[00:00:00] Speaker 03:
Our next case is number 21, 1946, Aquila Innovations Inc.

[00:00:05] Speaker 03:
versus Advanced Micro-Devices Inc.

[00:00:09] Speaker 01:
Okay, Mr. Shearn.

[00:00:12] Speaker 01:
Good morning, Your Honors.

[00:00:13] Speaker 01:
May it please the Court?

[00:00:15] Speaker 01:
The Board's determination that claim one is obvious over the combination of Ober and Nakazatov should be reversed for three reasons.

[00:00:24] Speaker 01:
First, the Board's findings were not supported by the evidence in the record.

[00:00:29] Speaker 01:
Second, the board implied an incorrect and inconsistent construction of the ordinary operation modes in claim one.

[00:00:38] Speaker 01:
And finally, because the board relied on a new theory of invalidity that was not presented in the petition, namely modifying Ober's microcontroller with the speed control circuit of Nakazato.

[00:00:59] Speaker 01:
Turning to the board's findings, the board's finding that first committed error by finding that OBOR alone teaches or suggests the claimed ordinary operation modes of the 519 PAT.

[00:01:19] Speaker 01:
The record shows that OBOR does not disclose the claimed ordinary operation modes, OBOR only

[00:01:28] Speaker 01:
discloses a single mode, run mode, in which Ober's state machine in the microcontroller, the low-speed clocks are false.

[00:01:35] Speaker 01:
While Ober's microcontroller is active, there is only one CPU frequency that is provided to that CPU.

[00:01:45] Speaker 01:
What the board pointed to for the purported suggestion was the reduced clocks in idle or sleep mode.

[00:01:53] Speaker 01:
But there's no evidence on the record

[00:01:56] Speaker 01:
support such a suggestion.

[00:02:00] Speaker 01:
Dr. Albanese testified, AMD's expert testified, that a similar technique could be used for sleep mode as in run mode.

[00:02:09] Speaker 01:
First, that implies that that shows that Ober doesn't explicitly disclose these divided clocks in the run mode.

[00:02:18] Speaker 01:
But moreover, the record evidence simply doesn't support that such a suggestion or that a person with a new skill in the art would find that obvious.

[00:02:27] Speaker 01:
First, we introduced evidence that using the technique of writing to unused bits in the register would not result in the control of the clock frequencies as AMD asserts.

[00:02:41] Speaker 01:
And in fact,

[00:02:45] Speaker 01:
The data sheets on the record show that writing to unused or undefined bits, whatever you call the blank fields in the register, would produce undefined, unpredictable or undesirable behavior.

[00:02:59] Speaker 01:
It would not result in what AMD asserts is the ordinary operation modes operating at different frequencies.

[00:03:10] Speaker 01:
I'm just wondering if your honors have any questions on these points so far.

[00:03:18] Speaker 03:
Just go ahead.

[00:03:21] Speaker 01:
Okay.

[00:03:21] Speaker 01:
The board also aired in its analysis of the motivation to combine.

[00:03:33] Speaker 01:
AMD asserted that the references themselves motivated such a combination and that

[00:03:40] Speaker 01:
because Ober didn't explicitly disclose how to modify these, reduce the CPU clock frequency during normal mode, that a person of skill in the art would have been motivated to go to Nakazato.

[00:03:54] Speaker 01:
But there's simply no evidence on the record that shows this.

[00:03:57] Speaker 01:
And in fact, what we showed was that Ober, the principal purpose of Ober was to

[00:04:07] Speaker 01:
introduces a microcontroller that is able to independently control the power consumption of the peripheral subsystems within the microcontroller, because as Ober teaches the conventional way of reducing or stopping the system clock to reduce power,

[00:04:27] Speaker 01:
was undesirable because it did not optimize the power consumption of the peripherals.

[00:04:33] Speaker 01:
And because of that, it introduced this decentralized approach.

[00:04:37] Speaker 01:
And this was also the subject of claim construction when the OBOR patent was in litigation.

[00:04:43] Speaker 01:
The district court in that case also construed and interpreted OBOR to espouse a decentralized approach.

[00:04:56] Speaker 01:
That evidence shows that one would not, personal skill in the art would not use OBERS to reduce the system clock frequency because overall that would cut against OBERS stated purpose of controlling the power consumption in the peripherals.

[00:05:17] Speaker 01:
On the second point, the board also aired

[00:05:25] Speaker 01:
in its construction of ordinary operation modes.

[00:05:30] Speaker 01:
The board's construction, we submit, was inconsistent.

[00:05:34] Speaker 01:
The board found that the preamble of claim one, a system LSI having a plurality of ordinary operation modes and a plurality of special modes in response to frequencies provided to the CPU, did not require the ordinary operation modes to have different frequencies in the preamble.

[00:05:55] Speaker 01:
But later on in the body of the claim for the limitation of first memory storing a clock control library for controlling a clock frequency transition between said ordinary operation modes, the board acknowledged, yes, there the ordinary operation modes do have different frequencies so that there can be a transition.

[00:06:16] Speaker 01:
And that's reflected in the rest of the claim where the clock control library writes a value into the register of the system control circuit, which is the actual hardware that actuates

[00:06:25] Speaker 03:
When you look at that argument that you just made, aren't you arguing or aren't you inviting us to import a limitation from another claim element into the preamble that you're addressing?

[00:06:40] Speaker 03:
I don't believe so, Your Honor, because it's... You agree that that would be incorrect to do that?

[00:06:44] Speaker 01:
I agree that.

[00:06:47] Speaker 03:
I'm not sure what your... That would be error to import a claim limitation from another claim into the preamble of another claim.

[00:06:55] Speaker 01:
That would depend.

[00:06:57] Speaker 01:
Here we are talking about the same term used in the preamble and in the body of the claim that's been given two interpretations by the board.

[00:07:05] Speaker 01:
First, in the preamble where the board said that, well, the plurality of ordinary operation modes in response to clock frequencies supplied to CPU don't require different frequencies.

[00:07:18] Speaker 01:
And then later in the body of the claim, the ordinary operation modes

[00:07:21] Speaker 01:
do require, do have different frequencies because there has to be clock frequency transitions between said ordinary operation modes.

[00:07:29] Speaker 01:
And the body of the claim there, the said ordinary operation modes, as the board acknowledged in the final written decision, refers back to the ordinary operation modes of the preamble, which the board agreed is limiting.

[00:07:43] Speaker 01:
So we submit that the board in saying,

[00:07:47] Speaker 01:
The ordinary operation modes don't have to be the same, don't have to be different frequencies in the preamble, but have to be different frequencies in the body of the claim.

[00:07:55] Speaker 01:
That was incorrect, and that was error.

[00:07:59] Speaker 00:
Mr. Chung, you made a point that I'm not sure I quite got.

[00:08:05] Speaker 00:
Something about how a modification to OBOR would not be undertaken because something about undermining

[00:08:15] Speaker 00:
Ober's effort to save power with respect to peripherals.

[00:08:19] Speaker 00:
Sorry for articulating it so badly, but my question is, can you say that again so I might understand it better?

[00:08:26] Speaker 01:
Of course, Your Honor.

[00:08:27] Speaker 01:
The theory in the petition was that a person of ordinary skill in the yard would take Ober's microcontroller and then load Nakazato's power saving driver into Ober's memory banks, execute it, and arrive at the invention claimed in the 519 patent.

[00:08:45] Speaker 01:
The power saving driver of Nakazato.

[00:08:47] Speaker 01:
Nakazato teaches that its power saving driver and utility writes to a register in its speed control circuit, which is off chip and then a separate chip from the CPU, which then will reduce the system frequency on the board that's provided to the CPU.

[00:09:06] Speaker 01:
And so based on AMD's sort of combination was that combining these two would, you know,

[00:09:14] Speaker 01:
microcontroller by installing Nakazato software on it would result in the invention claimed in the 519 patent.

[00:09:21] Speaker 01:
However, and one of AMD's assertions in support of this was that the technique of reducing SysFrequency, CPU frequency to supply power was known.

[00:09:34] Speaker 01:
And the 519 patent specification does cite to, for example, the SC7 microcontroller, the RM920T microcontroller,

[00:09:41] Speaker 01:
some examples of conventional microcontrollers at the time that did use this technique.

[00:09:46] Speaker 01:
But the data sheets that we submitted of those microcontrollers say that you can't write to the blank registers without the microcontroller performing unpredictably and undesirably, or that the output would straight up be ignored.

[00:10:08] Speaker 01:
So what AMD is saying would be the combination would actually not result in the claimed invention of the 529 pattern.

[00:10:16] Speaker 01:
Simply installing the software, writing to a blank field in the system over a power management block would not result in over the run mode CPU frequency being reduced, as AMD is assert.

[00:10:35] Speaker 01:
And finally, I would like to address the new theory that the board relied on in the final written decision.

[00:10:46] Speaker 01:
As I've mentioned several times before, the combination in the petition was to use Ober's hardware, Nakazato software, put them together, and there you go.

[00:10:56] Speaker 01:
We argued below that, as I've said, that that combination wouldn't do what AMD says it does, and that further actual hardware modification that was not asserted in the petition would be required.

[00:11:08] Speaker 01:
In the final written decision, the board said that we find that a personal steel yard would take Nakazato's speed control circuit and just slap it onto Ober's microcontroller, and that would achieve the combination.

[00:11:24] Speaker 01:
And this is something that had not appeared

[00:11:26] Speaker 01:
at any time before in this proceeding.

[00:11:29] Speaker 01:
It wasn't in the petition.

[00:11:31] Speaker 01:
It wasn't raised in the reply to the patent owner response.

[00:11:36] Speaker 01:
It wasn't raised at the argument.

[00:11:37] Speaker 01:
We simply had no opportunity to respond to this new theory that the board put forward.

[00:11:43] Speaker 01:
And so we submit that that was a violation of Aquila's rights under the Administrative Procedure Act.

[00:11:50] Speaker 02:
All right.

[00:11:51] Speaker 02:
Do you want to save the rest of your time for rebuttal?

[00:11:52] Speaker 01:
I would, Your Honor.

[00:11:53] Speaker 01:
Thank you.

[00:11:53] Speaker 01:
OK.

[00:11:53] Speaker 01:
Thank you.

[00:11:56] Speaker 03:
Mr. Claus.

[00:12:01] Speaker 02:
Thank you, Your Honor.

[00:12:01] Speaker 02:
I'm going to please the court.

[00:12:03] Speaker 02:
I'll just address the points raised by my colleague.

[00:12:06] Speaker 02:
I'd like to start with the question asked by Judge Toronto about this issue of being able to use over, which teaches, as is shown in the final written decision, everything about this claim except using software to control the register that changes the speech.

[00:12:26] Speaker 02:
That's really all this is about.

[00:12:27] Speaker 02:
Everything else in this claim is admittedly by Aquila taught by Ober.

[00:12:35] Speaker 02:
Ober and Nakazato are compatible references.

[00:12:40] Speaker 02:
They both teach the same advantage of if you lower the speed, you lower the power.

[00:12:45] Speaker 02:
That's an important advantage in CPUs.

[00:12:47] Speaker 02:
Every CPU

[00:12:48] Speaker 02:
that goes in one of our laptops, the company who's building it cares a lot about how much power it draws because that's battery life, that's how marketable it is, that's how well it does when you and I make decisions about which laptop to buy.

[00:13:01] Speaker 02:
So this is not a minor issue.

[00:13:03] Speaker 02:
This is not a minor motivation.

[00:13:05] Speaker 02:
So now let's talk about this particular issue about the data sheets and whether or not a particular data sheet says, well, this register bit is unused.

[00:13:14] Speaker 02:
That's right.

[00:13:15] Speaker 02:
It is unused.

[00:13:16] Speaker 02:
That's the whole point, was it was available.

[00:13:18] Speaker 02:
And the experts explained exactly how, or AMD's expert explained exactly how, that available, currently unused, but available register bit could be used with the software taught by Nakazato to control the system speed of the CPU and therefore lower the power usage of the overall device.

[00:13:40] Speaker 02:
And the board specifically looked at all of the evidence in rejecting this argument.

[00:13:44] Speaker 02:
And I would turn, Your Honor, to page 61 of the appendix.

[00:13:48] Speaker 02:
The board walks through evidence, including the 519 patents background, the over-reference specific parts of it, Nakazato specific parts of it, and specific parts of both experts' testimony, and concludes, these documents provide ample evidence that a system CPU can operate and behave properly

[00:14:11] Speaker 02:
at multiple and lower frequencies.

[00:14:14] Speaker 00:
Again, I'm afraid I don't think I've wholly internalized it.

[00:14:19] Speaker 00:
I think I heard Mr. Sheng said that maybe something in the data sheets said, don't use unused registry bits.

[00:14:32] Speaker 00:
And are you saying that

[00:14:33] Speaker 00:
First, I don't know if that's accurate.

[00:14:35] Speaker 00:
But if it's accurate, are you saying that there was contrary evidence and the board credited the contrary evidence?

[00:14:41] Speaker 00:
Or have I correct anything that needs correction there?

[00:14:46] Speaker 02:
So first of all, what the data sheet says is it's unassigned.

[00:14:50] Speaker 02:
So Ober, and the chip in question,

[00:14:54] Speaker 02:
It's unassigned, right?

[00:14:55] Speaker 02:
So as built, it's not performing a role.

[00:14:58] Speaker 02:
That doesn't mean it can't perform a role if you rebuilt it based on the teachings of Nakazando, but as built, it doesn't perform that role.

[00:15:06] Speaker 02:
But this is not an anticipation case.

[00:15:09] Speaker 02:
This is an obviousness case.

[00:15:10] Speaker 02:
And this court has said time and time again, it's not about whether you can cut out figures from the two different references and glue them together.

[00:15:19] Speaker 02:
It's about whether the teachings of the references, not the specific inventions that have been claimed, but the teachings in the hands of a person of ordinary skill and the art, that person would be motivated to combine them and would have a reasonable chance of success in combining them.

[00:15:35] Speaker 02:
And here, the board identified a wealth of evidence backing up both of those findings.

[00:15:42] Speaker 02:
And we didn't hear anything about the board's evidence.

[00:15:45] Speaker 02:
We heard about the evidence that Keela put in about how this district court construed Ober and about this data sheet.

[00:15:52] Speaker 02:
That's fine.

[00:15:53] Speaker 02:
That would have been fine if Aquila had won below.

[00:15:55] Speaker 02:
But here, we're not looking at that.

[00:15:57] Speaker 02:
We're looking at whether the fact finder, the board, had its own evidence.

[00:16:01] Speaker 02:
And it did.

[00:16:02] Speaker 02:
It identified it.

[00:16:03] Speaker 02:
And we haven't heard anything about that evidence, either in the briefs or in the argument.

[00:16:06] Speaker 02:
It's basically ignored.

[00:16:09] Speaker 02:
Does that answer your question, Your Honor?

[00:16:10] Speaker 02:
I think so.

[00:16:10] Speaker 02:
Thank you.

[00:16:11] Speaker 02:
So then I just moved to the other points that were discussed.

[00:16:15] Speaker 02:
So for example, he criticizes the board in saying that over

[00:16:19] Speaker 02:
that had ordinary operation modes, but it's important to listen to his language carefully.

[00:16:24] Speaker 02:
He says, the board was incorrect that Oberd disclosed ordinary operation modes.

[00:16:28] Speaker 02:
That's not what the board said.

[00:16:30] Speaker 02:
The board said it suggested ordinary operation modes.

[00:16:34] Speaker 02:
And the board cited to column 10, lines 1 and 2 as supporting that.

[00:16:39] Speaker 02:
And the language there says, with regard to this dividing clock, and also may provide a divided clock signal to

[00:16:49] Speaker 02:
the subsystem during a normal mode.

[00:16:55] Speaker 02:
So all the board was saying was, look, there's a suggestion of that, not necessarily a disclosure.

[00:17:00] Speaker 02:
And that was the whole point.

[00:17:01] Speaker 03:
What would you say is the difference between, in the Penn Law context, a suggestion and a disclosure?

[00:17:09] Speaker 02:
Well, Your Honor, I would say that the teaching here, a teaching, of course, is just a broad technology.

[00:17:16] Speaker 02:
A disclosure is enough so that someone can make it

[00:17:19] Speaker 02:
without substantive, what's the rule under enablement, that you're able to make it without undue effort, right?

[00:17:26] Speaker 02:
So that would be a disclosure.

[00:17:28] Speaker 02:
And we presume that prior art in the form of US patents does meet that level of disclosure.

[00:17:34] Speaker 02:
A suggestion is, here's something you can do, but doesn't necessarily give all the details so that the person of ordinary skill in the art could meet the enablement.

[00:17:42] Speaker 02:
It would be my view of how those play out.

[00:17:46] Speaker 02:
But the bottom line is that the board said suggestion and proved suggestion and had substantial evidence for it.

[00:17:53] Speaker 02:
With regard to the district court construed over during litigation, as I said, that may have been evidence Aquila put in, but that in no way shows that the evidence the board relied upon was unsubstantial.

[00:18:04] Speaker 02:
That's not the test on it.

[00:18:07] Speaker 02:
Now let's talk a little bit about this issue with regard to whether or not there was inconsistency in the board's claim construction.

[00:18:15] Speaker 02:
This is not an inconsistency case.

[00:18:18] Speaker 02:
Simply put, the analogy would be a preamble that says a chair with a further limitation that says a chair having four legs.

[00:18:26] Speaker 02:
The fact that the board would say, well, a chair isn't required to have four legs just from the term chair.

[00:18:32] Speaker 02:
It's required to have four legs because you included the language with four legs.

[00:18:37] Speaker 02:
That is exactly what this court has done many times in saying where there's additional language, you don't read the initial term as having that limitation.

[00:18:46] Speaker 02:
That would be a surplus problem.

[00:18:48] Speaker 02:
And that's exactly what happened here.

[00:18:49] Speaker 02:
You had ordinary operation modes.

[00:18:52] Speaker 02:
And then in a later limitation, you had ordinary operation modes with this transition of frequency.

[00:18:59] Speaker 02:
And the board recently enough said ordinary operation modes doesn't require that.

[00:19:03] Speaker 02:
But later on, there's a limitation that does.

[00:19:06] Speaker 02:
and it used the proper evidence at each time to say, here it is for the preamble where it doesn't require it, here it is for limitation 1.1 where it does.

[00:19:15] Speaker 02:
Fairly normal claim construction.

[00:19:17] Speaker 02:
No problem with ordinary operation modes being interpreted differently because it was the additional language that caused the change in the interpretation.

[00:19:27] Speaker 02:
Finally, we have this issue of the argument that there was a new theory because the initial theory was just hardware plus software.

[00:19:36] Speaker 02:
And I turn the court to the reply brief, the yellow brief, and specifically in that brief on page six.

[00:19:46] Speaker 02:
Because yes, there's an attorney argument there where Aquila says, AMD admits that the petition's theory was Ober's hardware combined with Nakazato's software.

[00:19:56] Speaker 02:
But that attorney argument survives for roughly six lines.

[00:20:00] Speaker 02:
Because right underneath that, Aquila gives a quote about Ober

[00:20:06] Speaker 02:
which, so this is like three lines underneath that 2A heading, over-discloses a system LSI with a clock subsystem that supplies various frequencies of clocks and a system control circuit that can switch between clocks frequencies via a register.

[00:20:25] Speaker 02:
Two lines later, Nakazato,

[00:20:28] Speaker 02:
disposes user-selectable speeds in a utility that uses a driver to control CPU frequency with a register.

[00:20:35] Speaker 02:
There's no dichotomy between hardware and software between these two references.

[00:20:40] Speaker 02:
If there were, where are you putting register?

[00:20:42] Speaker 02:
Clearly both references understood that there was hardware involved and taught the combination of software and hardware.

[00:20:49] Speaker 02:
So the very dichotomy that they premised this new theory argument on, even when they cherry-picked their best sentences from all the briefs and the decisions, they can't maintain it because it wasn't there.

[00:21:02] Speaker 02:
Both references taught that you could use software and hardware.

[00:21:07] Speaker 02:
And the 591 patent itself in the background taught that it was known in the art in the ST7 to be able to modify the frequency at which the CPU ran in order to save power.

[00:21:21] Speaker 02:
This is a garden variety obvious in this case.

[00:21:23] Speaker 02:
The board had the evidence it needed.

[00:21:25] Speaker 02:
There's certainly no claim construction problem with regard to what was raised today at argument.

[00:21:31] Speaker 02:
So unless the panel has additional questions, I ask you to affirm and give up my time.

[00:21:37] Speaker 02:
OK, thank you.

[00:21:38] Speaker 02:
Thank you, Mr. Horowitz.

[00:21:41] Speaker 01:
Sure.

[00:21:41] Speaker 01:
Thank you, Your Honor.

[00:21:42] Speaker 01:
Addressing this issue that's come up several times now about what exactly the petition has asserted, I would direct the court to Appendix 132 for really what the gist of AMD's theory is.

[00:21:57] Speaker 01:
There, AMD says that combining Ober and Nakazato would yield predictable results because Nakazato's driver and utility would merely be executed

[00:22:08] Speaker 01:
from Ober's memory banks to just Ober's CPU clock control register resulting in an integrated system for addressing CPU clock.

[00:22:15] Speaker 01:
And my colleague is correct that physical combinability is not required, but this was their theory.

[00:22:21] Speaker 01:
They relied solely on Ober's hardware and Nakazato's software.

[00:22:25] Speaker 01:
If there is a dichotomy, it exists in the petition's theories, which we had to respond to.

[00:22:31] Speaker 01:
If you look further, the court looks further at

[00:22:35] Speaker 01:
Appendix 133 in the section that follows, you see that the petition relies solely on OBOR to teach or suggest the preamble.

[00:22:49] Speaker 01:
AMD did not bring in Nakazato in hardware or software form to teach the preamble, and that is where, that's kind of where the issue is here.

[00:23:01] Speaker 01:
I also wanted to address this characterization that OBOR

[00:23:05] Speaker 01:
discloses everything but using software to control the register.

[00:23:11] Speaker 01:
That is actually one of the big points of contention between parties that we do not admit that that is the case for Ober.

[00:23:17] Speaker 01:
In fact, one of the big arguments is that Ober doesn't disclose in run mode using reduced frequencies because its power management state machine, everything that Ober's microcontroller in the patent defines is premised on

[00:23:32] Speaker 01:
run mode a specific state in the state machine having low speed clocks equals false.

[00:23:38] Speaker 01:
And so there's no teaching that overdoes that.

[00:23:43] Speaker 01:
Now onto this thing about a suggestion.

[00:23:45] Speaker 01:
That's where I mentioned earlier that AMD's expert testified that, well, yeah, we can use a similar technique to just write

[00:23:55] Speaker 01:
available registers, bits in the available, in the blank fields in the SFR62, and that would result in the resulting clock frequency reduction.

[00:24:05] Speaker 01:
But as we've just discussed in the data sheets, that that's actually not the case.

[00:24:09] Speaker 01:
And the data sheets don't call them, you know, regardless of how you want to characterize them, unused, undefined, they're blank, nothing's being written to them.

[00:24:18] Speaker 01:
And the data sheets say clearly, don't do anything with them.

[00:24:22] Speaker 01:
And so where the point is being made that, well, if you just rebuilt the microcontroller to do it, that's not the theory that was presented.

[00:24:29] Speaker 01:
The theory was merely executing Noxado software from Oberst hardware, and the rest would follow.

[00:24:37] Speaker 03:
OK.

[00:24:37] Speaker 03:
Thank you, Mr. Chairman.

[00:24:38] Speaker 03:
Thank you.

[00:24:39] Speaker 03:
Both counsel cases submitted.