Clever indeed, sure! Some reading: Spectron Digital Audio Amplifiers
And about patents? You can even patent the word "Monster" (Google for it) 😀 B.t.w. I'm not discussing Bruno's skills, which I regard high, but the history/origin of class-D etc.
And about patents? You can even patent the word "Monster" (Google for it) 😀 B.t.w. I'm not discussing Bruno's skills, which I regard high, but the history/origin of class-D etc.
Last edited:
Yes the Spectron is just a development of the mid 70's Infinity SW(itching) AMP. Which might have been just a development of 1964 Sinclair, which was a development of a 1958? patent ?
Everyone knows class-D has been around since the late 40s
Switching Amp Patents has just some of the patents around the technology.
The fact remains that Hypex is well known in stereophile recommended components. ICEpower maybe should have been, but they clearly didn't sell or license right. To develop a state of the art implementation AND get big names using it to me is still genius.
Mod86 still performs better within its power envelope though.
Switching Amp Patents has just some of the patents around the technology.
The fact remains that Hypex is well known in stereophile recommended components. ICEpower maybe should have been, but they clearly didn't sell or license right. To develop a state of the art implementation AND get big names using it to me is still genius.
Mod86 still performs better within its power envelope though.
Don't forget about Tripath and its patented version of class D (US 5,974,089) which has shipped a lot of amplifiers, gotten its own article in Wikipedia, managed to go bankrupt in ten years, and have its Intellectual Property purchased by Cirrus Logic for pennies on the dollar.
A message from the marketing department 😛 But maybe your are right...
Nevertheless nested feedback loops is also ages old. Baxandall described it 50 years ago for audio amps. It's no rocket science, just do your homework about control theory.
The US has two different types of patents:
1] To keep it simple, lets call these inventions.
2] Design patents, somewhat similar to a copyright. You can patent the design of your new refrigerator, so no one else can manufacture a refrigerator that looks like yours.
I am not a marketing dept. But I can look at 2 graphs next to each other.
No one said otherwise, but you are becoming very Troll-like. Please try to move comments back on subject. The subject not being class D. 🙂
I've never claimed it was new or rocket science. You may refer to Post #1 of this thread for some references on the topic.
It takes a bit more than "just doing your homework", though. In particular the idiosyncrasies of the LM3886 are interesting to deal with in a composite amp.
I second Bill's notion that you may want to adjust your communication style a bit. You're coming across as someone who is looking for a fight rather than as someone looking to engage in a meaningful dialog.
Tom
Exactly. It was novel, non-obvious, and reproducible by someone skilled in the art. That's what a patent is. At least that was the standard I was held to as a co-inventor on US Patent #8,446,193.
Post-filter feedback is how Bruno gets the stellar performance in his circuits. That's his secret sauce. While the THD of his circuits is quite a bit higher than that of the Modulus-86 (based on my measurements of the NC400 vs MOD86 Rev. 2.0) Bruno's amps are the best performing Class D amplifiers I am aware of. They're quite nice. Closing the feedback loop around the output filter is definitely not trivial. That's why Bruno is fighting a 5th order control loop. That's also how he was able to do something novel and non-obvious and get a patent. Good for him.
Tom
I am not blaming you Tom. But stating a Mod86 still performs better within its power envelope than UcD, say a nCore 500, is just a plain bold statement without evidence. So maybe Bill can adjust his communication?
I've finished last year my own LM3886 amp supported by extra loop gain of an op-amp. An I know about the implications (using the chip for more than 20 years in various projects). The amp sounds way better and more refined than just a plain LM3886. So I've no doubts yours will do also.
This amp has nothing to do with your modulo amps. End of May I have probably an opportunity to compare one of your modulus builds, nCore and my own amp.
Last edited:
US Patent Full-Text Database Manual Search agrees! I asked it for a list of all US Patents with an inventor named "Christiansen, Tom" (and Thom and Thomas), and the one above was the only electronic circuit design patent it found.
While on that site I checked my own name too, and found that my earliest US Patent is 4,523,110 and my most recent is 9,214,243 with a bunch of them in between. The one that paid for my house was 5,101,117 🙂
You assumed I hadn't bothered checking all published result for the Hypex amplifier before I made that statement. The 500 is hard to get comparative figures for as you can only see complete amplifiers where some OEM has added their own secret source (read higher distortion) input stage. I would be happy to be proved wrong, but I can't find any parameters other than efficiency and power output where the hypex wins.
Now if you need 400W into 4 Ohms it's the go to puck of goodness. I don't..yet, but may well do in the future and will be top of my list.
What if you "normalize" by power consumption though?
What I mean is, if you define performance strictly in terms of all the various noise, THD, etc, graphs, that's one thing. But if your definition of performance includes something along the lines of noise per watt of power consumed, a solid class-D amp might take the lead. (By virtue of class D typically (always) having dramatically better power (and heat) efficiency compared to class AB.)
Not trying to say one is the correct definition of performance, different strokes for different folks of course. But if you are (for example) battery or solar powered, you might be willing to sacrifice some ultimate specs for more runtime. Would you rather have an hour of absolutely pristine music, or a whole day of really good music? 😉
On the other hand, at typical listening levels---say typically pushing ~5W or so to the speakers---the efficiency delta becomes smaller.
At 1 Watt the difference is small. Hypex has about a 5W idle dissipation. 3886 has 50mA quiescent. Not much in it. By 5W the hypex is probably ahead, by 50W no contest. Not sure how a blameless would stack up in that.
But if you were running off batteries would you chose a 200W amplifier?
In our modern green age it's worth considering. Also worth noting that Mod-86 max dissipation in recommended configuration is 40W. Compared to the highly class A biased behemoths some people build it saves polar bears and plants rainforests.
My off grid system would be econowaves and a Tripath I was given.
I would be interested in multi-channel version of Modulus-86/Power-86. Would it be interesting to see an integrated 5-channel Modulus-86/Power-86 as a single PCB? For example, it can be used to drive either active crossover speakers or 5.1 speakers. Considering such a good performance, Modulus-86's effective power (i.e. max power at a certain THD threshold) is likely to far exceed big brands' separate multichannel amplifiers.
I'm mulling over building a stereo board as I have a customer/co-conspirator who's looking for a multi-channel amp. I'm leaning in the direction of a stereo amp board that allows for stacking. You could then stack as many stereo boards as you can pack in on your heat sink.
This would need to fly in a production environment, so there are some things to think through on connector selection and such. It would also be an SMD build to the extent possible in order to facilitate external assembly.
At this point I'm at the "wouldn't it be cool if..." stage or "product definition" stage I guess is the proper phrase. Nothing on paper yet.
Tom
This would need to fly in a production environment, so there are some things to think through on connector selection and such. It would also be an SMD build to the extent possible in order to facilitate external assembly.
At this point I'm at the "wouldn't it be cool if..." stage or "product definition" stage I guess is the proper phrase. Nothing on paper yet.
Tom
It would be. On several occasions have thought about lobbing a couple of mod-85 in a 2U case and making an 'ultimate' low power integrated amp. But I always got stuck on layout and thermal management and realised that the QUAD 405 was a sensible layout.
Maybe the simplest physical arrangement would place the PCBs perpendicular to a single shared heatsink, rather like PCI daughter cards plugged into a desktop computer's motherboard, perpendicularly. You'd probably want a piece of flat aluminum bar attached to each PCB; the chipamp IC's flange is greased and insulated and bolted to the aluminum bar in the factory by trained workers. (not in the field by untrained customers). Then the aluminum bar is bolted to the heatsink when the customer assembles an N-channel amplifier.
The aluminum bar gives an opportunity to mount a thermal sensor in relatively intimate contact with the chipamp IC, and send temperature data to a master controller / whizzo front panel display if present. A TC74 chip ($1.37) mounts extremely easily since it's in a TO220 package with a bolt hole and big flat thermal contact surface. Or to save BOM cost the MCP9700 ($0.20) temperature sensor is in a TO92. Epoxy it to the aluminum bar, good to go. People who want it, get it. People who don't want it, pay a couple bucks extra for a feature they disregard.
This gives the possibility to build temperature controlled fan cooling, if desired. It also lets you individually shut down each amplifier PCB at whatever temperature YOU like, rather than at some arbitrary fixed temperature the chipamp IC designers chose. You have the option to shut it off and keep it off / let it cool down to XX degrees of your choosing, before restarting.
Vertical (or perpendicular) arrangement seems to be reasonable for thermal management as well as modular design. Emotiva's XPA GEN3 design looks really nice for modular / multi-channel amp. Is Modulus-86 working fine with SMPS (i.e. no significant performance degradation)? If so, multi-channel implementation would be more compact.