Hybrid Power semiconductors in development

Q: Will those products retro-fit the present GenII & III vehicles? If not, they are absolutely NO good to current Prius owners.

 

It's all ready been discussed here: New Chip: 5-10% Better Hybrid MPG | PriusChat

 

I suspect that is why this is in News, not in a Gen 2, Gen 3, or Gen 4 forum. It is expected to get to production in 2020.

 

One thing gleaned from the article was the 5% improvement in efficiency of their current prototype vehicles over current GenIII cars. Earlier some sources had stated ~10% gains. Bummer that we might have to wait until the GenV for the new chips, but by then, I'd bet most mfgr's will have developed something similar, if not better.

 

That is their goal, not the results so far.

Purely a guess, while the new semiconductor needs less cooling and space, I bet the prototypes are in a gen 3 Prius that was designed for a larger and hotter Inverter. So in 2020 when it goes in a production car,THAT car can be designed not to need as elaborate cooling for the inverter as current Prius do.

 

There are more than one semiconductor material around that has promise for fast switch times and low, ON-resistance:
http://powerelectronics.com/site-files/powerelectronics.com/files/archive/powerelectronics.com/images/pdfs/PETInnovations.pdf

• Silicon - existing, well known, highest ON-resistance
• Silicon carbide - refractory, lower ON-resistance
• Gallium nitride - lowest ON-resistance, very fast

There are two efficiency metrics that are important, switching speed and ON-resistance. But to understand them, let's start with Ohm's law and the definition of electrical power:

• Ohms Law has three variations
  • Volts = Resistance * Amps
  • Volts / Amps = Resistance
  • Volts / Resistance = Amps
• Electrical power
  • Watts = Volts * Amps (heat!)

Switching speed is how fast a semiconductor goes between ON-state and OFF-state. Switching between ON and OFF, the voltage and resistance changes lead to brief but very high power consumption:

• OFF - voltage high, 600V, but no current, 0.000001 amps
  • 600V * 0.000001A = ~0.006 W # cool
• ON - voltage low, 0.6-1.2V, during high current, ~10 amps
  • 1.2V * 10A = ~12 W # warm
• SWITCHING - voltage intermediate * current intermediate
  • 300V * 5A = ~1500 W # HOT! HOT! HOT!

The faster the switching time, the less time spend in the heat generating, SWITCHING state. It is why 'over clocking' a CPU often requires more cooling and power saving mode in many laptops run at a slower, cooler speed.

The ON resistance is the other critical metric. Notice that Ohms Law and the power definition have both Volts and Amps in common. So we can take Volts out with algebra:

• Watts = Volts * Amps AND Volts = (Resistance * Amps)
• Watts = (Resistance * Amps) * Amps
• Watts = Resistance * (Amps * Amps)

Decreasing the Resistance directly reduces the Watts generated in the semiconductor. Since we want to switch lots of Amps, reducing the ON Resistance lets the part run cooler, a good thing!

So the primary metrics we look for in power semiconductors are:

1. Low ON-resistance - so it won't generate a lot of heat when conducting
2. Fast switching - to minimize the heat lost when going between ON and OFF

Don't get lost on the specific semiconductor material. A lot of materials have semiconductor properties but the important metrics are ON-resistance and fast switching.

Bob Wilson

ps. I wanted to go over the basics so more folks might understand what this buzz is about.

pps. I did not cover operating temperature because it has nothing to do with efficiency. It can make the semiconductor cooling easier, higher operating temperatures are good, but efficiency is the loss of power during the switch between ON and OFF states.

 

These new electronics will not do much more then 5% in themselves, but 5% in just ONE ITEM is a big deal nowadays. Find another 5% that is cumulative and you are near 60mpg. If these electronics are at the top end of efficiency the water cooling loop could be totally eliminated saving weight and cost.

 

I'm surprised this innovation is coming from Toyota instead of National Semi or Maxim (power IC experts). Nevertheless:

Cool.

 

I worked in high power electronics for a decade. Every year one or more specialty manufacturers would visit with the promise of Silicon Carbide based transistors to revolutionize the industry. The key benefit for us being that they could more easily be made in single devices that could work up to 1000 Volts or more and could operate at higher temperatures. Regular transistors required many of them in series to work at the necessary voltage with larger, heavier and more expensive cooling systems to keep them from frying. They just never seemed to ever get them to market, at least not in quality, quantity or pricing that was ever competitive. It's been a while, so maybe Toyota has finally realized the promise for making them cost effective and reliable enough for automotive applications?

 

Confirmed by EE Times



News & Analysis

Power Week: Toyota Develops SiC Transistor, Structural Supercap & More

Rich Pell​

5/22/2014 01:10 PM EDT​

​

Toyota has developed an SiC power semiconductor for use in hybrid vehicle (HV) power control units (PCUs). The company hopes the device (shown below) will improve HV fuel efficiency by 10% and reduce PCU size. PCUs contain an inverter and voltage converter and account for about 25% of HV electrical power loss -- most of it associated with the power semiconductors.
Toyota developed the SiC device in house to ensure that it met automotive requirements. The company expects to begin test driving vehicles with PCUs containing the device within a year.



Hybrid vehicle PCUs fitted with SiC power semiconductors (right) are expected to be much smaller than those using
traditional silicon power semiconductors (left).
(Source: Toyota)

Source

 

A 25% improvement in power efficiency is possible by reducing the intrinsic forward-conduction voltage-drop of diodes/rectifiers from 0.70V to 0.53V...basically, the difference between Silicon and Germanium (and Schottky configutation) semiconductor materials.

 

True, over the diode, but we are talking over a hybrid. The bulk of the power goes from gasoline, through the ice, through the psd to the trans axle. We do have power flowing between mg1 and mg2, but IIRC this is greater than 85% efficient. Even if you could make the inverter 100% efficient most of the losses will be in the mgs, and it still wouldn't increase system efficiency by 10%. That leaves us with power between mg1 or mg2 and the battery and accessories. I don't have a good feeling for this but imagine the power electronics are normally measured in hundreds of volts, making a 0.17 difference in voltage drop small. Not saying its insignificant, but I am trying to understand the big chance for improvement.

 

Just like the fact that we have so much experience with Otto Cycle engines makes it hard to introduce a new engine technology, (Wankel, Atkinson, Turbine, Electric, etc) the fact that we have so much experience with Silicon makes it hard to introduce new Semiconductors. (GaAs, GaN, SiC, etc.)

Do not expect this to be instant or easy, but there are no instant or easy ways to save 10% (or even 5%) in the mechanical side of a new Prius either.

Austin I think this technology will help getting electricity into or out of the battery (DC) to where we need it in an M/G (AC) do not expect it to change the relations between M/Gs. Part of how the current 5% might become 10% is in software. There is a 'map' of where gas is more efficient than electric power. If they can change power losses to/from the battery, they can change that map. There is currently no mechanical reason we switch from electric to gas at 42 MPH except that if the electricity is just stored gas power, it is more efficient to just run the gas engine. The same hardware switches at 62 MPH when the electricity came from a plug, not conversion of gas. If this does widen the scenarios where electric drive is more efficient than gas power, the software can be taught to make better choices. Increasing the times that the engine is running at minimum grams per HP and decreasing the times it is not optimal.

 

<GROAN>So when is someone going to pop-in with high temperature, superconductors, as the next greatest efficiency gain. . . <GERRRRR>

Bob Wilson

 

When hades freezes over (double entendre' intended)?