Thanks for that. I think it's important to compare Fig 10 and 11.
Fig 10 shows distortion levels at 100 db and Fig 11 shows distortion levels at 109.6 db. At the higher power level the distortion level is described as "unusable", so this is the max that this system can put out.
The distortion level of the 2nd harmonic rises from 37.2 in Fig 10 to 67 percent in Fig 11. (Untreated by push pull or MF.)
The distortion level of the 3rd harmonic rises from 12 to 26.9 percent.
If there was another chart showing distortion levels at 90 db, the distortion levels would be VERY low, as the distortion is ramping up at exponential rates with increased power level.
Does push pull and MF help this situation? Sure it does. They claim they can drop that 67 percent distortion down to 10 percent.
You know what else helps this situation? Don't abuse your drivers in the first place, keep them happy at lower power levels and well within their linear region.
How does MF fix these issues? It adds MORE power to keep the drivers linear. At 109.6 db these drivers are probably well into thermal compression already, is adding MORE power really the wisest thing to do to fix the problem? The drivers will likely sound quite nice. Right up to the point that the voice coils melt.
Fig 10 shows distortion levels at 100 db and Fig 11 shows distortion levels at 109.6 db. At the higher power level the distortion level is described as "unusable", so this is the max that this system can put out.
The distortion level of the 2nd harmonic rises from 37.2 in Fig 10 to 67 percent in Fig 11. (Untreated by push pull or MF.)
The distortion level of the 3rd harmonic rises from 12 to 26.9 percent.
If there was another chart showing distortion levels at 90 db, the distortion levels would be VERY low, as the distortion is ramping up at exponential rates with increased power level.
Does push pull and MF help this situation? Sure it does. They claim they can drop that 67 percent distortion down to 10 percent.
You know what else helps this situation? Don't abuse your drivers in the first place, keep them happy at lower power levels and well within their linear region.
How does MF fix these issues? It adds MORE power to keep the drivers linear. At 109.6 db these drivers are probably well into thermal compression already, is adding MORE power really the wisest thing to do to fix the problem? The drivers will likely sound quite nice. Right up to the point that the voice coils melt.
Another interesting point in that patent that hasn't been brought up here is the non linearity of air and the effects of compression. It's not a huge problem until you really start to abuse the situation but it is a significant problem if you intend to force your 10 pounds of stuff into a five pound bag and push the drivers to their limits.
So now we have a situation where MF is expected to correct for thermal compression, non-linearities caused by pushing the driver's suspension to it's physical limits (where it obviously becomes stiffer and doesn't want to move any more) AND the non linear compression of the air spring, which becomes a very strong spring the harder you try to push it.
This becomes a recipe for disaster. Keep adding MORE power and MORE power to a driver that's already thermally stressed to push things beyond their physical limits.
From my perspective this is the definition of bad design. I've always like horns and my design philosophy is that distortion should be controlled by good design, high efficiency, large enclosures (when necessary), quality drivers and low power (when possible). Not by abusing drivers with extreme power levels to make them behave.
This seems pretty similar to working with horses in the old days. You can use a team to get the job done right or you can use a single horse and whip it to make it behave. The harder you push it the more you have to whip it and sooner or later you end up with a dead horse.
So now we have a situation where MF is expected to correct for thermal compression, non-linearities caused by pushing the driver's suspension to it's physical limits (where it obviously becomes stiffer and doesn't want to move any more) AND the non linear compression of the air spring, which becomes a very strong spring the harder you try to push it.
This becomes a recipe for disaster. Keep adding MORE power and MORE power to a driver that's already thermally stressed to push things beyond their physical limits.
From my perspective this is the definition of bad design. I've always like horns and my design philosophy is that distortion should be controlled by good design, high efficiency, large enclosures (when necessary), quality drivers and low power (when possible). Not by abusing drivers with extreme power levels to make them behave.
This seems pretty similar to working with horses in the old days. You can use a team to get the job done right or you can use a single horse and whip it to make it behave. The harder you push it the more you have to whip it and sooner or later you end up with a dead horse.
Best quality bass I've ever heard in home setups:
Martin Logan Descent.
Paradigm Servo 15
My Labhorns.
All of them absolutely blew my 3x sealed Tempests out of the water for sound quality. Obviously the tempests would go louder than the servo subs but at normal music listening levels the quality of the bass from them was excellent. They definitely sounded 'cleaner', 'tighter' 'less boomy' etc than my Tempests.
The Servo15 subs have limiters which cut the low bass at high spls to keep them from being 'blown up' so I'd imagine you'd need multiples if you are into dvd's at reference levels.
Martin Logan Descent.
Paradigm Servo 15
My Labhorns.
All of them absolutely blew my 3x sealed Tempests out of the water for sound quality. Obviously the tempests would go louder than the servo subs but at normal music listening levels the quality of the bass from them was excellent. They definitely sounded 'cleaner', 'tighter' 'less boomy' etc than my Tempests.
The Servo15 subs have limiters which cut the low bass at high spls to keep them from being 'blown up' so I'd imagine you'd need multiples if you are into dvd's at reference levels.
Seems pretty clear that all those who have lived with MF of whatever strategy are effusive in their praise and consistent in what aspects of sound they find better, like RobWells.
While those who haven't, are sure it couldn't possibly help (and they don't seem to have the basic empirical skills to notice that all the users are effusive)
When you say, "blown up" it refers to a situation where the feedback goes off the rails. That happens when you drive the speaker beyond its design limits and the MF can't process properly the phase and gain.
To avoid feedback hell, your Servo15 manufacturer has wisely designed limitations; other manufacturers just keep the feedback low overall (which saves them money but compromises the benefits of MF) or otherwise shape the feedback perhaps with DSP (which adds complexity but looks like the best direction for the future*).
But a DIYer can have both substantial feedback and still keep out of grief because they know what they are doing (or who to blame afterwards).
Ben
*the future is likely to include bigger pieces of our systems being digital and with wideband Class D amps which seems pretty natural for MF - and with feedback finally covering the final frontier
Last edited:
Maybe you should read his post again, you missed something pretty important.
My Labhorns.
Those aren't even a particularly good design, but he did say they were in his top 3. They definitely don't use MF, but they do use distortion reduction principles by SIZING THE SYSTEM PROPERLY.
He also mentioned that the Servo 15 had a limiter so it choked at high output. Again, sizing the system properly is key whether you have MF or not. You can gloss over that all you like but it's not going to go away. using antique woofers with 2 mm xmax in tiny sealed boxes or in hybrid OB is a source of problems that MF can't fix.
Besides, no one said MF "couldn't possibly help", in fact I've said several times that it does help, but not much, IF you have a good design, a properly sized system and a good room. At this point there's so much evidence backing this up that only the strongest confirmation bias could blind you to those facts.
Just in case you missed this very important paragraph let me quote it again for you, this is really all you need to know.
If you want to continue pushing ever forward (but not actually getting anywhere) with your flawed conception of the benefits of MF that's up to you but the flaws in your logic will continue to be addressed.
Not likely. "Blown up" as it refers to drivers means thermally stressing them to the point of failure or mechanically stressing them to the point of failure. It's most likely the manufacturer was trying to prevent these issues, not MF instability.
I can assure you, the Labhorn is not a particularly good design, especially by today's standards. I've talked about this extensively in the past, I can show you why.
The original unfolded Lab was a thing of beauty. But when Danley folded it he made a huge mistake, the sealed chambers ended up way smaller than he planned for. This was later discovered and never fixed. From Danley's own Lab design notes:
This little bit of damage control was not really true. The effect on the model is not small and stuffing the chamber doesn't help, in fact it exaggerates the thermal problem.
And the chamber may be even smaller than that quote implies. Parham's sims suggest the chamber is only 34 liters. PiSpeakers Forum - 12 ? observations and comparisons with other designs - Wayne Parham, July 08, 2005 at 03:53:47
That's a huge difference from the ideal 50 liters, and when using 34 liters in a sim the simulated response matches the measurements of real Labhorns quite well.
The Labhorn was designed to be thermally limited, not excursion limited (as long as a proper high pass filter is used). But this caused problems too, since the thermal limit of the drivers happened WAY before the mechanical limit so there was no hint of distortion to warn that you were about to fry the drivers. The Labhorn is one of the most famous driver cooker designs for this reason. The tiny chambers the drivers are stuffed into just make the situation worse. And if you put stuffing in the chambers as the quote mentions the situation will be worsened again.
Show and tell - here's the measurement that used to be on the Labhorn plans download page. That huge notch right above the low knee is due to the undersized sealed chamber.
An externally hosted image should be here but it was not working when we last tested it.
And a sim (I had to do my own, as no one else simulates the flare even remotely accurately). Notice it matches the measurement quite well.
An externally hosted image should be here but it was not working when we last tested it.
And some more info -
Row 1 - single lab cab frequency and excursion, ideal 50 liter chamber size overlaid on the actual chamber size
Row 2 - 8x Lab cabs with actual chamber size, showing that dip doesn't go away no matter how many cabs you stack
Row 3 - single Lab cab with high pass filter - all the low bass goes away
Row 4 - 8x Lab cabs with high pass filter - even with this many cabs the low knee is weak
In the very first image (top left) you can see what the response would have looked like with the ideal 50 liter sealed chamber. It's much better. It's really a shame that this mistake was never fixed.
An externally hosted image should be here but it was not working when we last tested it.
NOTE - all these sims are shown at 1625 watts per cab. This is more than 2x more than these drivers are rated to handle, which is why so many Lab 12 drivers are blown up. With any kind of demanding signal it's not possible to get anywhere near xmax with this design (when using a high pass filter). When stacked, the design needs even MORE power to reach xmax, as shown in the last row of the image above in which the drivers are more than 2 mm shy of xmax with 1625 watts.
Modern drivers are capable of handling up to 4x more power, they would be much less likely to fail in a design like this.
Considering the undersized sealed chamber and the very small power handling of the drivers, the fact that this design can't ever reach xmax with a high pass filter before the drivers melt, this is not a particularly good design. It's a lot better than a lot of the garbage that was around and certainly a lot better than a Klipschhorn, but it's not ideal. Understandably, it's a great cab for home use, especially if you have a few of them, despite it's flaws. But in real world pro use there are serious problems with frequency response and power handling.
Last edited:
Hi JAG,
Even though we're veering way off the path of motion feedback, I was around when Danley first released this plan, and am a good friend with someone who built one of the first ones measured (Curtis-TooTall-List).
One of the main design criteria was a specified and predetermined size of the enclosure to fit into a truck's cargo area. Does your revised 50 liter chamber fit inside the original overall size? Just wondering...
Speaking of size, one of the main wishes, proponents of MF are interested in is, "small, low, and loud" (plus lower distortion and accuracy)...including me.
Voice coils melting is just collateral damage, that's what warranties are for... Drivers are not living things like the horses you mentioned upthread either.
Another thing, MF is a method of getting maximum efficiency from a driver/enclosure, as you also talked about upthread, that's the way I think of it anyway..
Just my 2 cents on a grumpy back-to-work-tomorrow Sunday.
-wreck
Even though we're veering way off the path of motion feedback, I was around when Danley first released this plan, and am a good friend with someone who built one of the first ones measured (Curtis-TooTall-List).
One of the main design criteria was a specified and predetermined size of the enclosure to fit into a truck's cargo area. Does your revised 50 liter chamber fit inside the original overall size? Just wondering...
Speaking of size, one of the main wishes, proponents of MF are interested in is, "small, low, and loud" (plus lower distortion and accuracy)...including me.
Voice coils melting is just collateral damage, that's what warranties are for... Drivers are not living things like the horses you mentioned upthread either.
Another thing, MF is a method of getting maximum efficiency from a driver/enclosure, as you also talked about upthread, that's the way I think of it anyway..
Just my 2 cents on a grumpy back-to-work-tomorrow Sunday.
-wreck
Yes, one of the main goals was to have truck pack dimensions.
It's not "my revised 50 liter chamber", Danley's own design plans specified a 50 liter chamber size. And then he screwed up the folding and ended up with something much smaller.
The difference between the original unfolded 50 liters (25 liters per chamber) and what Danley mistakenly ended up with after folding (43 liters by his estimation, 34 liters by Parham's estimation) is only an error of 7 - 16 liters, depending on whose estimation you believe. (Or you could CAD it out and find the actual volume of the chambers.)
To fix the chamber size would require a refold of the entire horn, and if that were done it would be quite easy to "rob Peter to pay Paul" by taking the few liters the chambers need to get up to 50 liters from the flare. The flare can easily spare a few liters without changing anything. On the other hand, the effect on the chambers by losing that few precious liters was very dramatic.
So yes, a refold could be done that keeps the original folded truck pack dimensions and fixes the sealed chamber size.
I can send you Danley's Labhorn design notes if you like, it details pretty clearly what the original goals were, and how it went off the rails when it was folded, although he did publish the quote that I quoted in the last post to do damage control instead of having to refold it properly.
Warranties don't last forever, and warranties rarely cover thermally destroyed voice coils. Warranties cover manufacturer defect and blowing up speakers is user abuse, not manufacturer defect.
A design that allows for the possibility of damage from normal use is a terrible design. That's why there's limiters in the Servo 15 Rob was talking about. If you are using a design that can be damaged in normal use or uses limiters to prevent damage in normal use, your system is undersized.
Replacement parts don't last forever either. What do you do when you blow your driver a few years past the purchase date and are unable to get a replacement?
Small size is a compromise in direct violation of performance goals. If that's a compromise you want to make, that's fine and completely up to you. MF isn't a cure, more like a band aid used to make the most of a bad situation when small size is the number 1 priority.
MF can't improve efficiency. Efficiency is a determined by the driver and enclosure. MF can give you slightly better performance from a given driver/enclosure system but it can't improve efficiency.
Last edited:
Nothing in your post was new news to me - the thermal handling was discussed extensively on prosoundweb years ago. I also remember reading when Wayne released his labhorn knock offs complete with the ally heat sinks to fix the 'problems'.
Mine have been going for 11 years now and so far have failed to catch themselves alight. If they ever do I'll be sure to let you know.
Yes please, thanks
My comment was somewhat tongue in cheek, partly because you keep talking horns (large), and poo-pooing motion feedback in a motion feedback thread..lolz..
As far as the warranty goes, if someone doesn't realize they are going to damage the speaker for the first, and especially a second time... well.. they don't have much common sense.. I don't anticipate getting even close to system max, especially with two of them in this 17'x15'x8' room filled with all your typical living room stuff.
You can call MF a terrible design, but myself and a lot of people do not. It's just a electronic way to getting big sound out of a small enclosure, which is a noble goal as far as I'm concerned, plus as ben states, everyone that tries one usually loves the sound quality.
In a classic definition sense you are right, but again, I still call it efficiency by proxy, using the electronics to achieve it.
Like I said, I was just having a bit of fun with most of my comments anyway.
cheers
-wreck
Fun ?!?! What is this fun you speak of?
With how quickly some threads spiral into hyperbole and "reductio ad absurdum", I'd say inject some fun whenever you can.
I think you missed the point of the write-up, and why I pointed you to it. Analog MFB at best will result in a minimum phase system. Without LF extension to DC it will never achieve pretty looking square waves because of the effect of changing phase which extends 2 – 3 octaves beyond cutoff. Then, throw in the mix the phase changes from the crossover between subwoofers and mains. The only way I know of to achieve your goal of pretty waveforms is using DSP to linearize the phase. Analog MFB can’t do this.
Here are a few screen shots from a square wave calculator I made to help visualize response magnitude and phase effects. It lets you define different LP or HP filters, parametric EQ, AP filter(basically phase response for a perfectly implemented crossover), whether you want minimum or linear phase, and what frequency of square wave you want to look at.
Attachment #1 60Hz Square Wave, for essentially perfect full range response with LF corner at 20hz. Note that minimum phase results in significant waveform distortion. This is what you will see from a theoretically ideal analog MFB setup.
Attachment #2 60Hz Square Wave, for essentially perfect subwoofer response with LF corner at 20hz and HF corner at 100hz. As I mentioned to you on the planar forum when discussing 10kHz square waves and ESLs, when you are bandwidth limited just above the signal frequency, you will mainly see fundamental sine wave.
Attachment #3 60Hz Square Wave, for essentially perfect subwoofer + mains response with LF corner at 20hz and crossover at 100hz. This is what most of us listen to all the time. The only way to fix the waveform distortion is by linearizing the phase with DSP.
If anybody is interested in a copy of this spreadsheet to play around with, let me know and I will see about posting a cleaned up version. You will need Excel 2007 or better.
With sealed boxes, resonances can be dealt with by MFB, or by designing proper damping into the woofer-box system, or negating their effects with EQ like the Linkwitz Transform circuit. Non-linearities can be diminished by MFB for a given woofer-box, or minimized by adding enough woofers for a given SPL, so that each woofer is operating below the desired distortion level.
I wasn’t trying to be clever (ie a distinction without a difference). Force applied to a woofer is only one factor in determining the resulting velocity of the cone. Apply the same force to a cone, and then operate the woofer in free air, sealed box, ported box, etc. The resulting velocity of the cone will be different as will the SPL. This is why velocity feedback can be used directly by MFB to flatten the response, and current feedback cannot.
Attachments
I know everyone will be impressed with Sy who posted a far more erudite reference to Thomas Aquinas before me and with reference to Bertrand Russell. Call it back-EMF or variable reluctance or whatever, the motion of the voice coil generates the signal. After I believe, you can process the signal in simple analog ways that can end up equivalent whether talking velocity or acceleration. And you have choices for how you want to amp to behave.
B.
Last edited:
I'll work on getting that to you. It's a small text file but I'm not sure if attachments are allowed in PMs or not, and I'm pretty sure it's too big to simply copy and paste into a PM. I might need an email address to send it to but either way we can move this to PM.
My design philosophy is clearly different than yours. That's fine. If you read over my posts you will find that I find the proper design philosophy (IMO of course) to be -
- use good drivers
- use good enclosure design
- size the system properly so that even at the loudest spl you will ever use distortion is low
- treat the room
IF those thing are done benefits of MF will be negligible.
If you don't care to do any or only some of those things MF may well be the ticket to making your compromised system acceptable.
I myself may try MF at some point, but only after all the above problems are sorted, and fully aware that once those problems are addressed MF won't offer much benefit.
I never called MF a bad design. After all, MF in itself is a concept and can be implemented in several different ways and to several different degrees. AND I have stated very clearly several times that it CAN help, moreso in some situations than others.
When I talked about bad design, I was very specifically talking about a situation in which you choose a small enclosure ON PURPOSE, as a primary design goal and then implement MF as a means to drive the system components to their bleeding edge of destruction; a system in which damage may be caused by normal use, or in which limiters are required to prevent damage from normal use. That's very different than saying "MF is a bad design".
What you call a "noble goal" I call an unnecessary compromise, and again, that's just a difference in design philosophy. And there's nothing wrong with having different opinions on which goals matter. For the most part I am concerned with absolute performance and size doesn't enter the equation as long as enclosures fit through doorways and the whole system fits in the room. You have different goals, that's fine, but make no mistake, when small size is a priority, absolute performance is compromised. Always. Every single time. Whether you use MF or not.
It would help if people would read exactly what is written, in that case explanations like this wouldn't be necessary, as I've said all this several times already. And in fact other people, people that are industry professionals and people that like MF as a concept, have literally said a lot of the same things.
What's the relevance of squarewaves in relation to musical reproduction? This is a novelty....
Ability to play square waves (or not) and what the system actually plays when you feed it square waves can tell you a lot about the system, which I think he succeeded in pointing out in detail. Remember, all those colorful graphs are inter-related. The frequency response, phase, group delay, impedance, etc.
As far as ability of a system to create square waves as a prerequisite for a pleasant listening experience - not necessary. But neither is pretty looking tone bursts, and some people seem to be laser focused on that as a system goal for some reason.
A reasonable dispersion pattern, on the other hand, is an absolute requirement for good sound (even though some people don't seem too bothered by obnxious tweeter bloom) unless you are listening outside with no boundaries close by. That is something that 1 square meter ESL panels will never be able to do.
Last edited: