Jan,
I just noticed the thread so haven't read past the first few pages. The thought or concept that any electrical corrections can remove all of the mechanical phenomena that was just listed is plain silly. You can't electrically correct for these physical actions, that will never happen. I understand the premise of giving the amplifier a flat impedance curve and a conjugate network or networks can do that, all else being the same. I think all these other so called noises in a speaker would have to be attacked in the original speaker design, adhesives, cone material, motor design, etc. need to be looked at from the clean sheet thinking of the design before you ever get to the cross-over design.
I'm trying to come up with a term that could have been used from the beginning of the thread besides noise to describe what is attempting to be corrected. I'd rather call it impedance/phase angle correction, that seems to be more appropriate than "noise".
It would just make sense that an amplifier would be much happier seeing a flat impedance curve and no extremely low impedance dips. Sort of like driving a power resister!
I just noticed the thread so haven't read past the first few pages. The thought or concept that any electrical corrections can remove all of the mechanical phenomena that was just listed is plain silly. You can't electrically correct for these physical actions, that will never happen. I understand the premise of giving the amplifier a flat impedance curve and a conjugate network or networks can do that, all else being the same. I think all these other so called noises in a speaker would have to be attacked in the original speaker design, adhesives, cone material, motor design, etc. need to be looked at from the clean sheet thinking of the design before you ever get to the cross-over design.
I'm trying to come up with a term that could have been used from the beginning of the thread besides noise to describe what is attempting to be corrected. I'd rather call it impedance/phase angle correction, that seems to be more appropriate than "noise".
It would just make sense that an amplifier would be much happier seeing a flat impedance curve and no extremely low impedance dips. Sort of like driving a power resister!
Last edited:
I appreciate all that. I was trying to cut through the fog. In the beginning of the thread Joe R explained that he had developed an 'improved crossover' to combat 'noise' in 'normal' crossovers. (This is not a quote but I believe that was the gist of it).
However, neither Joe R nor anybody else was able or willing to explain what this 'noise' was. The thread then derailed, the reasons for that are not important right now. If you can stomach it you can read it.
So now this posts comes up with a list of these 'noise' issues. The list itself has no new elements, serious speaker designers are well aware of them.
But I really wanted to know if this was the answer to the question which 'noise' issues the 'improved crossover' was going to solve.
I do agree with you that 'noise' was a non-sense term in this context, and I can only speculate why it was used, why nobody explained it, and why nobody said sorry guys, bad term, I really mean xyz. Unfortunately it is still used in the post above. Why??
Your suggestion that the improved crossover would seek to flatten the impedance curve makes sense.
Jan
However, neither Joe R nor anybody else was able or willing to explain what this 'noise' was. The thread then derailed, the reasons for that are not important right now. If you can stomach it you can read it.
So now this posts comes up with a list of these 'noise' issues. The list itself has no new elements, serious speaker designers are well aware of them.
But I really wanted to know if this was the answer to the question which 'noise' issues the 'improved crossover' was going to solve.
I do agree with you that 'noise' was a non-sense term in this context, and I can only speculate why it was used, why nobody explained it, and why nobody said sorry guys, bad term, I really mean xyz. Unfortunately it is still used in the post above. Why??
Your suggestion that the improved crossover would seek to flatten the impedance curve makes sense.
Jan
Last edited:
Jan,
Were on the same page. I'll have to go back and read through more of the thread. Sy questioned very early the use of the term noise to describe anything to do with a cross=over in the first couple of responses.
I am aware of the concept of driving the speaker with a current source, that seems to have been lost somewhere in translation here.
What I have never been able to comprehend is what changes in the speaker device itself is usually being talked about. The only true answer I have ever been able to see that would make this realizable would be to increase the voicecoil impedance much higher than a standard speaker, say something like 100 ohms or such on the voicecoil. I don't see any other mechanical changes that would satisfy that notion in the speaker itself.
Were on the same page. I'll have to go back and read through more of the thread. Sy questioned very early the use of the term noise to describe anything to do with a cross=over in the first couple of responses.
I am aware of the concept of driving the speaker with a current source, that seems to have been lost somewhere in translation here.
What I have never been able to comprehend is what changes in the speaker device itself is usually being talked about. The only true answer I have ever been able to see that would make this realizable would be to increase the voicecoil impedance much higher than a standard speaker, say something like 100 ohms or such on the voicecoil. I don't see any other mechanical changes that would satisfy that notion in the speaker itself.
Last edited:
Inductor,
First, this copy/paste is way out of topic. X-over is the topic. What you copy/paste is irrelevant.
Second, the above source is very hard to be quoted here as technical reference, even if the source is a book. Pay attention that the book did not present a scientific study of the author nor is quoting any sort of sources at this level in these directions (what you selected for copy/paste).
The book has its value. But it is improper - in my opinion - to be used as a technical reference for many topics. Actually is very dangerous and detrimental to education, yours and DIY Forum in general, if you use it for such purpose. My recommendation is to use this book with some discerning thinking, as much as possible.
Best regards,
Neither do I. To be honest, I'm not sure even that changes to the driver were contemplated at all, other than the desire to throw interesting-sounding multi-syllable words around 😉
Jan
Jan,
What I am seeing is that the current gets dumped into the crossover with the series and shunt resistances, not sure how that helps the speaker hanging on the end of the chain? It truly is that voicecoil moving in the gap where you would want to have that series resistance and an even series resistance at that. Not many speaker designs could ever be considered to have a flat series resistance in the voicecoil itself. The flux modulation in most speakers would seem to discount that notion.
What I am seeing is that the current gets dumped into the crossover with the series and shunt resistances, not sure how that helps the speaker hanging on the end of the chain? It truly is that voicecoil moving in the gap where you would want to have that series resistance and an even series resistance at that. Not many speaker designs could ever be considered to have a flat series resistance in the voicecoil itself. The flux modulation in most speakers would seem to discount that notion.
Well a flat impedance response is not supposed to help the speaker, is it? It is supposed to help the amplifier. Not sure it does, though.
If we hang on to that thought, let's look at some scenarios:
1 - the speaker has a wildly varying impedance response, but the amplifier is well able to cope with the minimum value. I do not see any value in flattening out the rest of the response, as it would only require more current from the amp at frequencies where it otherwise would not.
2 - Same wildly varying response but the amp is not able to cope with the minimum value. You have a choice of getting a better amp or inserting some sort of series impedance between the amp and speaker, for some part of the spectrum where the dip occurs. I'm not a speaker expert but seem te remember that in most speakers the minimum impedance is anywhere between a few 10's and a few 100 Hz. Inserting a series impedance there also degrades the damping factor.
I know that damping factor is a very contentious issue and I have no intention to get buried into that one. But be that as it may, it is trivially easy to show that a series impedance between amp and speaker change the frequency response of the combination. Hard to predict how in detail, but in general it will increase the response ripples that are already present in the speaker.
So I guess I my suggestion in scenario 2 would be: get a better amp 🙂 and then I see no advantage in flattening the impedance response.
But, as always, YMMV.
Jan
For a commercial product, it could help to have the amp source impedance not cause a frequency response change. Trying to engineer out a big impedance dip (I'm thinking of Wilson) by impedance compensation networks is doomed to failure, but if a crossover starts out being reasonable, then flattening the impedance curve with compensation to the minimum won't do much harm, other than wasting a bit of power. That's the trade-off. Having that flat impedance means that you can drop in a current source amp, a voltage source amp, and anything in between and get the same frequency response.
From the driver point of view, I don't see this as any sort of help. It may reduce the umbrella effect (elucidated by Don Morrison a few years back), which is analogous to crossover noise.
From the driver point of view, I don't see this as any sort of help. It may reduce the umbrella effect (elucidated by Don Morrison a few years back), which is analogous to crossover noise.
It is supposed to help the resulted sound. Does it not, with flat impedance load? Let's see.
Foremost to say that I do not see your reasons behind dismissing spectral regions of High impedances. Why - do they sound better? Not at all. Or else said, at minimum load with good amplifier: does it help to have more curent in voicecoil? Again, sure No. Resume a bit: With VFB a High-Z load imposes a low current and a low-Z imposes a high current, this is known as trivial. But also trivial is to understand that SPL wise you do not obtain a flat response on VFB. Do we agree on this for a start?
Second, my statement would be that any local(!) low_Z (or high-Z) region in spectrum calls for phase inversions. This is the very nature of resonances, either parallel or series, damped or not. So besides the local highs or lows of SPL we will have to accept bunching and decoherence too.
Well anyone can try to experiment and reconstruct the audio program after X-over, just for fun. Those equipped with hardware will be less successful than those equipped with educated ears (because brain is a better computer and can handle much larger correlation times). But this is relative. There is a point where even our brain fails to correctly interpret the sound as having "musical meaning". Does not correlate with any instrument, strings or anything, the musical meaning is lost.
Fazit: low correlation time "typical electronic noises" (="white" thermal, "pink" shot and many many other types) + middle/long correlation time "typical ambiental noises" (from LP scratch to birds, wind and neighbors) are good examples of accepted noise sources.
But noise can be music too. When? If author (composer/interpret) wants this!
So, I would propose the definition of audio program noise to be something like: whatever "sounds" that the composer did not intended.
This definition calls to extend the above list of noise sources: a) speaker noises (aka. motor resonances and all troubles they create for you to understand the program), b) HD + ID etc, and.... tam tam... c) X-overs induced "noises" (aka.- sounds that have lost their meanings and correlations for our brains.
I invite all interested for a debate to report here public their square wave response after a multi-way X-over (aka not RMS SPL spectrum but the actual waveforms after summing up the N-way outputs of X-over). Here it is meaningful to see the waveforms before the speaker!
Therefore I partially agree with Joe seeing in that graph that speaker contributions are eliminated, indeed only for the amplifier. This we have seen above: it has advantages amplitude-, delay- and phase-wise!
Still, it does not prove the acoustic advantages. So I agree with critics too. But we must remember to read well: this was not the stated goal in the first place. The KL reference puzzled me too a bit; but not because it does not exist (oh IT DOES... and I know that being a microwave educated). I would choose to explain it via resonances: stored energy is build up and released with inverse or normal exponential time functions, both linear in Q-factor parameter. I mean Q-factor of each resonance processes that we have in a speaker and which is able to effect the musicprogram energy with substantial contributions which are frequency and power dependent (e.g. stirring of ferrofluid around coil-former is maybe relevant to HT at 40Khz, but not at 12KHz). ETC.
I am not sure if a flat-Z X-over is efficient to help sinking the internal "noises" "from the driver", but in my opinion it MIGHT acoustically help by providing a much cleaner path for the signal program from amp "to the driver".
It might. Without exact TS parameters the schematics of X-over are absolutely useless.
Thanks.
Last edited:
I guess the important factor that comes out of this is not to design a speaker and crossover design that dips extremely low in impedance and most amplifiers are going to work well. I don't understand why anyone would build a speaker system that dips as low as 2 ohms impedance, that just seems to be a poor design chose. It seems as long as we use something practical that doesn't dip below 4 ohms we should be fine and have many fine amplifiers to chose from. Yes I know some amplifiers can function as low as 2 ohms but why do it, to what advantage would that ever be?
The cynic in me says that it allows the speaker maker to claim that the speaker is particularly revealing of amplifier differences.
Ionmw,
My basic work is on the device itself. I am working to reduce the distortions from the speaker itself, before any corrections are needed from any electronics. Making a speaker is a series of trade offs, but that doesn't mean you shouldn't get things as good as you can. There are many aspects of the design that can be improved. The fact is that most commodity speakers just don't address those points, that cost money and means you don't use the most cost effective, meaning mass produced parts. Whether that is using higher grade magnets, cone materials or even hard to apply adhesives all these things increase the cost of a commodity speaker, that is the bottom line in 99.9% of all speakers produced in the world. We are in a rarefied part of the industry where we are not controlled by costs and the bottom line while installing millions of speakers whether into a TV set or a cell phone or most any other device that uses a speaker.
Sy,
I assume you are pointing to that extremely low impedance in the Wilson speakers. They have always done some strange things for no other reason than to make you spend more in the audiophile market to get what they market as the best. I have never been a fan of anything they have done beyond making boatloads of money doing it.
My basic work is on the device itself. I am working to reduce the distortions from the speaker itself, before any corrections are needed from any electronics. Making a speaker is a series of trade offs, but that doesn't mean you shouldn't get things as good as you can. There are many aspects of the design that can be improved. The fact is that most commodity speakers just don't address those points, that cost money and means you don't use the most cost effective, meaning mass produced parts. Whether that is using higher grade magnets, cone materials or even hard to apply adhesives all these things increase the cost of a commodity speaker, that is the bottom line in 99.9% of all speakers produced in the world. We are in a rarefied part of the industry where we are not controlled by costs and the bottom line while installing millions of speakers whether into a TV set or a cell phone or most any other device that uses a speaker.
Sy,
I assume you are pointing to that extremely low impedance in the Wilson speakers. They have always done some strange things for no other reason than to make you spend more in the audiophile market to get what they market as the best. I have never been a fan of anything they have done beyond making boatloads of money doing it.
Last edited:
To be cynical as well: it means "you say", right?
Seeing your web site invites all to see that you are purist DIYer and also inventor. Respect from my side.
But I choose to read better and respect any work if well done within purpose, especially if shared fully. I still wait 7days (read-only ban time of Joe) to see the TS parameters. If he gives them and if my analysis align with his, I have no choice but to indulge his own style and finally to send him my respects.
Best regards
I hope you are not manufacturing 15k speaker cables 🙂
Myself a lucky one too, at work and personal business (both have 0% collision with audio).
But I am not different as anyone here: DIY is really a personal business and one can choose to spend his time transforming a 99.9% cheapo into a 99kEur treasure.
...Or the amplifier, or the X-over for the sake of topic.
Real value lies in work and education. Buying the best of the best does not get you there. I know this for sure but of course your opinion is as important as mine.
best regards,
Ionmw,
No I am working on a self powered speaker for the consumer market that if it was an audiophile product I couldn't and wouldn't pay for myself. I am attempting to bring a very high quality product to market without the high dollar cost Not an easy thing to do. I've been down the audiophile market approach before and the lesson learned was that I wasn't trying to charge enough and that didn't sit well with the market sellers at the time. Marketing costs and budget for that were more important than the actual design, that still seems to hold true today. With what I am doing all you would need besides my speakers would be a source holding your music. Plug and play basically.
No I am working on a self powered speaker for the consumer market that if it was an audiophile product I couldn't and wouldn't pay for myself. I am attempting to bring a very high quality product to market without the high dollar cost Not an easy thing to do. I've been down the audiophile market approach before and the lesson learned was that I wasn't trying to charge enough and that didn't sit well with the market sellers at the time. Marketing costs and budget for that were more important than the actual design, that still seems to hold true today. With what I am doing all you would need besides my speakers would be a source holding your music. Plug and play basically.
Sadly True. Charge more next time.
Unless you invented the speaker-headphone you will need help from room acoustic too.
To be short: I want to stick on-topic - perhaps you care to share a set of T-S and your best choice of X-over for them (full details) for 2-3 way? This will explain your position more than 1000 words. I understand the info to protect is component related (speakers) so sharing T-S and X-over will not be disclosures or risks of any kind. It will be just a kind of analytical play on-topic, or even reciprocal help.
SY: Indeed, extremely good speakers are insensitive to amplifier, which 99.99% of time is also very good I suspect. But from physics I know this comes with ... lousy efficiency. The more perfect the speaker is the less the efficiency. Else, it will come down to tradeoffs and therefore by necessity of physics laws a difference between driving by CFB or VFB will exist.
Last edited:
I fully agree, but I think there is also a marketing slant to it. Expensive items that are difficult to use and/or setup do have an exclusivity air around them which may appeal to some. In the high-end industry, very little happens by accident!
jan
Ionmw,
I'd have to dig out some paperwork on the T/S values. At the same time that really wouldn't say much as the devices are specific to the design, they are not something you could place into another box. they are designed as frameless, the enclosure creates the mounting surfaces for the surround and the spider and hangs the motor. I am also going to be using a dsp/dac setup so it will be specific to the enclosure box parameters. The cone itself is not round, it is an elliptical shaped cone and will require a custom designed surround to work with it. I have done the original design work with round metal frames and used a wood enclosure to do the initial work, that would be a very different animal. I may offer those to the diy crowd, they do have some strange T/S values though, it is not a traditional motor design. As you just said to Sy I went for best audio quality and gave up much efficiency to do that, chose your poison they say. This is an extreme long excursion design with a very long gap and short coil. You can't do that and have high efficiency. I'll be supplying the power for the devices so I can match the power requirements without worrying about running out of usable output. My aim is a max of 107db with very low distortion, not that anyone needs that output in a normal room!
I'd have to dig out some paperwork on the T/S values. At the same time that really wouldn't say much as the devices are specific to the design, they are not something you could place into another box. they are designed as frameless, the enclosure creates the mounting surfaces for the surround and the spider and hangs the motor. I am also going to be using a dsp/dac setup so it will be specific to the enclosure box parameters. The cone itself is not round, it is an elliptical shaped cone and will require a custom designed surround to work with it. I have done the original design work with round metal frames and used a wood enclosure to do the initial work, that would be a very different animal. I may offer those to the diy crowd, they do have some strange T/S values though, it is not a traditional motor design. As you just said to Sy I went for best audio quality and gave up much efficiency to do that, chose your poison they say. This is an extreme long excursion design with a very long gap and short coil. You can't do that and have high efficiency. I'll be supplying the power for the devices so I can match the power requirements without worrying about running out of usable output. My aim is a max of 107db with very low distortion, not that anyone needs that output in a normal room!
I would say SY was right on the nail. If you look at the number of manufacturers whose success was based on some ornery requirements vs well engineered and sensible products that just fizzled out there is some correlation.
I have an irrational fondness for the designs of the late Jason Bloom. But there are a couple of his products I would only give house space to just for the masochistic challenge of trying to drive them.
Now in the 70s there was maybe some excuse for some of the insane products produced as people were pushing the limits of what could be done with the technology of the time. These days its all about pushing high end into a 'lifestyle' market which has created boutique audio with price rises way above inflation.
The good news is that, if you can ignore the fairy dust and loony beliefs you can have a high end system without spending a fortune.
Your system does not qualify for normal analytical play🙂 But if you like another set of drivers+Xovers Nway, of standard type, please look up for TS, impedance plots (phase too) and send with Xover schematic.
Success with your RnD. It seems well thought to me, technically.
Long gap (ferrites, a la 70 revised) seems a first good choice for performance, much more homogeneous field and less temperature dependence. High impedance voice coil (100ohm is not a bit too high though, amp wise??) It calls for small L, small coil, high-tech custom membrane, suspensions and serious know-how. Non circular long excursion for reasonable bass and resonance control seems reasonable choice. But it seems expensive to produce. 107dB seems your up limit, but is it enough for market?
Do you know your market?Rest of 95% is details. Good luck and PM me when you need.
Last edited:
- Status
- Not open for further replies.