not being writable. The fluid level instance is also refered to as the tank instance or data instance. The other instance, Device instance, was working fine and changeable.

    Changing the fluid level is necessary on older Raymarine Axioms, since they have a limitation that the fluid instance needs to be same or lower than 5. We're not aware of any other situations where changing that instance is required; but if you do pls leave a comment below, so we can add it into the documentation and make everyones life wrt dealing with instancing easier.

    Lastly, Thank you DevoDave for reporting!

  • Add transmission of battery selected as the System battery in GX settings, with a fixed device instance, as well as battery instance of 239. The aim of sending this PGN is that there is one instance, always the same, for the main battery in the system. Instead of a system with a Lynx Smart BMS using instance 0, and a system with a SmartShunt using different instances.
  • Update DC Detailed Status PGN: Add Amp Hours field, complies with NMEA200 v2.000.
  • Fix DC Detailed Status PGN, Time Remaining field: sometimes it showed zero when it shouldn't.


Touch disable function (read further down to learn what that is):

  • As additional measure to prevent lockout, re-enable touch if the IO that was configured for this is disabled or re-assigned.


Developers / internal:

  • Serial number was added in a TXT record in the mDNS response, for better "pairing" with other devices such as EV Charging Station.
  • Update VE.Bus inverter/charger driver, adding VE.Bus BMS Pre-Alarm data and some more; to be used by gui and VRM in the future.
  • Identify/recognise Pylontech Pelio-L battery. Just recognition for now, a future update will take care of automatic setting of DVCC and SVS and such.


Changes v3.00~26 -> v3.00~27


  • RV-C protocol, related to VE.Bus Inverter/charger DGNs

    • Add the ability to change the maximum charge current using the CHARGER_CONFIGURATION_COMMAND DGN. This is a volatile setting and will be reset to the value the inverter/charger was configured with once the unit restarts.
    • Add support for CHARGER_CONFIGURATION_STATUS to report the maximum charge current.
    • Report maximum charger current as percentage in CHARGER_CONFIGURATION_STATUS_2 instead of the fixed 100%.


  • Node-RED (part of Venus OS Large)
    • node-red-contrib-victron v1.4.32

    • Move ActiveSocLimit to new path
    • Fix pathData undefined error
    • Added EnableDisplay and Scheduled mode (evcharger) (#169)
    • Rename VE.Bus System reset to VE.Bus system restart (#170)
    • Updated documentation
    • Add L3 current for input-grid meter
    • Update dependencies


Changes v3.00~25 -> v3.00~26
All rather minor:

General

  • Recognise Batrium BMS (this actually wasn't included in the previous version). No functional changes.

Signalk-Server (part of Venus OS Large)

  • minor change: call sync after modifying files before starting the server, to make sure they're written to disk immediately.

Changes v3.00~24 -> v3.00~25

Signalk-Server (part of Venus OS Large)

  • On start-up, check if any of the now default installed plugins was installed manually, and remove that one.


Changes v3.00~23 -> v3.00~24

General

  • Recognise Batrium BMS; no functional change


Signalk-Server (part of Venus OS Large)

  • Add and default-enable plugins sending NMEA data out on TCP, includes AIS data
    • With this change, the GX device is a LAN and wireless AIS and navigation server for popular apps like Navionics, iSailor, iNavX, and Aqua Map on phones and tablets. This blog post by S/V Renaissance explains it nicely; but ignore all explanations about configuring plugins: that is all already done. Two examples: (1) Aqua Map App (link to Wifi connections page), (2) Navionics Boating App (link to AIS feature page).
    • This feature requires a NMEA2000 connected AIS receiver on board. No internet is needed.
    • The data is available as NMEA0183 packets on the default TCP port (10110), as well as signalk messages on the default websocket port.
    • Note that this is not tested. If you have a navigation system, please test and let us know if it works; and if indeed no further configuration was required.
    • All powered by the opensource signalk-server.
  • Fix pre-installing venus-signalk-plugin, it was missing. Unknown since which version; on v2.93 it was missing; most likely many prior versions as well.
  • Fix disabling updates of the pre-installed plugins
  • Known issue with signalk-server: mDNS advertisements most likely don't work.


Changes v3.00~22 -> v3.00~23

  • Internal changes


Changes v3.00~21 -> v3.00~22

  • RV-C out: increase timeouts on some battery and tank messages to match the maximum send interval. The previous timeout was too strict.


Changes v3.00~20 -> v3.00~21

Developer / Venus OS large related:

  • Node-RED (v1.4.29 & v1.4.30)
    • Fix default for the new "Only changes" setting, it now defaults to off.
    • Add option to round float values
    • Filling out service and path of a node is now required
    • Fix reporting relay state changes when the "only changes" option active. As well as other changes that use the dbus PropertiesChanged instead of ItemsChanged signal.
    • New paths:
      • /Settings/SystemSetup/AcInput1 and 2 for input and output-settings
      • /Ac/Frequency for input-gridmeter
    • Update documentation
    • Thank you @mr-manuel for your help with above!
  • MQTT: Improve response when receiving a R/<portalid here>/system/0/Serial message. Instead of just activating the keep-alive, it will now always also respond by publishing the VRM Portal id.


Changes v3.00~19 -> v3.00~20

  • Fix the issue that was introduced in ~19.


Changes v3.00~18 -> v3.00~19

  • Internal changes only
  • Has known issue related to MQTT and also the HTML5-App and possibly VRM real time controls. A new version, v3.00~20 is expected later today.


Changes v3.00~17 -> v3.00~18

  • Node-RED: fix the Node-RED Victron control nodes. The previous version (v3.00~17) broke all control nodes.


Changes v3.00~15 -> v3.00~17

General

  • ESS: Add peak shaving option (by observing the AC input current limit using PowerAssist)
    • Peak shaving already worked in Keep batteries charged mode; no changes there, other than making it more obvious by adding in the Peak shaving menu entry.
    • Peak shaving did not work will in the Optimised modes. It did work as long as battery SOC was above the configured ESS Minimum SOC level, but once discharged there the system would not assist the loads. This is now solved: use the new peak shaving option in the ESS menu, to let the system keep PowerAssisting when needed. And as soon as the peak is over, it will recharge the battery using power from the grid, while prioritising solar. Note that there is a 5% hysteresis on that: lets say Minimum SOC is set to 80%, it will then start recharging back to that 80% once (by peak shaving) the battery dropped to 75%.
    • The default setting, when using the Optimised mode, is off. To not change behaviour of running systems.
    • Warning: this works for the critical loads only. Not by energy meter.
    • For further details, see screenshots below.
  • GX Touch: add option to disable the touch input. See settings -> I/O -> Digital inputs for the new feature. This is intended for systems where the GX Touch is wanted to show the users what the system is doing; but nothing else. The status of the touch (enabled/disabled) is toggled by pulling the gpio down to ground. More information about locking a system down is here; that text was recently updated, amongst other things a "Hardening a GX device" chapter was added.
  • EV Charging Station: show progress indication during a firmware update.
  • GUI, in the Multi & Quattro advanced menu, reword Reset to Restart. Less ambiguous in relation to a factory reset.
  • VRMLogger: add various new fields to be sent to VRM to improve the dashboard and dashboard controls.
  • Support the Fronius Tauro via SolarAPI: This is a rare edge case. Customers should use Sunspec instead of SolarAPI whenever possible, as per documentation. This change allows using SolarAPI in cases where Sunspec cannot be used. For example where modbus-RTU is in use on the DataManager and modbus-TCP cannot be used.
  • Energy meters:
    • Fix issue with EM24 reported values being more erratic since earlier v3.00 releases. Now the behaviour is back to how it was pre v3.00
    • Fix issue with EM540 phase sequence check for three phase systems. It reported the phase sequence as being incorrect while it wasn't.


Few changes that were also made available in v2.93:

  • Fix a problem where Remote Console wouldn't work, and required a factory reset to be repaired. Any systems now having the issues will be fixed automatically once updated to v2.93 or v3.00~16.
  • Fix timezone bug related to ESS Scheduled Charging: after changing the time zone, a reboot was required to make that use the new time zone. Bug was introduced in v2.80. Thank you Patrick M. for your help on this!
  • Add the new Scheduled mode and more for the EV Charging Station. Fixes "unknown" showing when the EV Charging Station is configured for that mode. For more information on that, read here.


Developer related

  • RaspberryPis: fix bluetooth connection to VictronConnect using the rpi-built-in bluetooth radio. This was broken during v3.00 development.
  • Node-RED, update node-red-contrib-victron from v1.4.26 to v1.4.27:
    • Added option to input nodes to only broadcast changed values (see issue #147)
    • Added missing input-vebus paths: /Bms/PreAlarm and /VebusChargeState
    • Update documentation
  • Update the uPnP data (see this commit)
  • Update PHP from v7.4.28 to v7.4.33; next step is no longer using PHP at all.


Internals

  • Update MK3 firmware from v215 to v216. If already on v215, this update will happen automatically. If running an earlier version, the update will need to be triggered manually, as explained below.
  • Prepare VE.Bus firmware updater for some internal changes.


Changes v3.00~14 -> v3.00~15

General

  • add logging of generator runtime to VRM Portal (actual visualisation might still need work)
  • ESS: add support for all EM300 meters
  • ESS: add option to scheduled charging to allow discharging the battery (if SOC is below the configured minimum) while in the window; and more.
  • EV Charger: various features and fixes
  • Add support for Pylontech batteires having 16 cells in series (rather than usual 15 cells). Thank you Brian Finley!
  • Force good settings for Hubble batteries (DVCC on, STS off, SVS recommended off)

Protocols

  • Lots of improvements concerning RV-C; to be detailed
  • Modbus-TCP: fix issue related to inverter/charger state vs “ext control”.


Changes v3.00~8 -> v3.00~14

General

  • Fix Zigbee (DRF2685C) detection. (also available in Venus OS v2.93)
  • VE.Bus BMS v2: Improvements to remote firmware update feature
  • VE.Bus BMS v2: Add sending the pre-alarm status to VRM.
  • Bluetooth improvements for Cerbo GX with serial-nr HQ2208 and later, as well as all Cerbo-S GX production. The temperature issue which makes the built-in bluetooth unreliable/unusable no longer applies:
    • Per mentioned serial numbers, a second Bluetooth radio has been added to the hardware. And per v3.00~14, Venus OS also uses that one; rather than the unreliable one. In more detail:
    • For VictronConnect connection, new radio is used.
    • For Ruuvis, all radios are used including any externally inserted ones (USB).
  • Fix bugs in the GUI that could cause it to show glitches and/or get stuck.
  • Add Thai language


Modbus-TCP

  • Fix register 31, the /State path for the VE.Bus inverter/chargers. Since v2.89 it returned value 252, Ext Control, for systems where the Inverter/charger is externally controlled. Instead of 3 (Bulk), 4 (Absorption) and so forth. Breaking customer integrations. (also available in Venus OS v2.93)


Venus OS Large

  • Update signalk-server from 1.44 to 1.46.2


Changes v3.00~4 -> v3.00~8

General

  • Multi/Quattro: Fix issue causing a repetitive low battery alarm in case the battery is disconnected
  • ESS: Fix bug introduced in v3.00~2 related to external control mode.
  • Add progress indicator (0 to 100%) to Venus OS firmware download.
  • Fix tank temperature unit (Fahrenheit)
  • Add new VE.Bus product ids (2681, 2723, 2766, 2776)
  • Add diagnostic fields for Multis/Quattros (uptime counter per unit, terminal voltage, Vsense voltage for L1 master and DC ripple)


Modbus-TCP

  • com.victronenergy.battery: Add /Dc/0/Power and /Mode, add missing fields to /State enum
  • com.victronenergy.digitalinput: Add "Generator" to /Type enum
  • com.victronenergy.multi (applies to RS products): fix in enum /State


Venus OS Large


Changes v2.92 -> v3.00~4


  • For systems having multiple BMSes connected, allow selecting which one should be used for DVCC. It also allows the use of a BMV for SOC tracking -- by selecting BMV as battery monitor -- while still using the BMS for DVCC. A bit of a niche issue for special systems, more technical background here: https://github.com/victronenergy/venus/issues/901 (but please don't start posting on our github - thanks).
  • Detect Hubble batteries, untested as of yet.
  • Add Polish translations, thank you Jakub T for helping with that!
  • ESS: increase control speed of systems using todays supported meters (ET112, ET330, EM24, and also the ABB ones) a little. But, for faster meters, such as the EM540 which is not available yet, increase it very significantly. Requires updating the built-in MK3-chip, see next bullet.
  • Include a newer version of onboard MK3-chip firmware, v215. Updating that has a 1 to 10% chance of a short system outage (Multi/Quattro shuts down in VE.Bus Error 14, restarts after 30 seconds). And therefore the update needs to be initiated manually from within the menu. Note that this is a reversible action, no need to worry about being unable to roll back. After updating, and then rolling Venus OS back to v2.92 or some other earlier version, the MK3 will automatically (and silently) be downgraded to the for that version of Venus OS required MK3-firmware version.
  • Improve text for tank sensor name in pump configuration.
  • ESS: Fix bug where PV is not used for loads when scheduled charging to 100% during daytime
  • Fix bug where a PV-inverter on AC-in-2 would not be shown in the ESS overview. The workaround was to configure the PV inverter as being on AC-in-1. Not needed anymore.


Venus OS Large


Under water / developer

  • Various small under the hood changes, mostly resulting in small reductions of CPU load
  • Replace Hiawatha webserver with nginx; which is better kept up to date (security)
  • Include various OE Dunfell fixes
  • DVCC: simplify transmission of the charge voltage setpoint by sending it always, rather than only if devices that work with it are detected. This won't make a difference to any commonly known system type.


MK3 firmware update related screenshots

On systems not updated yet, you'll see this (after going to the Device list, and then into the MultiPlus, Quattro or EasySolar listing):

1666825009093.png


And for anyone wanting to make sure its updated, in that same menu, scrolling all the way down to the Device submenu, and going in there, and scrolling down again, you'll find this listing:

1666825094337.png

Wherein the 212 digits are the version number, and thats the old one. And here is what you see after the update:


1666825165306.png


Note that this is necessary only for this one update. Hereafter, updates will be silent again, like they used to be; since there is no longer a risk of a short system outage.



Prior threads:

Venus OS
7 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi! v3.00~32 is now available; mostly minor changes. happy easter.



0 Likes 0 ·
koro avatar image koro mvader (Victron Energy) ♦♦ commented ·

Hi Victron Team.
Today I've made the update to
Venus OS v3.00~36. Since this time, the bluetooth menue is no longer available.

Greetings from Austria

Klaus

0 Likes 0 ·
maddog75 avatar image maddog75 mvader (Victron Energy) ♦♦ commented ·

Hi, any chance you could upload the previous v3.00~32 cerbogx sdcard install file to the repo please? It would be great to test the ESS fast control loop with my ET112.
You only seem to keep the latest version in https://updates.victronenergy.com/feeds/venus/candidate/images/cerbosgx/

Thanks

0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·

Hi all, FYI Venus OS v2.94 was released officially today:


https://professional.victronenergy.com/news/detail/196/


0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi, v3.00~36 is now available for testing. Quite a major change - and some might call it a set back, for ESS control speed: except for systems using the EM540 meters, the control speed aka response speed is back to more or less the way it was. More information on that in above changelog and an already existing thread below.
-1 Like -1 ·
juf avatar image juf mvader (Victron Energy) ♦♦ commented ·

@mvader (Victron Energy)

For the EM24, ET112, ET330 and ABB meters, the ESS control loop is reverted to be slower and (thus) more stable. The response times are now similar as they are in v2.93 / before working on improving the response speeds. The ESS response speed for the (faster metering) EM540 is still the same: much faster than the other meters. Its actually even a bit faster than it was before v3.00~35.

this are a horrible decision because the EM540 only have RS485 and so it is not usable if my Solarsystem and grid-meter will only have network connection. so pls give the user a chance to decide by his self.


regards

4 Likes 4 ·
crash-override avatar image crash-override mvader (Victron Energy) ♦♦ commented ·
Hi mvader,

I use a custom dbus service for my smartmeter, because I installed it years before and dont want to change something. It is connected to a plc and then polled by Cerbo via modified dbus python script.

Update rate is nearly 2Hz and the faster ESS control loop works perfectly, no grid feed in anymore.

How can I prevent my custom service to get the lower ESS control loop speed, can I mark my custom service as EM540 (which number to change?) so I can update my Cerbo without loosing fast loop speed.

thanks

Carsten


0 Likes 0 ·
28 Answers
hominidae avatar image
hominidae answered ·

..upgraded from b26 to b29.

There are "gllitsches" in the Cerbo GX Remote Console, where thhe tile/icon of my (SMA) grid tied PV inverter disappears from view. Although battery charging remains, as if data is still available.

Here's a pic from VRM short after that event

1680435371697.png

...still missing, while on GX, via local LAN it reappeared 3 mins ago:

1680435387675.png

...the remote console via VRM is also fine (again).

Edit: Sorry, I did not manage to take a screenshot from the remote console, with the PV tile gone missing.

What happened was, the tile was showing fine, then it disappeared. I checked on VRM, where the tile was still there, but without a value (as shown). Battery was still charging. After going back to the remote console, the tile was back fine (as shown, I think, I reconnected to it). VRM data was back approx. 10mins later. Logs from my wallbox, which also integrates with the Cerbo, as well as with the SMA inverter directly, did not show a problem with the data itself.


1680435371697.png (39.2 KiB)
1680435387675.png (43.6 KiB)
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

frederic avatar image
frederic answered ·

Currently testing peak shaving in Belgium in a setup with 3xMultiPlus-II 3kVA, or total 9 kVA.

Firmware on Cerbo is v3.00~29

In Belgium if you don't want to pay extra on the electricity bill, the continious power taken from the grid needs te be lower then 2500W. I think it would be easier if you could set a peak shaving value '2500W' like the 'grid setpoint' in the ESS menu also has a Watt-value that can be set.


Because this is not possible (for the moment?) we have to limit the input current accordingly.

Imax = 2500 W / (sq(3) * 400 V) = 3,6 A.

In the MultiPlus, the maximum is set at 3.5A.


In the following situation with 3011W from the grid (adaption time > 1 min) power is taken from the grid on L2 and L3 and passed on DC-bus through L1 to not discharge the battery to much (I guess? Desired battery SoC is 35% in this setup)


A few solutions could be implemented here where +-3600W is needed on L1 (3845 load - solar 267W on L1):

Solution 1: Take 2500W from the grid and assist 1100W on L1, L2 and L3 did not need any action in here. By far the easiest solution in my opinion.

Solution 2: Take 2500W from the battery and assist 1100W from the grid. Important sidenote: this would be unnescessairy discharge from the battery.

Solutions 3 (now): take energy from the grid on L2 and L3 and convert AC energy in DC energy and back to AC energy. But even now, the grid power was more then 2500W. So the installation did not need to take more energy from the grid on L2 and L3.


For the moment the peak shaving is helping already, but could still improve heavily.

Hope this helps for coding.foto-1.jpg



foto-1.jpg (165.3 KiB)
2 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

frederic avatar image frederic commented ·

This is the second situation where improvent would be possible with a total "peak shaving" value in Watt.

Too much energy from the battery is drawn (also with adapation time > 1 min)

foto-2.jpg

Foto 2.jpg

0 Likes 0 ·
foto-2.jpg (84.1 KiB)
foto-2.jpg (84.1 KiB)
frederic avatar image frederic commented ·
@mvader (Victron Energy) can this be looked in to also? To change the peak shaving from an Amp value to a Watt value over the 3 phases? Ideal this setting would come under ESS
0 Likes 0 ·
frederic avatar image
frederic answered ·

Another issue with the same installation, beta ~29

When the input current is limited to 3.5A like described above, this should only be for taking energy from the grid. Today the batteries were 100% full and I guess that the MultiPlus tried to limit the SMA inverter (8kW), but couldn't.

What happened is that the voltage on the DC bus (with Pylontech US2000) got up to 54V and shut down the MultiPlus. 1680538155404.png


After the restart the voltage was rising again so we went back to the latest non-beta version. Voltage was stable there at 52.3V when batteries were 100% full.



1680538155404.png (26.3 KiB)
3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

patricknl avatar image patricknl commented ·
Could it be the MP wanted to limit the feed-in to the grid? Don't know if this is a thing in Belgium...?
0 Likes 0 ·
frederic avatar image frederic patricknl commented ·
Yes, although there is no limit I think that’ s what the Multi’s were trying to do. Limit the power due to the new factor peak shaving (which does not need to work in the way giving power to the grid, peak shaving needs to work only when taking power from the grid) and then overcharged the batteries.
0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ frederic commented ·
hi, this is just a note that I've seen this; it will be looked into this soon.
0 Likes 0 ·
Jan Wellergård avatar image
Jan Wellergård answered ·

NMEA 2000

Just a comment on data instances. Having a Seatalk NG (N2K) displays , I would say that Raymarine generally use data instances as the way to distinguish between different units and not Device instances for all types of units (batteries, tanks). Hence , the user needs to do the rather complicated configuration using the command line. It would be very useful to be able to change /update data instances via the GUI.

Instance 0 is the first instance, 1 the second. So battery 1 ( house) is data instance 0. One can’t randomly choose data instances.

10 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·

Hey @janwellergard , I thought the same. A few weeks ago one us spent several days setting up tank monitoring on a Raymarine, Garmin, Navico and Furuno too iirc.


And I expected the outcome to be that we need to make a UI to change those data instances.

But to, my surprise, newer Raymarines no longer need that.

We tested on a Raymarine Axiom+ with latest firmware (LH4.3.54).

So perhaps a Raymarine update on your system solves the same?

The other limitation older Raymarines have/ older Raymarine firmware has is this maximum number, and tank instance of 5 tanks.

Long story short, making a page to change those tank instances is a real effort for us; not something small to put in between jobs, its now even more low prio: not sure if we’ll ever get to it.


Documentation on how to otherwise change tank instances as well as other data instances is here:

https://www.victronenergy.com/live/ve.can:changing_nmea2000_instances#changing_nmea2000_instances.


0 Likes 0 ·
Jan Wellergård avatar image Jan Wellergård mvader (Victron Energy) ♦♦ commented ·

Unfortunately I don’t have a Axiom but their instrument i70(display). I think that is on their latest firmware.

Saw the same issue on YachDevices WiFi bridge. They have web gauges that display data. (The data flickers between different values. )

I understand that the customer base here having this is small compared to other users and there is a work-around doing it via the cmd line. So I’m fine with that as long as the changes are kept (which they were last summer).

0 Likes 0 ·
kris avatar image kris Jan Wellergård commented ·

I confirm the problem with Yacht Devices, I have 2 x YURI-04 in the system and a very extensive Victron system and this is the problem you are writing about.

0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ kris commented ·

Ok I'll look into the Raymarine i70 display as well; good to know how that works here also for support.


For the YachtDevices WiFi bridge: that I don't understand yet. In our NMEA2000-out feature, the fluid instances are all unique.

Here is what either is or will be in the GX manuals about that:

"The GX device automatically numbers each tank with a unique device instance and tank instance. They are made the same. This automatic numbering is done specifically, and only, for tank levels; to make the process of showing them properly on all different brands and types of MFDs as simply as possible.".

And I checked just now: thats nothing new, this was from day one, Venus OS v2.60 released in Sept 7th 2020 on which we first added tank levels to NMEA2000-out.


So, I don't fully understand yet what happened with your YachtDevices WiFi, ie. why it didn’t work out of the box, since all fluid/tank/data instances are unique out of the box. As well as the n2k device instances.



ps, for anyone else reading this, good to know that besides using the commandline to change the NMEA2000 fluid instances, you can also use an Actisense USB/NMEA2000 interface device.


2 Likes 2 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ mvader (Victron Energy) ♦♦ commented ·

Hey both, to add:

We tested with a Raymarine i70s (not the same as i70), with 6 tank senders, and all showed up Ok.


It was a bit of a hassle to get it shown, but that was mostly due to getting familiar with the menus of the i70s.

Then, in an attempt to find some authorative text from Raymarine about this, I found this link in Google:

https://forum.raymarine.com/showthread.php?tid=2543.

Which FAQ PDF says this: “for [example] fuel tanks we can support up to 5 tanks on our Dashboard data page and will look for instances from 0- 4.”

But that is out of date. I’ll see if we can get or find updated information.

Still the summary remains the same: in combination with Raymarine, most of the pain was due to the bug on our side (n2k fluid instance not configurable) in combination with the max 5 limit on older Raymarine devices/firmware. Which is now solved per v3.00 and for which we’ll release a v2.94 as well.

1 Like 1 ·
Jan Wellergård avatar image Jan Wellergård mvader (Victron Energy) ♦♦ commented ·

Do you know which software version the i70s was on?Mine is not on the very latest.

0 Likes 0 ·
Jan Wellergård avatar image Jan Wellergård mvader (Victron Energy) ♦♦ commented ·

I had the same issue with battery data, (values flucuating as it showed data for multiple data instances for the same instance. To my memory alla device instances had the same data instance. I gave them unique data instaces and then it worked. I attach my current config below.

0 Likes 0 ·
Jan Wellergård avatar image Jan Wellergård mvader (Victron Energy) ♦♦ commented ·

But if you have 1 fuel, one water, the water tank would be given data instance 0 right?

0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ Jan Wellergård commented ·

Hi @janwellergard was this solved for you eventually?

Sorry to reply only now. I missed your earlier answers.

0 Likes 0 ·
Jan Wellergård avatar image Jan Wellergård mvader (Victron Energy) ♦♦ commented ·

The solution (for me) is to change Data instances and that works for v 2.9x


Tank instances:

  • i70 don't show non-fuel tanks and I only have one fuel tank so I can't validate that multiple tanks work
  • I use Yacht Devices to show tank status. Yacht Device also only use Data instances to separate different tanks and they seem not to honor the type of tank.


Battery (or rather DC sources):

  • i70 uses Data instances to separate DC sources/uses.


When v.3 is released, I'll reset the NMEA2000 setting to see where I still need to use specific data instances for DC and tank. For example, I've stopped using Yacht Devices to display tank data as I now as a Victron display, so I might be able to stick to default for tanks.


0 Likes 0 ·
djbower1 avatar image
djbower1 answered ·

@mvader (Victron Energy) Will the new ESS peak shaving option give the ability to use a small generator on the Quattro's?

4 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

nickdb avatar image nickdb ♦♦ commented ·
You can’t have both a generator and grid working simultaneously. AC1 has preference, but it is one or the other.
0 Likes 0 ·
djbower1 avatar image djbower1 nickdb ♦♦ commented ·

Let me clarify....In an ESS setup you would you the grid on AC2 and Genset on AC1 which always takes priority. It is recommend the genset is capable of matching your max load for obviously reasons.


My question is. In the event of grid failure. Genset starts up but is only capable of 3kw but you have a 6kw load. If my interpretation of peak shaving is correct, Then the inverter should reduce charge current if needed and cover the loads?



0 Likes 0 ·
Kevin Windrem avatar image Kevin Windrem djbower1 commented ·
This is the way the basic Multi/Quattro works. Loads always have priority over charging. It has nothing to do with peak shaving which is only for grid on ESS.
0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi, the feature we have for that is called Power Assist. Which is not related to ESS or the new ESS Peak shaving function.


all Multi and Quattro models have a minimum AC input current limit; and that number defines the minimum size generator.

Please contact an experienced dealer to help select a proper one, to prevent having a too small one that doesn’t work with your Quattro.

0 Likes 0 ·
Tom Büschgens avatar image
Tom Büschgens answered ·

Version 3.00~29:

With starting solar production the ESS starts to constantly draw ~80-120W from the grid, even though plenty of battery is still available. This continues even during the charge of the battery. Shifting the "grid setpoint" manually (e.g. to -120W) works as desired, though.


openwb.png


See image for some details (from openWB - red shows grid consumption, yellow battery, green solar production). Clearly shows that a solid red part appears shortly after sunrise, 7:10ish on the diagram, even though plenty of battery is still available (roughly 50%, minimum SOC set to 10%)

Here also a snapshot from VRM showing point in time - still drawing ~100W or more while in parallel charging the battery with no significant load changes during that time:

vrm.png


Unsure whether this might be related to peak shaving settings or not (haven't found a switch to completely disable it, btw).

Behaviour was not apparent with VenusOS 2.93

Hardware: Multiplus II 48/5000, Rasp 2B with VenusOS.

Happy to test or to provide more details if usefull.

Best, Tom


openwb.png (60.9 KiB)
vrm.png (121.1 KiB)
3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi Tom, the next beta version will have some changes; can you please look out for those and let me know if all is well then?


Thank you.

1 Like 1 ·
Tom Büschgens avatar image Tom Büschgens mvader (Victron Energy) ♦♦ commented ·
Sure. First thing I'll do once I see ther announcement.
0 Likes 0 ·
Tom Büschgens avatar image Tom Büschgens mvader (Victron Energy) ♦♦ commented ·
Upgraded from ~.29 to ~.36.


Behaviour unchanged, still a continuous grid consumption as soon as solar production starts.

Besides that everything is running smoothly

0 Likes 0 ·
derrick thomas avatar image
derrick thomas answered ·

3.00~32 large causes ve.bus to disconnect for a second every few minutes.

my setup...parallel 24/3000/2x120, cerbo gx, daly serial bms using louis driver, 100/50 mppt on ve.direct, 100/30 mppt on victron ve.direct to usb adapter, guimods (same behavior with or without guimods installed), ESS with no grid feedback, dvcc enabled. Node red enabled.

Problem is not present on 2.93. I have went back and forth between 2.93 and 3.00~32 at least a half dozen times and the problem is persistent every time on 3.00~32. Every time I revert back to 2.93 the problem goes away and does not reappear.

there are no errors coming up when this occurs. The ac input and ac output tiles simply disappear for a second, then come right back. If I go to the inverters in device list, ve.bus shows disconnected for a second then comes back, agreeing with the gui display tiles.

If not for me using node red to enable/disable one of the relays based on ac input availability I would not have even known about this issue since it does not throw any errors and you would miss it if you weren't staring at the gui right when it happens.


edit to add a few more points of information:

I have went so far as to disable nodered and uninstall louis' dbus-serialbattery driver. I have also removed guimods as well as any other mods to the system. With clean 3.00~32 firmware running ESS and dvcc enabled Ve.bus still continues to disconnect periodically. sometimes it may be an hour, other times it may happen a few times within just a few minutes. A stated before, if I revert back to 2.93 ve.bus is stable with no disconnects.


and yet even more testing...

A little more info today after some more testing....I removed ess assistant and the problem persisted. I did however discover that node red does seem to change things a bit. Before migrating any of the flows victron nodes to the newer settings, enabling/disabling node red had no effect on the frequency or duration of ve.bus disconnects. However, after migrating all the flows victron nodes to the new settings, the frequency and duration of the ve.bus disconnects increased dramatically; even to the point that it would stay disconnected for several seconds at a time and reconnecting for just a second before disconnecting again. This behavior continues until disabling node red in settings. This could be reproduced every time after re-enabling node red simply by changing switch positions on the inverter tab in the gui. dbus round trip times would increase into the hundreds every time ve.bus would disconnect. I don't know if this is a result of the ve.bus disconnects or possibly a contributing factor. It seems possibly there is a timeout issue or something is hogging system resources. Thats not my area of expertise but I wanted to provide as much information as possible.

8 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

nickdb avatar image nickdb ♦♦ commented ·
Have you updated the mk3 in the inverter? This is listed under the inverter tab and iirc advanced or device tab. There was a note about this in an early RC. Not sure if it’s related but worth a try.
0 Likes 0 ·
derrick thomas avatar image derrick thomas nickdb ♦♦ commented ·
Yes when I originally went from 2.92 to 3.00~32 it prompted me to update the mk3 driver, which I did.
0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ derrick thomas commented ·
hi @derrick thomas , thank you for the report. Is it OK if we login and have a look?
0 Likes 0 ·
derrick thomas avatar image derrick thomas mvader (Victron Energy) ♦♦ commented ·

@mvader (Victron Energy) Yes absolutely. How do I set that up? I am currently on 2.93 I will need to switch back over to the RC.

I have since updated back to 3.00~32 and will leave it there until you have a chance to look at it.

0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ derrick thomas commented ·
Hi, for us to be able to access the device, you need to enable Remote Support access.


I’ll ask a colleague to send you an email, that will probably be Monday.


Have a good weekend!

0 Likes 0 ·
derrick thomas avatar image derrick thomas mvader (Victron Energy) ♦♦ commented ·
Remote support is enabled. Anything else I need to do to be ready for you? Guimods is currently installed but I did do a fresh install with clean os and no mods to verify the issue still existed before reinstalling guimods.
0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ derrick thomas commented ·

hi @derrick thomas , to confirm - the issue with ve.bus disconnecting for a seconds every few minutes is solved per v3.00~35 which is available for testing per just now.

Thank you for reporting!

1 Like 1 ·
jetronic avatar image jetronic mvader (Victron Energy) ♦♦ commented ·
Typo: v3.00~35


0 Likes 0 ·
t81s avatar image
t81s answered ·

Updated from 2.93 to 3.00~32.

Since the update opening the Remote Console in VRM fails.

img-4243.png

Disabling the option „Enable on VRM“ in „Settings“ -> „Remote Console“ will remove the Remote Console entry in the VRM navigation bar but after enabling it again, it will cause the same error when trying to access it.

Edit: Venus OS is running unmodified on a Raspberry Pi 3


img-4243.png (154.0 KiB)
1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

t81s avatar image t81s commented ·

Ok, after @David.Renaud comment, I restarted the Pi and now it works.

0 Likes 0 ·
henrik avatar image
henrik answered ·

Does anyone know if the v3.00~32 release includes a fix for the non-working Broadcom/BRCM drivers on Raspberry Pi Zero 2 W Rev 1.0? I had to replace /lib/firmware/brcm with the contents from the Raspberry Pi OS Lite (Legacy) image to get it working, as well as symlinking BCM.hcd to BCM43430B0.hcd.

3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi Henrik, no thats not included. We have some in progress work on that, but it needs more work (breaks the build).
1 Like 1 ·
henrik avatar image henrik mvader (Victron Energy) ♦♦ commented ·
@mvader (Victron Energy) Thanks for the swift reply! Is there an issue tracker I can follow for this?
0 Likes 0 ·
Show more comments
l81ker avatar image
l81ker answered ·

BUG REPORT? DIGITAL INPUT 0.

@mvader (Victron Energy) V3.00 ~32. I've added two door alarms a couple of weeks ago, and this week my 3 Ruuvi tags arrived so I have been setting those up. I noticed that there is now an additional digital input for which there is no physical connection on the Cerbo. My door alarms on Digital Input 1 and 2.

A new input "0" (Zero) has appeared and which shows as 'Unknown' - is this a bug, or is there something new coming that's related to movement sensors on the RuuviTags? I can't be sure when this digital input appeared, but for sure it was not there 10 days ago!


I dare not 'mess' with it by clicking on any radio button, but none of the radio buttons are selected which likely causes the state 'unknown' . - screenshots as below.

1681895706243.png


1681895790642.png


1681895706243.png (82.2 KiB)
1681895790642.png (32.4 KiB)
4 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

l81ker avatar image l81ker commented ·

I also noticed there are now 4 additional relays shown (but there is no control capability and no physical relay present in the Cerbo for these, either).


1681896428934.png

0 Likes 0 ·
1681896428934.png (33.4 KiB)
Kevin Windrem avatar image Kevin Windrem l81ker commented ·
This somehow has to do with GuiMods. I'll look into it because it should only show relays that exist in the system.
1 Like 1 ·
Kevin Windrem avatar image Kevin Windrem l81ker commented ·

GuiMods Relay menu displays relays for which a /Settings/Relay/n/CustomName exists. I checked my Cerbo and it had settings for all 6 relays (n = 0 to 5).

I deleted all Relay settings and rebooted and only settings for the first 2 relays came back.

I switched back to v2.94 then back to v3.00~32 and the additional relays did not appear.

None of my mods creates relay settings however RpiGpioSetup does modify the gpio_list file but this is only for Raspberry PI platforms.

I have no idea what might have created the additional relay settings unless a previous Venus OS (beta) version did that.

I also can't think of a better test for relay existence.

1 Like 1 ·
Rob Duthie avatar image Rob Duthie Kevin Windrem commented ·
Hi

Re relay test for v3.00.32 tests OK, you can just you can use a RPI GPIO LED test breakout PCB this will test all IO etc. very handy for confirming issues.

Regards

Rob D

NZ

0 Likes 0 ·
davidrenaud avatar image
davidrenaud answered ·

Hello,

@mvader (Victron Energy)

I'm running 3.00~32 large with no issue (Victron team deserve that we mention also when it run smooth ;o))).

Setup is a grid connected ESS with MII + Fronius in MG with 2wires BMS + CerboGX. I'm using AC ignore setup with Victron connect (when ACignore is manually set with assistants + ESS, Fronius is immediately disabled by frequency sifting when ACignore conditions occurs).

I switched another installation to 3.00~32 normal (CCGX). It is an MII with MPPT and an ET112 placed behind grid meter due to presence of AC load on ACIN. Reason to test a release candidate on a remote installation is the known issue with very slow response to grid setpoint,(https://community.victronenergy.com/questions/51697/victron-ess-et112-slow-response.html). (note that I'm reluctant to install in parallel with ET112 an additional external current sensor as proposed in the thread).

Grid set point response is now much better. Switching off a 1800W load takes now 7s to reach Grid set point (set at 100W). This may still looks a lot, but we see a steep correction occurring very quickly (following the 400W/s) until it reach -100W. then it is very noticeable that process become much longer as the algorithm apply minor corrections (10 / 20W) until it reach grid set point.

There is a logic here, corresponding to a "PID" style controlled process and I guess you’ve already tuned I&D optimally.

As these shorts but numerous grid feed-IN become an issue with powerful appliances switching ON and OF each 5 second ( Steam iron,…): I would recommend to adapt the logic to the “No feed” need.

You may allow a first overshoot in the correction of the grid setpoint (at 400W/s). When Grid set point is seen overshot (like +300W): Then the fine correction mechanism could be started to reach the +100W target.

Positive short bleeding is less an issue than negative ones.

I wish you my best regards,

@T81s , on both intsallation with 3.00~32, Remote console is accessible on VRM.


7 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi @David.Renaud , thank you ; good to hear all works well on that one system.


With regards to ET112: be advised that the next candidate will remove the new “fastness” for all systems except for those using the new EM540 grid meter.

So ET112, ET330, EM24, ABB: they all go back to how it was. Which is a major pity, but there still are systems for which we can’t get it fast as well as stable with those slow meters, and its really time to stop and get v3.00 released.

I’ll make sure the developer reads your tips - thank you for those.


0 Likes 0 ·
athlonfx avatar image athlonfx mvader (Victron Energy) ♦♦ commented ·

@mvader (Victron Energy) does Victron plan to add „fastness“ for the mentioned „slow meters“, especially the EM24, in a later version (3.XX)? I‘d really appreciate it.

0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ athlonfx commented ·
Hi @AthlonFX , the honest answer is maybe or maybe not. And if not, then the solution is to replace the meter for a faster one.


Nobody is helped by something that is faster, but then if you, for example, buy a new fridge the system becomes unstable. Not the end-user, not the installer (who will get the first phone call) and not the distributor either.


Future will tell; that is all I can say right now; I hope you understand.

0 Likes 0 ·
l81ker avatar image l81ker mvader (Victron Energy) ♦♦ commented ·

Thanks for the update @mvader (Victron Energy) that is a real shame, I hope Gavazzi soon release a faster single phase meter to update the ET112, as the EM540 is 3ph but only 65A and not well suited to 100A UK domestic installations. So far, the ET112 is the only candidate.

https://www.gavazziautomation.com/nsc/HQ/EN/search?q=Single-Phase%20Energy%20Transducer

0 Likes 0 ·
davidrenaud avatar image davidrenaud mvader (Victron Energy) ♦♦ commented ·

@mvader (Victron Energy) Bad news for customers who just paid for ET 112. I'll take it back to play around on less sensitive use case.

Thanks for considering adapted logic of the Grid set point control loop.

I would have tons of other proposals, but that's not the way for KAIZEN improvement. If I may advise, I would recommend that you indicate to the community what are the subjects Victron want to consider for next version. This could be done when announcing first candidate which usually take care of bug fix). So, we'll all focus on the same region of interest and every body will save time.

.... We still have the "features request" to steer next version.

Best regards.

-1 Like -1 ·
awalkerinuk avatar image awalkerinuk davidrenaud commented ·
I know their are reasons behind this realease now not speeding up the ET112 reactions. But its very disappointing have been watching and waiting on this update for a while. My system is less than a year old. So will not be replacing the meter any time soon. Only went with a ET112 as it was part of a recomended system.


I do appreciate all the work you have put in. Just hope the update gives us ET112 owners some thing.

Not wanting to appear ungrateful just disappointed.


0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ awalkerinuk commented ·
I understand fully; its also quite disappointing for us! Lets see what time brings
1 Like 1 ·
ikeakayke avatar image
ikeakayke answered ·

Just here to announce I just updated from 2.94 to to 3.00~32 (MK3 also) because I just installed an EM540 and all works great, blazing fast updates from the meter, no more L1-L3-L2 phony reporting (although the meter itself still has the alarm on its display when at 0 kw, but not under any loads), faster compensation to grid set points when loads suddenly on/off so I'm very happy about it, a great job by Victron!

Since lots of features added I have some to propose as well (I did in other posts but noone replied anything) so here they are:

- perhaps hub4control can also update inverter setpoints faster than 1/s? if it makes sense considering faster grid meter and systemcalcs?

- an option "no ac sources or consumers between grid meter and inverters" so that dc power can be visible in vrm and cerbo dashboards while using external grid meter? I did some local hacks and made gui show it, also made AC loads stay at dead 0, but I can't seem to do it for vrm, and then I'm afraid that even if I did it might register wrong grid metering instead of the nice plan C I have on EM540 (I had ET340 before)

- a 3rd mode for ESS multiphase compensation called "mixed" that would only compensate potential overloads on inverters. For various reasons I can't/shouldn't keep multiphase on so that I mostly act just like a normal consumer so I need to keep on individual. probably not that different from multiphase to compute and trigger with an "if overloaded by ..then...), probably also not that different from the peak shaving. Eg: 3Pn quattro 10kva, loaded at 7-7-7kw each. then sudden load on L1 say 9-7-7 which means 1kw from grid on L1. just for overloads it would be nice to compensate on one of the other phases with -1kw.


Thanks!

3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·

Hi,

Thank you for the suggestions!

First of all: to get really fast response and a stable system, fast meters are needed. And any work around is bound to be either very difficult, unstable, a lot of work, or all of the above.


Idea 1) its already doing that.

Idea 2) that might help a bit indeed

Idea 3) I’m not 100% sure what you’re suggesting.

1 Like 1 ·
ikeakayke avatar image ikeakayke mvader (Victron Energy) ♦♦ commented ·

Hi,

Thank you for reply!

Yes, I already have EM540 because I read here about the fast response, so I already purchased and installed, all if good with this RC firmware.

1) great news! I though that it is slow because in dbus-spy the individual setpoints in hub4 are updated by hub4control only 1/s, unlike the very fast values coming from the new em540. But maybe these in dbus are just for reference and hub4control talks directly with the inverters and I got the flow wrong

2) Thanks, can't wait for it for vrm especially. for gui it was as easy as manually keep the layout on Hub4 :), one line.

3) the purpose of the setting 'individual phase compensation' is to keep at gridsetpoint on each phase, presumably because you want to feed in the grid as little as possible, if at all, to avoid as much as possible any negative currents on any phase. With the mutiphase option it constatly keeps all 3 inverters at the same inverting power so that there could be phases with negative loads. I got this right so far?

But if I need to feed-in as little as possible (so not talking about sum total, but literally, as little negative energy on any individual phase) I need individual phase compensation, but this way I lose the benefit of total inverting capacity. With my ideea above, only when one/some inverters are overloaded so they can't keep the setpoint defined, which draws positive power from that utility phase, I need the other inverters to compensate with negative power.

So that the only time there is ever considerable negative power at the utility grid is only when one inverter is overloaded beyond it's capacity.

I hope I was more clear this time and that I expressed myself correctly.


ps: also the little hint in the gui when changing to invdividual is a bit confusing and misleading:

'each phase is regulated to individually achieve the grid setpoint'. this would mean if gridsetpoint is 30, each inverter will try to reach 30 individual when in fact hub4control splits that 30 to all phases so each actually gets 10. it should be refrased to:

'each phase is regulated INDIVIDUALLY TO achieve (1/3 of) the grid setpoint' so with those 2 words reversed and perhaps the 1/3 added as well.

0 Likes 0 ·
ikeakayke avatar image ikeakayke mvader (Victron Energy) ♦♦ commented ·

Hi again,

Regarding 1) I just did a little experiment:

- turn off hub4control at which point com.victronenergy.vebus.ttyS4 Hub4/L1/AcPowerSetpoint no longer gets updated (in this RC it still updates 1/s)

- I then manually change Hub4/L1/AcPowerSetpoint with say -1000 and that inverted instantly obeys it.

From the above I understand that hub4control calculated what needs to be done and outputs the individual setpoints on the dbus from where the vebus process gets them and sends it to the inverters.

I just did a 100ms python loop (which it can barely keep up) to output the values of Hub4/L1/AcPowerSetpoint and guess what? it is repeated the exact value several times per second so either I'm missing something or hub4control simply does not update dbus as fast as it should considering the new em540 fast updates.

Of course if hub4control updates dbus 10x/s then the vebus should also send it 10x/s to the inverters ( I don't know yet how that works)

0 Likes 0 ·
davidrenaud avatar image
davidrenaud answered ·

Hello,

Using V3. 00-32 with ACignore, I discover that led's bulk - absorbtion - float do not show up on the console nor on the multi ll (cerboGX).

I appears on second installation where ACIn is accepted. I'll investigate further.

BR.

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

juf avatar image
juf answered ·

Hi,

I‘ll do a fresh installation with 3.00~32.
My Rpi have full access to the internet but the time won’t sysncronice!
Put it manually in but thats not the way or?

Pls give us the chance define one Lokal NTP Server.

regards

jens

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi Jens, could you make a separate question for this in the modifications section? thanks!
0 Likes 0 ·
resuepeter avatar image
resuepeter answered ·

Is this a old problem coming back ?

I upgraded one of my MPPT Charger to a newer version
allconnected via VE Direct tot Cerbo S-GX

The VRM number and MODBUS are now the same as my SMART Shunt. Instead of the old BlueSolar Charger MPPT 75/10. We using the same cable, same port on de Cerbo GX

Cerbox S-GX (C014) v. 3.00-32 MODBUS ID 100
Smartshunt 500A (v4.14) VRM 279 MODBUS ID 226
BlueSolar Charger MPPT 100/15 (v1.61) VRM 279 MODBUS 226
BlueSolar Charger MPPT 100/15 (v1.61) VRM 278 MODBUS 100

The old MPPT-Charger had MODBUS 223

Is the only solution a total factory reset ? Or is there a quick fix.


1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·

hi @resuepeter having a SmartShunt and a Solarcharger at the same VRM Instance, and thus same modbusid is not a problem: the ModbusTCP registers of solar chargers do not overlap with modbustcp registers of smartshunts or battery monitors.


Hope thats clear like that.

To discuss that further , please make a new post in the modifications section .

0 Likes 0 ·
resuepeter avatar image
resuepeter answered ·

Feature request: Configuration Backup en Restore for Cerbo-S GX for after a factory reset ?



1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi @resuepeter , backup & restore of settings is on the todo list; thanks!
0 Likes 0 ·
hominidae avatar image
hominidae answered ·

@mvader (Victron Energy) thank you for letting us know about the downgrade of the ESS control loop for the older grid meters in v3.00b36 Two additional questions arise:

  1. is there a risk that you'll have to downgrade for EM540 model as well? I am on v3.00b29 still with an EM24 and it seems stable - in fact the savings from the faster control loop are quite impressive, with consumption from grid almost reduced by 50%, which would actually pay-off for a switch to EM540 in a short period of time by given energy prices from the grid utility (my estimate is that investing in an EM540 will pay off within 12month, so worth considering isn't' it).
  2. will a setup be stable in a mixed setup (like switching the grid meter to an EM540 and recycling the EM24 to measure another consumer (like heatpump, PV, AC-Out, ...) in parallel?

TIA,

hominidae

4 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi!


There is absolutely zero risk that we’ have to downgrade for the EM540. Also we’re working to support more (fast) meters. Including ones with Ethernet. I expect to have more info on that in three to five months.


> 50% savings with faster ESS response

That is a lot!


> will a setup be stable when mixing EM540 and EM24

Absolutely. As long as you use the EM540 for the metering the utility connection its fine. That is the only meter used for ESS.


Ps. For anyone that for now wants to keep the fast version, use the Backup firmware feature to revert to v3.00~32, and disable online updates. You are a bit stuck then ofcourse but at least the option exists.

1 Like 1 ·
hominidae avatar image hominidae mvader (Victron Energy) ♦♦ commented ·

@mvader (Victron Energy) thank you for the fast response.

Yes, the faster control loop showed a lot of potential savings already. Comparing with the MID utility meter, with V2.92 the ESS control loop was still causing a consumption (and feed-in at the other end of a consumer load cycle of 1.6-5kWh per day...with v3.000b29 this is down to 0.6-2kWh...so on average a saving of approx 1.2kWh/day....which sums up to approx 160EUR/year at actual insane energy prices here in Germany.

I am running some very dynamic high loads, 1ph and 3ph (heatpumps, induction stoves, intelligent EV charging stations) and with some clouds moving over the PV array, I can see 10-12kW change in grid power within a second easily.

So every Wh not consumed from the grid counts..
Thanks again, I ordered an EM540 right away ;-)


Edit: just as an example, these are captures from a duty cycle of my 3ph heatpump alone, during night:

...v2.92 (EM24)

ess-v292-em24-example.png


...v3.00b29 (EM24)

ess-v300b29-em24-example.png

Edit may17th23: ...v3.00b29 (newly installed EM540)

1684275570676.png

..an even smoother control and grid consumption is reduced by another 50% compared to EM24 (with fast ESS-loop in 3.00b29).

Well done @mvader (Victron Energy)

regards,

hominidae

2 Likes 2 ·
athlonfx avatar image athlonfx mvader (Victron Energy) ♦♦ commented ·

@mvader (Victron Energy) is it still possible to download and install the firmware ~32 if it wasn’t installed before the release of a more recent version?

1 Like 1 ·
rickc avatar image rickc mvader (Victron Energy) ♦♦ commented ·

Running v3.0.0~36 using an EM530 Grid Meter (same as EM540, but with external 5A CTs). This is a US split-phase dual quattro 5k setup. Nice to see the meter updating at what looks like 10Hz in the cerbo menu, and on dbus-spy. Everything appears to be running very smoothly in Keep Batteries Charged mode.


One caveat: I’ve set the baud rate on the EM530 to 115200, and have connected it to a Waveshare RS-485 to Ethernet DIN rail device in transparent mode. I’ve created some custom services using the socat command to emulate a pty on the cerbo, and have directed the cgwacs grid meter there. So it looks like a serial port to the cerbo. I’m not sure that a direct Usb-485 adapter running at 9600 baud would be fast enough.

0 Likes 0 ·
dansonamission avatar image
dansonamission answered ·

Is it possible to add an option on the Cerbo to allow us to switch between slower and faster ess control loops? For those it works for can then take advantage of this.


The ET110 was mentioned in the 3.36 release notes, did you mean ET112.


2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

gazza avatar image
gazza answered ·

I have just rolled back to Ver 32.(Multiplus 2 GX with ET112 grid meter). For many the "Most" important issue is that the ESS limits inverter overrun to the minimum. Ver 35/36 appears to degrade that performance. I accept that there is a technical limitation in ramping Up/Down the inverter but there was an improvement noticed in ver 32 (and earlier) in the snappiness of the system. Deliberately slowing down the performance of the ET112 data in the system is a retrograde step in my opinion....

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

derrick thomas avatar image
derrick thomas answered ·

Maybe not necessarily related to this release?? but an ongoing issue none the less.

When the cerbo reboots, often times the bms (connected through a USB to uart adapter) gets assigned a different instance number (I think that's the terminology I'm looking for). This obviously sets off battery alarms and mppt alarms. It then requires me to go into the settings and reselect the bms as the monitor and as the dvcc control. I also have to go into node red and reconnect the effected nodes. Can something be done about this? Just for clarity there are a few other devices connected via USB as well. (USB to be.direct for a mppt, tank 140, and a rs485 relay board)

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ commented ·
Hi, that sounds like some non-officially supported BMS that is installed using an add-on to Venus OS? This problem can be solved, by the software developers of that add-on; which is not Victron.


Please keep any further discussion on that outside of this beta testing post - thanks!

0 Likes 0 ·
Daniele Calabrese avatar image
Daniele Calabrese answered ·

I constantly lose gps connection and i need to reboot. i have USB external device. Now running beta 36. Not sure why with 2.9 all was working well. I have a USB 3.0 hub

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Peter Buijs - NL avatar image
Peter Buijs - NL answered ·

Why doesn't my Raspi Large Version 3.00~29 update to 3.00~36?

Trying Online update, it starts downloading and installing... but no update ... it is still 3.00~29

Trying SD/USB image update, it finds the correct image, starts installing but nothing happens ... still 3.00~29.

I never experienced these update problems before.

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Kevin Windrem avatar image Kevin Windrem commented ·

I just successfully installed the large version of v3.00~36 from the normal version of v3.00~36. I'd give that a try.


Edit: that is install v3.00~36 normal first then try to install the large version.

0 Likes 0 ·
cyberb avatar image
cyberb answered ·

Good day,

On the newer Venus OS (2.90 and above) batteries discharge themselves, for example, I have my ESS set to "Keep batteries Charged" after a loadshedding period my battery will recharge to 100% but then starts discharging itself. It once went down to 82% whilst I had adequate PV and the grid connected and it was still discharging. Slowly, but it was discharging. I had to keep changing to (Optimized without BatteryLife) and back to Keep batteries charged to have it recharge.


I started to notice this back on Venus OS 2.90, before this I always used Venus OS 2.73 as its the most stable in my opinion, this never occured, but I also noticed on this lower firmware when a Freedom Won or SolarMD is detected "Shared Voltage Sense" is forced off, so I tried turning it off when using the higher firmwares as its not forced off on them and this solved it.



My equipment:


Victron Multiplus-II 48/3000

SolarMD 3.7 / Freedom Won Lite 5/4 - This problem occurred on either battery

BlueSolar MPPT 150/35

ET112 Energy Meter

RPI Zero 2W also have a NanoPi, I kinda designed and built my own 2 different GX using these, Im running the higher Venus OS on the Zero 2W and im running 2.73 on the NanoPi. Im using a WaveShare RS485 CAN HAT on the Zero 2W and I have a USB-CAN converter for the NanoPi.


Kind regards


2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

stenis avatar image
stenis answered ·

1682358192239.png

Blue = Grid
Red = Load - Load change: 1300W Heater
Green = Batterie

time = from right to left (total measure 111 sec)

Configuration: MPII 3000, 3 x Pylontech, EM24

I switched now back to V3.32 ...


1682358192239.png (10.6 KiB)
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

blackmarlin avatar image
blackmarlin answered ·

I totally understand that stability of the system is the most important goal for any changes concerning the ESS control loop.

But please provide an expert option to manually activate the fast control loop as implemented in versions <3.00~36. I am using an EM24 and the fast control loop resulted in noticeable less consumption from grid without any issues concerning stability in my case.

Best regards

Michael

11 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

peregrines avatar image peregrines commented ·

An optional "fast control loop" switch would be a perfect solution!

3 Likes 3 ·
olafd avatar image olafd commented ·

The same here - EM24 with RS485 and v3.00~14 - no issues.

With the "fast control loop“, the Multiplus II (48/3000/35-32) only feeds in very little - less than 3 kWh for nearly half a year (ESS with 'feed in excess = inactive').

In the 3 months before the “fast control loop“, the feed in was 20 kWh - that’s more than 10 times less Power-Feed-In.

Additionally the 'Grid setpoint’ could be lowered from 50W to 30W.

Therefore the MP2 could maintain the lower '30W-Grid-Set-Point' more than 10 times more accurate.

Congrats! VictronEnergy - That’s quite a feat.


@mvader (Victron Energy) An optional "fast control loop" switch would be very appreciated for Systems with 'no feed in allowed' or if one chose so.

2 Likes 2 ·
gazza avatar image gazza olafd commented ·
Same experience here with ET112. Ver 3.00~36 is a massive backwards step! (I have rolled back to a previous release and will have to stay with that if Victron decides to stay with their current trajectory).


The 'fast control loop' switch idea is a great idea that I would support fully also.!

0 Likes 0 ·
maddog75 avatar image maddog75 gazza commented ·

Hi Gazza,

Do you or anybody have a copy of the Venus OS v3.00~32 sdcard install file for cerbogx? I would love to test the faster ET112 control loop and see if my setup is stable with it.
Thank

0 Likes 0 ·
gazza avatar image gazza maddog75 commented ·
I don't have it on an SD card I am afraid, I used the benefit of the roll back to last version available in the Remote Console> Firmware. I did have a look on the Victron software downloads and cannot see any of the 3.00~xx versions?
0 Likes 0 ·
maddog75 avatar image maddog75 gazza commented ·
Thanks for the reply, I really appreciate it. My electric oven pulses on and off and currently ESS with a EM112 monitoring the Grid is just too slow to catch the load! Hopefully Victron will consider revisiting the fast control loop again!
0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ maddog75 commented ·
Hi all, totally understood and noted. We'll look into it and see what the best option is.


It will take a while before I have a plan (weeks), as for now priority is to finally finish other open things for v3.00; since it has lots of valuable features which have been pending in beta too long. Matthijs
1 Like 1 ·
olafd avatar image olafd commented ·

Do your EM24 have RS485 or Ethernet ?

0 Likes 0 ·
olafd avatar image olafd commented ·

A Shelly 3EM with a Raspberry PI 2 with v3.00~14 (a test system) works very well too, with the “fast control loop“.

For Shelly-3EM-Driver-installation :

https://community.victronenergy.com/questions/174631/howto-install-venus-driver-for-shelly-energy-meter.html

0 Likes 0 ·
ojack avatar image ojack commented ·
I have no issues with the EM24 RS485 working with the fast control loop v3.00~32.

I think especially with fast changing loads like induction hobs it works very much better then the older slow control loop.

A parameter to switch between the two loops whould be a great solution I think.


0 Likes 0 ·
dirk-s avatar image dirk-s ojack commented ·
For me the same. But I have bought now a EM540 and after installing I will sell the old EM24. I think it will be the best solution at the end. And in total it will be an investment of around 50 Euro.
1 Like 1 ·
ojack avatar image
ojack answered ·

I do not know if this is the right place for my question.

Would it be possible to add an option to change the phase where an ET112 is installed?

At the moment the ET112 ist permanently set to L1 but in some installations it would be better to install an single phase inverter at L2 or L3 to reduce unbalanced grid load or to

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

blackmarlin avatar image
blackmarlin answered ·

Hello,

is there any archive for the previous release candidates?

Where can I find the version v3.00~32 (large image for raspberry2) or can anybody please send it to me? I want to do some more testing concerning the fast control loop with EM24.

Best regards

Michael

3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

athlonfx avatar image athlonfx commented ·

@mvader (Victron Energy) I‘m interested in this, too. But for the Multiplus-2.

0 Likes 0 ·
mvader (Victron Energy) avatar image mvader (Victron Energy) ♦♦ athlonfx commented ·
Hi both, we’ll have to build it again - I’ll think about how to go about this
1 Like 1 ·
marcus67 avatar image marcus67 mvader (Victron Energy) ♦♦ commented ·
Hi,

I would be interested in that version as well.

I would need it for a Multiplus II GX.


Thanks, Marcus

0 Likes 0 ·
hotbiker avatar image
hotbiker answered ·

Same here, it would be great, if the fast Version could be somewhere for download.

Thanks
Gerhard

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Related Resources