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

[garygid]

If you already have CAN on your control board,
would it be better, and perhaps more noise
resistant, than using serial RS232?

Are you thinking of TTL level RS232, or plus/minus
12 volts, which takes more chips/hardware?

Have you tried the "Serial" I/O on the Due
(not the USB virtual Comm Port)?
I have not.

 

[garygid]

If you can produce a power supply that is
sufficiently "performance-capable" for use
in a mini-QC EV-charging application, then
you would not be actually selling any
mini-QC controller at all, right?

Is there some open standard interface for
controlling power supplies, or do we need
to create another non-standard standard?

People might obtain a mini-QC experimental
controller kit for personal, non-commercial use.
They might try it with your power supply, or
perhaps some other power supply.

Is that your goal, to keep yourself out of the
QC end of the business, and avoid "commercial
use" of the mini-QC type of technology?

You might sell an optional touch-screen user
interface (UI) in your power supply, which could
control the power supply in various more or less
manual modes, that have nothing to do with
the actual internal control of the power supply,
and no code for mini-QC operation.
This Manual user-interface control code could
be for the Arduino Due and Open Source.

If others happened to adapt the code to use
the CAN (and other unused I/O hardware) on
your UI-board to implement some version of
mini-QC charging, and they kept that code
entirely for private, non-commercial use...
that might be a fortunate circumstance, da?

 

[garygid]

Valery,

As soon as you can suggest the control interface
details, hardware and speed, and command
data, I will look it over to see if it appears sufficient
for Manual Control, and future more-automated uses.

Once we agree on a "standard", I will try to include
that in my Due Sketch. Then you could use it for
exercising the power supply, and logging data,
in the Manual Mode.

Once we see that the power supply seems to
work well enough to try a mini-QC control,
we can see if somebody is able to produce
an experimental version of the UI code,
to help facilitate testing, and find someone
with a LEAF who is willing to risk their car
testing something like the mini-QC control.

 

[valerun]

noticed that small 5uH inductor's core was overheating at 150A primary current - calculation showed almost 100W core loss due to flux swings. Replaced with an air core 3.5uH inductor - gained another 0.3-0.5% efficiency! This is becoming addictive ;-)

 

[garygid]

Valery,
The Control Port:

In order to debug the "Valomo" control mode,
we really should have a power supply simulator,
since we do not have the iso-power supply
itself. Indeed, to debug the control on the
power supply, you would want a User-Interface
simulator, or a "real" first-version of the UI.

Since portable computers (laptops and such)
do not have real serial ports anymore, I guess
we could use a second Due, programmed to
act like a "Valomo" power supply.

Presumably we need a 3-wire connection,
Ground and Tx and Rx for a full duplex
connection, probably 3.3 volt compatible
signal lines?

Or, do we need to add a level shifter chip?

Do you already have the necessary hardware
for this "serial" control port on your control
board?

We need to get the chicken in front of the
wooden-wheeled conveyance, right?

How are you planning to allow for connection
of your power supply to a Jolomo plug, for
experimental purposes, of course.

All this would be for dry, indoor use only,
it seems, right?

The cable needs 2 power wires, perhaps rated
for 80 or so amps, a much lower rated protective
ground, and around 8 signal or low power wires,
typically for 12 volt use, with at least one pair
carrying a "relay-operating" (not yet measured)
current to/from the QC-type vehicle.

Any sources for suitable cable?

Presently, we use several wires/cables for our
experiments.

 

[garygid]

What is this "Nissan Hackathon" in January?

I do not find any info about it.

If I had a working iso-power supply, or perhaps
a working power supply iso-stack to supply the
DC voltage, and an output buck regulator
stage to control with the same PWM that
we use now, then maybe we could demo
something. But, getting something new
working would be difficult, I suspect.

Or, get a $500 isolation power transformer
from Chicago, and use it to "isolate" our
present 12 kW (240v 50 amp AC in) box?

 

[garygid]

Due to a change in circumstances, we are losing
our instrumented vehicle. That alone will set our
project back substantially. There might be a new
owner who would want the logging-mods, but
we would need to know soon, before these
modifications are removed from this 2011 LEAF
SL with QC port and the Rear-View package.

Also, we need to find someone to "buy" our
interest in the only working, non-commercial,
experimental non-iso mini-QC device known
to exist. If it stays here, or goes to someone
nearby, we might be able to continue to aid
the mini-QC development effort.

However, since I will not have one of the
new iso-power supply to experiment with,
my involvement might become just awkward
untested software support.

At least, I am considering ways to properly get
the results of our investigations into the public
domain, although perhaps "for private, non-
commercial use only" might be required.

We are considering options, offers, and suggestions.

 

[garygid]

I suspect that the person buying the LEAF will
have the first opportunity to obtain the
non-isolated mini-QC, if they have sufficient
skills and attitude to be able to use it safely.

A high bid would likely be a priority event,
but with the same conditions, mostly safety.

We are not selling the mini-QC technology,
but only selling the highly modified and
now-assembled kit, which just happens
to contain the AV-CAN based custom controller
that we have been using to explore the mini-QC
charging process and associated events.

Given a 240v AC 50 amp circuit, and a "willing"
LEAF with a QC port, we should be able to
demonstrate the mini-QC process.

In our explorations, we successfully charged
four different LEAFs. However, YMMV.

We continue to work toward getting the
results of our explorations into the public
domain.

The custom firmware is free, not perfect,
with no warranty or guarantee of any kind,
and for private, non-commercial use only.

 

[garygid]

As soon as Valery gets a control interface defined,
I will attempt to code a UI to make requests of
his Valomo power supply, and code a Volomo
Power Supply simulator, to help test the UI code.

The UI code will help Valery debug his command
interface, and refine the functionality of his
non-isolated power supply.

When those code additions are working, I will
attempt to add the mini-QC control signal
handling, and the associated CAN communication.

Then, with two simulators, and the User Interface,
I should be able to perform a simulated mini-QC.
The Valomo Power Supply, the UI, and the QC-Port
simulator.

Of course, all suggestions welcome.

 

[valerun]

If you can produce a power supply that is
sufficiently "performance-capable" for use
in a mini-QC EV-charging application, then
you would not be actually selling any
mini-QC controller at all, right?

Is there some open standard interface for
controlling power supplies, or do we need
to create another non-standard standard?

People might obtain a mini-QC experimental
controller kit for personal, non-commercial use.
They might try it with your power supply, or
perhaps some other power supply.

Is that your goal, to keep yourself out of the
QC end of the business, and avoid "commercial
use" of the mini-QC type of technology?

You might sell an optional touch-screen user
interface (UI) in your power supply, which could
control the power supply in various more or less
manual modes, that have nothing to do with
the actual internal control of the power supply,
and no code for mini-QC operation.
This Manual user-interface control code could
be for the Arduino Due and Open Source.

If others happened to adapt the code to use
the CAN (and other unused I/O hardware) on
your UI-board to implement some version of
mini-QC charging, and they kept that code
entirely for private, non-commercial use...
that might be a fortunate circumstance, da?

Hi Gary - thanks for your questions and comments.

Our company's end-state objective is to have a NRTL-approved Quick-Charge product that works with cars supporting CHAdeMO and J-Combo fast charge standards. That could take a while (e.g., 9-12 months).

Our company's interim objective is to offer a Quick-Charge product that works with CHAdeMO cars. It will be sold for experimental use and will be explicitly marked a not certified by any authority. It will be sold in a kit or complete form. I am hoping this could be achieved in the next 4-8 weeks. We have definitely invested into the hardware part of it over the last 2 months and made great progress as you can see. So I am optimistic.

To maximize the rate of progress towards that goal, I have suggested a modular approach to insert an API layer between power module and QC control module. We might or might not integrate these back together into a single control board later depending on what you and I agree on.

My intention is to publish the design as open source hardware and software, free to use for non-commercial purposes. I understood from the earlier posts in this thread that your intentions are the same, correct?

On the manual vs. Auto control: I would not characterize our product as a power supply. I would characterize it as a Li-ion battery charger. The product going on pre-production sale next week will be fully equipped with manual and automatic non-QC control in the same way our current 12kW units work. This means that if you have access to your battery, you will be able to charge your car without any additional software and hardware. You can configure the unit for your battery specs and it will be fully automatic, including J1772 compliance and all. Basically, all functionality you can see at http://emotorwerks.com/tech/electronics" onclick="window.open(this.href);return false;, only in a fully isolated, 20kW package. The only thing it will be missing is QC control which would allow to charge your car even *without* direct access to the battery. This is what I am hoping to work with you on.

Hope this makes sense.

Valery.

 

[valerun]

What is this "Nissan Hackathon" in January?

I do not find any info about it.

If I had a working iso-power supply, or perhaps
a working power supply iso-stack to supply the
DC voltage, and an output buck regulator
stage to control with the same PWM that
we use now, then maybe we could demo
something. But, getting something new
working would be difficult, I suspect.

Or, get a $500 isolation power transformer
from Chicago, and use it to "isolate" our
present 12 kW (240v 50 amp AC in) box?

Gary - we will loan you one of our first units provided you still have the ability to test on a real vehicle. Please email me.

Thanks,
Valery

 

[valerun]

Presumably we need a 3-wire connection,
Ground and Tx and Rx for a full duplex
connection, probably 3.3 volt compatible
signal lines?

Or, do we need to add a level shifter chip?

Do you already have the necessary hardware
for this "serial" control port on your control
board?

yes, 3-wire, 3.3v compatible.

I agree that eventually we will want to just use CAN on our new Due-based control boards. But we are not quite there yet, with the iso-stage hardware development taking precedence over the last few weeks.

Re output power cable, we will hack something together (similar to your multi-cable approach). The biggest problem is the plug which I am hoping we can get from Joel.

 

[garygid]

It appears that the Due hardware includes two
3.3 volt serial ports, so that ground, Tx, and Rx
would be sufficient. Not true RS232, but...

I will start with 115200 baud, 8 none 1, with
no flow control, just full duplex.

Perhaps to get started, I will use ASCII characters,
commands starting with 'c', and each line of
communication ending with a LineFeed character.
Perhaps requests for information start with
the '?' character (or maybe 'r').
Setting parameters might start with an 's' character.

For example, perhaps:
(commands marked (*) might be optional)

?vm<>
Request the Top (highest) voltage that the power
supply can output. (*)

?vmz<>
Request the Top (highest) voltage that the power
supply can provide at zero amps out. (*)

?vmt<>
Request the Top (highest) voltage that the power
supply can provide over the Max-Amps range.

?vmr<>
Request the Top (highest) voltage that the power
supply can provide over the MaxOut Amps Range. (*)

Similar commands for amps:
?am<>
?amz<>
?amt<>
?amr<>

sam298<>
Set MaxOut Amps to 29.8 amps.

ca312<lf>
Output 31.2 amps, unless limited by MaxOut Volts.

svr243<>
Set voltage ramp to appropriately 243 volts per second.

sam001<>
Set MaxOut Amps to 0.1 amp.

cv452<>
Ramp the output voltage up to 452 volts,
unless limited by MaxOut Amps.

more later...

 

[garygid]

The possibility of getting a unit for testing
is exciting, thanks. I do still have my own
LEAF, with the QC port. I will work towards
getting suitable circuitry installed, although
I suspect that 50 amps is the best that I
will be able to install. At our "other" location,
we have 50 amps available now.

Will the 12 kW and the 20 kW charger
be the same unit, at least for now?

Is the 20 kW the AC in, or the DC out?

Presumably the user will have the ability to
connect your 20 kW charger to an AC Power
Source of less than 100 amps?

How does the AC in-amps get "set" before the main
power is used to charge the first two stages?

Yes, we are eagerly watching your progress.

 

[garygid]

If we wanted to modify our present AVR-CAN code
to not control the charger hardware, but just do the
Serial "sequencing" of your new charger, and handle
the mini-QC interface, I think that we could have
something working within a week or so of establishing
the command details, provided that we had a charger
with sufficient capabilities to test with.

In any case, we would need to build another control
board, and build one for you to use for testing as well.

Maybe we could send a schematic of the circuit that
we use now, and it would be easy for you to make
a board to populate?

I will try to work toward getting the circuit that we use
drawn out, and broken into two parts, the part for
charger control, and the part for the QC interface.

 

[valerun]

Will the 12 kW and the 20 kW charger
be the same unit, at least for now?

Is the 20 kW the AC in, or the DC out?

Presumably the user will have the ability to
connect your 20 kW charger to an AC Power
Source of less than 100 amps?

How does the AC in-amps get "set" before the main
power is used to charge the first two stages?

Yes, we are eagerly watching your progress.

20kW is DC out. AC is in 21.5kW or so.

20kW system is a completely different architecture. It does reuse some building blocks from the 12kW but only at a rather low level of IGBT drivers, control chips, etc. 20kW isolated system is current 2x 10x10x8" air-cooled boxes connected together (DC power and a few signal wires). Each ~20lbs.

It can of course work at any power below 20kW and I would advise initially use it at below 15kW for testing QC.

The AC power is not set directly. You set the DC output level and that defines AC power drawn. Given the efficiency ranges only between 91-94%, the error is very minimal. This also saves at least $20 from BOM. Just trying to make it as affordable as possible for you guys to build and as easy as possible to debug. Trust me on this.

Thanks,
Valery.

 

[valerun]

If we wanted to modify our present AVR-CAN code
to not control the charger hardware, but just do the
Serial "sequencing" of your new charger, and handle
the mini-QC interface, I think that we could have
something working within a week or so of establishing
the command details, provided that we had a charger
with sufficient capabilities to test with.

In any case, we would need to build another control
board, and build one for you to use for testing as well.

Maybe we could send a schematic of the circuit that
we use now, and it would be easy for you to make
a board to populate?

I will try to work toward getting the circuit that we use
drawn out, and broken into two parts, the part for
charger control, and the part for the QC interface.

this is an awesome idea, Gary. let me go through a few latest posts and compose a Google Doc where we can collaborate on the spec instead of posting here sequentially

 

[garygid]

So, power out is approximately Input Amps / five.

30 amp in for about 6 kW out.
50 amp in for about 10 kW out.
100 amp in for approximately 20 kW out.

Breakers and circuits need to be derated by 20%,
if the load is continuous. Since some QC is under
two hours, it is technically not "continuous", I
believe, but the Charging EV part of the national
electrical codes defines all EV charging as "continuous"
loads, which would require 40, 75, and 125 amp
circuits and breakers, right?

A 40 (75, 125) amp breaker debates to 32 (60, 100) amps,
allowing output of about 6.4 (12, 20) kW output, I suspect.

My 50 amp breaker would derate to 40 amps, or about 8 kW.

For short (non-continuous) testing, one might get 10 kW
out of a 50-amp breaker and circuit.

All just approximate, mostly for planning purposes.

 

[garygid]

It has been suggested that the QC cable come into
the mini-QC charger, and the HV (plus and minus)
connect directly there, with the output diode (or other
backflow prevention) heat-sinked in the charger.

We need the actual output current and voltage, on
the outside of the diode.

The control or interface board you already have planned
(or are using) might already have provisions for handling
the 4 QC control lines. In which case, only the two CAN
bus lines, DC Power, Ground, and the two 5v TTL TX and Rx
lines need to come to the AVR-CAN board. These 6 wires
can be wired to a couple of connectors that would push-on
to header/connectors already on the AVR-CAN board, which
already has the CAN transceiver chip.

If you do not already have provisions to handle the
QC control lines, adding them to your control board
might be easier than making another board, just for
the "temporary" AVR-CAN board, and get you closer
to your end goal of a nicely integrated system.

However, the main issue is how to get something
working quickly, to log the performance of the charger,
check that the charger has sufficiently good behavior
so that we would risk connecting it to a LEAF, and then
test and log that the charger's capabilities appear to be
better than marginal for the QC task.

 

[garygid]

Oh, in addition to telling the charger when to
activate the 2 (or 3) QC control lines, and asking
for the status of the other control lines that you
are "sensing", it will have a DB9 connector to
use a Serial-to-USB adapter cable to have a PC log
the data, for better analysis of the QC interaction,
and improvement of the interface, and the charger.

Other status the we would need from the charger
would be the stopping or starting status of the
"red" and "green" buttons of the mini-QC charger.

The Google doc to get the Valomo command set
specified and well defined... should work.
Initially limiting the write/change access to only
you and me would be wise, I suspect.

Cheers, Gary