February 4, 2009 at 10:11 pm #15165
Experiments with amplifiers and loudspeakers, all part of the work at the Sound House to improve on systems of stereo sound from a single loudspeaker enclosure, have thrown up some fascinating results and may have produced a major breakthrough in audio reproduction.
I have worried for years over the extraordinary difference between loudspeaker performances compared to amplifier performance…. Modern power amplifiers are close to perfect in that they are capable of amplifying a voltage and then supplying it at almost any current, into a loudspeaker. But the loudspeaker is a dreadful hotchpotch of electro-mechanical hit and miss, that converts that amplifier output voltage (current actually) into sound. It does it extremely badly with pathetic efficiency and with distortions hundreds of times greater than the distortions of the amplifier.
Many years ago at ‘Alice’ we developed a little power amplifier specifically designed to get best performance from a talkback loudspeaker in a console. The amplifier worked in what we called ‘current mode’…. It was basically a constant current amplifier.
And here is the paradox: A conventional amplifier works best with a very high damping factor, in other words, with a very low output impedance. The output of the amplifier looks like a short circuit to the loudspeaker, and so the ‘speaker is heavily ‘damped’ because stray movement of the voice-coil is inhibited by a short circuit.
A ‘constant current’ amplifier has, by definition, high output impedance, so a loudspeaker connected to it would be completely undamped and so would resonate and generally sound horrible…… wouldn’t it?
The paradox is that while the constant voltage amplifier controls the loudspeaker moderately well, the constant current amplifiers we use nowadays control it even better!
Of course, the paradox, as is so often the case, is in our own heads and has its roots in our insistence on simplistic labelling.
Our ‘airDRIVE’ amplifiers operate as constant current in that their output current is exactly proportional to their input voltage, but they control the action of the loudspeaker completely by including it as a part of the amplifier electronics so that within wide limits, the actual pressures in the air are proportional to the input current.
In a conventional system the amplifier creates an accurate audio signal at very low impedance, the loudspeaker reacts to that signal being damped and controlled by the damping factor. This is valid except that it assumes that the loudspeaker is a single entity, behaving as a unit, when in reality it is a coil with parameters that change throughout its length so that the true amount of damping and control is very much less than theory would forecast; and even worse, the amplifier/coil interactions create momentary changes to the coil impedance altering significantly the linearity of the audio.
In an airDRIVE system all those non-linearities are swept away; the current flow in the coil is totally controlled and therefore the pressures on the air are much more accurate.
airDRIVE is a huge improvement to fidelity, however it is novel, not easy to implement, and can be downright dangerous with its potentially high voltages.
At the Sound House we shall continue to experiment, and it’s likely we shall have a production airDRIVE available before too long.August 25, 2015 at 8:29 am #15444
This is a very interesting idea. It makes total sense to control the current in the speaker of course – I know the idea has been around a while but this is the first I heard of someone actually making it work.
Out of curiosity- how high the voltages get at the amp output? I guess it depends to some degree on the speaker, so the amp and drive unit come as some kind of “matched pair”?
The different waveforms that would appear at the amp output are probably interesting too, as I guess they tell you quite a bit about what the speaker impedance is doing in real time?August 25, 2015 at 3:05 pm #15445
Yes, it’s all very interesting, and the waveform at the amplifier of course gives a perfect picture of how the loudspeaker is performing, but as with so many of these sorts of ideas, the world has moved on in other directions and modern (2015) ‘neo’ loudspeakers can perform remarkably well with voltage drive.
There is also a fallacy that I uncovered a couple of years ago; the current drive is OK and works perfectly except at the point of cone resonance, there, the theory seems to go out of the window and it’s possible to get some strange ‘undamped’ effects. But on balance, it’s still my preferred way to drive a loudspeaker with an analogue amplifier. Unfortunately, it’s not at all easy to implement it with a digital amp!
Nowadays we are using some very fine PWM (so called ‘digital’) amplifiers because they make economic sense….. ah well!
😥September 8, 2015 at 7:20 am #15446 I wonder if this technology of yours would scale to PA applications, Ted? I was at a small festival the other day watching an acoustic jazz group. They had the usual arrangement with a couple of speakers either side of the stage, and as usual, the sound was OK near the sound man, but quite poor off axis. Acoustic bass suffers most, I guess because of phasing issues between the two sides.
Amplifying small acoustic groups in open air seems to be about the worst case scenario of sound reinforcement, and I don’t think I’ve ever heard it sound good. So much so that people pretty much seem to expect it to be bad. I wonder if using a single sound source would help significantly.September 8, 2015 at 8:13 am #15447
Late last year I was talking to the sound man at the Albert Hall during a Mark Knopfler concert. He had finally given up the pretence of trying to get left and right images; the big (and extremely powerful) system was totally mono and he had concentrated on achieving good coverage and fine time alignment from the system. The Albert Hall used to be the most disappointing place to listen to bands, but his system sounded sweet and clean.
In the early days of airSOUND development we carried out many experiments with live shows and the results were spectacularly good; the arrangement works particularly well with quiet acoustic type bands, this is not unexpected as one of the features of airSOUND is that the spatial information is actually quite low in level and can get drowned out by too much reflection of the main signal from walls.
So the answer is yes, it’s an area that requires more research and probably a lot more time, but eventually most sound systems will work the airSOUND way, or at least the designers will understand the physics and be able to overcome the awful inadequacies of 2-speaker systems.
All that was about airSOUND and not about airDRIVE. The amplifier system airDRIVE could become a dead-end
😥The reason is that it makes sense to continue improving efficiency of power amplifiers, and there is no denying that analogue amplifiers are horribly inefficient. The very latest digital amplifiers are both inexpensive and offer superb performance…. and are unsuitable for modification to ‘current drive’. I have spent some time trying, but with limited success….. in addition to that, the physical design of loudspeaker driver that is suitable for current drive is a different beast….. but that’s another story.September 8, 2015 at 8:46 am #15448 well, regardless of the amplifier story – having better live sound at acoustic concerts would be absolutely fantastic! It is one the banes of my life, as a musician AND and as a listener! So many times this has been a huge frustration.September 24, 2015 at 7:54 am #15449 Ted, I’m looking to have a play with these class-T amps (I’m rather behind on this technology) but Tripath were bought by Cirrus, of course. Can I ask you what supplier you have had good results with? I have someone asking me about making some battery amps, and they might be just the job.
DanielSeptember 24, 2015 at 3:52 pm #15450
Can’t help much on this one. The idea of the ‘T-amp’ sounded good but I suspect that they got overtaken by quality improvements by the big suppliers and conventional class-D amplifiers.
We use very ordinary amplifiers in the Orbitsound soundbars….TDA8932 which are inexpensive but handle quite high rail voltages. The weak link is always the loudspeaker driver.
For experiments with current mode, I used LM3886 which is a rugged class A/B amplifier that is relatively easy to convert to current mode.September 29, 2016 at 5:58 pm #15451
To bring this subject up to date…..(End of September 2016)
Here at Orbitsound we have started a project to produce the ‘finest loudspeaker possible’ with little influence from the economics of the parts.
My job has been to investigate and prove the best possible amplifier/driver combination and it might not be surprising that I have finally chosen analogue amplifiers that have 80% current drive capability….. I have developed an amplifier with extremely low distortion and noise that is also ‘current drive’. The acoustic performance totally outclasses any conventional amplifier system, analogue or digital. One of the big challenges was the treatment of the audio drive at frequencies around the resonance of the driver; that was a problem but is now overcome and we have obscenely flat responses and virtually no distortion. How does it sound?? Well, I refuse to get into the business of Hi Fi overblown language, let’s just say it’s superior!September 29, 2016 at 6:11 pm #15452
that sounds extremely interesting Ted! and like a phenomenal design challenge too!March 20, 2018 at 10:32 am #15453 AS a further update (March 2018), We now have transconductance amplifiers (current mode, airDRIVE) in production and actually in stock at major retailers here in the UK!
But, there is no ‘song and dance’ about it yet, we introduced a version of a new design of mine which succeeds in converting a class ‘D’ power amplifier into a transconductance amp. We put this amplifier into the ‘main’ signal path of our Orbitsound P70W loudspeaker, but without telling anyone!
The intention was to see if the Great British Public would spot the difference…. it is quite startling, and to gain experience with these amplifier types in production numbers.
The results are not exactly shattering….. the problem is that our airSound loudspeakers sound pretty good anyway and the improved fidelity, although really noticeable, can’t be discussed openly with the customer or we would give the game away.
So to round it off, what I originally called ‘airDRIVE’ is actually a version of a transconductance amplifier, modified to overcome a couple of foibles of the system, namely a tendancy to some nasty peaks at resonance and a naturally rising HF response. All Orbitsound P70W units have this amplifier.
It’s very likely that we shall introduce these amplifiers generally throughout future product ranges; the advantages are very real.
😮March 20, 2018 at 2:11 pm #15454 A word of caution to anyone interested in the details; the marrying up of a transconductance amplifier with a loudspeaker is far from simple, The acoustic loading of the driver in its enclosure has a significant effect on the instant impedance of the voice-coil; a factor to remember is that sound is produced as a function of acceleration of the cone movements which is a result of changes in the currents in the circuit. all this means that the voltages appearing on the output terminals of the amplifier are not representative of the acoustic output!
On a previous post I inferred that a loudspeaker driver is ‘damped’ by a low amplifier output impedance (damping factor). Just to be clear, this is only the case at or near the resonant frequency of the driver in its enclosure. The requirement for a super-low damping factor is one of the myths of the audio business.
- You must be logged in to reply to this topic.