Full Range Drivers- Yes!
However....I have a problem with what is supposed to be a fairly good fulrange driver- Fostex FE 206 e. Since I couldn't wait for horn enclosures, I temporary mounted Fostex in a sealed enclosure. The result-horrible
It sounded thin and harsh, with peaking midrange and beamy highs. I felt as If I was listening to a cheap PA speaker
Then I added a BSC circuit: 3 mH inductor-10ohm 10 w resistor and a zobel. Much better- the midrange peak was gone. Unfortunately, it still sounded thin and harsh. Then I tried another thing: I connected a 33 ohm 15 W resistor in series with the driver. WOW! The result was much better-at a cost of a greatly reduced efficiency. So I could listen to a low volume only...
I also borrowed a pair of B&W DM 601 from a friend. Actually, they sounded much better than Fostex....
I don't want to get rid of Fostex, nor I like them as they are now. Is there any solution? Will horn enclosure cure these flaws? Any comment? Nelson? Kuei?
p.s. The amp is ZEN V3
However....I have a problem with what is supposed to be a fairly good fulrange driver- Fostex FE 206 e. Since I couldn't wait for horn enclosures, I temporary mounted Fostex in a sealed enclosure. The result-horrible
It sounded thin and harsh, with peaking midrange and beamy highs. I felt as If I was listening to a cheap PA speaker Then I added a BSC circuit: 3 mH inductor-10ohm 10 w resistor and a zobel. Much better- the midrange peak was gone. Unfortunately, it still sounded thin and harsh. Then I tried another thing: I connected a 33 ohm 15 W resistor in series with the driver. WOW! The result was much better-at a cost of a greatly reduced efficiency. So I could listen to a low volume only...
I also borrowed a pair of B&W DM 601 from a friend. Actually, they sounded much better than Fostex....
I don't want to get rid of Fostex, nor I like them as they are now. Is there any solution? Will horn enclosure cure these flaws? Any comment? Nelson? Kuei?
p.s. The amp is ZEN V3
Konnichiwa,
Predictably. One part is the severely tipped up frequency response (whch can be equalised), the other are fundamental problems with drivers like the Fostex.
Lets start with the first set of problems. Fostex's published response looks like this:
As you can see, the response in the range from 250Hz to around 1.5KHz is pretty flat at around 95db/2.83V/1m. Above that the response is shelved up on axis by around 5 - 6db with the worst on axis peaks being nearly 10db above the mean response.
The off axis response is flatter, I cannot tell at which angles this is taken but it may very well be 30/60 Degrees, which suggests that most of all the drivers should be listened to far off axis or you need to shelve the range above 1KHz down by around 6db across the range with an attenuation of around 10db at 10KHz.
One way to equalise this response is to do this acoustically by using a suitable front horn, which will boost the output below around 1KHz, such as the Oris 150 or Azurra Horn. In fact, ONLY EVER in such an Horn (and NOT in a rear only horn) can the range between 250Hz-10KHz ever be balanced using this driver and avoiding equalisation.
At low frequencies another problem makes itself known. If we place the driver ino a 30 Liter sealed box with 30cm baffle width and place it optimally near a corner for LF boost (that means using the golden ratio for hight above floor, distance to rear wall and distance to sidewall) we get an LF response looking likely a lot like this:
You need to "splice" the LF response from directly above to the HF response above 1KHz from the Fostex Datasheet to get a goo idea what is going on.
So, at 10KHz our SPL will be around 105db and at 100Hz our SPL will be around 85db. In other words, from 100Hz onwards the frequency response is boosted with 10db per decade so that at 10KHz there is 20db more SPL than at 100Hz!!!!!
There is simply no way to equalise that completely and thus an LF horn MUST be employed if any LF reproduction is desired.
The use of current drive BTW may very well bring the LF output in line. For fun here is what happens if the source has an internal impedance of around 33 Ohm:
(note, I scaled Re to "fudge" the effect of a High Z source with 2.83V applied, not easy to do in the spreadsheet without programing Re' - that is any source impedance in to be included but compensated)
We now have a balanced LF response and if we deliberatly roll off the treble a little, either with a parallel network or with an active EQ we should have a balanced sound.
As you can tell from the above, the answer is:
Not completely! You can use a rear horn to get the low frequencies balanced but you will still have problems with the upper midrange and treble.
I guess you have to buy a "FirstWatt" amp from Nelson or build something similar or combine a rear horn and RLC EQ.
Sayonara
Vix said:However....I have a problem with what is supposed to be a fairly good fulrange driver- Fostex FE 206 e. Since I couldn't wait for horn enclosures, I temporary mounted Fostex in a sealed enclosure. The result-horrible
Predictably. One part is the severely tipped up frequency response (whch can be equalised), the other are fundamental problems with drivers like the Fostex.
Lets start with the first set of problems. Fostex's published response looks like this:
An externally hosted image should be here but it was not working when we last tested it.
As you can see, the response in the range from 250Hz to around 1.5KHz is pretty flat at around 95db/2.83V/1m. Above that the response is shelved up on axis by around 5 - 6db with the worst on axis peaks being nearly 10db above the mean response.
The off axis response is flatter, I cannot tell at which angles this is taken but it may very well be 30/60 Degrees, which suggests that most of all the drivers should be listened to far off axis or you need to shelve the range above 1KHz down by around 6db across the range with an attenuation of around 10db at 10KHz.
One way to equalise this response is to do this acoustically by using a suitable front horn, which will boost the output below around 1KHz, such as the Oris 150 or Azurra Horn. In fact, ONLY EVER in such an Horn (and NOT in a rear only horn) can the range between 250Hz-10KHz ever be balanced using this driver and avoiding equalisation.
At low frequencies another problem makes itself known. If we place the driver ino a 30 Liter sealed box with 30cm baffle width and place it optimally near a corner for LF boost (that means using the golden ratio for hight above floor, distance to rear wall and distance to sidewall) we get an LF response looking likely a lot like this:
An externally hosted image should be here but it was not working when we last tested it.
You need to "splice" the LF response from directly above to the HF response above 1KHz from the Fostex Datasheet to get a goo idea what is going on.
So, at 10KHz our SPL will be around 105db and at 100Hz our SPL will be around 85db. In other words, from 100Hz onwards the frequency response is boosted with 10db per decade so that at 10KHz there is 20db more SPL than at 100Hz!!!!!
There is simply no way to equalise that completely and thus an LF horn MUST be employed if any LF reproduction is desired.
The use of current drive BTW may very well bring the LF output in line. For fun here is what happens if the source has an internal impedance of around 33 Ohm:
An externally hosted image should be here but it was not working when we last tested it.
(note, I scaled Re to "fudge" the effect of a High Z source with 2.83V applied, not easy to do in the spreadsheet without programing Re' - that is any source impedance in to be included but compensated)
We now have a balanced LF response and if we deliberatly roll off the treble a little, either with a parallel network or with an active EQ we should have a balanced sound.
Vix said:
As you can tell from the above, the answer is:
Not completely! You can use a rear horn to get the low frequencies balanced but you will still have problems with the upper midrange and treble.
I guess you have to buy a "FirstWatt" amp from Nelson or build something similar or combine a rear horn and RLC EQ.
Sayonara
The FE206E anf FE166E are my favorites of the Fostex product,
but they do require some work.
I recommend that you look at the article at www.firstwatt.com
in reference to the kind of network that will do the job. Note
that it is not essential to run a current source amp, rather you
want to have an amp with a finite impedance whose value is
optimized around the type of enclosure, it's location in the
room, and your taste. You also want to try a parallel RC network
whose value depends on the source impedance and your
taste.
but they do require some work.
I recommend that you look at the article at www.firstwatt.com
in reference to the kind of network that will do the job. Note
that it is not essential to run a current source amp, rather you
want to have an amp with a finite impedance whose value is
optimized around the type of enclosure, it's location in the
room, and your taste. You also want to try a parallel RC network
whose value depends on the source impedance and your
taste.
Here is the biggest problem with single driver. Dopler distortion. If you have a 3" speaker that has to modulate 3mm just to produce audible bass (due to its limited surface area) the midrange and treble will now be prduced from a moving source. even worse, the smaller the driver has to be to accuratly produce treble the more excursion there is to produce, say ,50hz. therefore the only way around this problem is to use many drivers to increase surface area. but if not done carefully you get the same comb filtering. but now its at all frequencies above the wavelength of the array. That is why its generally better to take your chances whith phase cancelation at one frequency like in a 2 or 3 way system.
wow. this means a drivers like the jx92 and 167e require some serious cabinet design.
maybe one can avoid doppler as well as phase by using bending wave transduces like Audio Physiks DDD driver or Manger or maybe even using 2 widerange drivers (one choice I am considering is the Fostex FF85K with a yet unchosen 6-10" woofer) with a 1st oder XO.
What if one were to take a 166/167/206/207 type driver load it with a ML-TL like Bob Brines, GM, Abpea and co do and place the tower against the rear wall to further boost the bass?
maybe one can avoid doppler as well as phase by using bending wave transduces like Audio Physiks DDD driver or Manger or maybe even using 2 widerange drivers (one choice I am considering is the Fostex FF85K with a yet unchosen 6-10" woofer) with a 1st oder XO.
What if one were to take a 166/167/206/207 type driver load it with a ML-TL like Bob Brines, GM, Abpea and co do and place the tower against the rear wall to further boost the bass?
doppler distortion? I think not.
phase shifting with resulting time-domain alterations, yup! but this is an area where the average fullrange driver is better than other traditional solutions (i.e. vs. your typical 2-way).
you can get around the time domain problem by utilizing the high excursion driver axially pointing upward or downward (and have a polar response for the driver that is very close to being omni).
oh and since the 206E has come up.. check out this design:
http://www.audio.co.jp/speaker/sou.htm
phase shifting with resulting time-domain alterations, yup! but this is an area where the average fullrange driver is better than other traditional solutions (i.e. vs. your typical 2-way).
you can get around the time domain problem by utilizing the high excursion driver axially pointing upward or downward (and have a polar response for the driver that is very close to being omni).
oh and since the 206E has come up.. check out this design:
http://www.audio.co.jp/speaker/sou.htm
FE206E cabinet
On the topic of a good cabinet for the FE206E...
Are you familiar with the Decware HDT? (High Definition Tower)
http://www.decware.com/HDT/hdt.htm
TL with 2 passive radiators each, ~ 35 Hz to 18 kHz, 96 dB efficient, recommend 1.8 ~ 30 watt amp. The cabinet design has deep, tight, accurate bass from “below 40 Hz”, “flat out to 18KHz.” Plans, pre-cut panels and empty enclosures are available.
On the topic of a good cabinet for the FE206E...
Are you familiar with the Decware HDT? (High Definition Tower)
http://www.decware.com/HDT/hdt.htm
TL with 2 passive radiators each, ~ 35 Hz to 18 kHz, 96 dB efficient, recommend 1.8 ~ 30 watt amp. The cabinet design has deep, tight, accurate bass from “below 40 Hz”, “flat out to 18KHz.” Plans, pre-cut panels and empty enclosures are available.
Thank you!
I appreciate the detailed and comprehensive reply
Even though my ZEN V3 is recently built, I already started thinking about building a current source amp (but I'll have to wait a year or two till Nelson publishes the schematics for F1 )
In the meantime, I will have to play with ZEN and correction circuits...
BTW, the upward-firing Fostex looks interesting, but I don't understand any Japanese
ScottG, can you possibly find that in English?
regards
Vix
I appreciate the detailed and comprehensive reply
Even though my ZEN V3 is recently built, I already started thinking about building a current source amp
(but I'll have to wait a year or two till Nelson publishes the schematics for F1 )In the meantime, I will have to play with ZEN and correction circuits...
BTW, the upward-firing Fostex looks interesting, but I don't understand any Japanese
ScottG, can you possibly find that in English?
regards
Vix
has http://www.audio.co.jp/speaker/uatau.htm or
http://www.audio.co.jp/speaker/yu.htm
ever been done using a fostex 6" (166/167/168) in maybe a cylinder that is 20cm dia or a 20cm square base?
also how does the cone that is reflecting the sound colour the sound?
http://www.audio.co.jp/speaker/yu.htm
ever been done using a fostex 6" (166/167/168) in maybe a cylinder that is 20cm dia or a 20cm square base?
also how does the cone that is reflecting the sound colour the sound?
I assume that this is something you've read somewhere because I've never found anyone (on forums or in 'real life') who has actually experienced Doppler distortion.eric180db said:
If you have though, please report back and let us know precisely how it sounds. Doppler distortion is an interesting-sounding, mythical creature. I doubt if it actually exists.
maybe the effects are so subtle that we cant hear it...
maybe we are so used to this we dont recognise it...
maybe our ear-brain actually hears it but corrects for it...
i do believe that we at present are not able to measure enough parameters in the audio domain to allow the measurements to determine if a product sounds good or not. this means we do not understand the educated ear-brain enough. maybe doppler (if it exists) is one of those missing parameters.
maybe we are so used to this we dont recognise it...
maybe our ear-brain actually hears it but corrects for it...
i do believe that we at present are not able to measure enough parameters in the audio domain to allow the measurements to determine if a product sounds good or not. this means we do not understand the educated ear-brain enough. maybe doppler (if it exists) is one of those missing parameters.
Konnichiwa,
I wish that was the biggest problem with single drivers. Sadly, it is not the biggest problem.
In fact, it's not even a big problem. It's not even a small problem. It's problem that actually is a VERY, VERY small one. It's so small, for all intents and purposes, it can be safely ignored.
So, please ignore the little man behind the curtain.
BTW, reproducing the midrange through a driver that has it's cone moving significantly still causes problems, but they are down to magnet design and other such electrical and electromechanical non-linearities, not down to "doppler distrtion".
Sayonara
eric180db said:
I wish that was the biggest problem with single drivers. Sadly, it is not the biggest problem.
In fact, it's not even a big problem. It's not even a small problem. It's problem that actually is a VERY, VERY small one. It's so small, for all intents and purposes, it can be safely ignored.
So, please ignore the little man behind the curtain.
BTW, reproducing the midrange through a driver that has it's cone moving significantly still causes problems, but they are down to magnet design and other such electrical and electromechanical non-linearities, not down to "doppler distrtion".
Sayonara
Konnichiwa,
Well, an enclosure like this will not solve many of the problems (there will remain a 10db tilt upwards at high frequencies) but it will now introduce at least one major one.
Namely a severe LF power handling problem, making the use of a Highpass at around 80...100Hz mandatory. The highpass needs to be probably at least 2nd order, so now we have the 6th order Highpass function around 80Hz. That is most definitly NOT GOOD.
If you wher eto use a small sealed enclosure and a 2nd order Highpass with some boost at the corner frequency you would have a much better result while keeping the rolloff order at a sensible point.
I would still think that using a larger sealed enclosure (say 40 Liter floor standing column) together with Nelsons suggested principle of a high output impedance amplifier makes a lot more sense, as it will actually turn the system into a true fullrange system with pretty good overall response.
While obviously reviled among the Pass Labs Boards contributors, the kind of Chipamp often known as "Gainclone" can be used to make a transconductance amplifier (current output) on a shoestring and there it is also possible to incorporate the EQ networks into the feedback loop, making the whole excercise quite inexensive to try.
Sayonara
Geek said:Try this
Well, an enclosure like this will not solve many of the problems (there will remain a 10db tilt upwards at high frequencies) but it will now introduce at least one major one.
Namely a severe LF power handling problem, making the use of a Highpass at around 80...100Hz mandatory. The highpass needs to be probably at least 2nd order, so now we have the 6th order Highpass function around 80Hz. That is most definitly NOT GOOD.
If you wher eto use a small sealed enclosure and a 2nd order Highpass with some boost at the corner frequency you would have a much better result while keeping the rolloff order at a sensible point.
I would still think that using a larger sealed enclosure (say 40 Liter floor standing column) together with Nelsons suggested principle of a high output impedance amplifier makes a lot more sense, as it will actually turn the system into a true fullrange system with pretty good overall response.
While obviously reviled among the Pass Labs Boards contributors, the kind of Chipamp often known as "Gainclone" can be used to make a transconductance amplifier (current output) on a shoestring and there it is also possible to incorporate the EQ networks into the feedback loop, making the whole excercise quite inexensive to try.
Sayonara
Kuei Yang Wang said:
At Elliot Sound Products page there is an interesting example of
variable impedance amplifier
Thanks, but I would rather prefer a sealed than ported...
In fact, as neither of them are suitable for FE 206E, and a horn can be quite difficult to build, I am wondering if the transmission line will work? Some folks have built Martin J. King's "Virtual Fostex FE-206E and FE-207E ML TL Design" and they seem to be satisfied with the result...
I did not mention it in the first post: I am also using two DIY sealed subwoofers, each using 12" driver, whith the active, equalized crossover at about 65 Hz. In fact, it is controlled by Rod Eliott's "Subwoofer Processor-P.48". So it seems that in this arrangement Fostex can work in a big sealed enclosure...
Regards
Vix
In fact, as neither of them are suitable for FE 206E, and a horn can be quite difficult to build, I am wondering if the transmission line will work? Some folks have built Martin J. King's "Virtual Fostex FE-206E and FE-207E ML TL Design" and they seem to be satisfied with the result...
I did not mention it in the first post: I am also using two DIY sealed subwoofers, each using 12" driver, whith the active, equalized crossover at about 65 Hz. In fact, it is controlled by Rod Eliott's "Subwoofer Processor-P.48". So it seems that in this arrangement Fostex can work in a big sealed enclosure...
Regards
Vix
Konnichiwa,
Not much better. The problem is the Driver.
That particular design, if driven from a Voltage source will be balanced ONLY EVER on a Front horn or with severe EQ.
The TL can offer a little more bass than a sealed box by virtu of adding resonant reinforcement of the sound, but it can only do so much.
No, the Fostex cannot work in a big sealed enclosure as there will remain a 20db difference between the level at 100Hz and at 10KHz. Once you have taken measures (whatever they may be) to equalise the response between 100Hz & 10KHz you can think of extending the LF with a subwoofer. But the main task remains to address a 20db level increase between 100Hz and 10KHz!
Sayonara
Vix said:
Not much better. The problem is the Driver.
That particular design, if driven from a Voltage source will be balanced ONLY EVER on a Front horn or with severe EQ.
The TL can offer a little more bass than a sealed box by virtu of adding resonant reinforcement of the sound, but it can only do so much.
Vix said:
No, the Fostex cannot work in a big sealed enclosure as there will remain a 20db difference between the level at 100Hz and at 10KHz. Once you have taken measures (whatever they may be) to equalise the response between 100Hz & 10KHz you can think of extending the LF with a subwoofer. But the main task remains to address a 20db level increase between 100Hz and 10KHz!
Sayonara
"but I don't understand any Japanese"
I don't either.. it was more of a visual representation to what I was suggesting.
But just from the look of it - it appears to be a TQWT design because of the "pyramid expansion". In fact if your NOT shooting for a low freq. response then a TQWT should be able to lift the lower-end of the baffle loss region similarly to a back-horn because the bandwidth of the TQWT's resonance is broader than something like a pipe or a passive radiator. Additionally there should be a freq. balance shift just by firing the driver upward that would decrease the driver's spl as freq. increases - assuming your listening at close to 180 deg.s off axis then I would suspect that "shift" occurs near the baffle-loss region. Of course this is all speculation, BUT if you have a 206E then you can always load the thing in a cheap sealed box, aim it upward and measure it to see how it affects the freq. balance. (..of course without the reflector the high freq.s will be significantly reduced above 1kHz.)
I don't either.. it was more of a visual representation to what I was suggesting.
But just from the look of it - it appears to be a TQWT design because of the "pyramid expansion". In fact if your NOT shooting for a low freq. response then a TQWT should be able to lift the lower-end of the baffle loss region similarly to a back-horn because the bandwidth of the TQWT's resonance is broader than something like a pipe or a passive radiator. Additionally there should be a freq. balance shift just by firing the driver upward that would decrease the driver's spl as freq. increases - assuming your listening at close to 180 deg.s off axis then I would suspect that "shift" occurs near the baffle-loss region. Of course this is all speculation, BUT if you have a 206E then you can always load the thing in a cheap sealed box, aim it upward and measure it to see how it affects the freq. balance. (..of course without the reflector the high freq.s will be significantly reduced above 1kHz.)
what about using an ELS in a hybrid system has many commerical units ?
the crossover point can be made really low ( sub 200hz )
and the esl makes for a "full range" driver from 200hz to supersonic freq. ??
i also guess that an esl doesn't suffer from much of the cone drivers problems ? aside from dipole wich can be work around with a semi closed box ( wings? )
hope i'm not too off subject here
the crossover point can be made really low ( sub 200hz )
and the esl makes for a "full range" driver from 200hz to supersonic freq. ??
i also guess that an esl doesn't suffer from much of the cone drivers problems ? aside from dipole wich can be work around with a semi closed box ( wings? )
hope i'm not too off subject here
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.