To be frank, there is nothing to write about at this moment. Everything is in Rob's papers and on MFBLABS. Some trickery we can't publish because it is also used in the Grimm sub. At this moment it's refining and getting the sensor in mass production at a reasonable price.
But, there is a very big BUT. Loopshaping with high feedback levels is tricky. To prevent a total support meltdown, loopshaping will not be available for the end user. External parties will supply drivers ( and plates) with the sensor mounted with a dsp file for a given volume and are responsible for support.
Launch? Munich 2021
But, there is a very big BUT. Loopshaping with high feedback levels is tricky. To prevent a total support meltdown, loopshaping will not be available for the end user. External parties will supply drivers ( and plates) with the sensor mounted with a dsp file for a given volume and are responsible for support.
Launch? Munich 2021
Sounds like OEM, not for end-users like us. Which, sadly, makes good sense for Hypex.
Feedback systems are capable of destroying drivers if something goes amiss with the phase. So Hypex would ask you to sign a liability disclaimer longer than the ones that come with software apps.
But... I wonder if there are design solutions, DSP or old-tyme analog, to prevent such catastrophes? Other than using a 10 watt amp on a 50 watt driver. Or fused for 3 watts on a 50 watt driver.
B.
Feedback systems are capable of destroying drivers if something goes amiss with the phase. So Hypex would ask you to sign a liability disclaimer longer than the ones that come with software apps.
But... I wonder if there are design solutions, DSP or old-tyme analog, to prevent such catastrophes? Other than using a 10 watt amp on a 50 watt driver. Or fused for 3 watts on a 50 watt driver.
B.
The issue for me is that objectionable noise from sub bass systems is primarily non harmonic distortion which these MFB systems don't help with and could potentially make worse if the plant dynamics change at the limits of operation cause loop instability. Unless especially space constrained the cost of the system would be better put into more driver area for the DIY user. MFB systems are however are coming for mobile phone speakers (one of the very space constrained systems) and are deployed in some high end subwoofer PA systems as part of the IPAL system (although this uses differential pressure not motion directly).
Maybe a (re)read of the papers published by Rob. Your first sentence clearly states that you missed something...
Nice but all turn-key ready not DIY. Of course, the box has some influence on cone motion unless a sealed box is specified in the agreement. And thinking of the worst, that kit may not prevent reverse wiring of the drivers or sensors, unless it is all tied together like a plate amp.
Yes, strange what kipman725 posted.
B.
B.
Yes, turnkey product makes sense but what about DIY options.
Speaking for DIYers, buying a la carte: a driver w/sensor, just sensor, just electronics, etc.
That way Hypex does not have responsibility for how the DIY person screws up... and no obligation to offer advice to individuals.
B.
You don't have a clue how "Klippel Controlled Sound" is working. Please stop embarrassing yourself.
Yes, see Fig. 2 on the right side, next to the loudspeaker:
Klippel Controlled Sound (KCS) - Controlled Sound Technology for Nonlinear Compensation of Loudspeakers | audioXpress
Actually that is a very similar scheme as Mr. Erath used in his patent many moons ago.
Once I finish mine I will see if they would allow a comparison to see just how similar they are.
Kind of like the diagrams telephone systems engineers prepared, intended to clarify for you as little as possible and confuse you as much as possible. Except telephone engineers weren't motivated by making money in doing so (I think their purpose was to prevent amateurs from fooling with phone circuits in their houses).
Looks like the good old current sensing resistor in series with the driver. On the other hand, MFB systems to not have boxes called "Memory" or need transducer models. The feedback circuit is the model, not the imperfect driver.
So conceptually, the diagram could be almost anything from a fancy T/S Klippel model inverted into the signal path (my previous post) or plain old current sensing MFB based on driver back-EMF (and no independent sensor at all).... which most aficionados today find a pretty poor way to do MFB.
My best guess is that it is current feedback corrected by a model of the driver. Which is an inflated belief that Klippel really has a really perfect knowledge of the driver and uses this really perfect knowledge to "correct" the sensor feedback.
B.
Last edited:
The diagram is self-explanatory, but only if one has enough engineering knowledge. Otherwise... well, someone will be confused. Don't blame engineers for that.
Looks like you didn't notice the voltage sensing arm in the diagram, which together with the current sensing arm can track every change of woofer impedance - correcting, for example, the woofer resonant frequency change due to ageing, or voice coil temperature change.
The diagram is a simple explanation of the quite complex Klippel system. To grasp the whole picture, don't guess - read, watch and hear carefully everything here:
KLIPPEL CONTROLLED SOUND
There are many ways to skin the cat, some are more effective than those of most aficionados.
Hearing is believing, hear yourself what can Klippel Speaker Control do:
404 Not Found
Last edited:
I wish some new wise person would add their post here.
BTW, I think it really is the same old current sensing system that Erath, ACE bass, and almost everybody before cheap accelerometers started being used plus, and here I am guessing, some trick parametric correction from Klippel based on prior testing of the driver.
The voltage sensing circuit is the same as in every low output impedance audio amp in the entire known universe (as far as I can understand his inspirational picture). It is feedback that corrects the amp (and not incidentally, to render the amp a low output impedance source of power). It doesn't sense anything about the driver. In fact, as it is widely understood, its role is to nullify any fluctuations in the speaker such as back-EMF, inductance, hitting max excursion, or even noisy dirt in the voice coil.
Klippel's diagram - and other stuff of his I've read - is much distorted by commercial motivation and the need to fool the patent office into thinking you are innovative. But Klippel doesn't blush as much.
I think your personal insults are out of place here. Please stop it.
B.
Last edited:
Thanks. Cap. 4 does explain everything (I've been re-reading that paper every once in a while for a long time and now understand 10% on a good day), including Klippel's mysterious interest in sensing temperature.
Yes, Klippel pre-distortion based on - I assume - prior measurement of the specific driver unit might be a good concept. But as Munnig Schmidt points out, if you need a sensor to sense instantaneous changes ("adaptive learning"), why not just use a sensor for MFB and be done. Why doesn't Klippel simply install an accelerometer? Can't patent it?
Feedback cycle "settling time" is a not an issue with subs unless you want to have DSP* in your loop.
B.
*I bought a spare Behringer DCX2496 to put in the loop to experiments but it seems to be too slow.
Last edited:
There is of course pros and cons with accelerometer feedback.
Fiddling and removing dustcap could be hard if you are not steady on your hand.
But when it´s properly and stiff installed, measuring the voice coil movement, and you decrease distortion with 15 dB and you can tune fs and Qts with a knob, thats amazing!
I can´t really find any reason for a DIY audio soul not to try.
There may be other ways to reduce distortion, but sensor feedback wins, if the sensor is good enough. ACH-01 works like a sharm...
Fiddling and removing dustcap could be hard if you are not steady on your hand.
But when it´s properly and stiff installed, measuring the voice coil movement, and you decrease distortion with 15 dB and you can tune fs and Qts with a knob, thats amazing!
I can´t really find any reason for a DIY audio soul not to try.
There may be other ways to reduce distortion, but sensor feedback wins, if the sensor is good enough. ACH-01 works like a sharm...