Mini-QC Rapid-Charger (RC) Project for LEAF QC Port

[valerun]

The latest mods to the mini-QC unit include:

1. user-settable maximum output current to better match
the input power source being used.
2. added an external J1772 socket to receptacle box to allow
powering the mini-QC from a J1772 source.
3. modified the inrush startup control to allow for the longer
time needed to get the PFC stage powered up on 208v input.
4. replaced the output stage discharge transistor with a relay.
5. reprogrammed the user display for better information utility.

Yes, it still lacks an isolated output, but Valery's latest work
might have solved that problem.

Hi Gary - great to see you again here!

Is this based on a Due? Also, what is the power stage it controls now?

Thanks,
Valery

 

[garygid]

Valery,
Our mini-QC is still AVR-CAN based, using your original display in the
non-isolated PFC plus Power two-stage charger kit, now rather modified.

Between doing taxes and recovering from a whiplash, I have not gotten
much done recently.

Your success charging a Rav4EV with injected regen (when ready to
drive) has been noted, and we need to find enough data on the Rav4EV's
two CAN buses to well determine when and how to manage the tapering
off of the regen as the battery pack gets fuller. Have you found that data?

Any ideas on where to put a QC socket, to charge the Rav4EV from a real QC?

Cheers Valery, hopefully we can meet sometime.

Our 3D printing friend from Canada should be passing through here on Easter Sunday.

 

[valerun]

Valery,
Our mini-QC is still AVR-CAN based, using your original display in the
non-isolated PFC plus Power two-stage charger kit, now rather modified.

Between doing taxes and recovering from a whiplash, I have not gotten
much done recently.

Your success charging a Rav4EV with injected regen (when ready to
drive) has been noted, and we need to find enough data on the Rav4EV's
two CAN buses to well determine when and how to manage the tapering
off of the regen as the battery pack gets fuller. Have you found that data?

Any ideas on where to put a QC socket, to charge the Rav4EV from a real QC?

Cheers Valery, hopefully we can meet sometime.

Our 3D printing friend from Canada should be passing through here on Easter Sunday.

I have a long-overdue trip to LA in plans so maybe can time it to meet you guys in person? Maybe I even take our modified RAV4EV ;-)

I don't think you should care much about what the exact stock profile is for recharging. Li-ion battery is a Li-ion battery. CC-CV and go. You know the CV points. To be on the safe side, I would probably start tapering current at Standard mode CV point. That would correspond to a classic 80% SoC. We know that Tesla-style batteries take 2C+ from Superchargers up to that point. Worst case, BMS will shut you off. Then, once current drops to the stock onboard charger level (~30A), I would move a CV point to Extended CV voltage, run 30A until then and then taper. Final cutoff seems to be done at ~10A battery side.

Anyway that's how I would approach it. Gets you a working system in a day. Then you can optimize & iterate

 

[jclemens]

from a real QC?

Cheers Valery, hopefully we can meet sometime.

Our 3D printing friend from Canada should be passing through here on Easter Sunday.

I have a long-overdue trip to LA in plans so maybe can time it to meet you guys in person? Maybe I even take our modified RAV4EV ;-)

Yes, now is a good time to chime in. I will be taking a vacation to California to visit. Starting in San Fran to crash with a childhood friend of mine, and then drive down to LA with my GF so she can visit her old friends, then down to San Diego for a day (monday) at LegoLand with her little one.
San Francisco day will be the 17th, the next morning we will check out the coast and head down to LA, Will be on the way to Carlsbad by (Easter) Sunday, and we can stop by Laguna Hills for the "EV celebrities". Flying out of San Diego on Tuesday morning (22nd))
Not sure if we will have time to check out emotorwerks while in the Bay area, but I'd thought I'd mention I'll be nearby.

We are still working out the specific schedule/details.

 

[TonyWilliams]

Any ideas on where to put a QC socket, to charge the Rav4EV from a real QC?

For the JdeMO project, the CHAdeMO charge socket goes under the hood.

The connection for DC to the battery is at the battery mounting lugs, where the fuse is located.

The taper profile for the battery is really easy to figure out, as Valery mentioned above.

 

[garygid]

Joel, are you in SF on the 14th, not the 24th of April?

 

[jclemens]

Joel, are you in SF on the 14th, not the 24th of April?

aah, i was a week off, our day in San Fran is actually the 17th. (land on the 16th)
We will be in San Diego to fly back on the 22nd. So both the 14th and the 24th are wrong. Sorry about the confusion.

 

[garygid]

Joel, sent you a PM about meeting on Easter Sunday.

 

[valerun]

Joel, sent you a PM about meeting on Easter Sunday.

perhaps we can coordinate via email - I will send to whose emails I have - could you pls add the rest?

 

[valerun]

Hi guys - would love your feedback on something.

As you know, we have successfully implemented full CHAdeMO control on Arduino Due and demonstrated charging of a number of vehicles from a number of sources. You also know that we are in the process of integrating all this into our chargers so that one does not have to have a separate CHAdeMO controller box. In doing so, we are realizing that a Due board as-is is quite an overkill. Also, ideally we would use the same MCU across all our products (JuiceBox, our DC chargers, CHAdeMO controllers, BMS, etc). But due to the size and cost of a Due, it is hard to integrate into some of those products.

So we are considering building our own board based on the same chip as Due (ATSAM3X8EA ARM with 512KB flash, 96KB RAM, etc). We would like to shoot for a form-factor similar to the Pro Mini or Teensy 3.x (https://www.pjrc.com/teensy/teensy31.html" onclick="window.open(this.href);return false. I think this is possible if we...:
* Use LFBGA package of the same 144-lead chip: http://www.digikey.com/product-detail/en/ATSAM3X8EA-CU/ATSAM3X8EA-CU-ND/3128690" onclick="window.open(this.href);return false;
* break out the programming circuit based on ATmega16u2 into a separate board - we wouldn't actually design / build that one - we would just buy a programming adapter from someone like http://www.geeetech.com/iduino-due-pro-board-with-due-pro-usbserial-adapter-p-655.html" onclick="window.open(this.href);return false;. Note that these guys use a similar approach that I am proposing but their main board is still too big - largely because they are using LQFP package that is 5-6x the surface footprint and they have all the Due pins in connectors.
* limit the number of external pins to just ~45 vs ~90 on a Due. We would have: 8 ADC, 16 digital, 4 CAN, 6 UART, 4 SPI, 6-8 programming / power. Some of those could be done as pads on the bottom - Teensy-style.

What do you think? This would have ~1/8th of the footprint of a Due and would be completely indistinguishable from a Due to Arduino IDE. I think this could be an interesting board in its own right...

Valery

 

[Berlino]

We know that Tesla-style batteries take 2C+ from Superchargers up to that point.

Are you talking about tests that have been done on Tesla-style batteries?

In actual recharging at Superchargers, I don't think Model S drivers have ever seen anything above 1.7C (102kW with the 60kWh battery).

I suppose that could have change recently with the software upgrade to 5.9.

 

[Kris1]

I'm in the same situation as Martin, access to 23kw 3 phase 415v~ here in Australia. If a PFC corrected 3 ph 415v~ 22kw Chademo did emerge one day what sort of weight & price might we be looking at? The Leaf is nice but the 150km apron strings are cramping my style. A portable chademo that charged in an hour over lunch would be real nice.

 

[valerun]

Hi guys - would love your feedback on something.

As you know, we have successfully implemented full CHAdeMO control on Arduino Due and demonstrated charging of a number of vehicles from a number of sources. You also know that we are in the process of integrating all this into our chargers so that one does not have to have a separate CHAdeMO controller box. In doing so, we are realizing that a Due board as-is is quite an overkill. Also, ideally we would use the same MCU across all our products (JuiceBox, our DC chargers, CHAdeMO controllers, BMS, etc). But due to the size and cost of a Due, it is hard to integrate into some of those products.

So we are considering building our own board based on the same chip as Due (ATSAM3X8EA ARM with 512KB flash, 96KB RAM, etc). We would like to shoot for a form-factor similar to the Pro Mini or Teensy 3.x (https://www.pjrc.com/teensy/teensy31.html" onclick="window.open(this.href);return false. I think this is possible if we...:
* Use LFBGA package of the same 144-lead chip: http://www.digikey.com/product-detail/en/ATSAM3X8EA-CU/ATSAM3X8EA-CU-ND/3128690" onclick="window.open(this.href);return false;
* break out the programming circuit based on ATmega16u2 into a separate board - we wouldn't actually design / build that one - we would just buy a programming adapter from someone like http://www.geeetech.com/iduino-due-pro-board-with-due-pro-usbserial-adapter-p-655.html" onclick="window.open(this.href);return false;. Note that these guys use a similar approach that I am proposing but their main board is still too big - largely because they are using LQFP package that is 5-6x the surface footprint and they have all the Due pins in connectors.
* limit the number of external pins to just ~45 vs ~90 on a Due. We would have: 8 ADC, 16 digital, 4 CAN, 6 UART, 4 SPI, 6-8 programming / power. Some of those could be done as pads on the bottom - Teensy-style.

What do you think? This would have ~1/8th of the footprint of a Due and would be completely indistinguishable from a Due to Arduino IDE. I think this could be an interesting board in its own right...

Valery

In the meantime, a couple of snaps of the latest boards (still based on a Due). On the photos:
* All sub-components separate - from the top and bottom
* fully assembled system - 128x128 color LCD, wifi, BT, wireless remote, and dual CAN (in addition to all other standard periphery required for CHAdeMO control and control of our DC chargers)

V

 

[valerun]

I'm in the same situation as Martin, access to 23kw 3 phase 415v~ here in Australia. If a PFC corrected 3 ph 415v~ 22kw Chademo did emerge one day what sort of weight & price might we be looking at? The Leaf is nice but the 150km apron strings are cramping my style. A portable chademo that charged in an hour over lunch would be real nice.

~15kg, $4000-$5000.

 

[TonyWilliams]

We know that Tesla-style batteries take 2C+ from Superchargers up to that point.

Are you talking about tests that have been done on Tesla-style batteries?

In actual recharging at Superchargers, I don't think Model S drivers have ever seen anything above 1.7C (102kW with the 60kWh battery).

I suppose that could have change recently with the software upgrade to 5.9.

135kW is the new Supercharger maximum.

 

[GregH]

We would have: 8 ADC, 16 digital, 4 CAN, 6 UART, 4 SPI, 6-8 programming / power.

And that's not overkill?

 

[valerun]

We would have: 8 ADC, 16 digital, 4 CAN, 6 UART, 4 SPI, 6-8 programming / power.

And that's not overkill?

not really. note that numbers are PIN counts, not port counts. E.g., 4 CAN pins are just 2 CAN channels, 6 UART pins are 3 UART channels.

In our chargers, 5 analog pins are used in a basic configuration (input voltage, output voltage, output current, heatsink temp, inductor temp). Plus 8-10 digital pins are used. So with 8 analog and 16 digital we are just building a small amount of headroom.

re overkill on a chip side - CAN + WiFi + basic charger operation takes ~120KB in flash footprint already - and that's before we handle any bidirectional comms and user auth workload...

 

[Kris1]

What you've got already looks fairly neat to me already Valery. Is the size reduction your chief motivation for rolling your own processor card? Won't that mean you have to redo all the controller boards it plugs into as well? I guess if you have the volumes it could be worthwhile?

Programmers are a pain for customers, wired or wireless serial firmware upgrades would be a worthwhile feature IMHO. Also I'm intrigued with your use of Roving Networks modules, I've never had an RN174 remember it's config for more than 48 hrs, flaky & poorly documented in my experience.

Does the $4k include the chademo plug?

 

[valerun]

What you've got already looks fairly neat to me already Valery. Is the size reduction your chief motivation for rolling your own processor card? Won't that mean you have to redo all the controller boards it plugs into as well? I guess if you have the volumes it could be worthwhile?

Programmers are a pain for customers, wired or wireless serial firmware upgrades would be a worthwhile feature IMHO. Also I'm intrigued with your use of Roving Networks modules, I've never had an RN174 remember it's config for more than 48 hrs, flaky & poorly documented in my experience.

Does the $4k include the chademo plug?

Hi Kris - yes, I believe we have a basic product already (just need to sort out the reliable plug supply - working on this with Joel). It would be further simplified if we had a powerful processor card that can handle everything. But this is just 20% of motivation for creating such a card. The main motivation comes from the fact that we are running out of memory on our current card for all our other products (our regular chargers, EVSE, etc). So we need something more capable. And the logic was, if we are investing into making a card anyway, why not go for the most powerful chip that Arduino supports.

RN-171 works great - one of our units is working great for 4 months already, pushing data every day.

$4-5k would include everything.

V

 

[Kris1]

So early adopters would need to buy a separate programmer for firmware upgrades Valery?