Archived Forum

Aye, not easy without an additional layer of battery storage. I have a Mitsubishi PHEV which will charge at either 10A using the portable EVSE or at 16A using the normal wall charger. I have a timer with a manual override at the moment and have been thinking the same thing.

My thoughts were that the manual override is just a plain 240V switch - no load, just a switch. Using some sort of remote control should be possible for me to replace this with an automated control. The logic I would apply would be along the lines of a level being reached for a sustained period (3kW for 5 mins) and then turning on for a minimum period (say 15 mins) regardless of subsequent solar output and finally just applying the same rules continously - so that the "on" time always exceeds the test time and hence the power will stay on until the solar PV stops generating a high enough current for a while. I have not yet written this logic out, but I hope you see what I mean?

My finger-in-the-air hi/low watermarks are based on day/night rate differences (16.5p/7p for me) and a small "virtue bonus" of using solar where available.

I would also note that the wall mount EVSE (an ABL unit) can be programmed via some serial interface to supply 6/10/16A/32A max to the car - it would be cute to be able to drive that, but that's outside my thoughts this year.

In fact, the WiFi MQTT relay in the online store might just work for me out of the box, so to speak. Hmm.

Thanks for your input Peter.

We are retired so the car spends a lot of its time in the garage during the day. We don't have Economy 7 or 10, or a smart meter at present.

One of the options I'm looking at is to use a LightwaveRF relay or remote control 13A socket. I have other LightwaveRF equipment in use.

The Zoe takes 2.3 kW (10A nominal - no idea of the inrush current) when using its "Granny cable" (occasional EVSE cable) . The solar can deliver up to 3.6 kW in summer but typically 1.5 to 2 kW spring to autumn. When time matters we can use our 7 kW charger.

The LightwaveRF relay or socket can, in theory, be controlled using MQTT from the EmonPi, though I might use my Raspberry Pi instead to avoid the risk of messing up the EmonPi.

I'm having a tough time doing the simple things with Emoncms at the moment, I've forgotten an awful lot in the 25 years since I was last building and programming hardware, hence my interest in learning from others.

As you've said, its not a simple case of turning the supply on and off according to the instantaneous solar output. As far as the car is concerned once plugged in and charging any break and resumption in the charging process without being unplugged and plugged in again (at the car) is regarded as an electrical system fault, and the car's computer sends  out worrying messages. Other models may not be as fussy.

Fred 

Well, now prompted by you I have ordered the WiFi relay controller from the store and I will wire it up "in parallel" with my manual override switch and then at least I have the hardware to le me think through the issues of limits and timing. I have never played with node red, but it does look interesting and a probable candidate for this kind of control.

Hi Fred

You're wasting a lot of energy when charging the ZOE at this low amperage.  I measured a lot of wasted energy when I charge my ZOE below 20A per phase. I have a charger where I can setup charging on 1-3 phases from 6-35A so I tested a lot with it.

With 16A 3phase loading I had to invest 31kw to load a almost fully empty battery. And the battery is around 24kw. With 20A it was just 27kw to load it and with 30A it was almost spot on.

I just can propose to use a charge controller which uses a 0-10V input which lets you actually change the charge instantaneously. The Zoe allows that without stopping charging (when not going under 10A). This charge controller is also mounted on my Bettermann B3200 charging box. 

You can find something here: http://www.ladesystemtechnik.de/ev-plug-technologies/ladesteuerung-mit-potenziometer-37

​Here is the data sheet for it: http://rtp-service.de/wp-content/uploads/2014/09/32A-ICC-Box-Datasheet.pdf 

 I think this way is a little bit nicer instead of cutting of power. I of course understand that you want to shut it off completely but you have to check carefully if the cars are not seeing it as "powerloss" moment and then are sending errors. When I disrupt my ZOE during power by switching of the box it goes into an error mode and will not leave it until I disconnect and then reconnect the charging cable. You have to check that first. With above solution you can actually use a wide power band.

Hi Andi

I'll have a look at your links as soon as I can.

I'm aware 10A charging of the Zoe is less efficient, but if we are charging entirely or substantially from the Solar then we get paid the same whether we use the energy or feed it to the grid, so I'd like to minimise how much energy we buy in to charge the car.

Ideally I would control the car remotely so that it switches charging on and off, but that's another story.

Thanks

Fred

I've been playing with Nodered and what I thought was simple to express took quite a lot of fiddling. My own fault for not keeping up with what the kids are doing nowadays :)

If you want to play see the attached flow. Obviously the MQTT as source needs changing and at the moment all it does is output debug numbers. My local copy also sends me an e-mail every time it switches the output from 0 -> 1 -> 0 etc.

The bit I spent ages pondering over and realised was that you need two triggers for each on high and low watermarks. The first trigger triggers the latch and then the latch keeps the output the same for a period unless reset.

This is the start of how I hope to control my EVSE via the WiFi relay.

Comments welcome.

I did a board about a year ago to do PV variable charging based on the emonrx test sketch and an AD5206

The code is here which you can use in a spare emontx2 if you like, you will need a rolec/mainpine charger

https://github.com/dodavies/Solar-EPC-Controller

​

Diverting solar power to charge an electric car appears more complicated than diverting for water heating.

The fundamental problem is that battery chargers are not well suited to using only power that would otherwise be surplus.  They like their source of power to be continuous.

I started this thread a while ago in order to develop such ideas, but am not aware of anyone following it up.

https://openenergymonitor.org/emon/node/5391

Hi Robin

Thanks for your reply and the interesting link. As a young design engineer 45 years ago I would have been keen to solve all my own problems, but I'm older and a teeny bit wiser now, and I am grateful to be able to learn from others. 

Unlike some DC charged EVs the Renault Zoe has a computer controlled AC/DC converter and the computer is fussy. The minimum charging period is 15 minutes, and if there is a break in the charging period between plugging in the car and unplugging the car it decides the electrical system is faulty and sends out messages to "check the electrical system" (worries the wife a lot). So you can't have an on-off on-off on-off charging situation.

So the extra considerations include:

Minimum charge period 15 minutes.
One charge period
Charge between hours when the sun is high enough in the sky
Detect sunshine intensity
Detect cloud cover and try to estimate probability of continuing sunshine
Is other equipment already using solar output?

Some of these requirements may be impossible to meet, particularly cloud cover prediction since it depends on wind speed and direction at cloud, not ground, level. Can one really trust a weather forecast?

Unfortunately, the only way to remotely control the in-car charger is via a Renault service, where a charge period is set in their database using a calendar via a ZE services App. There appears to be a restriction to one charge period a day. Its also not totally reliable.

But only by analysing a problem do we ever find solutions.

Fred

FB: Thanks for your reply and the interesting link.

The scheme that I've described in that thread is just a logical framework.   Within that framework, rules that are appropriate for the application need to be added. 

Having your charger remain on for a minimum of 15 minutes at a time is simple; deciding exactly when the charger should come on is less so.  The more sensors that you have, such as "sunshine intensity" and "height of sun in sky", the more complex those rules will become.  The coding of complex rules is not difficult, but getting those rules to behave as intended and be bug-free may be challenging.

The system that I've proposed has only one input, namely CT1 which allows the overall energy state of the premises to be monitored.  When the energy state rises, there must be some surplus power.  Apart from the instantaneous energy state, the control logic also knows about the state of each load (but not whether it is actually consuming power) and the passage of time.

By "one charge period", I take it that the car's inbuilt control system has one pre-set charging period with the 24-hour day, similar to the charging period for an overnight storage heater.  When taking an electric vehicle on a long journey, multiple charging sessions may be needed (each one being at least 15 mins in duration), so I would have thought that having multiple charging sessions to suit the weather conditions should be OK too.  

On a "sunshine & showers" day, my suggested logic would detect X amount of surplus power and then apply full power to the car's charging system for at least the next 15 minutes, whatever the weather.  At some stage, power would be removed because PV power was too low.  Later again, when conditions improve, the next burst of power would commence; and so on while surplus power was available.

Having small loads come on first is a convenient way of determining when surplus power is sufficient to warrant firing up the first "continuous" load.

The restriction to one charge a day is for remote control of charging via the Renault services app. You can manually charge as many times as necessary. Sorry if that was not clear. 

Fred

If the supply that you use for charging goes 'on' for at least 15 minutes, then goes 'off' for a while before returning to the 'on' state for another 15+ minute period, will that be acceptable to the vehicle's diagnostic system? 

Or is the process of doing a manual charge more complex than just switching the power supply on and then off 15+ minutes later.

No and Yes:

Manual charging in the intended way is simple:

1) open charge flap using remote or in-car switch
2) plug in charger (car locks connector in place)
3) car identifies charger type (2.3/7/22/43 kW) and begins charging
4) if 100% charge achieved car stops charging, else manually stop charge by switching off supply (some charge points may cut supply after a predetermined time eg. 30 minutes)
5) use remote or in-car switch to release connector
6) unplug charger
7) close charge flap

If for any reason the supply stops and restarts between (3) and (4) the car will issue "Check Electrical System" or, worse, "Battery Charge Impossible". The former will go away if the car is unplugged and left for a short while, whereas the BCI message has given some people real problems, the car still refusing to charge having been unplugged and taken to a another charger and plugged in. 

I've known the car carry on charging when an interrupted supply is resumed even though it presents the CES message.

Timed Control of charging assumes (1) and (2) have been performed. 

TC can be set from the car using the touch screen or via an Android or iOS app which communicates with the Renault ZE services web site. In the latter case Renault transmits the start and stop times to the car via the O2 network. It isn't 100% reliable. If you are not in an O2 area it doesn't work.  If the car has charge times set it won't charge outside the set period even though you carry out the manual procedure. The charge timer settings have to be canceled first. This has caused some anxious moments for those who use the timer to regularly charge overnight on economy 7 if they have needed to charge in the daytime.

It is the TC method that specifies only one charge period per day with a minimum of 15 minutes. I haven't used it yet.

At the start of charge time (3) will be performed.
When full charge or end of timed period (4) occurs.
Some time after the end of charge (5) (6) and (7) have to be performed before driving (obviously).

Renault monitors charging of the car and the use of the battery. Start and end of charge are notified to the user by email. Charge state is monitored by Renault every 20 minutes when the car is plugged in.

Hope that clarifies. Renault are not very forthcoming on detail and rather slow to remedy bugs in their software. Sometimes translation errors add to the fog.

Fred

The key point to establish is whether it's OK for your charger's power supply to be activated for periods of some minimum duration, and deactivated without warning in between.  From the info that you've provided, it seems that this would be OK.  In which case, the logic that I've outlined in my "smarter loads" thread may be of use to you.  

Send me a PM if you would like any practical help with this.

Is it possible to turn the charging Off or On via a webpage by maybe setting and cancelling the timer? If it is you could use something like Openhab to periodically take snapshots of surplus power and then send a new timer schedule out via the internet to turn the charging on once a threshold is reached. You could then run a 15 minute timer before getting it to check that the Grid power draw is not excessive, and if it is turn the charger off. This of course assumes it doesn't get upset when the timer is cancelled and recreated.

I use something similar albeit a simple On and Off command to operate my upstairs cooling in summer, it works well.