DIY Electric Car Forums banner
21 - 26 of 26 Posts

·
Registered
Joined
·
70 Posts
Discussion Starter · #21 ·
I think the key to using this type of technology will be to hide it from
the user behind a dc-dc converter having it's own inductance and
capacitance.
A very high efficiency bidirectional dc-dc converter (QR-ZVS?) would
provide something like a 300V, -100 to 400 amp interface .
This can then feed an AC 3phase hbridge or a dc motor. Maybe the dc
link voltage would be controllable and be part of the control or maybe
it is just dumb, acts like a battery and lets us use whatever controller
we want. At 300:3500 we are already at 12:1 voltage ratio in the dc-dc.
That is pushing the limit on practical high power dc-dc converters,
especially if we are asking to have a dynamic range of 12:1 to 2:1

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter · #22 ·
Jeff Shanab wrote:
> A very high efficiency bidirectional dc-dc converter (QR-ZVS?) would
> provide something like a 300V, -100 to 400 amp interface. This can
> then feed an AC 3-phase h-bridge or a DC motor. Maybe the DC link
> voltage would be controllable and be part of the control or maybe it
> is just dumb, acts like a battery and lets us use whatever controller
> we want.

Yes; that makes sense.

> At 300v:3500v we are already at 12:1 voltage ratio in the DC-DC.
> That is pushing the limit on practical high power dc-dc converters,
> especially if we are asking to have a dynamic range of 12:1 to 2:1

Not really. Any PFC boost converter has a fixed output (about 400vdc)
and an input that varies from 0v to the peak of the AC line (370v for a
240vac line). These are mass produced by the millions and have
efficiencies of 95% or so.

Wide-range DC/DC step-down converters are also common. Universal input
switching power supplies will deliver 12vdc with a 90-265vac input
range, and there are many DC/DCs with a 4:1 input range. A 6:1 input
range is a little more difficult, but straightforward.

--
Ring the bells that still can ring
Forget the perfect offering
There is a crack in everything
That's how the light gets in -- Leonard Cohen
--
Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.net

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter · #23 ·
> > At 300v:3500v we are already at 12:1 voltage ratio in the DC-DC.
> > That is pushing the limit on practical high power dc-dc converters,
> > especially if we are asking to have a dynamic range of 12:1 to 2:1
>
> Not really. Any PFC boost converter has a fixed output (about 400vdc)
> and an input that varies from 0v to the peak of the AC line (370v for a
> 240vac line). These are mass produced by the millions and have
> efficiencies of 95% or so.

Except that in a PFC boost converter, the power drawn is proportional
to the square of the input voltage, so there is low current at low
voltage. For an EV, you want constant power, so current is inversely
proportional to voltage. I suspect that this will have a big effect on
efficiency.

Also, we're talking much higher power here. How many 20kW wall
adapters do you see?

> Wide-range DC/DC step-down converters are also common. Universal input
> switching power supplies will deliver 12vdc with a 90-265vac input
> range, and there are many DC/DCs with a 4:1 input range. A 6:1 input
> range is a little more difficult, but straightforward.

Most of these are low current; an EV would need 100 times more
current, which means 10000 times more resistive losses. You don't
really want to use 10000 times more copper to make up for that.

-Morgan

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter · #24 ·
Lee Hart Wrote:
>Not really. Any PFC boost converter has a fixed output (about 400vdc)
and an input that varies from 0v to the peak of the AC line (370v for a
240vac line). These are >mass produced by the millions and have
efficiencies of 95% or so.

>Wide-range DC/DC step-down converters are also common. Universal input
switching power supplies will deliver 12vdc with a 90-265vac input
range, and there are >many DC/DCs with a 4:1 input range. A 6:1 input
range is a little more difficult, but straightforward.


I was thinking in terms of High Power and Bi-directional converters.
Since each side needs to be sized for the max current and max voltage it
can produce, Isn't their a inheriant loss of efficiency when some of the
active components are significantly oversized for the main mode of use?

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter · #25 ·
Lee Hart Wrote:
>> Any PFC boost converter has a fixed output (about 400vdc) and an
>> input that varies from 0v to the peak of the AC line (370v for a
>> 240vac line). These are mass produced by the millions and have
>> efficiencies of 95% or so.
>>
>> Wide-range DC/DC step-down converters are also common... there
>> are many DC/DC's with a 4:1 input range. A 6:1 input range
>> range is a little more difficult, but straightforward.

Jeff Shanab wrote:
> I was thinking in terms of High Power and Bi-directional converters.
> Since each side needs to be sized for the max current and max voltage
> it can produce, isn't their a inheriant loss of efficiency when some
> of the active components are significantly oversized for the main
> mode of use?

Not so much a loss of efficiency. It's the price that goes up when you
demand a wider range of voltage and current. For instance, if you want a
1000w converter to work at 100vdc, its transistors need to handle 100v x
10a = 1kw. But if it needs to work from 10v to 100v, then the transistor
need to handle 100a at 10v, and 10a at 100v. Such a transistor is just
as efficient; but it costs a lot more.

--
Ring the bells that still can ring
Forget the perfect offering
There is a crack in everything
That's how the light gets in -- Leonard Cohen
--
Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.net

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter · #26 ·
Morgan LaMoore wrote:
> Except that in a PFC boost converter, the power drawn is proportional
> to the square of the input voltage, so there is low current at low
> voltage.

That is a consequence of the power factor control algorithm, which
forces input current to follow input voltage. But all the PFC converters
I've tested will run on a straight DC input. In this case, the input
current they draw is *inversely* proportional to the input voltage. I.e.
a 100 watt 90-264vac PFC power supply draws:

- 100v at 1a
- 200v at 0.5a
- 300v at 0.33a etc.

> For an EV, you want constant power, so current is inversely
> proportional to voltage. I suspect that this will have a big effect on
> efficiency.

It just means that the voltage and current ratings of the semiconductors
and boost inductor need to take this into account.

> Also, we're talking much higher power here. How many 20kW wall
> adapters do you see?

As a rule, cost per KW goes down as the power goes up. I.e. one 20kw
supply costs less than ten 1kw supplies.

>> Wide-range DC/DC step-down converters are also common.

> Most of these are low current

Of course; that's where the demand is. The market uses a million times
more little DC/DC's than big ones.

> an EV would need 100 times more current, which means 10000 times
> more resistive losses. You don't really want to use 10000 times
> more copper to make up for that.

No; it doesn't work like this. On a crude basis, you could get 100 times
the current by just using 100 little DC/DC converters; thus 100 times
the semiconductors, copper, etc.

But in practice, you'd use a much smaller number of semiconductors and
other parts. When you double the power of a part its size, cost, and
weight will *less* than double. A 0.3w transistor is $0.10; a 3w
transistor around $0.30; a 30w transistor $1; a 300w transistor is $3 etc.

At some point this breaks down, because nobody makes 30,000 watt
transistors. So you have to use many smaller parts in parallel. But even
then, the size/cost/weight go up linearly, not as a squared function.
--
Ring the bells that still can ring
Forget the perfect offering
There is a crack in everything
That's how the light gets in -- Leonard Cohen
--
Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.net

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 
21 - 26 of 26 Posts
Top