You will have lower efficiency in the bass/low mid comparing to the high mids. The point of transition is dependent of the bafle size and the drivers own transfer function. The 0.5 driver needs to be x-ed to compensate for this gradual roll off towards lower frequencies.
The fenomana is called baffle-step and baffle-step-compensation.
/Peter
The fenomana is called baffle-step and baffle-step-compensation.
/Peter
Are you using the w3-871s? If you are, there are 2 ways I've seen people do it. The 2 versions of the ELF use the same cabinet volume and tuning, but different XO.
CSS uses a 5.6 ohm resistor in parallel with a 1.5 mH inductor for the "bass" driver. (this is baffle-step compensation)
Jim Griffin uses a 4 ohm resistor in series with a 1.5 mH inductor.
I have only heard it the CSS way and it sounded great. BrianGT told me that he has heard both and prefers it the Jim Griffin way.
I say that because the only different is a single resistor, (which is cheap!) you should try it both ways and see what you like best.
Good Luck!
CSS uses a 5.6 ohm resistor in parallel with a 1.5 mH inductor for the "bass" driver. (this is baffle-step compensation)
Jim Griffin uses a 4 ohm resistor in series with a 1.5 mH inductor.
I have only heard it the CSS way and it sounded great. BrianGT told me that he has heard both and prefers it the Jim Griffin way.
I say that because the only different is a single resistor, (which is cheap!) you should try it both ways and see what you like best.
Good Luck!
Attachments
hi dave, you still recomending the rear woofer. 
i've been out of the loop for the last 3 months and a lot has changed here (country flags, everyones a member thank god).
seriously i have tried the rear woofer and it resolves 2 issues so it makes sense.
1. low pass roll off with a big inductor
2. push pull reduction in bass distortion (i tried it with 2 x 7" SS drivers).
cheers
i've been out of the loop for the last 3 months and a lot has changed here (country flags, everyones a member thank god).
seriously i have tried the rear woofer and it resolves 2 issues so it makes sense.
1. low pass roll off with a big inductor
2. push pull reduction in bass distortion (i tried it with 2 x 7" SS drivers).
cheers
Hello Dave,
I have been thinking about building a pair of speakers using the 871 myself. For some reason, i'd like to use 3 of them per enclosure. How about using 2 in front and one on the back. Would it work ? Would I need additional baffle-step compensation ?
Could something similar be done to the Alegria Isa (I love the needle design, what can I say ?) where there would be 4 speaker in front and 1-2 on the back ? At that stage, I guess it might be cheaper to use electronic components instead of the back speakers. What do you think ?
i am too toying with multiple 871 (more than 2) to improve SPL levels esp in the bass. however the issues I have are this:
1. if we use multiple 871 over the full range we encounter polar interference from freq above 3000-3500Hz.
2. if we use 1 871 for the fullrange and the remaining 871 for the bass we can have very low impednace if the 871s are all wired in parallel or find that the SPL limit is again the single 871 that is being run fullrange.
lets assume for arguments sake we use 5 871s. (might be a bit extreme for some). 1 run fullrange and the other 4 (operating via a low pass filter) wired in series parallel. the 4 871s will share less power than the single 871 hence the full range 871 will have "maxed out" while the 4 871s will be well inside their operating range.
if we assume we wire the 4 871s (low pass again) in series then the situation is the same as above. the only way the power will be shared somewhat equally (discounting the Low pass filter) is if all 5 871 are in parallel. that drops impeandace to 8/5 ohms.
if one were to use 3 871s the impedance would drop to 8/3ohms better than 8/5 ohms but still a handful for most amps.
any solution for this.
to improve the SPL of the system I am investigating using larger fullrange drivers (read as 6" Fostex). however there we face the issue of beaming (baffle step can be compesated by using a 6-8" woofer).
this is a catch 22. beaming or low SPLs.
to put SPLs in perspective I consider 95db / 1 W / 2 m (50Hz +) to be the max SPL I should need.
below 50Hz a sub will provide the required SPLs.
1. if we use multiple 871 over the full range we encounter polar interference from freq above 3000-3500Hz.
2. if we use 1 871 for the fullrange and the remaining 871 for the bass we can have very low impednace if the 871s are all wired in parallel or find that the SPL limit is again the single 871 that is being run fullrange.
lets assume for arguments sake we use 5 871s. (might be a bit extreme for some). 1 run fullrange and the other 4 (operating via a low pass filter) wired in series parallel. the 4 871s will share less power than the single 871 hence the full range 871 will have "maxed out" while the 4 871s will be well inside their operating range.
if we assume we wire the 4 871s (low pass again) in series then the situation is the same as above. the only way the power will be shared somewhat equally (discounting the Low pass filter) is if all 5 871 are in parallel. that drops impeandace to 8/5 ohms.
if one were to use 3 871s the impedance would drop to 8/3ohms better than 8/5 ohms but still a handful for most amps.
any solution for this.
to improve the SPL of the system I am investigating using larger fullrange drivers (read as 6" Fostex). however there we face the issue of beaming (baffle step can be compesated by using a 6-8" woofer).
this is a catch 22. beaming or low SPLs.
to put SPLs in perspective I consider 95db / 1 W / 2 m (50Hz +) to be the max SPL I should need.
below 50Hz a sub will provide the required SPLs.
I've found that using 4-871s and applying BSC to the lower two drivers works well. They sound fine when you're sitting down. When you stand up you lose some HF as the four drivers create a cylindrical sound field rather than a spherical one. An alternative to the BSC might be to use a 4" TB woofer (front and back if you like, right Dave?). As for center channel use I would recommend either 3 or 4 drivers in horizontal but angle off the two outside drivers to help the dispersion. Having them on the same plane results in too much beaming.
what is the height of the upper 2 drivers. ican assume that your listening height is between the upper 2 drivers.
the HF loss could be due to interference. 7V or dave might be better placed to offer a judgement on this.
if 2 x 871s can be run fullrange without interference loss (HF loss related to teh c-c distance) one could build a system using 4 x 871s 2 in front and 2 in the rear. ofcourse one still has some diffcult impedances. is teh BSC drivers are in series (16 ohms) and the full range drivers are in series the impednace of the system will be about 16 ohms till say 300hz (BSC freq) and drop to 8 ohms below BSC freq.
16 ohms is a high imedance and one looses 3-4db of sensitivity due to this.
alternately if one has all drivers in parallel you get 4 ohms for the full range except the impedance drops to 2 ohms below BSC freq.
alternately one can have the full rnage drivers in parallel and the 2 rear BSC drivers in series (this also means a large series inducrtor) and have a impedance of 4 ohms above BSC and 2.8ohms below BSC.
the HF loss could be due to interference. 7V or dave might be better placed to offer a judgement on this.
if 2 x 871s can be run fullrange without interference loss (HF loss related to teh c-c distance) one could build a system using 4 x 871s 2 in front and 2 in the rear. ofcourse one still has some diffcult impedances. is teh BSC drivers are in series (16 ohms) and the full range drivers are in series the impednace of the system will be about 16 ohms till say 300hz (BSC freq) and drop to 8 ohms below BSC freq.
16 ohms is a high imedance and one looses 3-4db of sensitivity due to this.
alternately if one has all drivers in parallel you get 4 ohms for the full range except the impedance drops to 2 ohms below BSC freq.
alternately one can have the full rnage drivers in parallel and the 2 rear BSC drivers in series (this also means a large series inducrtor) and have a impedance of 4 ohms above BSC and 2.8ohms below BSC.
When i think about 5 drivers, i use 1 as a midrange & 4 (in series/parallel) as woofer. XO at about the the baffle step frequency. Since the drivers have response beyond the XO you can get away with a simple 1st order (series?) XO, and since they are the same driver, they should mate pretty well.
dave
dave
planet10 said:
That would probably be an interesting setup.
I am just beginning to understand the principles behind BSC. Now I'll have to learn about XO... Is there such a thing as a XO software simulator ? Would the XO "module" in Winisd do a proper job or should I go looking somewhere else ? Any chance to find that kind of software running on a Mac ?
That's correct.
The Nonsuch has the same issue.
The Elf1.5 (2-871s) uses a BSC on one driver. In the 4 driver arrangement The upper pair and the lower pair are wired in parallel then both pairs wired together in series for 8 ohm nom. The BSC is wired in series to the bottom pair. I understand impedance will increase above the BS frequency but I don't think it's as high as suggested.
I think you've just described the Spires from CSS.
fireman said:
I wish. With X we might well start to see more of these kinds of things (being UNIX a lot of science types are happily dumping their NT boxes and moving to Mac)
Timn8ter said:
Basically... i've been thinking of it for long before the Spires thou (i pulled the idea from a 5 driver Magnat speaker made in germany -- actually 6 drivers, 5 4's and a tweeter). I have a box of a dozen 6" famous maker, fairly high efficiency (96 dB claimed) drivers waiting just such an implementation.
dave
planet10 said:
sounds the best suggestion by a long way to me, the
narrower the baffle and the higher the BSF the better.
(with a wide baffle a larger unit 6dB more sensitive is simpler)
I can't see a good arrangement with four drivers.
For 3 drivers, similar to 5, 2 parrallel for bass, 1 for mid/treble,
crossover at BSF, but has half nominal impedance of 5 driver.
For 2 drivers use a standard 0.5 way, again at BSF, I can't see
the point of the crossovers variations of standard 0.5 way posted.
sreten.
planet10 said:
all fine only since about half the power will go to teh fullrange and the other half shared between the 4 BSC drivers the fullrange is still the weak link. in fact the max SPL for such a sysmte will not be much more than one using only 2 x 871s (like in the elf).
what might happen though (in the 5 driver system) is that the quality of the bass might be better (if the very low bass is reduced for teh fullrange) since each of the 4 drivers will be moving a lot less and the voice coil will be in the gap. remember if the fullrange is product the full freq. range it's Xmax will be the weak link.
Timn8ter said:
in this case i was raising the question of interference which can be destructive that occurs when 2 drivers reproduce a sound that is of a freq. higher than the distance between their centers. in tis case 2 x 871s c-c distnace would about 4"(.1m for the metric crowd) which is .333 ft , speed of sound for ease of calculations can be approximated at 1200 ft/s (330m/s for the metric folk). theis means a freq above 1200/.333 (3600hz) there would be some interference. this being destrcutive woudl reduce the HF respone of the system.
not having any real world expereicne with this intereference, I would like to ask all those who have exprimented with such systems incl 7V (steve) what the real world effects are. whether a single larger driver (which could suffer from beaming) such as a Fostex 166e would be a better solution.
planet10 said:
Canton also had such systems but since they had dedicated tweeters the HF response was managed. what dia are the famous make drivers you got?
the way i see it we got 2 principles that are at odds with each other....
for better hf prodcution one needs a radiating area that is small, for better bass prodcution one needs to move air and hence needs a radiatig area that is large. the options are planar, leaf, ribbons, etc...
then there is a twist....to make things simple one would prefer a fullrange or 1.5 way system (with a single inductor)....aargh! such is life.
I need some help. serious help.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.