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

[garygid]

The car does not ask for more current than the QC advertises, but
sometimes asks for less, and the car certainty tapers off its request
as the battery gets closer to full. At the end, the car lowers its request
to zero amps.

The QC is expected to "behave", and follow the car's requests, although there
appears to be some latitude in how closely one follows the current requests,
and how quickly one follows the initial ramp-up.

But deviate too much for more than a short time, or respond to slowly,
and the car will feel mistreated.

 

[garygid]

When we can examine DTCs and reset the "hurt feelings" codes,
we might try to experiment more broadly.

 

[FalconFour]

I love the way you put all this. :lol:

And yes, I fully understand tapering while charging... otherwise I think the QC would be encroaching upon "rape" territory. I imagine the car doesn't like that. But that's why there's a central control for the mini-QC that handles the protocol interactions... it would control the charger(s)/power suppl(ies) independently, telling them what to do and monitoring the total output to the car as well.

The tricky part would be finding a basic power supply that's not tied to pre-existing charger logic (like the NetGain charger posted earlier would - presumably - be). Then, slap a few of those in parallel, then we've got this figured Right? Or are we already past that, and already playing with a test rig that's thrown a few DTCs already?

 

[garygid]

We have charged: tickled, wooed, mated, adventured, annoyed, and
downright insulted the car, but we still need a reasonable power supply.

Cost, performance, size, etc. as mentioned in previous posts.

 

[Cor]

We have charged: tickled, wooed, mated, adventured, annoyed, and
downright insulted the car, but we still need a reasonable power supply.

Cost, performance, size, etc. as mentioned in previous posts.

Since I expect that the Leaf will not ask for more power than the charger advertises, the first step in development would be to take *any* (even a low power) controllable power supply and advertise its capability to the Leaf, then follow the Leaf's requests to ramp up to the max power, charge and finally slowly ramp down to zero and finish. Once you have the control to do it slowly (I mean, at lower power) then you can invest in the high power controllable supply to do it fast. Possibly you can develop the SW (protocol) and the hardware (charger) in parallel. Other than expensive lab supplies with GPIB (IEEE-488) I do not know of off-the-shelf supplies that allow control of their outputs, so I suspect that you will need to build the supply from modules that you modify or a combination of standard modules and an additional external circuit that takes a fixed supply and has the capabilities to control its outputs in voltage and/or current. Most likely that is a DC/DC converter such as a standard buck or boost converter, with output control and communication, either digital or analog.

 

[garygid]

Yep, Lucy, you got some 'splainin to do...

We be doing basically that.

 

[DaveEV]

What about using a Brusa charger as a controllable power supply? Bonus - you can stack them too, though the NLG6 can do up to 22kW.

 

[TonyWilliams]

What about using a Brusa charger as a controllable power supply? Bonus - you can stack them too, though the NLG6 can do up to 22kW.

I would say that the target power would be either the 40 amps * 240 volts (9.6kW) at RV parks and welder outlets, or 80 amps * 240 volts for 19.2kW.

Unless you're planning some 3 phase connections? 61 amps * 208 volts * square root of 3 ?

 

[poweredbysun]

A small, inexpensive, light weight power supply typically is not isolated.
Isolated power supplies typically rely on an isolation transformer which is
heavy, and relatively expensive.
So the power supply "problem" as I see it, becomes one of "how does one isolate a
power supply without an isolation transformer", or "how does one make
an isolation transformer which is light weight and low cost?"
Since isolation is not optional, the problem becomes how to isolate a power
supply without drastically increasing size, weight and cost.
The simple solution would be to add an isolation transformer to the front end
of the existing, working, "prototype" charger. But this brings size, weight and
cost outside of the specified parameters for this project.
The Brusa is advertised as "fully isolated" and weighs around 25lbs, in one version.
The cost is outside of the hopes of this project, but size is not unreasonable.
Air cooled versions could possibly be considered "portable".
Jim

 

[mwalsh]

A small, inexpensive, light weight power supply typically is not isolated.
Isolated power supplies typically rely on an isolation transformer which is
heavy, and relatively expensive.
So the power supply "problem" as I see it, becomes one of "how does one isolate a
power supply without an isolation transformer", or "how does one make
an isolation transformer which is light weight and low cost?"
Since isolation is not optional, the problem becomes how to isolate a power
supply without drastically increasing size, weight and cost.
The simple solution would be to add an isolation transformer to the front end
of the existing, working, "prototype" charger. But this brings size, weight and
cost outside of the specified parameters for this project.
The Brusa is advertised as "fully isolated" and weighs around 25lbs, in one version.
The cost is outside of the hopes of this project, but size is not unreasonable.
Air cooled versions could possibly be considered "portable".
Jim

Well said, Jim. And this was EXACTLY (almost to the word, as it happens) the discussion we were having on Saturday morning.

 

[garygid]

The isolation is typically achieved not by a heavy 60-cycle transformer
on the input, but by using a much smaller, relatively light weight,
high-frequency transformer between the input stage and the output.

The input stage converts the AC input (single phase, sometimes
called split-phase) to somewhat filtered DC, providing a "pool"
of energy for the "switching" transformer to dole out to the
Output stage, providing both the isolation and the necessary
regulation of the output current and voltage.

To provide a good Power Factor (current drawn from the "wall"
mostly in phase with the input voltage), there is usually a "switch"
between the input rectifier and the DC "pool", filling the pool
more when the input voltage is high, and less when the input
voltage is low.

But, the difficulty is in the details, especially at higher powers,
where the losses must be minimized to reduce the amount
of cooling required. A design that allows the lossy components
to successfully operate at higher temperatures could allow
the use of air cooling.

Of course, we want to avoid using $4000 in parts, so that it
might typically be sold for $50,000. Instead, a design that
could allow a $2000 kit to be assembled is more in the
range of reasonableness.

Using single-phase 240v power, in the most available amperages,
is the design goal. Something to use at home, and portable to use
at other typical sources of 240v AC. The supply would typically
charge for less than one hour at full current, and less than two
hours at reduced current, so it might not be considered a
"continuous" load, allowing full use of 50 amps from a 50-amp
breaker (and circuit), like one might install at home, or find
at a dock or RV Park.

Allowing operation, at reduced output, from the typical L2
30-amp EVSE would also be convenient.

So, no heavy 60-cycle input transformer, that is "right out".
No 3-phase design, since the three-phase is rarely available "in the wild".

 

[garygid]

There are many small, isolated, switching power supplies,
in PC power supplies, and many "all-voltage" wall warts.
So, isolated is not difficult to do.

Designing the switching transformer for higher power,
low losses, and effective regulation is an important
"little detail".

We are trying to end up with a design that could be
scaled to 240v 100 amp, for more modern homes
that could safely be wired for such a breaker.

Of course, in some jurisdictions, the local
building codes and practices might not allow
the permitted installation of a 24 kW circuit.

 

[poweredbysun]

I asked on the EVTech email list for power supply help.
I gave a link to this thread. I hope Cor did not get scared
off because we need all the talented experienced EV friendly
engineers we can get working on this...
I guess I did not make it clear that the communications
and control aspects were more done than not at this point.
Sorry.
Jim
And, thank you for clarifying the "isolation" being on the
doling out, output stage. I know Gary
had told me this, but all the references I have found
for isolation seem to require an "isolation transformer"
on the input.

 

[mwalsh]

For a truly portable unit, that could be hooked up to a full-power (as we find them in the wild) J1172 or a 50A RV outlet, a 50-60A unit would be more than sufficient for me.

In most domestic instances, for me to safely use a 100A unit using my patented "raid the breaker panel" technique, I would likely have to disable some of the homeowners big draw items - pool pump, A/C, etc...

 

[DaveEV]

What about using a Brusa charger as a controllable power supply? Bonus - you can stack them too, though the NLG6 can do up to 22kW.

I would say that the target power would be either the 40 amps * 240 volts (9.6kW) at RV parks and welder outlets, or 80 amps * 240 volts for 19.2kW.

Unless you're planning some 3 phase connections? 61 amps * 208 volts * square root of 3 ?

Good target. Unfortunately looking at the specs of the NLG6 it only does 16A on single phase, 32A on 3-phase, so not a good candidate unless you have 3-phase.

 

[TonyWilliams]

I think a scaled down "Luigi" transformer would be the most likely candidate. The Andromeda weighs 160 pounds for up to 62.5kW performance. A huge portion of that weight is the aluminum block for dissipating heat from the high frequency switches.

For portability, I would recommend liquid or a freon based cooling, where the cooling hardware is in a separate container for easier transport. I remember my dad's welder that used simple liquid cooling, and it's nothing more than a simple transformer.

The switching transformers would just be on a relatively lightweight cooling plate / heat sink that is chilled remotely.

 

[garygid]

For now, we are trying to avoid the weight and expense
of a liquid cooling system.

So far, with our experience with a 240v 50 amp (about 11 kW output)
device is that air cooling is sufficient, with very modest, relatively quiet
fans. Hopefully we will not need a more elaborate system.

But, running in 120 degree ambient might be near the tipping point.

 

[Cor]

The isolation is typically achieved not by a heavy 60-cycle transformer
on the input, but by using a much smaller, relatively light weight,
high-frequency transformer between the input stage and the output.

While we are on the topic of transformers and weight, you can think of the transformer
as a "bucket" of energy that you can fill and empty in the same way as you can fill
and empty a bucket of water The size determines how much energy you can transfer
every time you fill and empty it. In a transformer the "fill" consists of magnetic energy
which is created by the electric current through the windings, but the idea is the same.

If you need to transfer 12,000 gallons of water per second and your bucket can only
fill and empty 120 times per second, you need a 100 gallon size bucket.
(I am sure you recognise the numbers for the charger you are designing)
If you have a much faster moving bucket that can fill and empty 12,000 times per second,
then your bucket size only needs to be 1 gallon, so 100 times smaller (and lighter and cheaper
in the case of transformer) and that is the reason that high frequent switching power supplies
with small transformers are so popular and still can be of isolated architecture.

The input stage converts the AC input (single phase, sometimes
called split-phase) to somewhat filtered DC, providing a "pool"
of energy for the "switching" transformer to dole out to the
Output stage, providing both the isolation and the necessary
regulation of the output current and voltage.
To provide a good Power Factor (current drawn from the "wall"
mostly in phase with the input voltage), there is usually a "switch"
between the input rectifier and the DC "pool", filling the pool
more when the input voltage is high, and less when the input
voltage is low.

This is a good description of the operation of a PFC (Power Factor Correction) network
that is typically used on higher power devices to take AC input and also force the
input current to follow the same AC (sinusoidal) shape to get maximum performance
from the AC power without introducing distortion.
The advantage of PFC is that buffer capacitors are somewhat smaller than if you
do the "cheap" thing and just fill them with a bridge rectifier. One consequence of
following the AC waveform is that the current into the capacitor varies, so there is
always a ripple on the capacitor, depending on how much current you draw from it.
This is usually not a problem, since the PFC is typically a non-isolating prep stage
and typical power supplies have a second stage that does the DC/DC including the
regulation of the (high voltage) ripple power into a stabilized, isolated output signal.
Almost all power supplies follow this architecture and it is also a good architecture
for a charger, although the double conversion introduces twice a loss, so for that
reason (as well as for size and cost) several EV charger designs omit the second stage.
Since you are already dealing with lethal DC voltages, they feel that the additional
safety of the grid isolation does not bring much benefit.
I, for one, am happy that I can touch my batteries *while* I am charging, since my EV
charger provides the isolation from the grid. Even though it is a heavy 60 Hz transformer in this case.

 

[Cor]

I asked on the EVTech email list for power supply help.
I gave a link to this thread. I hope Cor did not get scared
off because we need all the talented experienced EV friendly
engineers we can get working on this...
I guess I did not make it clear that the communications
and control aspects were more done than not at this point.
Sorry.
Jim

Hi Jim,
I am not scared off, just had some time off with family. I am also waiting to see if this group can get some consensus where they want to go with this design - I posted a few choices and did not see a clear reply, so not sure what is feasible and desired.
BTW, you say that communication and control are relatively clear - can you post a link to where I can look at it and give some advice what architecture for the high power can possibly work best with that?
I gave some suggestions as to how to setup a validation system using low cost power supplies from redundant (in multiple ways) servers. Since the question I saw was to try to find low cost off-the-shelf modules to build this. I am not aware of a low cost module that would allow you to do this. The EMW 12kW charger was discussed at last month's EAASV meeting where EMW presented the JuiceBox and they made it clear that they considered making it isolated, but found that the losses and heat produced doubled due to the secondary conversion, so to keep it compact, simple, efficient and cheap they chose to make a non-isolated charger, otherwise that would have been a great candidate.
With an idea of the control and communication that you want to use, I can see if there is a solution with low cost
modules that I know of can work. As I said on earlier posts, I have charged EV with different cheap solutions, ranging
from a stack of isolated 48V supplies plus one current limited lab power supply to charge a 312V pack to a non-isolated
charger made from two PFC boards, each modified to deliver constant current to the max voltage limit.

 

[Cor]

BTW, on EVtech mailing list I posted following reply a couple days ago:
=================================================================================
If money is a serious limit, then you *can* get creative using surplus "computer server" power supplies which typically have almost their full power available at the 12V output, often they go for $10-20 and are available in quantity. You just need to figure out the control line that enables the power supply (takes them out of standby) from the control interface and you can build a 400V 12kW power supply for about $500 using approx 33 of these supplies. For example:
http://www.ebay.com/itm/HP-DL380/230932790375" onclick="window.open(this.href);return false;
The link brings you to a G4 server supply rated at 575W (12V 47A) and you can get lower price by using G3 (400W) supply such as
http://www.ebay.com/itm/HP-DL380-G3-400W/190882489104" onclick="window.open(this.href);return false;
which would bring the cost down to less than $10 per supply unit or you can go to the G5 version that goes for about $28 each and that can deliver up to 82A to go even beyond the max utilization of a 100A circuit at 240V AC, because assuming 85% efficiency a 400V 82A supply will draw: 400 * 82 * 1/.85 = 38588W or 161A at 240V. This even maxes out a typical house's 200A service. Oh and it delivers 32kW DC. The Leaf's recharge time would then start approaching the Tesla Fast Charge (which is 100kW but into a much larger pack) but I do not want to think about battery heating and effects on life expectancy...
http://www.ebay.com/itm/HP-DL380-G5/370776509551" onclick="window.open(this.href);return false;

My own preference would be to go for G2/G3 supplies rated at 400W each, which will likely bring you at or close to the desired 30A at 400V for pennies (well, about $300) for the isolated power supply itself.
You might be able to find other open frame power supplies at more convenient output voltages (such as 80V or 150V) to lower the number of supplies, but I find it hard to get them below $100, especially for high current versions, so that is your choice.
Then the work is to make the control circuit to limit current and ramp it up as commanded via Chademo.
=================================================================================

The first link brings you to a cheap 575W power supply for a HP G4 server.
In the mean time I acquired one of these, the model is DPS-600PB as you can see in the Ebay auction.
I will open it to see if it can be easily modified to add control and/or change the output voltage.