Re: Patents
Well I might clarify that you don't need to infringe upon all claims in order to have a patent infringement. It is sufficient that every element of at least ONE claim is shown to be used. This is what so upsets people about the UcD patent. It squarely claims every amplifier that uses a single voltage loop taking off at the speaker terminals. The reason why the patent attorneys who wrote it, felt confident to do so, is that before nobody had actually realised it was at all possible to use only the last state variable in the amp to take feedback from, so there wasn't any prior art.
The above already settles the legal issue.
To say that the "new" feedback circuit function is "somewhat comparable", is the understatement of the year. Anyone equipped with even the most basic of high school maths can show that the impedance equations of both networks are identical. i.e. for each "UcD patent style" (R1//(R2+C)) there is a (R3+(R4//C)) circuit with the exact same impedance and vice versa. The two circuits are not "similar". They are equivalent.
For Lars to keep using this circuit, he will need to obtain a license from Philips. However, since he believes pre-filter feedback to be nicer sounding, I suppose he'll be just as happy just to delete it from the circuit board and the web site.
Lars Clausen said:
Well I might clarify that you don't need to infringe upon all claims in order to have a patent infringement. It is sufficient that every element of at least ONE claim is shown to be used. This is what so upsets people about the UcD patent. It squarely claims every amplifier that uses a single voltage loop taking off at the speaker terminals. The reason why the patent attorneys who wrote it, felt confident to do so, is that before nobody had actually realised it was at all possible to use only the last state variable in the amp to take feedback from, so there wasn't any prior art.
The above already settles the legal issue.
To say that the "new" feedback circuit function is "somewhat comparable", is the understatement of the year. Anyone equipped with even the most basic of high school maths can show that the impedance equations of both networks are identical. i.e. for each "UcD patent style" (R1//(R2+C)) there is a (R3+(R4//C)) circuit with the exact same impedance and vice versa. The two circuits are not "similar". They are equivalent.
For Lars to keep using this circuit, he will need to obtain a license from Philips. However, since he believes pre-filter feedback to be nicer sounding, I suppose he'll be just as happy just to delete it from the circuit board and the web site.
Hello Bruno! Nice to read your clarifications again. What about UcD like style control loop for not audio purpose, e.g. SMPS, motor control? I hope it's quite legal, without license from Philips?
IVX said:
The claim is then made that it's now become novel again since even though prior art is still power switching, it wasn't audio per se.
The obviousness of that particular discovery is rather arguable though, but Philips would be a big Bear, wouldn't it?
Surely this can be resolved in a mutually agreeable way?
I mean, really.....
http://www.semiconductors.philips.com/acrobat_download/literature/9397/75015527.pdf
and I quote "UcD Patent
The UcD principle has been patented
by Philips. The right to use the patent
is granted royalty-free to third parties who
use enabling semiconductor components
from Philips in their UcD amplifi ers.
For more details consult your Philips
sales representative."
Problem solved?
Chris: Thank You for bringing this up, as it turns out i actually use 11 semiconductors from Philips on my amplifier. 😉
Anyway like i said earlier, i have no intention of stepping on anyones toes, much less to enter any legal disputes with anyone. So i used last night to reconfigure the global control loop, to something else that works exactly just as well, and even has the advantage, of maintaining the frequency adjustment intact.
Here is the measurement.....
(10W @ 4Ohm Global Feedback Loop)
Here is the control loop:
And everybody, please feel free to use it as you like.... 😉
Anyway like i said earlier, i have no intention of stepping on anyones toes, much less to enter any legal disputes with anyone. So i used last night to reconfigure the global control loop, to something else that works exactly just as well, and even has the advantage, of maintaining the frequency adjustment intact.
Here is the measurement.....
(10W @ 4Ohm Global Feedback Loop)
Here is the control loop:
And everybody, please feel free to use it as you like.... 😉
Well, Philips is free to turn its patent portfolio into money in any way it likes. It does mean you'll have to use their FETs in the output stage and get an agreement signed. Just designing in a BC857C is not going to get you a license.
Luckily (for us) they're still using the same oversimplified basic circuit implementation described in the patent 😀
Luckily (for us) they're still using the same oversimplified basic circuit implementation described in the patent 😀
Thanks for this information Bruno 
However i really don't think it's worth the effort.
(How about an HEF4093 then??) [just kiddin'] 😀
However i really don't think it's worth the effort.
(How about an HEF4093 then??) [just kiddin'] 😀
That is quite fascinating, i think. I didn't think it was that easy (and free) to get that license. I wonder why this isn't used more often (probably because Hypex is so much ahead??).
Nils
Lars,
I'm sorry, but that pre/post-filter feedback schematic you just posted is not going to do anything to make the frequency response, and to a lesser degree the THD curve, invariable with the attached load. That's what a good class-D design should aspire to achieve; load independent frequency response and THD, if possible with as little phase shift in the audio-band as possible.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
I'm sorry, but that pre/post-filter feedback schematic you just posted is not going to do anything to make the frequency response, and to a lesser degree the THD curve, invariable with the attached load. That's what a good class-D design should aspire to achieve; load independent frequency response and THD, if possible with as little phase shift in the audio-band as possible.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
"The right to use the patent
is granted royalty-free to third parties who
use enabling semiconductor components
from Philips in their UcD amplifiers."
Ok, ok, but my question about SMPS, there is no amplifiers in my plans. E.g. nice and simple BUCK converter, with good dynamics+low impedance, so why not if it's not audio amplifier?
is granted royalty-free to third parties who
use enabling semiconductor components
from Philips in their UcD amplifiers."
Ok, ok, but my question about SMPS, there is no amplifiers in my plans. E.g. nice and simple BUCK converter, with good dynamics+low impedance, so why not if it's not audio amplifier?
29 minutes............
Sassen: D**m you are fast to build my circuit, and test it with different loads.
Or is your comment above based on something else? 🙂
Sassen: D**m you are fast to build my circuit, and test it with different loads.
Or is your comment above based on something else? 🙂
Lars,
I'm actually one of those people that can write a letter, build an ampilfier, play a game of tetris on my PC all at the same time (and pick up the phone if needed) 🙂 No, really, the fact that your circuit doesn't have the desirable features is simple math, no need to whip out the breadboard and build it, some math and a simulator is all that is required. But I'm not trying to rain on your parade, just providing some feedback, afterall, this is the DIY audio forum.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
I'm actually one of those people that can write a letter, build an ampilfier, play a game of tetris on my PC all at the same time (and pick up the phone if needed) 🙂 No, really, the fact that your circuit doesn't have the desirable features is simple math, no need to whip out the breadboard and build it, some math and a simulator is all that is required. But I'm not trying to rain on your parade, just providing some feedback, afterall, this is the DIY audio forum.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
I believe it could work if the distance between the switching frequency and the bandwidth of the filter it big enough.
I believe you also need some kind of voltage divider to control the output voltage to acceptable levels in the integration.
I believe you also need some kind of voltage divider to control the output voltage to acceptable levels in the integration.
lars, looks exiting to say the least.
too bad the amp will cost an arm and a leg as it seems.
i'l just have to talk with that chineese guy ^^
(a phun, i hate chinese products)
too bad the amp will cost an arm and a leg as it seems.
i'l just have to talk with that chineese guy ^^
(a phun, i hate chinese products)
I did some measurements of the new feedback configuration, to target the exact point mentioned by Sander.
It seems the bandwidth now extends beyond the capabilities of my PC based system. So i had to take out my ol' B&O TG7, that plays up til 1 MHz. But unfortunately it cannot generate a curvepaper that i can show here. 🙁
The bandwidth in Global Feedback mode is ...
8 Ohms: 2 - 160.000 Hz (-3dB)
4 Ohms: 2 - 160.000 Hz (-3dB)
Here is the signal at 100kHz, a bit haunted by switching residuals, as i am only playing 3 V rms.
It seems the bandwidth now extends beyond the capabilities of my PC based system. So i had to take out my ol' B&O TG7, that plays up til 1 MHz. But unfortunately it cannot generate a curvepaper that i can show here. 🙁
The bandwidth in Global Feedback mode is ...
8 Ohms: 2 - 160.000 Hz (-3dB)
4 Ohms: 2 - 160.000 Hz (-3dB)
Here is the signal at 100kHz, a bit haunted by switching residuals, as i am only playing 3 V rms.
Cheers Lars,
Appreciate the effort, but the idea is to plot the frequency against the output in a number of loads, ie. 2, 4, 8, 16-ohms. The resulting graph should be a straight line from 20Hz to 20KHz without any peaks or dips for all loads. That's a tell-tale sign of load invariant behaviour.
The same exercise should be repeated for the THD, but plotting THD vs. frequency in 2, 4, 8, 16-ohms at 1W, 10W and 1/2 full-power for example.
If you could round things off with an IMD measurement with 18 and 19KHz testtones we can get a good impression of the performance of your new amplifier in different loads.
Quoting frequency response is as meaningless as stating that an amplifier has 0.01% distortion. Because what kind of distortion does this number actually signify? THD+N? 2nd harmonic? IMD? You get the point I'm sure.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Appreciate the effort, but the idea is to plot the frequency against the output in a number of loads, ie. 2, 4, 8, 16-ohms. The resulting graph should be a straight line from 20Hz to 20KHz without any peaks or dips for all loads. That's a tell-tale sign of load invariant behaviour.
The same exercise should be repeated for the THD, but plotting THD vs. frequency in 2, 4, 8, 16-ohms at 1W, 10W and 1/2 full-power for example.
If you could round things off with an IMD measurement with 18 and 19KHz testtones we can get a good impression of the performance of your new amplifier in different loads.
Quoting frequency response is as meaningless as stating that an amplifier has 0.01% distortion. Because what kind of distortion does this number actually signify? THD+N? 2nd harmonic? IMD? You get the point I'm sure.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
- Status
- Not open for further replies.