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

[garygid]

A latching solenoid usually requires a pulse of
current to latch, and another to unlatch, so
there is no significant heating unless it is cycled
on and off frequently.

In the event of power failure, and the plug remains
latched, one probably wants the "unlatch" trigger
to override the latch-lock solenoid.

Very nice, revealing picture of the plug, pins, and
the handle that you are working on.

 

[garygid]

Valery,
Difficult to predict what the car would do.
Thus the reason for the output diode "requirement", I suspect.

Normally the charger would be in low-power mode, with
all controllable stages powered down, and just the controller
functioning.

The Final stage, with the PWM off, and after discharging,
would just remain at "zero" volts.

Possibly the discharge relays should be powered
to not discharge, so that a power-off would discharge
the internal caps?

I will try to access the documents again tomorrow.
After I login to Google, I see the rather large
"thumbnails", but how do I open or edit them?

 

[garygid]

Apparently the Google documents page is written such
that the IE7 browser gets an error when displaying
the Page with the document thumbnails.

So, I copied the URL to the Google Chrome browser,
and... it works.

I edited the States document to add some states for
the charger, and some for the vehicle's QC system.
Of course, all are just guess-o-states.

I created a document to specify the Val-Serial Commands,
Queries, and Settings messages to send to the Charger,
and the expected Responses from the charger.

I am adding code to my Due Script to test the charger
by sending it many commands, and logging the command
and the response. When complete, one should be able
to use this to go through the operations required to
start a QC session, and to "charge" into an an open circuit,
a dummy load, or an available attached battery.

To test this out, I am coding a companion function to
pretend to be a Val-Serial capable charger, at last to give
the expected responses in a timely fashion.

So, let's get the commands defined, at least a sufficient set
to exercise the charger.

Commands will expect a change in operation.
Power Up
Stop
Sleep
Request Volts out
Request Amps out

Queries just expect state or values back.
Get Max-Capable Amps and volts
Get present Amps and Volts being output
Get Max-Capable Temerature
Get charger Temperature
Get Charger State (what states?)
Get Charger Status (like HV leakage tests OK, etc.)

Settings expect to change the operating parameters of the charger.
Set Charger Max Voltage
Set charger Max Amps

 

[garygid]

Valery,
A thought:
Do you need to read the HV voltage on the capacitors of
the output stage to control the charger?

We need the actual output voltage, I believe, for the
QC process, to see when the vehicle exposes the
battery pack to the charger. That is why we
measure the HV voltage outside of our diode,
to get the earliest warning that the battery
is actually attached.

Testing for the HV leakage current when the
HV is ramped up is another important test.

 

[valerun]

Valery,
A thought:
Do you need to read the HV voltage on the capacitors of
the output stage to control the charger?

We need the actual output voltage, I believe, for the
QC process, to see when the vehicle exposes the
battery pack to the charger. That is why we
measure the HV voltage outside of our diode,
to get the earliest warning that the battery
is actually attached.

Testing for the HV leakage current when the
HV is ramped up is another important test.

Hi Gary -

We could work with either cap voltage or final output voltage. That said, we will have an ability to measure both before and after the relay.

HV leakage test seems like a car's job, no? Implementation of that is not trivial due to very low current to be measured and I would avoid in the first version unless absolutely required by protocol.

Another update on the production work.
After trying for a few hours to fit the iso stage into our standard 10x10x8 enclosure, we gave up and designed a separate 18x8x8 box for it. We could get away with 14x8x8 if we did not need the output relay and a bunch of measurement channels. Still, even at 18x8x8, we are at ~1.3kW/liter (or 22W/in^3) at design power of 25kW. Liquid-cooled version will likely be more like 18x11x4, or ~1.5x the power density of the air cooled version. The major side benefit of the new enclosure is much better cooling as we have moved the transformer from its current position behind the sink to the position on the side of the sink - with its own pair of fans and fresh airflow. This should eliminate any potential thermal issues.

V

 

[garygid]

Is the 18x?x? box the whole charger, or just
the isolation stage (one of 3 stages)?
It is not clear from the above post.

With no load, the current reading is not
very noisy, and at least check for no
detectable current flow. I believe that
the QC needs to represent (to the car?)
that the leakage test was OK, but I am
not sure.

I will plan to look at a HV leakage status
bit coming from the Val-Charger, and
check that it is still OK before continuing
after the HV ramp up, hold, and ramp
down again (during which time the
output relay would need to be closed).

The output diode that you supplied to
us seemed to work quite well, with
only about 0.4 volt drop at about 30
amps, as I seem to recall. That is
only 12 watts... humm... maybe it was
1.4 volts, for 42 watts????

If you use a relay, and the two voltage
measurements are not very well calibrated
and linear (expensive), then a voltage mismatch
of only a few volts can result is a rather large
current ZAP when the relay gets closed.

The published QC overviews all show the
use of an output diode. With the diode,
we ramp up until we get about 0.1 amp
to flow, and then we know that the voltages
are well matched, and we only need to
measure the output voltage.

So, using a relay seems to be the less
"reliable" way to do the "output" job,
at least to me. Possibly, one would
want to use two diodes in parallel?

Compared to heat-sinking 2000 watts,
another 84 is almost insignificant,
and a 75 amp DC contactor is not
totally inexpensive, right?

Just my thoughts, Cheers

 

[garygid]

Just the liability issue of "Your QC info
appears to suggest that you design with
a diode, and you chose to ignore that
advice... and we claim that my client's
vehicle was... damaged... ".

So, food for thought.

Ignoring heat production, which is
the "better" solution, for minimally
functional...?

 

[garygid]

Presently, we are adding a resistor and relay
to deal with the inrush situation.

The resistor is a pair of 100 ohm 100 watt devices,
wired in parallel to make 50 ohms at 200 watts,
at least for our first test.

Then, I think that we can eliminate the
"thermisters" on the input, which get
very, very hot during operation, and do not
limit the inrush sufficiently, especially
when they are already hot.

Trying to be helpful, Gary

 

[valerun]

Is the 18x?x? box the whole charger, or just
the isolation stage (one of 3 stages)?
It is not clear from the above post.

... I believe that
the QC needs to represent (to the car?)
that the leakage test was OK, but I am
not sure.

...humm... maybe it was
1.4 volts, for 42 watts????

18x8x8 box is for stage 2 of 2. Stage 1 is a controlled PFC which is now standing at 10x10x8.

QC can tell the car HV test is ok even if it didn't do it. That would be my MVP definition. I do not consider that unsafe given that the car will double-check anyway (the leakage detection is a full-time feature in Leafs).

Diode: 1.4v is more like it. At 50A will will probably see closer to 2V. At 20kW, its 0.5% efficiency or ~7% of the total loss in the stage. Also, heatsinking is only part of the issue - we also need space for it on the sink and kit-friendly assembly procedure. Given that we already have to have an input precharge relay, seems like a good idea to reuse the second pole of that relay for output precharge and save 100W. Let's go with this for now and if we have to, we will add diode later? When we do add the diode, we would still have the relay shunting it after the voltages are equalized.

Power relays normally are spec'ed to 80-90% precharge, meaning that they are ok to be closed when the voltages are within 10-20% on both sides. 10% difference is quite trivial to measure. Linearity of the chips we use for voltage measurement is 0.06% typical, 0.55% max per datasheet. So we should be fine on the linearity front. If anything, linearity depends more on the resistor divider thermal stability now rather than on an isolation chips.

Yes, we would get rid of the thermisters. They are now the most failure-prone element of the standard charger setup anyway.

V

 

[jclemens]

Just bought a couple of lock actuators. I suspect that the OEM plug has something similar or a residual magnetic latching solenoid.
There are instructions for all OEM plugs for in the case of a fault, there is a special procedure to unlock the plug from the car. This implies that even with the power off, the latch remains on.

There are many of these on ebay.http://www.ebay.ca/itm/Car-Auto-Pla...LH_DefaultDomain_0&hash=item4612a28761&_uhb=1

They are very cheap, and repeatable, I paid a bit more to buy locally so I can get them quicker, but still much cheaper than buying a latching solenoid from digikey!

I will re-design the plug to make v2.1
The in-line data-tap that I am planning on sending to gary will not have a locking mechanism simply because too many unknowns of enabling it based only on the 7 data pins that it connects to. However it will have a proximity switch that IF one were panicked enough to actually touch the thing while in operation, pulling the trigger will open the proximity switch, and will interrupt the charge. Insulating gloves are highly recommended as my units are not exactly rated IP67.

Valery, if you want a locking latch, you may have to wait a while longer, i bought 2 actuators for testing/designing with, but in the end, one would be for me, and the other would be for you... it is just going to take ... "a while longer" (I have no idea how long...)

 

[garygid]

Here is the door lock Actuator on USA ebay, but still from HK...

http://www.ebay.com/itm/Car-Auto-Plastic-Universal-Heavy-Duty-Power-Door-Lock-Actuator-Motor-2-Wire-12V-/300960352097" onclick="window.open(this.href);return false;

Or, slightly more money, but a pair shipped from Los Angeles...

http://www.ebay.com/itm/2-X-UNIVERSAL-HEAVY-DUTY-POWER-DOOR-LOCK-ACTUATOR-MOTOR-12-VOLT/161011781949" onclick="window.open(this.href);return false;

Are these 12v pulse On, and another pulse Off, as if a plate inside
is driven perhaps a quarter turn, either by a motor or a solenoid?

Any spec sheet or operating details?
Thanks, good find, even if it is a bit large.
Are you going to disect one?

Adapter for QC logging?
How is that coming, if at all, please?

 

[valerun]

just for laughs: we got a quote from Y a zaki on the 6-meter CHAdeMO cable with connector.

$2,112 per piece. In quantity of ONE HUNDRED! Damn, I just don't get why some Chinese guys haven't built a knock-off already and started selling that for $300 which is should cost...

 

[garygid]

It might be possible to "make" a reasonable cable by using
something like two 2 to 4 awg stranded for the HV wires,
an 8-conductor shielded signal cable, and an 18 or 22 gauge ground?
Or maybe better, use a separate 2-Wire shielded cable
for the CAN bus wires?

Probably use heavy-duty outdoor insulation?

What gauge wiring does the open Chademo literature suggest?

What gauge wire is used in the cable that you got a bid on?

Then, lay the wires and cables together nicely, and wrap
the bundle with a helix of plastic cable-bundling "stuff"
(what is it called)? Apparently "spiral cable wrap":

Something like:
http://www.delcity.net/store/Spiral-Cut-Wrap-Loom/p_800831.r_IS1013?mkwid=sY3TIvD84&crid=28718371349&mp_kw=spiral%20cable%20wrap&mp_mt=e&gclid=CJ6Yzp_EnbsCFSho7Aod3V0A1Q" onclick="window.open(this.href);return false;
or
http://www.panduit.com/wcs/Satellite?c=Page&childpagename=Panduit_Global%2FPG_Layout&cid=1345564328950&packedargs=classification_id%3D301%26locale%3Den_us&pagename=PG_Wrapper#2" onclick="window.open(this.href);return false;


Finally, one or two layers of heat-shrink tubing, for better
protection against dirt and moisture?

Then, for how, 3D-print a good copy of a Jolomo mini-QC
Plug connector, with some well-made silver-plated pins.

That is the best that I have to offer the DIY guy for now,
but it might not pass the crush test, or other specification.

Maybe try to get a bid on just the cable?

 

[garygid]

Time is passing, and there is no Valery-Serial Command
language yet for the Valery iso-charger. So, I will
post today what I am "using" so far, in the software
thread.

With the holidays coming, and tax paperwork coming
in January, it seems unlikely that we will be able to
test the new iso-charger in time for any January show
or event, but perhaps I will be pleasantly surprised.

Merry Christmas peace and happiness to all, Gary

 

[valerun]

Sorry guys for silence. Worked on proto #2, case design and integration with PFC stage.

We got the product announcement out for both non-isolated 25kW charger (basically a PFC stage of this isolated charger) and an isolated 20kW charger. Dubbed QuickCharge-25000 and ISOCharge-20000, respectively, these products are now available for limited volume order with ~4 week lead time from our site (http://www.emotorwerks.com/products/online-store" onclick="window.open(this.href);return false!

We also did an article in the EAA's Current EVents magazine on these. Take a look!

Gary - the holidays is best time for testing IMO ;-) We are planning to work through the end of the year on these. So yes, you may be pleasantly surprised ;-)

I also have posted a few comments into your Google doc on command set. Please take a look.

Thanks,
Valery

 

[garygid]

Suggestion:
QuickCharge is a substantial confusion conflict with
the Chademo Quick Charge (QC).

Perhaps NISO-Charge, or Nice-O-Charge, or whatever,
but not the term QuickCharge. Have you checked
trademark? Through public usage "Quick Charge" might
already be copyrighted.

 

[garygid]

For your ISOCharge-20000, where does the mini-QC
high-voltage power cable go, into your charger box?

Is there a hole for a suitable cable compression gland?

Is there a terminal strip, or space for a terminal strip
inside the charger for the 7 signal wires?

Assuming that your driver or control board does
not have the interface circuitry to the QC socket wires,
about 2 power and 8 signal wires would need to go from
your charger box to the mini-QC controller module :

5v and 12v power
ground and
3 QC handshake signals
two CAN lines, C-high and C-low
two serial lines, Tx and Rx

What is the connector and pinout on your charger?

Or, is there space for the Due, a proto-shield, a
display shield, and a Display board inside the Charger?

Not critical for your plans, just trying to understand
what you are planning, and how you envision the
wiring to be done.

 

[garygid]

I edited your document to ask many questions there, thanks.
Merry Christmas!

 

[valerun]

For your ISOCharge-20000, where does the mini-QC
high-voltage power cable go, into your charger box?

Is there a hole for a suitable cable compression gland?

Is there a terminal strip, or space for a terminal strip
inside the charger for the 7 signal wires?

Assuming that your driver or control board does
not have the interface circuitry to the QC socket wires,
about 2 power and 8 signal wires would need to go from
your charger box to the mini-QC controller module :

5v and 12v power
ground and
3 QC handshake signals
two CAN lines, C-high and C-low
two serial lines, Tx and Rx

What is the connector and pinout on your charger?

Or, is there space for the Due, a proto-shield, a
display shield, and a Display board inside the Charger?

Not critical for your plans, just trying to understand
what you are planning, and how you envision the
wiring to be done.

Hi Gary - yes, there will be 2 cable glands on the back of the box - rated for 18-25mm diameter cables.

There will be space on the front wall of the box (where the screen will reside, as well) for a Due-based controller.

We have not yet planned the exact terminal structure for signal wires. There is definitely space for it.

The whole thing will be optimized for power density as we crank these out. I think we will be able to further reduce the size by 20% or so if we try hard enough. But I wanted us to get the product out asap.

Valery

 

[GregH]

Time is passing, and there is no Valery-Serial Command
language yet for the Valery iso-charger. So, I will
post today what I am "using" so far, in the software
thread.

Is this public? I assume this is the ASCII based command structure you proposed last month? I understand the attraction of a string based approach if you're going to tap in a terminal to enter or log information...but from my point of view a non-ASCII approach would be a lot easier to code to.
Something along the lines of an 8bit command, optional 16 bit payload and 8 bit checksum.. response could echo the command (with perhaps a bit set) and deliver 16 bit payload with 8 bit checksum.. Just a thought. The spec should also detail the required periodicity before a failsafe condition is reached..
I might be interested in interfacing with the charger as well..