DC solar directly to EV (no AC!)

[summit]

You, you can change charge current live on J1772.
Here's a random day last month. I removed home loads and other things from the graph. The home base load is something like 500-700 watts, hence the difference between my solar production and EVSE/car charging.

This is using an OpenEVSE. Works great.

slightly Off topic, how is the OpenEVSE configured/connected to charge with only excess current ? it would need to know the available excess power then adjust the pilot signal duty cycle dynamically

 

[summit]

You do need something else to communicate available charge wattage to the EVSE. I'm using a Raspberry Pi, but you could go with anything that can send the correct MQTT commands. Open Energy Monitor is made by the same group as OpenEVSE, they're designed to work together.

ok, got it. Thank you, seems a bit involved; but the throttling is analog and not digital. Did you get them off-the-shelf ? the software stuff seems a bit scary

 

[Ampster]

May I ask, what do you use to throttle charging rate? Does that use the PWM pilot signal?

I don't do anything to throttle charging rate. It is all done in the communications between my Emporia Energy monitor and my Emporia EVSE. The energy monitor has a CT on the bus bar of the main service panel.
The J1772 protocol uses either the pilot signal or resistance to tell the onboard charge what current to draw based on communication from the Emporia Energy monitor. I had considered OpenEVSE but the Emporia system was seamless.

 

[zanydroid]

To clarify what I meant by Emporia possibly working better on-grid than off-grid.

So if you are off-grid and the system is curtailing production b/c storage battery is full, there is a big probability that Emporia will not know you have surplus PV.

If you are on-grid, then Emporia can easily detect that (I have a Vue and solar but not their EVSE).

 

[Ampster]

So if you are off-grid and the system is curtailing production b/c storage battery is full, there is a big probability that Emporia will not know you have surplus PV.

I do not use the Emporia EVSE when the grid is down because it is not connected to my critical loads panel. Even if I did, I would just take it out of excess solar mode, throttle it down and run it like a normal EVSE.
Technically, when the grid is down, there is no such thing as surplus PV because the inverter always has to match its output to the loads.

 

[zanydroid]

Technically, when the grid is down, there is no such thing as surplus PV because the inverter always has to match its output to the loads.

How do you determine how much unused PV can be applied to a dump load? Do you turn on dump loads until the battery starts to supply current?

 

[Ampster]

How do you determine how much unused PV can be applied to a dump load? Do you turn on dump loads until the battery starts to supply current?

I do not have a need for dump loads. As I mentioned earlier, when the grid is down the hybrid inverter modulates its output to match the loads. In that case there is unused PV capacity. If I wanted to use that capacity I would charge an EV, turn down the A/C or turn up the heat pump or bake some cookies. Of course that assumes that the solar has already charged my batteries.

 

[zanydroid]

In that case there is unused PV capacity. If I wanted to use that capacity I would charge an EV, turn down the A/C or turn up the heat pump or bake some cookies.

How do you know how many loads to turn on to fully use your PV capacity? Does the hybrid inverter provide an estimate somewhere of the unused solar? (probably can gather some hints from the MPPT sweep).

 

[Ampster]

How do you know how many loads to turn on to fully use your PV capacity?

During a power outage my first priority is charging my stationary pack which the SolArk manages. Next I then charge my EVs. I have enough PV capacity that I really don't care if I am fully using it. In some power outages associated with storms and cloudy weather, there might not be any surplus capacity.

Does the hybrid inverter provide an estimate somewhere of the unused solar? (probably can gather some hints from the MPPT sweep).

It is pretty easy to see on Solar Assistant when the loads exceed the solar generation and the battery starts discharging. There is nothing in the MPPT sweep that is useful for that exercise..PV Capacity is a function of season, time of day and weather. My goal is usually to get through the power outage rather than worry about unused capacity.

 

[zanydroid]

There is nothing in the MPPT sweep that is useful for that exercise

It can theoretically pull in just enough power to see what the current produced by the string is currently, and estimate the Vmpp. Without needing to pull in too much energy to do so.

 

[Ampster]

It can theoretically pull in just enough power to see what the current produced by the string is currently, and estimate the Vmpp.

Theoretically yes, but of no value that I can see. During power outages, especially ones created by storms, I have more important things to think about than my MPPT algorithm. In addition this digression has nothing to do with the title of this thread.

 

[Ampster]

The appeal of direct DC charging is quite apparent to me, due to "less hardware" and higher potential efficiency.

I think DC charging from home is over rated and I see no value in DC charging from home. On the road I love it. From what I have seen, DC charging requires more expensive hardware if one uses the CCS protocol. Not using that protocol risks battery damage or high voltage arcs. If Electrify America DC charging stations are any example, DC charging equipment requires a lot of maintenance, of which EA is apparently not very good at since many of their charging stations are reportedly out of service.

 

[summit]

I have an off-grid (outage) application needing to absorb excess solar. My roof top has both a DC-PV system and an AC-PV system with older micro-inverter, UL1741 without frequency shift. During outage, the DC-PV powers an AOI, charging a 48v pack and powers the AC critical loads. The AC-PV system can be AC coupled and produce AC power to the critical load, but there is no throttling capability thus a need to absorb excess solar, one way is to charge an EV. It sounds like the Emporia EVSE + Vue may work. I was thinking of hacking my existing EVSE to alter the pilot duty cycle.

 

[mark0]

I am considering a van or a light truck large enough to hold my tiny off-grid home and with the surface area to hold 6 kW of solar panels.

I would use the vehicle battery for both driving and powering my home needs. So I need to charge and draw power from the EV battery.
Most of my home power needs would be met by 12 VDC. An instant tankless hot water heater would required the HV. A smaller heat pump may work off 12 VDC but maybe a higher DC voltage would be required for a bigger unit. I would need to use 120 VAC only occasionally for appliances that don't come in a DC version.

When I mentioned using a buffer battery, I had in mind a relatively small battery that could trickle charge when solar power is low until enough power builds up to transfer to the EV battery. The buffer battery would also serve to collect solar power when it is not possible to simultaneously charge the EV while driving (or while shore charging).

Most light trucks seem to have very accessible electronics. Connecting directly to HV battery or BMS is no problem as can be seen in this example image...


Even though access is easy, the light truck electronics are proprietary and understanding their behaviour would the challenging part.

The Ford e-Transit has an integrated battery and BMS like most small cars. But, it seems the HV systems could still be accessed from points under the hood as seen here...



There are many companies that do EV conversions using Tesla battery packs and motors. Such conversions of Ram Promaster are common, and this would be a viable EV for my use case. This is the reason I had mentioned Tesla swaps before.

There seems to more information available on Tesla battery packs, so this type of vehicle could be more easily hacked.


I'm learning more about the e-Transit, so considering it as a candidate...

A) The EV is off and NOT connected to a charger.

The e-Transit has an option for a 2.4 kW inverter that can be used while stopped or driving. So it seems that the HV battery contactors could be kept closed all the time to access HV without some hacky trickery.

Because of this, I am feeling more confident that it would be possible to solar charge the HV directly while e-Transit is parked.

B) The EV is off and charging from shore power.

I'm gonna forget about trying to charge from solar and shore power simultaneously for now. It's not critical for me to be able to do so.

A mechanical switch to disconnect solar when the charge port is accessed would prevent this from happening.

C) The EV is on and possibly being driven.

I wonder how the vehicle control systems handle simultaneous draw from motor and with the unpredictable draw of up to 2.4 kW from the built-in inverter. For example, if the accelerator pedal is floored and a 2.4 kW appliance is simultaneous turned on, would the control system simply reduce power to the motor (thereby reduce acceleration rate) to comply with the max current draw reported by the BMS?

I also wonder if the control systems (as they are) would be able to handle unpredictable extra current from solar panels. For example, if the brake pedal would be floored and max regen power produced, say 60 kW, and suddenly you emerge from a tunnel into the sun and solar power jumps from 0 W to 6 kW, would the regen power be reduced to comply with the max in-current reported by the BMS?

Any thought?

Thanks!

 

[400bird]

ok, got it. Thank you, seems a bit involved; but the throttling is analog and not digital. Did you get them off-the-shelf ? the software stuff seems a bit scary

Uh, the entire system is digital.
The adjustments are made in 1 amp steps (at 120 or 240 volts)

The open energy monitor is available for purchase. But the Emporia system likely is more consumer friendly.

 

[summit]

I wonder how the vehicle control systems handle simultaneous draw from motor and with the unpredictable draw of up to 2.4 kW from the built-in inverter. For example, if the accelerator pedal is floored and a 2.4 kW appliance is simultaneous turned on, would the control system simply reduce power to the motor (thereby reduce acceleration rate) to comply with the max current draw reported by the BMS?

I also wonder if the control systems (as they are) would be able to handle unpredictable extra current from solar panels. For example, if the brake pedal would be floored and max regen power produced, say 60 kW, and suddenly you emerge from a tunnel into the sun and solar power jumps from 0 W to 6 kW, would the regen power be reduced to comply with the max in-current reported by the BMS?

can't address specifically to the e-transit, but in general the EV BMS can already handle these intermittent demands/regen, such as cabin heat/air-conditioning, foot-of-go-pedal etc...

 

[Ampster]

in general the EV BMS can already handle these intermittent demands/regen, such as cabin heat/air-conditioning, foot-of-go-pedal etc...

Yes, an EV BMS has to do a lot more work than one on a stationary pack. When the pack is full or temperatures are low, it has to limit regen. When the pack is low it has to go into turtle mode instead of just cutting out because disconnecting while going down the freeway could be catastrophic.
EDIT: To further clarify, the BMS communicates to the motor controller so that the motor controller can limit regen or go into turtle mode.

 

[mark0]

can't address specifically to the e-transit, but in general the EV BMS can already handle these intermittent demands/regen, such as cabin heat/air-conditioning, foot-of-go-pedal etc...

Just to test my understanding...

I believe that the BMS can only measure the current in and out of the battery. It can also disconnect the battery in case it detects dangerous levels of current in or out of the battery. But, the BMS itself cannot control this current directly, it only broadcasts (over CAN) the safe maximum current out and maximum current into the battery.

It is up to other control systems to comply with the limits determined by the BMS.

The BMS makes the determination of the safe current level in and out of the battery based on the temperature of cells. The BMS may allow larger currents for short amounts of time (order of seconds). I've read that some BMS's have even more sophisticated behaviour that consider the age and state of health of the battery and will adjust the current limits in an effort to extend battery life too.

Some other system controls the power into the EV motor, and the power out during regen braking.

Other power drawn from the EV battery can't be controlled: lights, radio, heating, cooling, 12V appliances, 120VAC appliances via inverter.

In theory a control system could reduce power to heating & cooling for a short period of time during hard acceleration, but I doubt that there is one that actual does.

Obviously, the control system is able to deal with these unpredictable changing demands just fine.
I wonder how. Does the control system just assume some maximum draw by other device and limit power to the EV motor to account for the worst case of other unexpected loads? Or, is the power to the EV motor adjusted dynamically? For example; the EV motor is drawing max power during a time of max acceleration and the user turns on the air conditioning – will the acceleration be decreased?

Now we are considering adding current into the battery that can't be controlled from solar panels.

The situation of concern in this case is having too much current into the battery with regen braking PLUS solar power.

It is important to understand the behaviour of the control systems in this case.

Would the control system already be able to dynamically adjust regen braking power in response to fluctuating solar power so that it does not exceed the maximum current into the battery as determined by the BMS at the time?
Could the maximum regen braking strength be lowered (by permanent configuration) to account for the worst case (maximum) extra current from solar panels?

 

[Ampster]

Just to test my understanding...

Yes, it is correct that the BMS communicates to the other systems which are the operative mechanisms that control motor output or current draw on the battery. Similar communications occur during charging.

Would the control system already be able to dynamically adjust regen braking power in response to fluctuating solar power so that it does not exceed the maximum current into the battery as determined by the BMS at the time?

It could. The amount of solar that could be placed on a vehicle of that size is insignificant compared to the amount of regen that could be generated when stopping at highway speeds. I have seen estimates as high as 40 kWs for regen breaking.
EDIT: @zanydroid made a similar guess that the solar would be insignificant.

 

[zanydroid]

Other power drawn from the EV battery can't be controlled: lights, radio, heating, cooling, 12V appliances, 120VAC appliances via inverter.

I believe in most EVs these are powered by a second battery (charged from the traction battery by a DC-DC stepdown + charger), and other than the HVAC the power levels are pretty small. The max total power and architecture are mapped out for the cars that are popular to work on (EG for Teslas there's a lot of detail). I think it's usually in the 150-300W range, excluding HVAC. And EVs power up to charge the second battery in a controlled way. So it can predict this load on the traction battery (though it's a small amount so I'm not sure it matters).

In theory a control system could reduce power to heating & cooling for a short period of time during hard acceleration, but I doubt that there is one that actual does.

I don't think it's worth it for the relative amount of power that can be delivered/absorbed during acceleration and regen. Quick google says Model 3 was tested at ~4kW of resistive heater when coming up to temperature.

My guess is that the BMS for supplying current will just have a cap based on the temperature, age, SoC, etc, and this cap would be updated slowly by a controller since the inputs change slowly. However the enforcement of that cap I would assume is done by some kind of fast acting "eFuse" (maybe CTs connected to a controller that opens the contactor, not like FET-based eFuse).

Did you look up the documentation on the Bolt BMS state machine? I believe a lot of booboo situations that cause it to lock down and need a diagnostic tool connected to clear are documented. That would give you some hints on what the BMS checks for.

The situation of concern in this case is having too much current into the battery with regen braking PLUS solar power.

I bet the solar power would be a fraction of the typical and worst case power that regen can feed in. Let's suppose you have 10kW of panels on this thing (seems like a lot, this is more than a lot of houses have on their roof in sunnier locations). That translates to 13.4 mechanical HP. That is much lower than the HP you need to do a regular stop. Quick google says 30kW max regen on old generation leaf and 60kW regen on new generation leaf. It should be proportional to the drivetrain power rating since a lot of the components are shared. So for the 30kW case, if you can put a panel that large, I think this is something to be concerned about. But for higher regen capability, probably not.

Did you write down yet the DC capability of panels you'll put on? I think having those real numbers in place will focus your thinking about potential risks and bottlenecks.