Hi guys, my first time posting here...
I'm a relative newbie to speaker design, i've built a couple of speakers before,
but they were all pretty basic, using small fullrange drivers - never had the
room or money (or tolerant neighbours, for that matter) to undertake the task of
building a more serious pair of speakers.
After spending the last few years setting up a pretty decent headphone system,
i've recently moved in a new home, and i'm itching to build something nice for
my living room!
I've wanted to try out some open baffles ever since i first learned about them -
i guess i find the concept kinda similar to using open-back headphones, which i
am in total love with, can't stand closed back headphones at all... Even on the
good ones i find the sound awkwardly resonant and the compressed soundstage
drives me nuts! Maybe I'm spoiled by my AKG K501s, haha...
So, to the point. I've had a pair of Fostex FE103E laying around, so i decided
to start off with them for mids and highs. After considering just building MJK's
project with the FE103 and the Alpha 15A, i quickly realized i wanted a bigger
challenge. So i went searching for another option for the bottom-end. My search
came to an end when i looked through Visaton's catalog - their WS25E looked
perfect on paper - it's smaller in diameter, which would probably let me use a
narrower baffle, and has a high Q factor - 1.4. Visaton themselves have it
featured in a pretty interesting ripole design, and the german hi-fi magazine
Klang&Ton built a completely baffle-less speaker with apparently some pretty
decent results. They were smaller than the Alpha15, but i figured i'd just use
two per side to compensate. So, the decision was made and i ordered 4 Visatons.
While i waited for delivery i went to use MJK's worksheets, spending many
sleepless nights trying to work out the details. Baffle size and driver
positioning was relatively easy, but crossovers? Having only dealt with
headphones and fullranges in the past, i'm a complete newbie to crossover
design. I've read a lot of info (especially in the last few days), but having
never actually put one together i still consider myself a total noob, and to be
honest, i'm a bit overwhelmed by the amount of details one has to consider...
I played around with different filter types and component values till i came up
with something that looked OK on the graphs... Or at least it looked that way to
me.
Basically, i have a few worries:
1. I don't trust my own judgement about the graphs. Most look OK to me, but i
may be missing something...
2. I need to pad down the FE103E a few db, which i did not know how to simulate
in the worksheets. The only thing that occured to me was that i could add a
single series resistor right before the driver (after the high-pass filter). I
could then add it's value to the speaker's Re in the worksheet. I came up with a
value of 8.2 Ohms for the resistor, added that to the driver's Re and sure
enough, the response curve reacted as expected, dropping down the desired amount
of db. Will this happen the same way in real life and are there any downsides?
3. I'm worried i'll be having a ~16ohm load for the mids and highs and ~4ohms
for the lows, wired in parallel... Doing the math, this will drop the overall
impedance quite low - my amp (NAD 3150) will probably still handle it, the
manual states it's good down to 2 ohms... But will this sound as OK, as the
simulation seems to suggest?
3. Again, I'm not entirely sure how to interpret the graphs under the "OB driver
impedance" section... Are these graphs for the entire system - baffle, drivers,
crossovers...? How are they looking, good, ok or horrible...? Trying different
order filters i managed to make these look curly as hell, so i guess these are
not too bad, are they?
4. Oh, another thing - according to the simulation, i need to reverse the phase
on one driver to avoid a dip at the crossover point. From what i understand this
is normal for a lot of speakers using 2nd order filters - to me this sounds like
it won't hurt performance in the other parts of the spectrum, correct?
Basically the Visatons have arrived and i'm itching to throw a quick prototype
together, but i'm not confident enough in my simulation to go out and buy all
the parts for the filters - i'm afraid i might have done something totally wrong
and i'd end up with a bunch of not very cheap parts that are useless...
I've exported my simulations as a HTML file, here's the link:
http://www.bikepornmag.com/agentorange/1xFE103E_2xWS25E.html
I'd appreciate someone more experienced giving it a look!
Cheers guys,
George
I'm a relative newbie to speaker design, i've built a couple of speakers before,
but they were all pretty basic, using small fullrange drivers - never had the
room or money (or tolerant neighbours, for that matter) to undertake the task of
building a more serious pair of speakers.
After spending the last few years setting up a pretty decent headphone system,
i've recently moved in a new home, and i'm itching to build something nice for
my living room!
I've wanted to try out some open baffles ever since i first learned about them -
i guess i find the concept kinda similar to using open-back headphones, which i
am in total love with, can't stand closed back headphones at all... Even on the
good ones i find the sound awkwardly resonant and the compressed soundstage
drives me nuts! Maybe I'm spoiled by my AKG K501s, haha...
So, to the point. I've had a pair of Fostex FE103E laying around, so i decided
to start off with them for mids and highs. After considering just building MJK's
project with the FE103 and the Alpha 15A, i quickly realized i wanted a bigger
challenge. So i went searching for another option for the bottom-end. My search
came to an end when i looked through Visaton's catalog - their WS25E looked
perfect on paper - it's smaller in diameter, which would probably let me use a
narrower baffle, and has a high Q factor - 1.4. Visaton themselves have it
featured in a pretty interesting ripole design, and the german hi-fi magazine
Klang&Ton built a completely baffle-less speaker with apparently some pretty
decent results. They were smaller than the Alpha15, but i figured i'd just use
two per side to compensate. So, the decision was made and i ordered 4 Visatons.
While i waited for delivery i went to use MJK's worksheets, spending many
sleepless nights trying to work out the details. Baffle size and driver
positioning was relatively easy, but crossovers? Having only dealt with
headphones and fullranges in the past, i'm a complete newbie to crossover
design. I've read a lot of info (especially in the last few days), but having
never actually put one together i still consider myself a total noob, and to be
honest, i'm a bit overwhelmed by the amount of details one has to consider...
I played around with different filter types and component values till i came up
with something that looked OK on the graphs... Or at least it looked that way to
me.
Basically, i have a few worries:
1. I don't trust my own judgement about the graphs. Most look OK to me, but i
may be missing something...
2. I need to pad down the FE103E a few db, which i did not know how to simulate
in the worksheets. The only thing that occured to me was that i could add a
single series resistor right before the driver (after the high-pass filter). I
could then add it's value to the speaker's Re in the worksheet. I came up with a
value of 8.2 Ohms for the resistor, added that to the driver's Re and sure
enough, the response curve reacted as expected, dropping down the desired amount
of db. Will this happen the same way in real life and are there any downsides?
3. I'm worried i'll be having a ~16ohm load for the mids and highs and ~4ohms
for the lows, wired in parallel... Doing the math, this will drop the overall
impedance quite low - my amp (NAD 3150) will probably still handle it, the
manual states it's good down to 2 ohms... But will this sound as OK, as the
simulation seems to suggest?
3. Again, I'm not entirely sure how to interpret the graphs under the "OB driver
impedance" section... Are these graphs for the entire system - baffle, drivers,
crossovers...? How are they looking, good, ok or horrible...? Trying different
order filters i managed to make these look curly as hell, so i guess these are
not too bad, are they?
4. Oh, another thing - according to the simulation, i need to reverse the phase
on one driver to avoid a dip at the crossover point. From what i understand this
is normal for a lot of speakers using 2nd order filters - to me this sounds like
it won't hurt performance in the other parts of the spectrum, correct?
Basically the Visatons have arrived and i'm itching to throw a quick prototype
together, but i'm not confident enough in my simulation to go out and buy all
the parts for the filters - i'm afraid i might have done something totally wrong
and i'd end up with a bunch of not very cheap parts that are useless...
I've exported my simulations as a HTML file, here's the link:
http://www.bikepornmag.com/agentorange/1xFE103E_2xWS25E.html
I'd appreciate someone more experienced giving it a look!
Cheers guys,
George
I would most recommend looking for OB kit with MiniDSP. e.g Musicanddesign's Nao or Nao Note.
1. Someone else had done the hard work of measurements and optimal shapes
2. With DSP it's future proof as better design appears.
The only compromise is that you'd need 3 or 4 amps (which are dead cheap nowadays).
If you want to keep using the Fostex/Eminence, MiniDSP is still a winner too. You will not care about impedance nor sensitivity and able to apply powerful EQ easily. Plus other goodies like time alignment.
1. Someone else had done the hard work of measurements and optimal shapes
2. With DSP it's future proof as better design appears.
The only compromise is that you'd need 3 or 4 amps (which are dead cheap nowadays).
If you want to keep using the Fostex/Eminence, MiniDSP is still a winner too. You will not care about impedance nor sensitivity and able to apply powerful EQ easily. Plus other goodies like time alignment.
Yeah I admit i've really thought hard about going active & digital, but i came to the conclusion i wanna avoid buying a second amp at this point... Although driving the Fostex with a tube amp will surely sound sweet.
The MiniDSP and other dsp's do look good, even just for the ease of use and flexibility, but i'd rather try out a passive crossover first. I've never designed one and i don't really wanna back out from that challenge haha.
Then, if this project turns out OK i can surely see myself going digital a bit further down the road, when i can afford a good tube amp for driving the Fostex...
The MiniDSP and other dsp's do look good, even just for the ease of use and flexibility, but i'd rather try out a passive crossover first. I've never designed one and i don't really wanna back out from that challenge haha.
Then, if this project turns out OK i can surely see myself going digital a bit further down the road, when i can afford a good tube amp for driving the Fostex...
George,
I can't read that html. There are no input values and diagrams in it, so I can't comment. But I recommend to use the H frame worksheet and to look at MJKs "Two driver" PDF for more info how to simulate two drivers in one H frame. A decent H frame for the WS25E would still allow a crossover at up to 400 Hz, but will narrow the efficiency gap between fullrange driver and woofers somewhat.
Rudolf
I can't read that html. There are no input values and diagrams in it, so I can't comment. But I recommend to use the H frame worksheet and to look at MJKs "Two driver" PDF for more info how to simulate two drivers in one H frame. A decent H frame for the WS25E would still allow a crossover at up to 400 Hz, but will narrow the efficiency gap between fullrange driver and woofers somewhat.
Rudolf
Rudolph, check that link, I made it into a PDF:
http://www.bikepornmag.com/agentorange/1xFE103E_2xWS25E.pdf
H-frames - thanks, i'll look into that option as well... The idea is to avoid having to pad down the Fostex, correct?
As for the crossover point - I've tried to keep it as low as possible, in fear of messing up the lower mid-range... Currently it's around 250Hz.
http://www.bikepornmag.com/agentorange/1xFE103E_2xWS25E.pdf
H-frames - thanks, i'll look into that option as well... The idea is to avoid having to pad down the Fostex, correct?
As for the crossover point - I've tried to keep it as low as possible, in fear of messing up the lower mid-range... Currently it's around 250Hz.
Thanks. Now I can see it. I'll come back to you with a H-frame proposition
There is no way to avoid it completely, if you insist on one amp for all drivers.
Do whatever keeps the Fostex in its linear range 🙂
This is my try with a H-frame for the WS25E:
http://rudolffinke.homepage.t-online.de/audio/Dipol/WS25E/H_Frame_FE103E_WS25E.pdf
You may change the listening distance to 2 m and the input power to 10 W (same for your OB design) to see the advantage of a higher crossover. Sadly I had to increase the resistor value in front of the Fostex even more. 😱
The SPL hump around 1 kHz will be reduced, if you don't listen on axis.
Rudolf
http://rudolffinke.homepage.t-online.de/audio/Dipol/WS25E/H_Frame_FE103E_WS25E.pdf
You may change the listening distance to 2 m and the input power to 10 W (same for your OB design) to see the advantage of a higher crossover. Sadly I had to increase the resistor value in front of the Fostex even more. 😱
The SPL hump around 1 kHz will be reduced, if you don't listen on axis.
Rudolf
Looks good, definitely more bottom end than the basic OB version! Thanks for the effort Rudolf, much appreciated. I'll try raising the input power in the worksheets tonight, thanks for the advice.
I got a response from MJK over at the yahoo discussion group for his worksheets that i can't really simulate a series resistor in the 3 driver OB worksheet, as the added resistance has an effect on the overall Q factor... I can see the formula is there on the H-frame worksheets though, maybe i can try working out adjusted Q for the Fostex from this worksheet?
I knocked together a quick prototype baffle with only the Fostex driver hooked up to check if the 1kHz hump will be noticable, it certainly isn't much of a problem in my room, response sounds relatively smooth, speakers are slightly toed in, but not much - the main listening position is a bit off-axis for both speakers.
The OBs have a very curious way of interacting with the room, much different to my box speakers. Several new issues have been noticed, but i'll try to finish up the prototype speakers before i start thinking about room acoustics...
I got a response from MJK over at the yahoo discussion group for his worksheets that i can't really simulate a series resistor in the 3 driver OB worksheet, as the added resistance has an effect on the overall Q factor... I can see the formula is there on the H-frame worksheets though, maybe i can try working out adjusted Q for the Fostex from this worksheet?
I knocked together a quick prototype baffle with only the Fostex driver hooked up to check if the 1kHz hump will be noticable, it certainly isn't much of a problem in my room, response sounds relatively smooth, speakers are slightly toed in, but not much - the main listening position is a bit off-axis for both speakers.
The OBs have a very curious way of interacting with the room, much different to my box speakers. Several new issues have been noticed, but i'll try to finish up the prototype speakers before i start thinking about room acoustics...
Ok, after a week of fiddling around with different designs in MJK's worksheets i came to this design:
http://www.bikepornmag.com/agentorange/1xFE103E_2xWS25E Copy.pdf
It's still open baffle (no H-frames or U-frames) with pretty much the same dimensions as my initial design, but the final response plot looks pretty flat, and the rest of the plots look decent as well...
The tricky part is that to get the drivers to sum flat in the final response curve i had to use a 3rd order high-pass filter instead of the 2nd order i was trying with before. I'm currently trying to read up on the topic of mixing up filter orders like that, but the simulations don't seem to suggest anything too bad besides a low-ish speaker efficiency, which i'm not too worried about, i don't listen at high volumes at all, and i'll be using a decent solid state amp which can supposedly drive low-sensitivity speakers with ease...
Of course, actual in-room response might be totally different, but at this point, if i haven't done something awfully wrong in the simulations, i'm willing to throw a couple of prototypes together and give them a listen.
Should i go ahead and try it or I'm missing something big here?
http://www.bikepornmag.com/agentorange/1xFE103E_2xWS25E Copy.pdf
It's still open baffle (no H-frames or U-frames) with pretty much the same dimensions as my initial design, but the final response plot looks pretty flat, and the rest of the plots look decent as well...
The tricky part is that to get the drivers to sum flat in the final response curve i had to use a 3rd order high-pass filter instead of the 2nd order i was trying with before. I'm currently trying to read up on the topic of mixing up filter orders like that, but the simulations don't seem to suggest anything too bad besides a low-ish speaker efficiency, which i'm not too worried about, i don't listen at high volumes at all, and i'll be using a decent solid state amp which can supposedly drive low-sensitivity speakers with ease...
Of course, actual in-room response might be totally different, but at this point, if i haven't done something awfully wrong in the simulations, i'm willing to throw a couple of prototypes together and give them a listen.
Should i go ahead and try it or I'm missing something big here?
Have you tried changing the input power in the worksheet to see how loudly it will play before you run out of either excursion or power?
Have you tried offsetting the Fostex to one side of the baffle? You may be able to reduce the ripple some that way. It might be faster to play with the driver position in The Edge, then transfer the results back to the worksheet.
Bill
Have you tried offsetting the Fostex to one side of the baffle? You may be able to reduce the ripple some that way. It might be faster to play with the driver position in The Edge, then transfer the results back to the worksheet.
Bill
Resurrection 🙂 Agentini, have you progressed on this design yet? Last night I was simming those WS25E's paired with Vifa TC9FD18-08, in MJK's sheets. Got a pretty decent results with two woofers and one FR. Unfortunately there is no sheet for "FR and two woofers in u-frame" to compare. Mock-up baffle is easy to build and adding a u-frame afterwords isn't hard to accomplish. Just not sure if should psuh the button and suffer (oh so much as) 70euros for pair of Visatons.
Look for http://www.quarter-wave.com/General/Two_Drivers.pdf
If your u-frame features identical conditions for both drivers, the advice in that pdf should work pretty well. 🙂
Rudolf
Yes, I went through those and found a way (halve Re and impedance, double Vas and Sd). In the first sims, it seems that open baffle behaves better, u-frame having few dBs lower output in the low end. I might have to run few more since those were sort of quick and dirty try outs.
For the price WS25e is tempting, though it's not as versatile as Peerless SLS 10. But sls10 seems to need a lot more boost inthe low end. At least according to sims. Unfortunately you can't have everything for price of few beer cases 🙂
For the price WS25e is tempting, though it's not as versatile as Peerless SLS 10. But sls10 seems to need a lot more boost inthe low end. At least according to sims. Unfortunately you can't have everything for price of few beer cases 🙂
- Status
- Not open for further replies.