question

durval avatar image

Phoenix Smart Inverter: very inaccurate VA values?!

SUMMARY: very innacurate VA values (33% less than the real value). How to improve?

I finally pulled the trigger and bought a Victron Phoenix Smart Inverter, exact model 12/2000 and with AC output of 230V. It has been automatically updated to the latest firmware version (v1.25) by the VictronConnect app, the first time I connected to it. I have installed it following the manual directions, and I specifically grounded it to my installation's ground (in my case, my vehicle chassis ground point). It's being fed by a single (so far, during the tests) Pylontech RV12100 12V 100 Ah battery, connected over the recommended 70mm2 battery cables with ~1m each.

I'm having some issues, which is kinda disappointing given Victron's price (about 50% more than the equivalent/comparable EPEver/EPSolar IP-Plus). I will post these issues in separate questions so as not to confuse the subjects.

Here in this post, I will specifically ask re: its VA precision: with a purely resistive load (ie, a 60W incandescent lightbulb with a power factor of exactly 1.00), I measure (using both my Uni-T UT210E DMM and a Poniee PN2000 Electricity Usage Monitor (both of which I have strong reason to believe are very precise), they both register 54W being drawn by the load; the VictronConnect app shows only 36VA (ie, about 33% or 1/3 less).

As the power factor is exactly 1.000 (as measured by the PN2000, and also as expected from a pure resistive load like the aforementioned incandescent lightbulb), the VA value shown by Victron should be exactly the same as the watts value shown by the DMM and the electricity monitor.

I believe the Victron inverter is under-reporting the VA value; as its voltage measurements were very precise (less than 1% difference when compared to the other two instruments), I understand this means its current (ie, A / ampere) measurements are way off.

This is very worrying, as not only I was counting on these VA measurements to be able to follow up on my electricity usage, but also because I understand its measured current values are used by the inverter itself for a series of other things (among them the ECO mode wake/sleep loads, and the Dynamic Cutoff value -- the latter of which I'm also having issues with, coincidence or not -- please see my next post).

Yesterday I went to the local Victron representative where I bought my inverter, and they were very helpful; they weren't aware of the issue, but they immediately pulled a brand new inverter of the exact same model from their inventory and together we were able to almost exactly reproduce the issue with it, with the only difference being that new inverter showed 39VA instead of 36VA (in the exact same test conditions). We also did a quick test (as it was getting late) with a much heavier resistive load (an air heater drawing about 800W/VA) and the VA measurement seemed to be much more precise, even if not exact either.

So, what is the expected accuracy of a Victron Phoenix Smart Inverter VA measurement? Is there any way to improve it (better grounding, parameter configuration, different firmware version, etc), specially with light loads?

Thanks in advance to anyone who can help.

EDIT: please see my posts regarding other issues I'm having with those inverters here:

Post #1: This very one.

Post #2: Phoenix Smart Inverter: issues with Dynamic cutoff

Post #3: Phoenix Smart Inverter: very little data being reported in the app


Phoenix Inverterinverter current draw
2 |3000 characters needed characters left characters exceeded

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

1 Answer
Alexandra avatar image
Alexandra answered ·

@durval

The heart of every inverter system is the battery. As an offgrid person (as i assume since you got a phoenix you are) i can say that you are all focused in the wrong place.

At the end of the day it is the battery you should be most concerned with. When I am pulling a load I can quite honestly day i look at the VA being pulled from my battery bank rather than the AC loads/ accuracy. I have a BMV in the system, have tweaked it for my bank and application and that is the heart of all my interests.

While i can see your concerns as well with your system, since you have purchased from the less expensive units of the range you will be experiencing what happens when you do so. Also nowhere in the documents does the victron claim to be a measuring instrument.

So my advice get a bmv and definitely move your plans up for a bigger bank. 100AH is not enough to feed the beast on full load.

I have a 1600va multiplus compact. Ran a short test with a 100W bulb I have no complaints with how the system measures. The 16W you see is the system load (mppts bmv colour control, etc) on the battery, we switched off everything and disconnected all other loads and ran the bulb on its own.

screenshot-20210721-021137.jpg
screenshot-20210721-021113-comhuaweihimovieoversea.jpg


3 comments
2 |3000 characters needed characters left characters exceeded

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

Hi @Alexandra! First of all, many thanks for your long and detailed response. I will try and answer in an equally detailed manner:


> At the end of the day it is the battery you should be most concerned with.

I'm aware of that. But in my case, the battery is connected only to the charge controller (an EPEver/EPSolar Tracer 30A which is working perfectly, including providing me with very precise current measurements -- among a ton of other data -- via its built-in RS485 port) and to the inverter.

So, as nothing else is neither supplying charge nor taking it from the batteries, my expectation was to use the Victron reported VA value (which, divided by the voltage, would give me the current), along with the charge controller's current measurements, to monitor the battery.

This approach has the advantage of not only cost, but also simplicity and robustness (less equipment to install, less things to maintain, and less stuff to break). And given my experience in other contexts, it would work if only the Victron gave me usable (as in, minimally accurate) current (or VA) values.


> Since you have purchased from the less expensive units of the range you will be experiencing what happens when you do so.

It seems a stretch to call the Victron Phoenix "less expensive". In fact, its price is over twice the price of the equivalent EPEverEPSolar IP-2000-12-Plus inverter which has very similar specs to the Victron Phoenix (actually, it offers a larger/longer peak capacity at 4000W for 5 secs) *and* has a RJ45 MODBUS port (which I haven't tested yet, but from which I would expect to obtain at least the same data, and with the same accuracy, I now obtain from their charge controller). The Victron Phoenix advantages (from comparing their spec sheets) are (a) its ECO mode, and (b) its Dynamic Cutoff, neither of which are offered by the other brand. (But then, I'm having issues with the Dynamic Cutoff feature too, please see my other post here if you haven't already).

So you would ask, why didn't I buy the EPEver/EPSolar? The reason is that I really, really like my local Victron reseller. They are very knowledgeable and helpful, and they sold me my EPSolar controller some months ago, among a lot of other stuff. But they don't sell the EPEver/EPSolar line of inverters -- they only sell Victron inverters. And then I thought, "OK, it's twice the price, but certainly Victron is a better brand and their inverter will be at least as good or better than the EPEver/EPSolar". At least in terms of accuracy, your answer seems to be proving me wrong.


> So my advice get a bmv

I think that even adding BMV (which alone costs over half the price of a complete EPEver/EPSolar IP-2000-12 inverter), I would not have direct info on the inverter, specially its temperature, which I think is really important to monitor (please see my other post here). Please correct me if I'm wrong.

[To be continued]

[continued from my previous comment]

> and definitely move your plans up for a bigger bank. 100AH is not enough to feed the beast on full load.

I'm well aware of that, it's why I stated on my original post that my inverter was "being fed by a single (so far, during the tests)" battery; my planned final setup is with 3 of those LiFEPO4 batteries in parallel, for a total of 300Ah.


> I have a 1600va multiplus compact. Ran a short test with a 100W bulb I have no complaints with how the system measures.

Very interesting! Thank you very much for taking the time and effort to do all those tests and photos, much appreciated!

From the pictures, I see you have the BMV (the little round thing on top, right?) *and* a color monitor which I guess is a different accessory. Please excuse my ignorance, but:

0) Are these values being obtained from the BMV or from your Multiplus Compact?

1) Which BMV model you have, exactly? My local reseller lists 3 different ones: BMV-700, BMV-702 and BMV-712, it would be helpful to know which one exactly you are using.

2) I see the values you mention (16W "empty" -- ie, only your accessories -- and 118W with the 100W lightbulb -- which would mean an additional 118-16= 102W just for the lightbulb, which is indeed very precise) are all being reported not on the BMV display, but in the aforementioned color monitor. Is the color monitor simply reporting what the BMV is measuring? In other words, can I dispense with the color monitor and obtain these values straight from the BMV, preferably via Bluetooth? (I want to capture it on my Raspberry Pi computer and store all the data and then display it with Prometheus/Graphana etc, plus activate alarms etc)

3) If the color monitor (or some other accessory) is required to be able to display/capture the data, what kind/model exactly would I have to buy in addition to the BMV?


Thanks again,

-- Durval.

@durval


The battery figures on the Color Control are obtained from the BMV 700 (you can use the Multiplus itself select-able in the menu). It is nice and basic (the other models have other features). They all use a shunt so very accurate measuring. I like the time to go feature on it, so we can "budget" the battery usage. You can use the Bluetooth dongle instead of the color control for data, or you can get the 712 it has Bluetooth already. This Blog explains the difference between models.

I love the BMV range, even install it for clients with non Victron systems for battery monitoring, as many inverters lack in this area. Have a few on the test bench right now going through their paces. I have used EPEver before, to be honest here where we are with all the monitoring found them just as costly as Victron for installs, so prefer Victron and for the same reason as you do, a nice friendly dealership close by.

With the BMV you can choose what is displayed on the BMV screen - at the moments of the pictures it was on Percentage. you can have time to go, volts, amps or rolling display.

The AC loads are read by the multiplus itself, reported by VE Bus (ethernet type cable) to the Color Control. In the Beginning we had just the inverter, batteries and a smart mppt, monitoring with a VE bus Smart Dongle. But as what always happens you end up wanting more and it grows from there.

I have played with the Pi and Graphana, load of fun if you like tinkering and plenty of resources online for that.