OpenEVSE - Open Source Charging Station

[GlennD]

If you download the source from lincomatic's github archive the RAPI commands are in the RAPI file as comments.

[stopeject]

That was my question- is anything more user friendly going to happen?

[GlennD]

I was asked how the OpenEVSE timer works. OpenEVSE is totally open source and Goldserve added the timer and Lincomatic enhanced it.

This requires an !2C timer chip like the DS1307 or even better the DS3231N. Originally separate timer boards were used but Chris at OpenEVSE added them to the display boards. Currently he offers a color board with the better chip. On some of the units a have built I have used VFD and Monochrome units with timer chips.

The OpenEVSE opens up a charging window with adjustable start and stop times. The defaults of 5 min after midnight and 5 min before 7AM work for most people.

With the timer enabled the display says waiting until the start time. Then it turns into an ordinary EVSE and it starts charging the car assuming it is plugged in. Most cars charge fully in less than 4 hours so the car tells the EVSE to open its contactor and the display switches to connected. At the stop time the window closed and it is back to waiting.

If you need energy right now a short press of the button returns the EVSE to normal. The operating window will stay open until the stop time rolls around and then it returns to waiting.

The whole point of the timer is to allow charging late at night when the costs are low.

With the timer disabled the display says sleeping instead of waiting. The button function stays the same.

[evtifosi]

I have had the 50A OpenEVSE for just about a year, and just started to use the on board timers to great effect.

For a few months I used the Leaf timers, but those became harder to work with when I needed to charge during the day for some reason. I eventually just turned the timers in the car off.

I started playing around with the current options to reduce the amount of charge I was getting, and then finally decided to just use the OpenEVSE timers. I have them set to start at 4:30am - and shut off at 6:55am. This will charge my 2015 SV from close to 20% to 100%, and if I am around 10% or lower it will typically charge it back to around 85%. It seems like a good setting for our limited use. I am on PG&E E-6 TOU rate plan and have Solar Panels...so this guarantees that we are charging off peak.

I got tired of remembering to go out to the garage after 9PM or plug it in when I woke up in the morning - so I love the timer function - just wish I had used it sooner!!!

edit: start time is 4:30 AM - so 27.5A for 1:25mins for most of our charging

[GlennD]

I have had the 50A OpenEVSE for just about a year, and just started to use the on board timers to great effect.

For a few months I used the Leaf timers, but those became harder to work with when I needed to charge during the day for some reason. I eventually just turned the timers in the car off.

I started playing around with the current options to reduce the amount of charge I was getting, and then finally decided to just use the OpenEVSE timers. I have them set to start at 4:05am - and shut off at 6:55am. This will charge my 2015 SV from close to 20% to 100%, and if I am around 10% or lower it will typically charge it back to around 85%. It seems like a good setting for our limited use. I am on PG&E E-6 TOU rate plan and have Solar Panels...so this guarantees that we are charging off peak.

I got tired of remembering to go out to the garage after 9PM or plug it in when I woke up in the morning - so I love the timer function - just wish I had used it sooner!!!

Yeah, the unfortunate thing about the Leaf timer is that the over-ride button charges to 100%. I charged both at work and at home at 80% so a timer on the EVSE was better for me. Now they removed the 80% option so it is not as important. As far as I can tell the eGolf also charges to 100%. The documentation on charging is lacking.

[GlennD]

Since I am retired and I no longer have a commute playing around with the window will not work for me. I will just charge at around 60% to 100%. I do not like having under a half charge.

My Mercedes normally charged to 80%. If you pushed the range button it charged to 90%.

[DfGriggs]

Regarding:
So perhaps I should build the 50A kit for nominally the same cost and gain that extra 3.5A?


I think the reasoning for choosing the 50A kit may include:
1) Future-proofing to some extent.
2) For Leaf's, the marginally quicker charge, as described.
The marginal cost increment then provokes the decision.

Of course, one must have a 50A circuit available -- so some folks will buy the 50A kit, but keep it turned down to 24kW max.

[GlennD]

A 50A EVSE is really a 40A EVSE (80% of 50 =40a) but unless you are planning on a Tesla derived car no other car draws over 30A. The cars's internal charger only takes what it needs and it ignores the rest.

If you have a Volt then your 50A EVSE will only supply 14A. An enhanced Leaf will draw 27.5A and ignore the rest. There is a great price delta going from a 30A EVSE and a 40A EVSE. Any car that needs 40A is also J1772 complyant and it will charge at 30A just 25% slower.

Basically the EVSE broadcasts the maximum current available and out of that the car takes what it needs and it ignores the rest. If the EVSE offers less than the car needs the car matches the current.

For example the maximum for a 30A circuit is 24A. A Leaf would set itself to 24A and it would be happy even though it could charge at 27.5A.

[GlennD]

Some thoughts on Barbouri’s OpenEVSE board ver 4.2. The source was available earlier and the board works for me. I think the current version is 4.23. All of his designs are Descrete vs Chris,s SMD design. I am an old fossil and I like repairable vs throw away

First the regulator IC is no longer available from Digikey so you may have to use my descrete 78L05 mod. Before it was was very expensive.

The most expensive chip on the board is the mid400 at nearly $5. It is now second sourced from China for less than $1. I was using 6N139’s but at this price why not use the real thing.

I use 47K 1/2W thin film resistors vs the original 28K’s. The board has a space for fuses for carbon resistors and I put in 1N4004’s instead since current firmware requires half wave sensing. It makes a good spot for the diodes. You can also use 1N4148 shunt diodes across the input but the fuse spaces are neater

The current board uses Nick Sayer’s GFCI test. It turns out that that is the same as RLY2 on the board. You just need 2 connectors and a 2 pin header at RLY2. The resistor value is the same.

There is no standard for the I2C pins and Chris has them swapped vs Barbouri so if you use a stock OpenEVSE display you need to swap the wires.

I do not populate the other resistors or headers since I do not use them.

Barbouri has gone hog wild with bypass caps but since the holes are there I use them. 1 chip caps are less than a nickel on eBay so it hardly adds an expense.

I use both Chris’s DIY ver 2 and Barbouri’s ver 4.2 boards. Both work well but Barbouri’s output is more elegant. With Chris’s boards I add 1N4148 shunt diodes to the mid400 inputs. Since the most voltage they ever see is the 3V LED they work well.

[GlennD]

By the way, The Chinese 100A 20MA coils work fine for the GFCI. They are less than $3 each. You have to add the test winding and a connector But for cheap me it beats Chris's retail price. Connectors and pins are on eBay as well as the coils. That same coil without the test winding's works for current also. I change the setting in the current section to 185 and it is well within Lincomatic's 1A spec. With a little patience you can calibrate it exactly.

I got to thinking about the GFCI coil. The JuiceBox uses the same coil for the GFCI and the current measurement. I experimented with cheap Chinese coils and this is what I came up with. In all I tried 3 or 4 coils with no success but these work fine.