US Tesla Wall Connector + 3P3W 240V Delta power - FAIL

[wk057]

The OpenEVSE board could probably be adapted to do the job, although if you went J1772, then other vehicles would actually be supported...Tesla isn't the world after all.

Have you contacted Tesla, yet? There may be an official way to support this configuration. The worst they can say is no.

Yeah, OpenEVSE's AC detection stuff would need modification (or deletion), since it relies on the neutral/ground bond.

I'm fine with J1772 protocol, but as far as I'm concerned Tesla is all that really matters on my end, so sticking with the Tesla connectors for sure. I've made a Tesla to J1772 adapter out of an old Tesla charge ports that I use to charge Volts at my house, though.

Edit: Oh, and I did reach out to Tesla. No response yet.

 

[petersv]

The EU version supports isolated terra, in Norway IT is the most common power grid, so maybe you could look at the firmware in the eu/Norway version, og swap the cable of one.

 

[wk057]

The EU version supports isolated terra, in Norway IT is the most common power grid, so maybe you could look at the firmware in the eu/Norway version, og swap the cable of one.

Didn't realize there was a market where this was common. Not a bad idea! Thanks

 

[mongo]

The EU version supports isolated terra, in Norway IT is the most common power grid, so maybe you could look at the firmware in the eu/Norway version, og swap the cable of one.

Didn't realize there was a market where this was common. Not a bad idea! Thanks

One thing to keep in mind is that the EU vehicle charger is different than the US. It natively supports three phase.

 

[wk057]

One thing to keep in mind is that the EU vehicle charger is different than the US. It natively supports three phase.

Yeah, but I wonder if it can be installed in a 1-phase situation and use the US cable? I'd really just need to mainboard probably.

 

[mongo]

Yeah, but I wonder if it can be installed in a 1-phase situation and use the US cable? I'd really just need to mainboard probably.

Maybe, put it seems like that might not work.
From the EU install manual at Installation Manuals - Wall Connector

• Nominal voltage:
• 230V AC single-phase: L1, neutral, and earth
ground
• 230V AC three-phase: L1, L2, L3 and earth
ground
• 400V AC three-phase: L1, L2, L3, neutral and
earth ground
• Voltage range:
• 180V to 264V AC between N and L1 terminals

The 3 phase no neutral is a Tesla specific mode that likely requires the 2 phase charger:

 

[wk057]

If not, a simple (non-code!) temp fix would be to take a small center-tapped 240v transformer (or one with two 120v windings) and connect it to your panel on the same 2 legs used for the HPWC. Use a 2-pole breaker sized for the transformer.

For example, get this one: ACME ELECTRIC TRANSFORMER 1 KVA T-53110 USED | eBay

Wire the 240v secondary wining to a 2-pole 5A breaker, connecting it's center tap to ground, and cap off the 600v primary. The transformer should not be dissipating any power. It will only if there is a ground fault somewhere else in the network, in which case the 5A breaker would trip letting you know.

Did this hack with that exact transformer from eBay (which showed up a few hours ago)... works for one of the wall connectors. It's happily charging a car at 80A, virtually no voltage drop too.

Thanks for the suggestion!

So... yeah, guess I need to think on this. Probably going to leave the one unit running like this until I get time to roll a replacement control board for the others.

Edit: I wired the 240V side to a 20A breaker, and the neutral leg to ground via a 5A fuse. I moved the breaker around to each of the 3 possible positions, and nothing blew up. Also don't see any current with my clamp meter. So seems good.

 

[eladts]

Did this hack with that exact transformer from eBay (which showed up a few hours ago)... works for one of the wall connectors. It's happily charging a car at 80A, virtually no voltage drop too.

Thanks for the suggestion!

So... yeah, guess I need to think on this. Probably going to leave the one unit running like this until I get time to roll a replacement control board for the others.

Edit: I wired the 240V side to a 20A breaker, and the neutral leg to ground via a 5A fuse. I moved the breaker around to each of the 3 possible positions, and nothing blew up. Also don't see any current with my clamp meter. So seems good.

Wouldn't connecting two phases via large resistors to the ground work just the same for converting the floating delta to "center-tapped" one? Also, if we do it to all three phases, wouldn't just essentially create a 240V/139V Y feed? I don't know if 139V to the ground is acceptable to the HPWC, but if it is this will allow you to use all the phases for charging.

 

[wk057]

Wouldn't connecting two phases via large resistors to the ground work just the same for converting the floating delta to "center-tapped" one? Also, if we do it to all three phases, wouldn't just essentially create a 240V/139V Y feed? I don't know if 139V to the ground is acceptable to the HPWC, but if it is this will allow you to use all the phases for charging.

Interesting. Not sure if it'd work or not, honestly. Might be worth a shot with sufficiently high resistance.

Did this hack with that exact transformer from eBay (which showed up a few hours ago)... works for one of the wall connectors. It's happily charging a car at 80A, virtually no voltage drop too.

Thanks for the suggestion!

So... yeah, guess I need to think on this. Probably going to leave the one unit running like this until I get time to roll a replacement control board for the others.

Edit: I wired the 240V side to a 20A breaker, and the neutral leg to ground via a 5A fuse. I moved the breaker around to each of the 3 possible positions, and nothing blew up. Also don't see any current with my clamp meter. So seems good.

OK, so... in hindsight, this is not going to work out.

Turns out that while this does make the one wall connector functional, it causes the other phase to become ~210V to ground.

I turned off the transformer, and measured current... each phase seems to allow about 10-20mA to ground with a direct short. Wasn't enough to light up the 240V bulb at all. Probably isn't a great situation, but there is some coupling to ground which would allow ground faults to be detected appropriately one way or another.

With the hackformer on, the 240V test bulb lights 50% ground to A and ground to B, as expected due to the center tap... and lights nearly full ground to C.

Long story short, I don't trust the fusing on the transformer enough to be the sole source of ground fault protection, so not going to use this solution at all moving forward, even temporarily. Good as a test, but doesn't seem like the best idea.

 

[eladts]

OK, so... in hindsight, this is not going to work out.

Turns out that while this does make the one wall connector functional, it causes the other phase to become ~210V to ground.

I turned off the transformer, and measured current... each phase seems to allow about 10-20mA to ground with a direct short. Wasn't enough to light up the 240V bulb at all. Probably isn't a great situation, but there is some coupling to ground which would allow ground faults to be detected appropriately one way or another.

With the hackformer on, the 240V test bulb lights 50% ground to A and ground to B, as expected due to the center tap... and lights nearly full ground to C.

Long story short, I don't trust the fusing on the transformer enough to be the sole source of ground fault protection, so not going to use this solution at all moving forward, even temporarily. Good as a test, but doesn't seem like the best idea.

I don't find the voltage on the wild leg surprising, this exactly what is expected in centered-tapped delta configuration which the hackformer emulates. Come to think of it, the HPWC is correct in refusing to work in a floating delta configuration, as it is impossible to detect a ground fault. As an EVSE is supposed to function as a GFCI, it cannot function correctly with a floating delta feed.

 

[mongo]

Interesting. Not sure if it'd work or not, honestly. Might be worth a shot with sufficiently high resistance.



OK, so... in hindsight, this is not going to work out.

Turns out that while this does make the one wall connector functional, it causes the other phase to become ~210V to ground.

I turned off the transformer, and measured current... each phase seems to allow about 10-20mA to ground with a direct short. Wasn't enough to light up the 240V bulb at all. Probably isn't a great situation, but there is some coupling to ground which would allow ground faults to be detected appropriately one way or another.

With the hackformer on, the 240V test bulb lights 50% ground to A and ground to B, as expected due to the center tap... and lights nearly full ground to C.

Long story short, I don't trust the fusing on the transformer enough to be the sole source of ground fault protection, so not going to use this solution at all moving forward, even temporarily. Good as a test, but doesn't seem like the best idea.

After @eladts 's comment, I ran a spice simulation with a Y set of inductors to a common ground point. It looks like it works (assumed 1 Henry for the inductors which could be transformers with open secondaries like are being used now). If you set this up with a three phase breaker on the inductor to delta feeds, it would give you a ground/ neutral path for an HPWC on any two delta phases at 137V which may be close enough to 120V.

The breaker is for a phase to neutral short. Personnel level ground fault would be handled by the EVSE GFCI.

 

[wk057]

My concern is with affecting more than just the EVSE with such mods. I don't want to cause any potential issues down the road, hence why I'm probably going to leave the electrical system as-is and just modify the wall connectors themselves at some point to work with what they've got.

Since, in my building, a short circuit condition from phase to ground (without changes) results in some ~10-20mA max to ground (enough for a jolt, for sure, and certainly enough to trip a 5-10mA GFCI), I'm comfortable that a modded unit with a standard GFCI setup would be perfectly fine and detect a ground fault as it should. I'm unsure exactly where this low current coupling between the phases and ground actually is, though, but a direct short only causes a max of about 20mA of current... and the power at 20mA is essentially nothing, also. So who knows. With no complete bond of any phase or neutral to ground in the first place, I feel like this should be fine since there should be no real chance of a ground fault to begin with.

 

[mongo]

My concern is with affecting more than just the EVSE with such mods. I don't want to cause any potential issues down the road, hence why I'm probably going to leave the electrical system as-is and just modify the wall connectors themselves at some point to work with what they've got.

Since, in my building, a short circuit condition from phase to ground (without changes) results in some ~10-20mA max to ground (enough for a jolt, for sure, and certainly enough to trip a 5-10mA GFCI), I'm comfortable that a modded unit with a standard GFCI setup would be perfectly fine and detect a ground fault as it should. I'm unsure exactly where this low current coupling between the phases and ground actually is, though, but a direct short only causes a max of about 20mA of current... and the power at 20mA is essentially nothing, also. So who knows. With no complete bond of any phase or neutral to ground in the first place, I feel like this should be fine since there should be no real chance of a ground fault to begin with.

Good plan. The leakage current may be parasitic capacitance from windings/ wiring to housings/ frame/ conduit.
(reminds me of the HV transmission line inspections where they have the long rod to bond the helicopter to the wires and get a huge arc)

 

[eprosenx]

My concern is with affecting more than just the EVSE with such mods. I don't want to cause any potential issues down the road, hence why I'm probably going to leave the electrical system as-is and just modify the wall connectors themselves at some point to work with what they've got.

Since, in my building, a short circuit condition from phase to ground (without changes) results in some ~10-20mA max to ground (enough for a jolt, for sure, and certainly enough to trip a 5-10mA GFCI), I'm comfortable that a modded unit with a standard GFCI setup would be perfectly fine and detect a ground fault as it should. I'm unsure exactly where this low current coupling between the phases and ground actually is, though, but a direct short only causes a max of about 20mA of current... and the power at 20mA is essentially nothing, also. So who knows. With no complete bond of any phase or neutral to ground in the first place, I feel like this should be fine since there should be no real chance of a ground fault to begin with.

Tuning in really late to this thread as I typically hang out in the Model 3 forums.

So yeah, ungrounded delta three phase is not exactly common these days! ;-) Interesting setup.

I was wondering: Could you have the utility convert this (or could you legally convert it yourself) to corner grounded Delta? Would just be as simple as bonding one phase to ground.

Then on two of the phase pairs you would have 240v from one phase to ground, and 0v from the other phase to ground. Then you could flip the dip switch in the Wall Connector (that was intended for 277v use on three phase 480/277 service). We know the Wall Connector is designed for and works in that configuration. This would basically be exactly the same except it would see 240v vs 277v (I don't think it cares in the slightest about the actual voltage, just that its ground tests pass).

Obviously this would only let you put Wall Connectors on two of the three phase pairs, but perhaps that would be good enough and you could put the other random loads (like lighting) on the last phase pair for balance?

Also, I am very curious how you are getting current flow to ground when you short each phase to ground? I wonder if turning off each thing in the shop one at a time to see where that is happening would make sense?

This really makes me wonder how the UMC's are operating? Are they depending on that current leakage to ground you describe (unintended) in order to work? Or do they just not do as extensive tests as the Wall Connectors?

Also, have you experimented with that dip switch on the Wall Connector that is intended for 277v operation? I wonder if it would allow the wall connectors on the other two phase pairs to work with your "hack" transformer active"?

 

[wk057]

So, touching base on this again.

I've been in contact with the utility, electrician, and Tesla on this. No good solutions. A bunch of below is c/p from my last email to Tesla.

I have no safe way to bond any phase to ground nor get any neutral connection from the utility company. This setup is not super common, but it is to-code. I confirmed this with both the utility company and my electrician.

I spoke with a rep from the utility company on Friday, along with my electrician. There's nothing they can do (easily) to satisfy the wall connectors. My only utility-side option to do so appears to be to change the entire building to a 3-phase wye setup with a neutral. That's not really an option, since doing all of the external and internal changes is estimated at $15k+. There is the other option of installing some large private transformers, so that there is a bonded neutral on the isolated side, just for the wall connectors... but that also is not really an option due to excessive cost.

I could replace the three wall connectors with non-Tesla hardware confirmed to function here without a problem, although that cost is still significant (~$3k).

The other option would be to convert the system to have a center tap from one of the transformers bonded to ground, and have a "wild leg" delta setup. Unfortunately there is equipment (existing ground fault detection equipment and private transformers) that are not compatible with this setup. Plus, this only "fixes" the situation for one phase, and we can't run three wall connectors from one phase with our current load profile.

At this point, my only sensible option is to replace the guts of the wall connectors with custom J1772 hardware so that it is compatible with the floating ground in our building. Estimated cost comes to a few hundred bucks plus some dev time in my spare time.

Overall, I mainly just find it super irritating that the Tesla wall connectors are the only piece of equipment here that has any issue with our electrical setup. The Gen1 Tesla UMC works fine (have never tried a gen2), three different J1772 hard-wired 240V chargers work fine, forklift chargers, and a host of other non-EVSE related equipment all have no problems functioning perfectly fine.

---

Here are some excerpts from Tesla's responses:

We’re not actually looking for 120V line-to-ground to confirm the ground connection. Here’s what the Wall Connector is looking for to confirm a good ground connection: To measure ground impedance, both L1 and L2/N are rectified to high voltage DC at very low current and injected into the PE connection. The voltage from L1 and L2/N to ground are measured to see if the PE connection “moves” with respect to L1 and L2/N, indicating it is high impedance and unable to sink a fault current.

The underlying problem is still the lack of low impedance Earth connection.

The blink code change when DIP switch one is up makes sense, it’s likely detecting the N/L2 as a higher voltage than L1 due to the floating ground. It will need to be in the up position when a proper ground connection is made, because the wall connector expects 120V L1 to PE and L2/N to PE when the switch is down. The installation with a proper ground should measure 240V L1 to PE and 0V L2/N to PE so the DIP switch should be in the up position to indicate it should expect a grounded neutral (typical for a 277V wye install). It's also important to note that L2/N will need to use the grounded phase.

---

Overall, not much to work with. The unit will not respect the existing non-neutral-bonded ground system, and all suggestions from Tesla involve modifying the building in some way that doesn't make sense.

So, just going to do the brain replacement route when I get time.

 

[JPoldo]

These workarounds appear very dangerous. I would ask the utility for standard 3-phase service with a neutral. It may be a simple re-wiring job at their transformer. Tell them you think the current service is a potential fire hazard. Dont be surprised if they do it for free.

 

[wk057]

These workarounds appear very dangerous. I would ask the utility for standard 3-phase service with a neutral. It may be a simple re-wiring job at their transformer. Tell them you think the current service is a potential fire hazard. Dont be surprised if they do it for free.

As noted in my post above, changing to 3 phase service with a neutral is a massive and expensive job that just isn't going to happen.

There is nothing unsafe about the existing wiring setup, it's just the wall connectors and Tesla's overly ambitious CYA coding that makes this an issue.

 

[davewill]

...
So, just going to do the brain replacement route when I get time.

I hope you document the process. Folks are definitely going to be interested.

 

[wk057]

I hope you document the process. Folks are definitely going to be interested.

My design time is super limited right now, so will probably be another month or so, but will eventually get to it.

 

[yuhong]

I wonder how well does most CHAdeMO/CCS chargers work with 480V delta.