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

[klapauzius]

IMO, the best approach is a 2-stage system I described above (the one we are working on now).

+1

I think for this purpose this is the only solution that could also be portable.

Also the failure risk for an n-stack of power supplies in series is n -fold compared to a single integrated design.

I think cost-wise, given all the extra time to put such a stack together, hack it and make it safe, the integrated design approach will be not significantly more expensive.

 

[garygid]

Great, Valery, thanks!
I await news of your initial testing.

As the frequency goes up, the hysteresis losses go up,
so a 40 kHz PWM is not all gain. Also, with the 84 MHz
clock of the Due, that would only give us adjustments of
one part in 2000, so staying below 20 kHz, or even near
14 kHz might be... needed, or something?

Tomorrow at our LEAF Gathering, I will show the Arduino Due
with the CTE32HR touch display and gradually improving
programming, powered by a standard little rectangular 9v battery.
Super portable for demonstrations.

I will add a CAN transceiver, and build a second unit, initially
a clone of the first, and add a CAN communications testing
function, a later an 8-wire QC-Port simulator function.

 

[jclemens]

This is exciting to hear about the progress in terms of the power supply.
Shifting topics a little bit, I need to re-visit the power pins.
I have started my next revision of the QC plug 3d printing, this includes strain relief testing with actual 2awg welding cable (insulation rated for 600V)

As it turn out, the multi-strand flex cable diameter does not fit the pins that Gary made to the spec below. The bore is too small.
While I can work around this for my own plug, I feel it is difficult to move forward without knowing what the final pin size is going to be.

again, here is the dimensions for our current power pin.
Is anyone working on getting these manufactured or otherwise searching for other 9mm power pins?

The diameter of the bore in the blow image needs to be increased, but if we increase it more, the diameter of the entire back-section of the pin will also need to be increased.

3d design for real-life is very much trial and error, which is why I don't mind giving my plugs away for free because they are generally flawed.
The one I printed recently will go to Valery once I am done my evaluation and photos. (still working out cable management)

 

[garygid]

What is the amp rating on 10 or 15 feet of stranded 2awg?

If we use that, what diameter hole do we need in the power pin?

Or, for a 30 amp application, or even a 60-amp version,
what thinner power cable is appropriate?

Do you still think that a snap-on C-ring in a shallow, narrow
grove would be a better solution for the collar?

With the snap-ring, the pin can be fully attached to the wire,
then slid through the backing plate, a have the snap-ring
installed in front of the backing plate, and the pins then
inserted into the front barrel section.

If these rings would work, that would be a substantial
improvement in assembly process, I think.

The only critical part of the power pin is the mating section,
slightly tapered or rounded at the end, 9mm in diameter,
and about 1-1/8 inch insertion length.

 

[garygid]

I assume that we could source 9mm copper rod, and just
taper the end, cut a shallow, narrow grove for the snap
ring, cut to length, and drill the large hole for the wire.

That would make the power pins much easier to construct.

Is your hole in the side for applying solder after crimping?

Or, do you feel that crimping is not necessary on these pins
at the lower current values (less than 40 or 70 amps)?

I suspect that the industry standard is to crimp, or in some
cases, crimp and then solder to keep the air out, but
probably NOT just solder alone.

 

[garygid]

Perhaps this 4 awg wire, rated at 85 amps would be sufficient?

http://www.lowes.com/pd_72610-295-20499001_0__?productId=3129539&Ntt=awg+wire&pl=1&currentURL=%3FNtt%3Dawg%2Bwire&facetInfo=" onclick="window.open(this.href);return false;

However, maybe the welding cable is two-conductor with
a much more durable insulation?

The hole in the back of the power pin should really be sized to fit the wire,
so that a good crimp can be made without too much deformation
of the back end of the pin.

With a fixed collar on the pin, crimping deformed the pin so that
getting it back through the backing plate was a substantial issue.
However, with a snap-ring as a collar, this problem does not exist.

 

[DaveEV]

What is the amp rating on 10 or 15 feet of stranded 2awg?

NEC rates 2AWG copper w/60C at 95A. I don't know if it needs to be de-rated 20% as EVSE use is considered a continuous load. (76A).

Wire Ampacity Tables

Or, do you feel that crimping is not necessary on these pins
at the lower current values (less than 40 or 70 amps)?

I suspect that the industry standard is to crimp, or in some
cases, crimp and then solder to keep the air out, but
probably NOT just solder alone.

Crimping alone is the industry standard. No solder. In a good crimp there is no room for air or solder.

Perhaps this 4 awg wire, rated at 85 amps would be sufficient?

http://www.lowes.com/pd_72610-295-20499001_0__?productId=3129539&Ntt=awg+wire&pl=1&currentURL=%3FNtt%3Dawg%2Bwire&facetInfo=" onclick="window.open(this.href);return false;

In my experience, the Home Depot / Lowes wire like the one you linked to is sufficient for wire that is not directly exposed and not subject to flex. For extension cable use it is not sufficiently flexible and the insulation is not thick enough.

 

[TonyWilliams]

I suspect that the industry standard is to crimp, or in some
cases, crimp and then solder to keep the air out, but
probably NOT just solder alone.

Industry would never crimp AND solder, and would likely never solder such a pin.

The hole is a "witness" hole, which is standard in crimped pins. It's for inspecting that the wire is actually in the pin, and not halfway in and crimped.

 

[TonyWilliams]

What is the amp rating on 10 or 15 feet of stranded 2awg?

NEC rates 2AWG copper w/60C at 95A. I don't know if it needs to be de-rated 20% as EVSE use is considered a continuous load. (76A).

You could do like Tesla does and run two parallel 4 to 8 gauge wires for each 100 - 150 amp pin.

 

[garygid]

This chart for welding cable chart seems to indicate 3 awg or 4 should be sufficient?

http://www.directwireusa.com/Img/Welding-Cable-Size-Amperage-Chart.jpg" onclick="window.open(this.href);return false;

 

[TonyWilliams]

This chart for welding cable chart seems to indicate 3 awg or 4 should be sufficient?

http://www.directwireusa.com/Img/Welding-Cable-Size-Amperage-Chart.jpg" onclick="window.open(this.href);return false;

I'm not sure what amperage you plan to use, but 50 amp breakers at RV parks and welding shops can pull 40 amps continuous per NEC, giving 9600 watts at 240 volts, and your DC charger at 88% efficiency will push:

(9600 * 0.88) / 500 volts DC max CHAdeMO spec = 17.28 amps to 21.77 amps at 388 volts in a LEAF. The specification goes down to 100 volts DC, I think, so that's 84.48 amps.

That's a huge swing. Are you planning on limiting the maximum amps?

By the way, I know somebody will claim that there won't ever be a 100 volt CHAdeMO vehicle, except there already is: the 2013 ZERO motorcycle... 100 volts, 100 amps, all DC through the optional CHAdeMO plug.

 

[wwhitney]

Perhaps this 4 awg wire, rated at 85 amps would be sufficient?

There are different types of stranded wire. Standard building wire is "Class B" stranded, for something like an EVSE or QC cable, you need greater flexibility and more strands. For more info, see, e.g.,

http://static.schneider-electric.us/docs/Circuit Protection/0515DB0301.pdf

As you can see in Table 2 of that document, in most cases the OD of the wire bundle increases slightly with increasing number of strands, although the effect is small, just a few mils.

The pin diagram earlier listed a hole diameter of 7.14 mm, which is 281 mils. Per the above table, that is never going to be sufficient for #2 stranded (292-296 mils OD), but would work for #3 stranded (260-262 mils OD).

Cheers, Wayne

 

[DaveEV]

This chart for welding cable chart seems to indicate 3 awg or 4 should be sufficient?

http://www.directwireusa.com/Img/Welding-Cable-Size-Amperage-Chart.jpg" onclick="window.open(this.href);return false;

Welding duty cycle is just about the exact opposite of EV charging.

When welding, you may be using the wire at a 50% duty cycle at most, and most likely far less giving the wire plenty of time to cool off.

If you get 90C rated wire, 3AWG is good for 110A and 4AWG is good for 95A. While the NEC considers EV charging a continuous load, I think you probably could get away without derating the wire for QC use.

 

[jclemens]

Please everyone, we do not need to think about which AWG cable to use, we are trying to emulate as closely as possible what is used in OEM-land.
See page 15 of this PDF
http://charge.yazaki-group.com/english/pdf/YPES-15-1197E.pdf" onclick="window.open(this.href);return false;

Yes, we are not going to use the maximum amperage of the plug at this current time, however, it adds some heat-sinking to the pins themselves, which are currently not OEM.

I can work with my current pins, but we need to ensure future pins are built correctly, especially if there is some money involved in getting them manufactured in any volumes greater than 2. (or OEM pins actually purchased from a supplier, that would be the most ideal)

Thanks for reminding me of the E-clip/ring idea. I will re-post a concept at some point.
for those who don't know ->

spec for e-clip: http://www.mcmaster.com/library/20110328/98317A231.PDF
I have a couple projects on the go right now, (j1772) plus an unrelated day job, I also manage to squeeze a life in there somewhere

There is no question on my end, I WILL be soldering/brazing the pins on. YES, crimping is the industry standard in OEM-land, for many reasons, lower labor time for one, repeatability... and if we had a way to get the proper crimpers to go with this that crimped at the appropriate force, then yes, crimping would be the way to go, however we are not doing this.
The main dis-advantage to soldering is that it creates a bend point on the wire which can lead to internal breaking of the copper, but if we have strain relief down the line a little bit (currently being implemented), then I feel this is a non issue.
Soldering is the only way we have access to, to create a reliable safe connection, we just need to ensure we are using strain relief.

The hole shown in my drawing is inspired from seeing j1772 pins, I read somewhere that this is actually for flowing fluid in and out of the bore hole when electroplating, however I solder my j1772 pins too and the hole works well for feeding solder wire into as it melts.

Now please note, when I say soldering, I could also be referring to brazing, like using a torch on water pipes, this is likely to burn/melt the insulation, but that can be covered in heatshrink. This is another reason why soldering is not commonly used in industry, because of quality standards (IPC anyone?), however with projects like this, we don't really care about quality as long as it is safe.

 

[valerun]

Great, Valery, thanks!
I await news of your initial testing.

As the frequency goes up, the hysteresis losses go up,
so a 40 kHz PWM is not all gain. Also, with the 84 MHz
clock of the Due, that would only give us adjustments of
one part in 2000, so staying below 20 kHz, or even near
14 kHz might be... needed, or something?

You're right. But at <100kHz or so, it seems to be more beneficial to increase the frequency as in a typical Ferrite core, the losses are proportional to 2.6-2.7 power of flux swing and only 1.5-1.7 power of frequency. So given same core and number of turns, raising frequency reduces loss (as flux swing is inverse proportion of frequency). It seems that after 100kHz or so, frequency matters more but we couldn't find consistent formulas around that.

Anyway, we have built the test rig for this isolation stage (pic below) and run it from a 100V supply, with 50V output at 20A (= 1kW). In a simple half-bridge setup (not asymmetric), we are getting ~75% efficiency at this power level, with majority of the losses concentrated in the secondary rectifier diodes. On the scope, it's clear that diode's switching losses contribute a big part of that so we will be trying some snubbing circuits and playing with transition timing on the primary side.

Also, as suspected, we have confirmed the need to rewind our transformer for lower leakage inductance ;-)

Will report later this week how things go.

V

 

[Cor]

I received my Olimex AVR-CAN board today.
Gary, what code are you running on that board?
I'll need to get familiar with Atmel Studio, their devt system.

 

[TonyWilliams]

Please everyone, we do not need to think about which AWG cable to use, we are trying to emulate as closely as possible what is used in OEM-land.

"OEM land" would absolutely crimp the pins. You hit on several of the reasons why soldering is not so smart, but still insist on soldering?

Again, I have the professional level crimping equipment to do these, and I have previously offered this service for this project.

Also, OEM land would have real specifications and use peer reviewed and tested data. Why not start with the specific end goal?

As I posted above, the amount of amps can very a WHOLE bunch, so let's start with one simple specification that directly relates to pins: how many amps will they use? This directly relates to your pin design and wire choices.

So, let's just start with max amps first.

EDITED REPOST FROM ABOVE:

My assumption is that the project would be a portable DC charger, therefore the most likely places are 50 amp breakers at RV parks and welding shops where it can pull 40 amps continuous per NEC, giving 9600 watts at 240 volts. You can't safely use more than this with these resources, and in addition, many 50 amp RV park breakers tend to be "weak", needed even lower than 40 amps continuous. Your DC charger at 88% efficiency will push out:

(9600 * 0.88) / 500 volts DC max CHAdeMO spec =

17.28 amps at 500 volts DC
21.77 amps at 388 volts DC in a LEAF
84.48 amps at 100 volts DC

That's a huge swing. Are you planning on limiting the maximum amps?

 

[jclemens]

My assumption is that the project would be a portable DC charger, therefore the most likely places are 50 amp breakers at RV parks and welding shops where it can pull 40 amps continuous per NEC, giving 9600 watts at 240 volts.

I am just trying to keep things simple. The (my) goal is a portable QC plug to work with the (my) Leaf.
There are j1772 public charging stations across Canada with 90 amp breakers. That means, they can do 70 amp continuous load, which happens to be the maximum common j1772 inlets can do (or the pins that go in them.) (I'm sure these charging stations exist in the USA too)
(google "longest greenest highway")
So, that would be designing for 240*70amps =~ 17kw with 100% eff. The leaf is 360V with a low battery, so it would only actually pull 46.6amps DC, again assuming 100% eff.

yes, 2awg is totally overkill for this, but it covers all bases.

I insist on soldering because the connection is reliable when using strain relief. We don't have pins that are meant for use with crimpers, we actually don't have pins at all. If you can provide pins, crimped with cables, then great, where can I buy some? because I sure have spent way too much time looking.
And the whole reason I am doing this is because I was told that the official yazaki chademo cable and plug is about 3000$, if you can provide a source and a price for something more reasonable, I'd love to see it posted here.

 

[TonyWilliams]

Ok, you want to use J1772 to power it; sounds great. I've got great news for you!!!! Save yourself a lot of headache, potential danger and money and just buy as many fully isolated, CAN controlled J1772 chargers as you want, bolt them on, some simple wiring, viola! You're motoring down the Sun Country Highway before you know it. No CHAdeMO required. Just plug into the 75 amp rated J1772 socket where the old 16 amp one was (under $200 for that, with crimped wires).

We don't have a full network of 70 - 80 amp charge stations in the USA, so we're stuck using RV parks and welding shops at 40 amps without J1772.

Let's assume those Clipper Creek 70 amp chargers eventually are mixed with 80 amp ones (the max of the standard). We have to also assume the max voltage of 240, plus 10% max tolerance, so 264 volts * 80 amps.

That's 21.12kWh * 0.90 efficiency = 19kW max DC charger output from public J1772.

38 amps = 19kW / 500 volts DC max CHAdeMO spec
49 amps at 388 volts DC in a LEAF
190 amps at 100 volts DC

So, that would be designing for 240*70amps =~ 17kw with 100% eff. The leaf is 360V with a low battery, so it would only actually pull 46.6amps DC, again assuming 100% eff.

The lowest I've ever seen CHAdeMO output on the LEAF is 388 volts, up to 395 volts. The LEAF battery will drop below 300 volts, however.

And the whole reason I am doing this is because I was told that the official yazaki chademo cable and plug is about 3000$, if you can provide a source and a price for something more reasonable, I'd love to see it posted here.

Well, they are expensive, for sure. All those pins for all the manufacturers are made for Japanese companies, probably in Japan. The liability on these things in insane, so I'm surprised anybody would want to... hence, few do, and they're grossly expensive.

Why not copy the Dyden exactly, including the pins? I can help you there.

 

[jclemens]

I didn't answer one of your earlier questions, I thought you already knew the answer... yes the current is going to be adjustable via the microcontroller, but that is not my department, I'm just working on the connector. I expect that I would pull up to any charging station, find out the max current I can draw, and if it is worth-while, lug out the portable chademo QC and plug it in.
The point of doing this is that it makes cross-country driving much more feasible. Charging at 16amps is simply not realistic for road trips. Yes, I am aware of the mods available to add bigger chargers, however I would like to keep my warranty. Charging at 70 amps or whatever the maximum is, IS feasible. That is what is motivating me to help out with this project.

Tony, I appreciate your help and input, but there are times where I feel your attitude or way of writing is condescending and it is not conducive to positive motivation. I feel de-motivated.

Moving on...

If I had a dyden or any other OEM'd plug in my possession, I'd definitely be emulating as much as possible. Perhaps I'd even copy it exactly, measure everything, but I just don't expect anyone (including you) to just hand one over to me so I can take it apart. I am in Toronto, and I don't want to fork out the $$ for shipping, or even the cost of the materials, so I wouldn't ask anyone to give it up, so I never did.

For now, it is trial and error, the cheap way. By creating it ourselves, we also avoid any copywrite stuff.

Here is my latest spin on the 9mm power pins. (for Gary/Rob or anyone else who wants to attempt to machine these or have them made otherwise)
It is using mcmaster carr e-clip part number 98317A227

You can view the 3d model here: http://www.thingiverse.com/thing:121581
It is the 3rd pin, click the blue thumbnail and click "thingview", I'm not sure if you need a browser plugin.