Yes, into a very complicated one stage (that contains actually 2 stages at least).
Yes, you can, if you find place enough, maintain cooling, and manage the interference. But this requires full redesigning.
Yes, you can do it, and a buck converter can operate at this freq. without problems. But a hard-switched offline converter at this freq would exhibit terrible EM emittion and/or poor efficiency. But this is only a sidenote, not relevant to the real question.
I'm working on a 3 level power stage now, with only 400 kHz effective switching freq, and the bandwidth (-3 dB) is 75 kHz. This is not a particularly problematic parameter.
The real problem is not here either. The basic problem is: an amplifier is a 4 quarter energy converter, not 1 or 2. It must be able to output positive and negative voltage and independently from the voltage negative and positive current. (Of course average output power is positive.)
Or to be precise 2 (for positive and negative voltage) synchronous (for reversing current), offline (for isolation) buck converter. And this is basically requires much more component than the traditional method, and produces much more EME and switching loss, not to mention the sound quality.
If there was a much more simple power topology without introducing other serious problems, please inform me!
For example the topic starter patent power topology could be modified to be almost as simple as an SMPS and a half-bridge, only by modifying the control, but the switching loss would grow to an unacceptably high level.
Maybe I wasn't clear at making my point.
It IS possible for someone to build a single box to encompass the power supply and amplifier. It COULD be a single stage device with line isolation, but the control loop would be very complex requiring a DSP and a carefully designed algorithm. It may not be the optimal, or most economical solution.
It is NOT possible to simply hack an existing SMPS.
The audio amplifying circuitry MUST indeed be a 4 quadrant device. It must source and sink current and voltage, often in opposing directions due to the phase angle of the speaker load. Many people often assume that a speaker is an 8 ohm resistive load while designing or testing an amplifier, but it is far from it, especially under complex DYNAMIC conditions since it is both an electric and a mechanical device, which converts electrical energy into cone motion, but also converts (uncontrolled) cone motion into electrical energy which can be opposing the amplifiers attempts to move the cone.
The modulated boost / buck converter that I designed some time ago needed to operate at 1MHz for several reasons, one being the bandwith and linearity of the modulation signal. Another was the extremely stringent requirements for suppression of the switching frequency components (carrier) because this went into a radio transmitter than had to pass FCC emission limits. The output filter was all passive lumped elements that would be carrying up to 10 amps. It indeed was not a 4 quadrant device since the "load" was an RF power amplifier (power mosfet) which always drew current.
At that time it required "leading edge" mosfets, fet driver chips, and magnetics. We even had to select specialized ceramic bypass capacitors in the filter that were non piezoelectric, with low dissipation to avoid heating losses. These are all common parts today, even if you don't need (or want) to operate at 1 MHz.
Dear friend, you were clear, and I agree with you, I just wanted to point out that a chain is as strong as its weakest link. If there are many things that can be done in this project, but there is only one that can not, then the result is "no". And one of the weakest link is power topology IMHO. Can be done, but in a more complicated way than the best existing one. Saying "one stage" doesn't help if that stage consist of more parts than the conventional design.
Pafi,
I wanted to thank you for the constructive insight and input. It encouraged me put the smps-amp on the project list for 2017 as a type 1 'new topology' development research. I normally do this when there appears to be wide general interest and an engineering challange.
Type 1 means there will be an extensive IP search and a series of mini-experiments on possible circuit candidates. Then all is evaluated and if promising, moves to Type 2 status.
There may be industrial applications beyond audio for such a device too.
So we'll see, I agree existing topologies are problematic.
Best Regards
Bruce
I wanted to thank you for the constructive insight and input. It encouraged me put the smps-amp on the project list for 2017 as a type 1 'new topology' development research. I normally do this when there appears to be wide general interest and an engineering challange.
Type 1 means there will be an extensive IP search and a series of mini-experiments on possible circuit candidates. Then all is evaluated and if promising, moves to Type 2 status.
There may be industrial applications beyond audio for such a device too.
So we'll see, I agree existing topologies are problematic.
Best Regards
Bruce
Last edited:
.
Exactly....SMPS's and Class D amps are both established technologies well down the development maturity / cost curve.
A new combination of the two, however it is designed will start at the bottom of the maturity curve, and the top of the cost curve. There might be a small market for a $1000+ box like this, but it will be competing for market space with under $100 boxes that are here today.
The Milbert / Berning tube amp / integrated SMPS is rather unique with no direct competition so he can sell them in the $1000+ market. Will there be a rash of clones when his patent expires? Who knows, but I rather doubt it.
I was a "Mr. Fixit" in a large factory circa 1974 to 1984. I can think of a few applications:
Many industrial strength ultrasonic cleaning machines already run directly off of the line voltage. A KW+ at say 40 KHz isn't uncommon. Note, the ceramic transducers (hockey pucks) can store a charge like a capacitor and knock you across the room when you try to take them apart with a metal wrench! You only need to learn this once.
Many giant vibration tables are already class D "audio" amps with a SMPS.....maybe there is a market, but some just use commercial audio amps. Our original table used some really big tube amps. Replaced in the 70's with solid state Crown amps during an upgrade. The whole system was swapped out with "digital" equipment in the early 80's. I suspect the only thing digital about it was the program storage. The program called "Military Jeep" set on about 5G's could disassemble an old trunk mounted two way radio in a few minutes.
Laser cutting tables often use a "voice coil" beam positioner or "galvo". Ours had a pair of 4 inch voice coils each moving a small front surface mirror. Again driven with a specialized audio amp (response to DC) in a feedback loop with optical encoders.
- Status
- Not open for further replies.