The shorting turn above the gap cannot be very effective at flux modulation (thermally it would work fine), because the fringing flux in that region is very small if the magnet structure is designed properly. The ring can only control that flux which passes through it and then only that portion which is normal to its plane. Thats a very small amount.
Almost all of the flux will go through the lower one however.
Almost all of the flux will go through the lower one however.
I dont understand why such expensive drivers are still used. Theres really nothing special with them compared to much cheaper drivers with shorting rings or more. I suppose its tradition.John_E_Janowitz said:
In fact I recall a JBL paper showing their motors exhibit heaps of power compression compared to any comparable ferrite driver because of the alnico.
tech.knockout said:
The magnet is not a major factor in power compression. The magnetic energy of ALNICO is completely stable with temp while the ferrit will vary with temp, but this is at most 10-20% at typical temps. The voice coil changes with temp can be on the order of several dB or in % terms 70-80%. The magnet is not the issue.
Sounds like you are talking about the Vented Gap Tech Sheet. They had a Tad in there as one of the comparisons. I seriously doubt one of their older alnico's would have done any better at the power levels they were testing them at. Here's a link.
Rob
http://www.jblpro.com/catalog/support/getfile.aspx?docid=297&doctype=3
Rob
http://www.jblpro.com/catalog/support/getfile.aspx?docid=297&doctype=3
As Earl mentioned it isn't really the magnet that is the issue with power compression. It is getting heat away from the coil in the first place. That is where the copper sleeve on the pole and the 2 alum rings on the OD of the coil in the Apollo really help out.
I guess I could see in a case where you can transfer heat away from the coil well that the alnico vs ferrite may be an issue. Not the magnet itself really, but just the thermal mass. Take the TD15M for example. The pair of ceramic 5 magnets weigh close to 8 lbs where the Alnico ring to get equivalent flux in the gap is about 2lbs. The mass of steel in the back plate and top plate is also significantly less in the alnico motor as it is only 4" diameter vs 6" diameter. An 8 lb motor structure is going to heat up much faster than a 22 lb motor structure. Again though this is only going to be a factor if you can get the heat away from the coil quickly and sink it into the motor.
John
I guess I could see in a case where you can transfer heat away from the coil well that the alnico vs ferrite may be an issue. Not the magnet itself really, but just the thermal mass. Take the TD15M for example. The pair of ceramic 5 magnets weigh close to 8 lbs where the Alnico ring to get equivalent flux in the gap is about 2lbs. The mass of steel in the back plate and top plate is also significantly less in the alnico motor as it is only 4" diameter vs 6" diameter. An 8 lb motor structure is going to heat up much faster than a 22 lb motor structure. Again though this is only going to be a factor if you can get the heat away from the coil quickly and sink it into the motor.
John
John_E_Janowitz said:
Hi John
There is one thing that you forgot in this analysis. The Alnico magnet will hardly change its properties at all when its heated while the Ferrite will. So even though the Alnico may get hotter faster, the ferrite will probably still change more. But again, thermal flux change is a secondary factor compared to the thermal effects of the voice coil.
There is a factor that most people don't understand very well and that has to do with thermal time constants. No matter how fast you can draw off the heat from the voice coil this is always going to be slower than the voice coils heating and cooling with signal, which it happens almost instantaneously. These temperature changes depend, for all practical purposes, only on the thermal mass of the voice coil, mostly the wire. The greater this mass the less the temperature fluctuation and the less the thermal modulation. This is a short time constant effect.
The sleeves that you guys use wil be very effective over a longer time scale to keep the mean temperture of the voice coil down thus lowering the long term power compression and increasing thermal capability before failure.
In essence the bigger the voice coil the lower the short term thermal modulation. This, to me, is why small dome tweeters don't have the same dynamics as compression drivers. The difference in the voice coil size in these two different designs is significant and so is their ability to thermally handle the signal dynamics. I think that this effect in woofers is less, but still not insignificant.
I recall reading discussions like these about power compression in the past with almost exactly the same information.
After deciding that:
- AlNiCo 's flux is less susceptible to fluctuation from heat, and
- A lighter motor structure will ultimately have lower long-term power handling unless cooled, and
- The mass of the VC largely determines short-term power compression
People decided that it'd be nice to devise ways to cool either the pole piece or the air in the magnetic gap using
- machined pole piece heatsink (pi speakers famously did this with apparently good results)
- forced air cooling (through rear pole vent = pi speakers tried this for us... of course it didn't work too well)
- Creative array utilizing heatpipe technology on the pole piece (I have read about this and independently thought about it, but never seen utilization)
- liquid cooling of some sort (I suppose a water jacket passed to a radiator and pump which are outside the listening area)
- There is a patent for a sealed air gap with cooling from both sides, but I do not know what this entails...
The general consensus is that most of the more active-cooling methods are too cumbersome or expensive to implement, or will have other negative effects far outweighing the benefits (or just that the benefits are far too small).
I figure then that the conclusion was to try to deal with power compression in other ways. AEspeakers/Lambda has addressed this particularly well, and so has piSpeakers I suppose.
Heh... I have an idea that I have not seen or even read about, but I'd have to try to visualize how it could work before I make myself look like an idiot. Dynamic signal-dependent cooling is what I'm after to buffer a CHANGE in VC temp, and not just to cool the assembly. I could care less about the rest of the motor... may as well heat it up to reduce electrical conductivity
BTW: I'm quite experienced with watercooling computers and even subzero cooling: the notion that there could be condensation on the tubes is incorrect. You'd need sub-ambient coolant temperatures, which is not possible running a simple radiator. The coolant used IS and SHOULD BE distilled water + anti-microbial/corrosion additives (I've seen a test of only ONE aftermarket coolant outperforming water, and it is $$$). The Heat-Exchanger on the CPU (or Waterblock) is now almost always pure COPPER, and the design of said block is now, after numerous generations of revisions, nearly at the point of diminishing returns. For reference, 3 of the best well-known blocks are: Swiftech Apogee GTZ, Dtek Fuzion v2, EK Supreme.
After deciding that:
- AlNiCo 's flux is less susceptible to fluctuation from heat, and
- A lighter motor structure will ultimately have lower long-term power handling unless cooled, and
- The mass of the VC largely determines short-term power compression
People decided that it'd be nice to devise ways to cool either the pole piece or the air in the magnetic gap using
- machined pole piece heatsink (pi speakers famously did this with apparently good results)
- forced air cooling (through rear pole vent = pi speakers tried this for us... of course it didn't work too well)
- Creative array utilizing heatpipe technology on the pole piece (I have read about this and independently thought about it, but never seen utilization)
- liquid cooling of some sort (I suppose a water jacket passed to a radiator and pump which are outside the listening area)
- There is a patent for a sealed air gap with cooling from both sides, but I do not know what this entails...
The general consensus is that most of the more active-cooling methods are too cumbersome or expensive to implement, or will have other negative effects far outweighing the benefits (or just that the benefits are far too small).
I figure then that the conclusion was to try to deal with power compression in other ways. AEspeakers/Lambda has addressed this particularly well, and so has piSpeakers I suppose.
Heh... I have an idea that I have not seen or even read about, but I'd have to try to visualize how it could work before I make myself look like an idiot. Dynamic signal-dependent cooling is what I'm after to buffer a CHANGE in VC temp, and not just to cool the assembly. I could care less about the rest of the motor... may as well heat it up to reduce electrical conductivity
BTW: I'm quite experienced with watercooling computers and even subzero cooling: the notion that there could be condensation on the tubes is incorrect. You'd need sub-ambient coolant temperatures, which is not possible running a simple radiator. The coolant used IS and SHOULD BE distilled water + anti-microbial/corrosion additives (I've seen a test of only ONE aftermarket coolant outperforming water, and it is $$$). The Heat-Exchanger on the CPU (or Waterblock) is now almost always pure COPPER, and the design of said block is now, after numerous generations of revisions, nearly at the point of diminishing returns. For reference, 3 of the best well-known blocks are: Swiftech Apogee GTZ, Dtek Fuzion v2, EK Supreme.
Archmage said:
OT, but would you also happen to be experienced with designing custom phase change cooling units? Always wanted one to play with, but they're so expensive.. I figure better could probably be had for less, but I've never seen anyone advertising prices for custom built units, so I'm not sure. Do you hang around XS forums? Any idea how much a simple setup would cost? Couldn't do it right now tho, I'd have to save some pennies first.
BHTX said:
Short answer: Yes, but you'd be better off addressing inquiries for purchasing custom setups to (I can't remember the name of a reliable builder because I have only considered DIY). I have *some* DIY phase change experience, but there are a few HVAC engineers on there who will be MUCH more helpful.
I've begun to prefer watercooling for simplicity, reliability, noise, and power consumption (less guilt).
Why don't I see the PM feature on this forum? I must be blind...
Tech.knockout wrote:
"In fact I recall a JBL paper showing their motors exhibit heaps of power compression compared to any comparable ferrite driver because of the alnico."
Several years ago I spoke with someone familiar with that test, and he said that JBL (deliberately?) overheated the alnico magnet of the TAD woofer, which caused a permanant loss of flux. If they hadn't damaged the magnet, it probably would have exhibited slightly less flux modulation than the shorting-ringed ceramic magnet. Going by memory here, the graphs showed something like 7 dB compression for the alnico magnet, and it did not recover over time - an indication that the magnet was indeed damaged (partially demagnetized) by overheating, which would not occur with a ceramic magnet.
As John Janowitz noted, I'm using the TD12M woofer in a "high end" (not prosound) application, and am very happy with it. I can post links to online comments if anyone wants me to. I bought a pair of TD15Ms years ago when Nick McKinney was building them, and they sounded great and were easy to work with.
That Supravox field-coil woofer is also very nice, and I like its parameters better for applications where subwoofers won't be used. I owned a pair, but frankly it's too expensive for anything I'm likely to be doing anytime soon.
In Earl's Summa, the B&C TB15X100 sounds great - but with a lesser crossover design I'm not so sure. In a woofer with parameters and response closer to the TD15M, the B&C 15NW76 looks pretty good.
Duke
"In fact I recall a JBL paper showing their motors exhibit heaps of power compression compared to any comparable ferrite driver because of the alnico."
Several years ago I spoke with someone familiar with that test, and he said that JBL (deliberately?) overheated the alnico magnet of the TAD woofer, which caused a permanant loss of flux. If they hadn't damaged the magnet, it probably would have exhibited slightly less flux modulation than the shorting-ringed ceramic magnet. Going by memory here, the graphs showed something like 7 dB compression for the alnico magnet, and it did not recover over time - an indication that the magnet was indeed damaged (partially demagnetized) by overheating, which would not occur with a ceramic magnet.
As John Janowitz noted, I'm using the TD12M woofer in a "high end" (not prosound) application, and am very happy with it. I can post links to online comments if anyone wants me to. I bought a pair of TD15Ms years ago when Nick McKinney was building them, and they sounded great and were easy to work with.
That Supravox field-coil woofer is also very nice, and I like its parameters better for applications where subwoofers won't be used. I owned a pair, but frankly it's too expensive for anything I'm likely to be doing anytime soon.
In Earl's Summa, the B&C TB15X100 sounds great - but with a lesser crossover design I'm not so sure. In a woofer with parameters and response closer to the TD15M, the B&C 15NW76 looks pretty good.
Duke
Given the significant 1.6 kHz peak exhibited by the 15TBX100 (within an octave of the crossover) and the contrasting smoothness of your frequency response in that region, I figured that you're using the compression driver's phase rotation to help cancel that peak. If so, it seems pretty sophisitcated to me.
Duke
Duke
That peak can be a problem, but in the Summa, the crossover point for the woofer, by the standard -3 dB specification, is about 700 Hz. The shallow rise in the woofer above this point brings the acoustic LP response to more like a 900-1000 Hz filter. But the LP is pretty standard third order. The tweeter is somewhat low at 1600 Hz so the net result is flat. I suppose, in essence, I am compensating for this peak in the waveguide and in that sense you are correct. But, to me, this is just good crossover design which uses the acoustic results, not the electrical transfer functions, to yield the correct end result. It is a typical example of where the use of blind electrical filter specifications (i.e. Linkwitz-Riley, etc.) for a crossover will not get the job done.
Then it goes back to what said previously. I see it as a fault instead of abuse.audiokinesis said:
I suppose in the previous posts I used the wrong term; it isnt power compression.
Going back to the topic of best 15" drivers; I wonder why 18 Sound hasn't implemented their Active impedance coil to other drivers besides the 10" and 12" midrange/midbasses (they use triple/single shorting rings instead). Their double spider, two sided voice coil technology and triple ring would probably rank among the best. If they use AIC more often it would make their drivers appear very high end and very novel/esoteric.
P.S. The test i was thinking of was indeed the Vented Gap sheet posted by Robh3606.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.