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

[garygid]

Reviewing the QC control lines:

GP = Ground, Protective
QR = QC Ready, QC applies 12v
EC = Enable Charger (permission to charge), car pulls to ground.
PP = Power, DC Plus, isolated
PM = Power, DC Minus, isolated
GD = Ground, Disable car motion, Plug Proximity
CH = CAN-High
CL = CAN-Low
BR = Enable Battery Relay closure, to begin charging, Ground from QC

It appears that there are only two outputs:
QR = QC Ready, QC applies 12v
BR = Enable Battery Relay closure, a Ground from the QC

and, only one input:
EC = Enable Charger (permission to charge),
car pulls to ground, QC provides the pull-up.

The Outputs:
1. The QR is either open, or we supply plus 12 volts.
(apparently the current is on the order of 20 ma, but we
do not know what other cars might require)

2. The BR "output" is either open or we Ground it,
which returns the current supplied by QR.

Note: We might also need to supply N volts to
another wire going to the QC plug, that activates
the latch-lock (if any) and an LED. (future issue)

The Inputs:
1. The EC input is either near ground, pulled down by
the car, or at whatever the QC pulls it up to, which
could be a convenient voltage, like 5v. Sensing high
or low is sufficient. So, a simple digit in port using an
internal pull-up would work, but better protection
of the uP input would be well advised.

Note: Will we need to read the state of the latch and/or
latch lock on the QC plug? (a future subject)

So, in Summary
Having a 12v output, a Ground "output", and one
"5v" sensor would seem to get us going.

 

[valerun]

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.

yes, about right (actually, your numbers are ~10% lower than actual as your nominal voltage is 240V). So I would not do any derating on the breakers considering all this is going to be <3 hour application.

Note that you normally would be able to get 1.3x the rated current out of the breaker for at least 10-20 minutes...

 

[garygid]

It would be nice to log the input AC Current,
and Voltage, at least for logging purposes,
but also to measure compliance and inrush.

It would be nice if the capability was on the board,
even if the circuits are not populated, to save
money. There could be an extra-cost option
with additional parts in the kit?

Just a thought.

 

[garygid]

European circuit breakers are different, I believe.
When the current goes over the rated value,
they trip immediately (so I have been told).

Even if you do not expect to seek UL or similar
listing, or just sell experimental kits, I suspect
that it would be wise to design such that the
finished product would be able to be Listed.

An On/Off switch or an Emergency Off device
(which only needs to function once)
might be a good safety feature to consider.

Having some specifications detailing the
operating envelope, design limitations,
and applicable safety procedures might
also be wise. For example, the max current
drawn when set to various max-power or
max-current settings.

Usage limitations like humidity, moisture,
and ambient temperature limits (something
like indoor, dry conditions, non-condensing,
115 degrees F maximum ambient, and such)
will help potential buyers evaluate
their planned operating conditions.

Operating and procedural warnings
would be wise, including noise into
the AC wiring, air, and the car.
The possibility of "bricking" the car,
requiring a tow to a dealer. The
possibility of damaging... something.

How do you handle these issues with
the other products that you sell?

In liability situations, you want to be able
to prove that your customer was very clearly
warned, and that the troubles were clearly
a result of the warnings being ignored.
Even then, liability cases can be a real pain.

Note that a QC vehicle wants a DC Current
delivery figure, and an operating DC max
Voltage value, and usually it initially asks
for the max current, sometimes even when
the battery is more nearly full. Generally,
the QC vehicle does not care if the charger
could deliver more current at lower
output voltages.

Again, just thinking out loud, and
trying to be helpful.

 

[garygid]

I will try and program a Due with a 3.3v
Tx and Rx serial port that would attempt
to emulate what a mini-QC controller would
ask of a charger, using your Valomo commands.

With that code and your Due and CTE Display,
and suitable level shifting, if necessary, you will
be better able to test your Valomo-capable
charger, and log its operation, both no-load,
and into a dummy load. A resistive load is
first, and then, if the logs look good, using
some high voltage battery as the load.

The logging will require a Windows PC running
my CAN-Do program (follow the links in my
signature).

Except for the high voltage battery test load,
we will attempt to be set up here to run the
tests and do the logging whenever we get
a Valomo-capable charger. We will share
our logs, if you are not able to make
your own logs.

When the Valomo command set and the
charger are as reasonably tested as we
can do here, we would hope to see some
of your logs using a test-battery load.

In parallel, we need to build a QC interface
card, hopefully partially populating
one of your boards, and connect it to
our AVR-CAN board. We might be able
to test the QC interface with a variation
of our present controller for our non-iso
charger, if we still have it.

Once that is all tested, we hope to attempt
to test our mini-QC AVR-CAN (with the new
Valomo-capable code) and your Valomo-
capable charger, very cautiously, with my LEAF.

Then, if you have a QC LEAF available, and
a QC cable and a suitable Plug, we might
be able to send you a programmed AVR-CAN
to use for your own testing.

That would be a great step forward.
Cheers, Gary

 

[jclemens]

Perhaps I should have asked sooner and in a different thread, but does Turbo3's Leaf Spy app have DTC reset capabilities? Especially the kind of DTC that one may accidentally cause while experimenting with the QC port? Being able to un-brick the car at home would be monumental. I am asking here because it is directly relevant and I am hoping Gary might know something or have some influence with Turbo3

 

[valerun]

Once that is all tested, we hope to attempt
to test our mini-QC AVR-CAN (with the new
Valomo-capable code) and your Valomo-
capable charger, very cautiously, with my LEAF.

Then, if you have a QC LEAF available, and
a QC cable and a suitable Plug, we might
be able to send you a programmed AVR-CAN
to use for your own testing.

That would be a great step forward.
Cheers, Gary

Sounds good. Also, yes, we have Leafs with QC ports here that we can use for testing any time.

PS. The unit was pushed to 20kW continuous into a 330V pack yesterday!

Valery.

 

[valerun]

Having some specifications detailing the
operating envelope, design limitations,
and applicable safety procedures might
also be wise. For example, the max current
drawn when set to various max-power or
max-current settings.

Usage limitations like humidity, moisture,
and ambient temperature limits (something
like indoor, dry conditions, non-condensing,
115 degrees F maximum ambient, and such)
will help potential buyers evaluate
their planned operating conditions.

Operating and procedural warnings
would be wise, including noise into
the AC wiring, air, and the car.
The possibility of "bricking" the car,
requiring a tow to a dealer. The
possibility of damaging... something.

How do you handle these issues with
the other products that you sell?

In liability situations, you want to be able
to prove that your customer was very clearly
warned, and that the troubles were clearly
a result of the warnings being ignored.
Even then, liability cases can be a real pain.

Note that a QC vehicle wants a DC Current
delivery figure, and an operating DC max
Voltage value, and usually it initially asks
for the max current, sometimes even when
the battery is more nearly full. Generally,
the QC vehicle does not care if the charger
could deliver more current at lower
output voltages.

Again, just thinking out loud, and
trying to be helpful.

Thanks Gary - good points.

Just a reminder though - we will be going after a Minimally Viable Product (http://en.wikipedia.org/wiki/Minimum_viable_product" onclick="window.open(this.href);return false for this initial release. As such, it will not have all the bells and whistles [yet]. I will be spec'ing out what will be delivered in that Gdoc I mentioned. Please understand that every change / addition of feature (such as another measurement channel etc) costs at least a few $100 in PCB reordering, possibly multiple hours in PCB adjustments, possibly multiple hours in analog circuit & software debugging, and possibly multiple hours in mechanical layout changes. This is very very different from the digital world where addition of a digital input is a matter of a few lines of code. Noise and filtering considerations are different for different types of sensing circuits, isolation is far from trivial (as I mentioned in my previous posts), and things like placement of the analog circuitry relative to power-carrying wires / traces matter - a lot of times in ways that cannot be predicted before you actually put everything together and ramp up to design power. Then, when something goes wrong, it's not just the matter of a simple reset button push - the failures at 20kW can be pretty spectacular ;-) All of this makes adding features very expensive from time and $ perspective. And, of course, this is why you don't see these things being open sourced ;-)

So let's try to keep these MVP concepts in mind when discussing features.

PS. Small update that also illustrates some of the above. We discovered that mounting the diodes is a bit cumbersome in the current design - requires precise sequence and if you mess up one step (which we did), you have to remove everything (including the gooey thermal pads) and start over. Fortunately, we were able to get around this by just separating secondary diodes into a separate PCB that will be mounted under the main PCB and connected to the main PCB with alum standoffs. So we don't have to redesign the main power board. Still, doing a small run of this new PCB (in 4oz copper and thick substate) just cost us $500 and ~5 hours design time. You get the idea. Of course, we took the opportunity and designed this board so that we can put more diodes in parallel if needed. My estimate is that it will result in ~0.25% improvement in efficiency (50W less heat to remove).

Thanks,

V

 

[garygid]

I do understand the difficulties.

We have been through all of that over the last
6 months as we reworked your non-iso 12 kW
kit hardware and completely rewrote the control
software to get the performance needed for
a working, though experimental, mini-QC process.

Our investment in time has been very significant,
accompanied by a considerable monetary
investment as well.

We still struggle to help, now with diminishing
resources and little potential benefit to ourselves.

So, you are not alone, and we wish you success.

 

[valerun]

Our investment in time has been very significant,
accompanied by a considerable monetary
investment as well.

We still struggle to help, now with diminishing
resources and little potential benefit to ourselves.

So, you are not alone, and we wish you success.

Thank you Gary. And Happy Thanksgiving to you all!

Just did the first low-power (1kW) test of the fully assembled packaged ISO stage (with the proper power PCBs, dual-processor controller, new transformer). All good so far! In addition to the simple stage operation (similar to the test bench version), tested stage control. As you might recall, we were planning to have an ability for narrow control of the iso stage - to potentially be able to filter out all current ripple if needed. Works, too.

The control board has 2 MCUs (our old friends Pro Minis) - one is a tightly controlled driver signal generator. disabled interrupts, timing control, etc. Second MCU is a state controller that monitors current / voltage / temp / serial commands and outputs analog signals to control PFC and iso stage. Can also play a role of a watchdog for the first MCU and shut everything down if detects problems. There is hardware overcurrent protection on the output (in retrospect, we should probably have put it on the primary side of the transformer, as well...).

Will be adding measurement and shunting board and testing the whole things at 10kW later today.

Happy holidays!

Valery.

 

[Cor]

I saw the question about the cable wires and remembered that I took a pic of the Leaf being Quick Charged by the AAA at the local EV ralley 2 months ago. That was a bit weird situation - the lady arrived too late (about an hour before the end) and did not want to wait for the Level 2 charger to give her enough charge, so she called the AAA truck with the generator to charge her Leaf. Talk about an environmental disaster...
Anyway - the racket from the genset caught my ear, so I walked over and watched how the truck put a relatively low (30 Amps at 390V) power into the battery, so the charger was running for quite a while (almost half hour) before she had enough charge. After the AAA guy had put on his insulating gloves to disconnect the charging plug (really!) I grabbed the plug and made some pics of it, including the cable.
The cable showed: 2x 2AWG + 5x 18AWG + 4x 18AWG 600V. 105 degC.
I am quite sure that this is over-dimensioned, you can send a much higher current through 2AWG, certainly over 100A!

 

[valerun]

I saw the question about the cable wires and remembered that I took a pic of the Leaf being Quick Charged by the AAA at the local EV ralley 2 months ago. That was a bit weird situation - the lady arrived too late (about an hour before the end) and did not want to wait for the Level 2 charger to give her enough charge, so she called the AAA truck with the generator to charge her Leaf. Talk about an environmental disaster...
Anyway - the racket from the genset caught my ear, so I walked over and watched how the truck put a relatively low (30 Amps at 390V) power into the battery, so the charger was running for quite a while (almost half hour) before she had enough charge. After the AAA guy had put on his insulating gloves to disconnect the charging plug (really!) I grabbed the plug and made some pics of it, including the cable.
The cable showed: 2x 2AWG + 5x 18AWG + 4x 18AWG 600V. 105 degC.
I am quite sure that this is over-dimensioned, you can send a much higher current through 2AWG, certainly over 100A!

thanks Cor. Yeah, this whole idea of a generator-powered quick charge is a disaster...

 

[Cor]

BTW, if you want to make sure you can safely charge from your current non-isolated charger, then you could simply buy a used transformer as used by utilities. Often you can find Ebay deals (or Craigslist) for only a couple hundred bucks to isolate power in the 20-25kVA range, for example near LA: http://www.ebay.com/itm/310595831429" onclick="window.open(this.href);return false;

 

[garygid]

For a fixed indoor installation, with sufficient air flow
to handle the cooling of the charger, the transformer
might be a reasonable solution. The weight of the
transformer does cut down on the desired
portability factor.

Hopefully Valery's new iso-charger will become
at least the first pass at a better solution.
It remains to be seen what ambient air
temperature will be required for running
it at 30 amps (and less) output for around
two hours to charge an almost "empty" LEAF.

A reasonable path for progress might be to test
the present non-iso experimental mini-QC with
a transformer, if the new owner can get one
to us, rewiring the floating Ground to the
Earth Ground, and running a few basic tests.

Then, the documentation of the hardware, including
the transformer, could be prepared for the owner,
and perhaps even released to the public.

It would be better if the transformer was one that
others could obtain from a reliable source.
I understand that there are some sources in the
Chicago area.

This first mini-QC unit evolved from a 12 kW
PFC charger kit from Valery, but there were
a number of handmade modifications, and
the controller "brain" was replaced and
reprogrammed to facilitate the functions
needed for the mini-QC operation and
testing.

The inrush current is presently not handled
properly or sufficiently, and it requires a
manual 120v AC pre-charge procedure.

Adding an inrush-limiting "resistor" and
a pre-charge relay, which would bypass the
resistor after the pre-charge is complete,
might be the best solution in this case,
but we have not yet tested or programmed
that solution to this very important problem.

Happy after-turkey day.

 

[jclemens]

I saw the question about the cable wires and remembered that I took a pic of the Leaf being Quick Charged by the AAA at the local EV ralley 2 months ago. That was a bit weird situation - the lady arrived too late (about an hour before the end) and did not want to wait for the Level 2 charger to give her enough charge, so she called the AAA truck with the generator to charge her Leaf. Talk about an environmental disaster...
Anyway - the racket from the genset caught my ear, so I walked over and watched how the truck put a relatively low (30 Amps at 390V) power into the battery, so the charger was running for quite a while (almost half hour) before she had enough charge. After the AAA guy had put on his insulating gloves to disconnect the charging plug (really!) I grabbed the plug and made some pics of it, including the cable.
The cable showed: 2x 2AWG + 5x 18AWG + 4x 18AWG 600V. 105 degC.
I am quite sure that this is over-dimensioned, you can send a much higher current through 2AWG, certainly over 100A!

I am building my plug for 2awg cable, it is not just for current capacity, but also a heat sink to pull heat away from the pins.

Can you post any photos you took? I'd like to see how you saw the individual cable sizes and what the connectors looked like... (and why the guy put on insulating gloves)

 

[valerun]

Gary, am programming the ISO stage controls now. Specifically pulldown control. I need to know the sequence / timing of events to manage precharge correctly.

My current understanding & implementation:
1. Car requests pull-down to zero
2. ISO opens output relay (to disconnect output caps from the output)
3. ISO closes shunt relay (small relay in series with 330R power resistor)
4. ISO checks for voltage decay on output (and throws an error if voltage does not decline fast enough - to prevent burning a shunt resistor)
5. Car requests <WHAT?>
6. ISO opens shunt relay
7. ISO closes output relay (reconnecting output caps to the output)

My main questions are:
1. what will the car do / request after doing pulldown test?
2. can I assume that the car will NOT connect the battery to my output right after the pulldown test is complete?

What I want to avoid is closing the main output relay when I have low voltage on my output caps but the car already connected the battery to my output.

Do you see what I mean?

BTW I have started a couple of google docs in a folder that I shared with your (using your email address from your prev messages) - one doc is detailing QC states and transitions (you will own that doc), another detailing our hardware design & implementation, including command set etc (I will own that one). Please feel free to edit yours and comment in mine. I will do same. This will help with version control.

Thanks,
Valery

 

[valerun]

also added a third doc on the universal Due-based power conversion controller. Please comment in doc: https://docs.google.com/document/d/1loHEL3U3F6QzSlJru7SjJD5GJK_A4rEJ8TwDGfeMvdM/edit?usp=sharing" onclick="window.open(this.href);return false;

 

[Cor]

I am building my plug for 2awg cable, it is not just for current capacity, but also a heat sink to pull heat away from the pins.
Can you post any photos you took? I'd like to see how you saw the individual cable sizes and what the connectors looked like... (and why the guy put on insulating gloves)

I have uploaded the pics I took: https://www.flickr.com/photos/33178064@N05/
Note that it is *my* (bare) hand holding the QC connector. No idea why the guy put on gloves, probably the same reason
that the EV/Hybrid servicing shops used to require that a mechanic who had to work on the electric part of the car, had to
have a rope attached to him and a second person holding the rope to pull him away in case something went wrong.
(besides the guy wearing high-voltage gloves, of course).
Why do you think that many dealers did not want to have/service EV or Hybrids? I hope that lunacy has changed now.
And no, it is *not* an old wives tale, this was reported by a reliable member of the Prius community!
Anyway, I have never worn an insulating glove in my life and I hate gloves except when riding bike while it is freezing outside.
And I have worked on both Prius (276V) and my EV (312V) packs, just by *knowing what I am doing*. OK, end rant.
I was tempted to push the plastic guards around the HV pins back (there was a spring behind them to avoid contact to the pins when unplugged, since the generator was already stopped, but the guy warned me that the pins might be hot (in temperature) from charging for half an hour, so I left that alone. You can't really compare how fat the cable was, other than that it passed through the handle in the back of the plug (which is in two of the pictures) although you can see the shadow of the cable where the plug is in the Leaf. The other pic shows the sise of the plug and handle compared to my hand and I do not have small hands (I am 6'3"). Let me know if you have questions about them. BTW, clicking on the pics twice will show a better resolution one.

 

[jclemens]

Thanks for posting Cor.

For everyone else, I have good news and some not so good news.
The good news is that my plug is comming along and I got to do some mechanical tests today.
I am very excited right now and can't wait to get all this documented... (probably won't finish that today)
I have a working plug with latch and strain relief, there is an audible 'click' when the plug is inserted 100% into the car.
The plug will not come out unless the trigger is pulled.

More high res photos here: https://skydrive.live.com/redir?resid=A24E22DEAC49B01F!1216

Now for the bad news.
I did some electrical testing on the solenoid I am using. (lowest power, cheapest cost).
Digikey PN 527-1000-ND, it is only 4W
The issue is NOT the strength, though if you pull the trigger hard enough, you can over-power it.
The issue is that even though this thing is rated for continuous duty, it gets really hot!!
I measured 40 degrees over ambient, that is hot enough to soften the PLA plastic that is encasing it, and to make things 'burny - ouch' feeling on mounting screw that happens to be on the handle where the palm of your hand goes.

I am not sure how to proceed, it seems that all the solenoids available will behave the same way, or be even hotter. I can make future grips out of ABS so it can take the heat, perhaps include some sort of heat sink... I dunno.
Any thoughts?

edit, I just googled it, and as I thought, the hot temperature is normal.
http://www.colehersee.com/home/faq/faq/solenoids/
however I am not familiar with latching solenoids or even if this would work with our charger. Digikey has them, but the datasheets don't explain everything, (such as how hot it will get). Latching solenoids cost more, which is why I didn't buy one to play with.

 

[valerun]

looks super-cool, Joel!

just a thought on the heating issue: holding current can usually be quite a bit less than engage current on those things - is there any way to reduce the current once the solenoid engaged? For example, I have seen a circuit with parallel resistor & cap connected in series with a solenoid - so on initial connection, solenoid sees full voltage as a cap behaves as a short for the initial time period (~RC) and then after a few RCs the solenoid is fed through a resistor. Normal recommendation for relays is that hold current is 60-70% of the 'make' current. So you would have R be ~1/2 of the relay coil resistance. Then calculate C so that RC>>make time. For power relays usually 20-30ms, so as long as RC>100ms, you should be fine. 4W 12V solenoid is ~40 Ohm, so your R could be 20R and C=5,000uF.