Deconvolve etc
@ RNMarsh
Have you seen/tried this ? Impulse response deconvolution tool software - Voxengo Deconvolver - Voxengo
Also you might be interested in this/these ScopeFIR: FIR Filter Design Software for Windows | Iowegian International Download Digital Filter Design and FFT Signal Analysis Software for Windows | Iowegian International
@ RNMarsh
Have you seen/tried this ? Impulse response deconvolution tool software - Voxengo Deconvolver - Voxengo
Also you might be interested in this/these ScopeFIR: FIR Filter Design Software for Windows | Iowegian International Download Digital Filter Design and FFT Signal Analysis Software for Windows | Iowegian International
I agree with this.
Everybody will take care with the power amps because they know the initial ratings were given for 25 deg C and after an hour, this is long gone, the amp has heated up inside and the power stage is God knows where in temperature terms.
Yet, so very few people even ask what about the loudspeakers? How do they change when heated up? How does this reflect on their nominal efficiency? They don't ask this even after they have seen ads, as I have, from Britain's B&W and France's Focal(JM Lab, clearly stating that nominal impedance was 8 Ohms, but minimum impedance was 3.1 and 3.3 Ohms respectively. Great! So what happens there when the speakers heat up? Are we now asking our amps to work into sub 3 Ohm impedances, which may well correspond with possibly large negative phase shifts, which will draw yet more current?
Considering not at all rare -60 degree phase shifts, we are asking for double the current, or from another perspective, the amp sees the real world impedance dip halved. 3.1 : 2 = 1.55 Ohms. What if it drops down further still when heated up?
All other things being equal, our only possible rdeeming value might be the speaker's efficiency, which will set the base point. If that's high enough, then some loss can be tolerated as they will draw less power from the start, but what if not? What if we start out with a modest 88 dB/2,83V/1m, which then lose say 3 dB effcicency at high SPLs, drop down to 3 Ohms normally and this coincides with a -60 degree phase shift?
And please, don't anyone tell me I'm a fatalist, this is in fact what did happen with some legendary speakers, such as AR 3a Improved, or Yamaha NS 1000. And I'm an old AR (Acoustic Research) man ...
Dare I even mention Apogees? Or Wayne's NOMINAL 1 Ohm speakers?As for the old and new standards Simon mentioned, if memory serves, IHF standards specify that an amp should be ran for 30 minutes at 1/3 of its nominal power before measuring its nominal maximum power. I seriously wonder how many of the current crop would even survive the first 30 minutes in these days of cutting costs?
Good points, Simon.
Perhaps not perfectly, but near enough.
For example, the reVox 740 power amp is rated at 100/185W into 8/4 Ohms (aka Studer A68 amp). Not precisely perfect, but near enough. I think the basic point is that it will do just fine driving 4 Ohm speakers.
Or, our own Pavel Macura (PMA here) rates his amp at 150/300W into 8/4 Ohms on his site. I don't know Pavel personally, but I have heard his work and know that he's a man who knows what he's talking about in terms of good sound. I have no reason to doubt him or his power ratings.
Of course, there's always the option that the power output lie applies to the 8 Ohm rating. So when someoney says his amp does 100/200W into 8/4 Ohms, it is in fact doing more like 115W/8 Ohms to satisfy the 200W/4 Ohm rating.
No Dan, I do not. I am suggesting that we don't KNOW what happens in a speaker when it has been pressed hard, or harder, for say 30 minutes, not until we have measured that specific speaker.
There are far too many variables working together for anyone to know what realy goes on until he measures it.
The impedance does not have to drop, but the negative phase shift may increase, which in turn effectively reduces impedance from the amp's point of view.
Can anyone here swear on his life that this does not happen?
It would seem the amplifiers Ed has to choose from for commercial installations have a different set of goals than DIY / Home audio reproduction.
Brands like Crown, Behringer, QSC, Cost - watt per dollar. Efficiency - operating cost, size - portability and travel expense, fibre optic signal distribution, digital control of phase in arrays, time delay, pattern control, crossovers, etc
switch mode power supplies, class D, amplifiers, surface mount construction
For $400 to $700 in 3 and 1/2 inches of height and 12 pounds you can get 2 x 2,000 Watts into 2 Ohms; 2 x 1,400 Watts into 4 Ohms; 4,000 Watts into 4 Ohms (bridge mode). Have you listened to these in the home audio system? Are we doomed to follow? ........ Nelson won't
Brands like Crown, Behringer, QSC, Cost - watt per dollar. Efficiency - operating cost, size - portability and travel expense, fibre optic signal distribution, digital control of phase in arrays, time delay, pattern control, crossovers, etc
switch mode power supplies, class D, amplifiers, surface mount construction
For $400 to $700 in 3 and 1/2 inches of height and 12 pounds you can get 2 x 2,000 Watts into 2 Ohms; 2 x 1,400 Watts into 4 Ohms; 4,000 Watts into 4 Ohms (bridge mode). Have you listened to these in the home audio system? Are we doomed to follow? ........ Nelson won't
Last edited:
Thank you.... I will read them and then see what I can do with the downloads. The idea to use all thier affects software means they can deconvolve accurately.
THx-RNMarsh
Last edited:
Hmmm, as far as I understand, hotter voice coils means increased VC DC resistance.
Negative phase shift cannot come from increased VC DC resistance, but, ok, what happens to suspensions and magnets with increased temperature.
Pro audio driver magnet assemblies can get too hot to touch with prolonged usage at high levels !.
Dan.
Hmmm, as far as I understand, hotter voice coils means increased VC DC resistance.
Negative phase shift cannot come from increased VC DC resistance, but, ok, what happens to suspensions and magnets with increased temperature.
Pro audio driver magnet assemblies can get too hot to touch with prolonged usage at high levels !.
Dan.
Dan, again, I suggest nothing but our ignorance at what's happening inside a home speaker once it is pushed hard.
I'm pleased to note that when I pushed my own speakers hard, the overall sound balance did not change at all. On the other hand, the amp peaked at around 70 W, which is a lot harder than usually, but is not too hard, given the driver rating at 150W nominal. And it lasted for some 20-25 minutes or so, I don't know what would have happened if that had been 1-2 hours. Or if I had pushed them harder still. The peak warning LEDs never once showed any sign of life.
I didn't because the in room SPL was already very uncomfortably high, and my window panes were beginning to rattle (big and 52 years old, about to be exchanged).
That is also my experience - at first I was quite surprised that speaker drivers could be 'pushed so hard' without any ill effect; but now I'm comfortably comfortable with the fact that even cheap and nasty drivers can be run at "deafening" levels almost indefinitely without any problem, provided the electricals driving them are completely clean in behaviour - my ears give up well before any system problems arise ...
Hey Jan, Your post was taken at face value, but I will observe that the post you responded to (which I missed) stated what I was thinking so I wasn't too far off base in responding. After the metaphysics response I realized that this is not the forum or audience.
Mike
Mike, your verbage seems to agree with jan's, from what I can see.
I agree.
I do not believe he goes far enough.
When I worked, tested, and troubleshot bare IC dice that were mounted into mil hybrids, I had to contend with the fact that they all are created (analog at the time) with one supply rail as the substrate. This is evident with power IC, where the backside is negative voltage (like the LM 3886). So, rail bounce was not insignificant. Even with DI tubs like silicon nitride (IIRC) isolating the individual elements from the substrate (as opposed to reverse biased PN junctions), there is still capacitance to deal with.
So while my goal is to "diss" JC whenever possible
, whenever he talks disparagingly about IC's vs discrete, I do wonder about the consequences of the entire gain circuit being built within microns of a supply rail. We have enough issues to deal with without adding the rail proximity into the equation, and the consequence of that rail pulling current and upsetting the entire circuit ground potential.In this case, my experience tends to lean towards JC's assertions. Sigh, such is life..
I recall reading maybe 20 or 30 years ago that the iron tracks the VC at about 30 degrees C behind it.Hmmm, as far as I understand, hotter voice coils means increased VC DC resistance.
Negative phase shift cannot come from increased VC DC resistance, but, ok, what happens to suspensions and magnets with increased temperature.
Pro audio driver magnet assemblies can get too hot to touch with prolonged usage at high levels !.
Dan.
I also suspect the iron's eddy currents will drop with temperature, and that the hysteresis curve will be different at temp. Suspicions of course, I've not examined data on the stuff I work on, our goal is to maintain a constant fixed temp. Not like me pushin Led Z into my speaks where the goal is maximum temp.
jn
I think its a strawman to claim that IC/op amp circuits have to be limited by what a single chip can achieve
Scott and Walt published multiloops with AD811 output just a bit ago - isn't a 1/4 century enough for our gurus to take in the concept that you just might have to change your topology to fit the devices strengths, weaknesses
of course augmenting op amps with discretes - inputs, output buffer, Vgain stages is as old as op amps - predates monolithic op amps
to see modern monolithic performance read the TPA6120 datasheet - lots of audio plots - you can see likely thermal effects in the 10s of ppm bumps in some plots otherwise single digit ppm distortion floor - while the chip is driving over a watt into 32 Ohm loads
Scott and Walt published multiloops with AD811 output just a bit ago - isn't a 1/4 century enough for our gurus to take in the concept that you just might have to change your topology to fit the devices strengths, weaknesses
of course augmenting op amps with discretes - inputs, output buffer, Vgain stages is as old as op amps - predates monolithic op amps
to see modern monolithic performance read the TPA6120 datasheet - lots of audio plots - you can see likely thermal effects in the 10s of ppm bumps in some plots otherwise single digit ppm distortion floor - while the chip is driving over a watt into 32 Ohm loads
Last edited:
Power audio ICs have their place in the market. I doubt that there are any technical issues that prevent a good design from performing very well, or that cause sonic issues because of the fabrication technology. I've certainly never heard of any bounce issues other than in high speed logic, or in switching power mosfets ( modern trench devices are quite capable of switching 10 A in 3 to 5 ns) - you just don't see these problems at audio where the bandwidths and rise times are orders of magnitude narrower and slower.
I have not stated that.
That timeframe should indeed. It does require understanding of any issue that may be present, but once identified, yes indeed the big guys can certainly attack it.
My example was just one specific part, and I chose that because anybody who used the non-insulated version can trivially measure v minus to mounting tab to verify that the chip substrate is indeed the negative rail.
While I haven't worked bare dice in two decades, I suspect the bulk of the IC's out there still have substrate as negative rail. Perhaps scott can elab.
While I used the term bounce, I understand it's a digital word for the most part. I worry about all aspects of the designs I create, as I'm kinda "nitpicking". Heck, I'd play with wirebond dress and layout if I had the option.
As I said, from my experience testing and troubleshooting the bare dice, I've noted my concerns. The fact that my concerns happen to seemingly support the general statements of JC, well I can't help that.. I will not censure my opinons because they support him.
edit: I also look forward to his IC tests. It may yet be that he finds things unexpected (or not), and that it may just be layout issues that bite him. Recall when he reported seeing wire distortion years ago??? That led me down the ground loop/EMC path a decade ago, and now look at me.. I can spell it..
jn
Last edited:
. . . I don't use 'em. Never will. But, I'll sell 'em
Sony, Panasonic and most of the lo-fi flashing light boom boxes use IC's. There were a few really notable devices out in the 90's from Philips that had ultra low distortion etc but that's all gone now. Almost everyone is onto class D.
I remember doing a 'sound taste' test with the S company in Shinagawa about 5 years ago in their sound room. Place had ML, the B&W Nautalis and an array of very high end gear from them and other Japanese and a Western high end companies. We all duly sat down in this room on the 23rd floor and they started to play this boom box - my god it was terrible. Then they switched it to another amp board. Just as terrible. Lots of nodding and breath being drawn through the teeth. Then onto the third amp board. Same story again. By this time I was numb and completely ear bashed with this class D crap. After some discussion, they declared board number 2 the winner. I could hear no difference other than it was crap.
But, they were making 15 million of those things a year . . .
Sony, Panasonic and most of the lo-fi flashing light boom boxes use IC's. There were a few really notable devices out in the 90's from Philips that had ultra low distortion etc but that's all gone now. Almost everyone is onto class D.
I remember doing a 'sound taste' test with the S company in Shinagawa about 5 years ago in their sound room. Place had ML, the B&W Nautalis and an array of very high end gear from them and other Japanese and a Western high end companies. We all duly sat down in this room on the 23rd floor and they started to play this boom box - my god it was terrible. Then they switched it to another amp board. Just as terrible. Lots of nodding and breath being drawn through the teeth. Then onto the third amp board. Same story again. By this time I was numb and completely ear bashed with this class D crap. After some discussion, they declared board number 2 the winner. I could hear no difference other than it was crap.
But, they were making 15 million of those things a year . . .
I am curious regarding this, many of the new packages are bottom terminated with the die mounted onto a copper interposter (thermal pad) that is usually tied to ground. These newer packages allow much better control of the device temperature, with the correct layout and use of thermal vias etc.
- Status
- Not open for further replies.