OpenEVSE DIY Tethered 3-phase

This is how threephase energy is calculated:

- P(3ph) = V(LN) * I * 3
- V(VL) = 230V (line to neutral voltage)
- V(LL) = V(LN) * √3 = 230 * 1.75 = 400V

Hi Glyn
Thanks for your patience.
Please take a look of my screen shot
Intensity = 12.76 A and Power = 2935 W
That means Voltage = 2935 / 12,76 = 230 V, and not 230V x 3
Then your software to calculate the energy is correct but there is a bug:
THREEPHASE is not active in the firmware and it continue to calculate as a Monophasic system.
Kind regards
Jordi

Glyn I have a second question:
In the formula we miss COS (FI)
Power = Voltage LN x Intensity x 3 x COS (FI)
do you know the value of COS (FI) in this charger?

probably TREEPHASE it’s not defined at open_evse.h

KR

Please try again with the FW attached, I have explicitly defined THREEPHASE

openevse-3ph.hex (85.8 KB)

If this doesn’t work you could try building the FW yourself. Three-phase calculation works for the EmonEVSE.

Hi Glyn
Thanks for your patience
Now it works
Thanks a lot
Jordi

Fantastic! Very sorry, for all the hassle. The mistake must have been at my end.

Hi,

I know that this is an old thread, but have a question about 3-phase installation and setting the THREEPHASE constant. Is it possible to change it on the fly? For example have some toggle in the web ui to be able to switch it to either 1-phase or 3-phase? Basically I have 2 cars (one has single phase charger and the other has 3 phase charger) and I would like to be able to switch the param on and off based on which car is currently charging. Ideal solution would be to add one more current meter (on L2 or L3) so that the software knows automatically if I’m charging 1 or 3 phase car.

Thanks in advance.

I’ve just had a 3 phase EmonEVSE installed and despite the firmware being reported as 7.1.3.3P, I too see only single phase energy calculations both in the UI and the MQTT feed to my EmonCMS.

I’m also unsure therefore if the PV divert is going to be observed correctly on the grounds that the value I’m returning about grid import/export is for all three phases, not just one.

I see the reference in this thread to the constant THREEPHASE being defined, but no detail on how to set this? I would also expect that to have been set correctly in this firmware?

The OpenEVSE UI displays real-time charging current, which is the same for both single and three-phase e.g 32A 1P = 7kW / 32A 3P = 22kW. However, the kWh/wh reported on the OpenEVSE UI and Emoncms will show the correct energy calculation for 3P.

Solar PV divert is based on current, therefore should work the same for both single and three-phase.

Hi Glyn,

thanks for replying.

I get the 32A bit. But on the emonCMS, there is only one energy calculation and it only shows the power for a single phase on my 3 phase unit. Please see attached screenshot showing the consumption at circa 3.7kW, the unit is actually charging at 11kW. This is the reasoning for my query.

I’ve had to add the energy feed to each of my 3 phase monitors to get my total usage correct.

Thanks for the screenshot, I can see the issue. The charging power shown on Emoncms is calculated in Emoncms input processing, the amp current Input is multiplied by 230V to display the power, this should be multipled by 3x for three-phase.

To change this, open up Emoncms Inputs and click the spanner icon next to openevse ‘amp’ feed:

Then change 230 to 230V x 3 = 690

The power will now be displayed correctly. Alternatively, for more accurate power reading, you could multiple by emonPi VRMS feed x3

The ev_energy Feed should be correct for three-phase since this energy is calculated on the openevse.

To make this easier for new users, I’ve added a three-phase device template which will do this automatically, but since you already have your Feeds setup it’s probably easier to adjust manually. Thanks for reporting this issue.

Are you suggesting “230” (or 3 × 230) to allow for a non-unity power factor - because the UK ‘centre’ voltage is likely to be closer to 240 V ?

You’re right, obviously a better option would be to

Like this:

I’m not sure how your supply compares to ours, but I see about a 5V drop when the EV starts sucking 32A. That’s measured at the fuse box so doesn’t include any additional drops in the run to the EVSE (single phase setup).

I also had a chance recently to capture the current signatures of the on-board chargers of a few different EVs at various power levels:


MG ZS charging at 30A (though 32A was on offer)


MG4 charging at 30A (though 32A was on offer)


Tesla Model 3 charging at 32A


Tesla Model 3 charging at 5A


EQC 400 charging at 32A


EQC 400 not charging, but pre-heating the cabin


Volvo XC40 at 32A


Volvo XC40 at 16A


Volvo XC40 at 6A

That’s about 360 - 370 kVA, so that’s about the same as an urban UK domestic supply. I can’t speak for a rural supply, but it could well be significantly lower.

How did you work that out? Is that some rule-of-thumb for distribution transformers based on the effective impedance?

Ohm’s Law. If you say the supply has a source resistance (not impedance), it’s going to drop a voltage depending on the load (again a resistance). You know both, so you can calculate the supply internal resistance (again, not impedance). That determines the prospective fault current, and that’s the maximum (or rather, more than the maximum) power you can get in the fault. It’s not accurate, but it gives a ball-park value.

Your supply fuses will operate long before that current is reached, but that’s where it’s heading.

I think the UK network is designed so that it’s around the 300 kVA mark for most consumers - but I’m really a drives & controls applications person, not a power systems one.

Nice, that’s fascinating. I think it would be worth creating a dedicated thread for these captures. It would be interesting to build up captures for lots more EVs. I would be interested to see how a Leaf and a Zoe compare, I’ve heard a Zoe has poor power factor.

I think the reason for the double spike on the Model 3 is because the model 3 has a 3x 16A chargers (three-phase 11kW) so to achieve 32A single-phase the car connects multiple phases of the onboard charger together.

Thanks for this. Since I have all 3 phase voltages measured elsewhere, would it be better (or worse) to create a feed with an average voltage calculated from the 3 voltages and then use that feed to multiple in this input processing, or is it “just not worth it”?