830875 (peerless) Fs 46Hz, Qts 0.36 VAS 21.78L
FF105WK (fostex) Fs 75Hz Qts 0.41 VAS 4.848L
Efficiency for both 88-89dB, I expect ~85dB after baffle step correction.
I'm building this for a friend. Exactly 67cm height is the only fixed part of the design (to match other furniture). This is a touch low for a floor stander, so most listening will automatically be a bit off axis (which works out well for this Fostex range, which all have an on-axis spike around 8kHz). She'd be happy if it looked something like this:
http://www.diyaudio.com/forums/multi-way/147000-another-peerless-830875-execution.html
I plan to use the Peerless below the baffle step (300-500Hz, depending on my baffle). I'd like to use a BR or some other form of vented enclosure.
The Fostex will probably be sealed ~2L, with minimal crossover (maybe a ~8ohm resistor paralleled by a small cap, to bring the 500-10,000Hz range down to 85dB dB).
The 67cm height gives me a lot of box to play with, far more than I need for a simple bass reflex. I was hoping someone could suggest a build (with dimensions) that would make good use of this size, e.g. a 2m long, 4 fold transmission line / TQWP, or possibly a big vent reflex - neither of which I've built before.
I have a multimeter, SPL meter, can generate test tones, have wool, tools, woodworking experience, a crossover parts pile, so I could measure and tweak port lengths, components and stuffing densities to get it right.
A big front port / BVR mouth is fine visually - in that case I'd possibly use a panel of (black) perforated steel as the front grill, to keep the presentation sleek and uncluttered (I'm considering it anyway, to get some scattering of HF from the grill).
FF105WK (fostex) Fs 75Hz Qts 0.41 VAS 4.848L
Efficiency for both 88-89dB, I expect ~85dB after baffle step correction.
I'm building this for a friend. Exactly 67cm height is the only fixed part of the design (to match other furniture). This is a touch low for a floor stander, so most listening will automatically be a bit off axis (which works out well for this Fostex range, which all have an on-axis spike around 8kHz). She'd be happy if it looked something like this:
http://www.diyaudio.com/forums/multi-way/147000-another-peerless-830875-execution.html
I plan to use the Peerless below the baffle step (300-500Hz, depending on my baffle). I'd like to use a BR or some other form of vented enclosure.
The Fostex will probably be sealed ~2L, with minimal crossover (maybe a ~8ohm resistor paralleled by a small cap, to bring the 500-10,000Hz range down to 85dB dB).
The 67cm height gives me a lot of box to play with, far more than I need for a simple bass reflex. I was hoping someone could suggest a build (with dimensions) that would make good use of this size, e.g. a 2m long, 4 fold transmission line / TQWP, or possibly a big vent reflex - neither of which I've built before.
I have a multimeter, SPL meter, can generate test tones, have wool, tools, woodworking experience, a crossover parts pile, so I could measure and tweak port lengths, components and stuffing densities to get it right.
A big front port / BVR mouth is fine visually - in that case I'd possibly use a panel of (black) perforated steel as the front grill, to keep the presentation sleek and uncluttered (I'm considering it anyway, to get some scattering of HF from the grill).
We have done a couple boxes with the smaller FF85wk and multiples of the smaller 830870.
Ideally you want the woofer(s) to be 3-6 dB more sensitive than the mid-tweeter, and given the XO point, the better performance of the smaller fostex in this range, and its lower sensitiviry you should consider switching from 105 to 85.
I haven't run any sims on the 830875, but the smaller 4" wants a relativily large reflex box (we used them in scottmoose quarter-wave lines), so when we were doing the small stand-mount we went with a pair of Silver Flute W14 to match with the FF85 (round2 of passive XO develpment on those)
WRT the height, we have added a 5 degree tiltback of the baffle to solve some of the issue of short floorstanders.
dave
Ideally you want the woofer(s) to be 3-6 dB more sensitive than the mid-tweeter, and given the XO point, the better performance of the smaller fostex in this range, and its lower sensitiviry you should consider switching from 105 to 85.
I haven't run any sims on the 830875, but the smaller 4" wants a relativily large reflex box (we used them in scottmoose quarter-wave lines), so when we were doing the small stand-mount we went with a pair of Silver Flute W14 to match with the FF85 (round2 of passive XO develpment on those)
WRT the height, we have added a 5 degree tiltback of the baffle to solve some of the issue of short floorstanders.
dave
Thanks for the reply Dave. I already have both drivers, don't wish to change up at this point. I appreciate that the smaller Fostex makes a better mid tweeter, but went for the 105 because it should be a lot more robust (higher excursion).
Re the tilt, I'm actually seeing the low height / being off axis as a win. They seem to measure better off axis than on. If I pad the Fostex down by 3 dB or more, and bypass that with a small cap, hopefully I can maintain some HF sparkle.
So it's just maximising the LF (and trying a new enclosure type) that I'm after.
Re the tilt, I'm actually seeing the low height / being off axis as a win. They seem to measure better off axis than on. If I pad the Fostex down by 3 dB or more, and bypass that with a small cap, hopefully I can maintain some HF sparkle.
So it's just maximising the LF (and trying a new enclosure type) that I'm after.
I'd still go for a bit of slant on the baffle, you can do the off-axis thing by playing with toe-in, much more flexible ultimately IMO.
I've a speaker based on the larger 8" Nomex, 830869, good midbass with good midrange too. I almost did something with the 6.5" as well.
As P10 said, the main issue will be the fact both drivers are about as sensitive. I never did the FR+midbass thing - would like to though - so could not advise on the best way to deal with this passively. I suppose a L-pad made of two resistors could be a start, I see you already planned to swamp the FF105WK to deal with its Z peak, so you'll have this there. P10 or others might advise better on this point.
Still, do make the crossover frequency coincide with baffle-step as much as possible. This would be 343/width/3, width in meter.
I've a speaker based on the larger 8" Nomex, 830869, good midbass with good midrange too. I almost did something with the 6.5" as well.
As P10 said, the main issue will be the fact both drivers are about as sensitive. I never did the FR+midbass thing - would like to though - so could not advise on the best way to deal with this passively. I suppose a L-pad made of two resistors could be a start, I see you already planned to swamp the FF105WK to deal with its Z peak, so you'll have this there. P10 or others might advise better on this point.
Still, do make the crossover frequency coincide with baffle-step as much as possible. This would be 343/width/3, width in meter.
Quick check in WinISD, the 830875's suggested QB3 looks pretty good to me right off the bat. ~15L tuned to 50Hz. I would not tune much above this, but 45Hz, or possibly down to 40Hz could be good depending on the taste of your friend. You could start with a 2" round vent, 8.75" long for Fb~40Hz. If too lean, trim to 6.625" for ~45Hz, or 5" for ~50Hz.
Tuning to 50Hz would be good for ~8W down to 30Hz. Tuning to 40Hz, ~16W down to 30Hz. SPL would be ~98dB and 101dB respectively. I don't think you'd want to take the FF105WK much higher than the high 90ies anyway.
Tuning to 50Hz would be good for ~8W down to 30Hz. Tuning to 40Hz, ~16W down to 30Hz. SPL would be ~98dB and 101dB respectively. I don't think you'd want to take the FF105WK much higher than the high 90ies anyway.
IME a swamping resistor is not going to be nearly enough to make a highpass cap work. Been there, failed that. It will still track impedance. An aperiodic enclosure of some sort will be necessary, or a full impedance compensation series trap. (expensive)
I agree that you would be better off with the FF85wk. If you're stuck on the FF105, then go up to an ~8" helper woofer. A 10 would not be unreasonable.
+1 on the 3-6db difference. Usually 4 or 5 is just right.
I agree that you would be better off with the FF85wk. If you're stuck on the FF105, then go up to an ~8" helper woofer. A 10 would not be unreasonable.
+1 on the 3-6db difference. Usually 4 or 5 is just right.
I comeup with 14 litre tuned to 60 Hz, with the bumpknocked down with a high ratio slot vent. Good to about 40 Hz F10.
dave
dave
Thanks all for contributing.
I feel like perhaps I didn't explain my intentions properly.
My plan is to optimise the Fostex for off-axis listening. This is to avoid its on-axis spike at 8 kHz*
Off axis, the FF105WK rolls off >10 kHz. I intend to use simple eq to pad it down 3dB (or more) below 10kHz.
Therefore the Fostex driver's midband efficiency in my application will be well below the Peerless' midband efficiency. Moving up to an 8" or 10" bass driver is tempting for extra LF extension, but not really needed for efficiency matching (and also bad for my budget)**
Details will be something like this:
Peerless woofer in vented box, lowpassed 300-500Hz (baffle step frequency).
--> starting point will be 5.6mH coil (because I have some already), 30uF cap, with expected passband efficiency ~85dB
Fostex in sealed box. No large cap to in the 300-500Hz range: the driver / enclosure will be rolling off anyway, plus the sealed box will limit its excursion. I think this is what Greg was thinking of when he referred to problems with a highpass cap.
As I stated in my first post, my starting minimal crossover will be a 8ohm resistor, paralleled with a small cap (maybe 1.5uF) to bring the Fostex's response down to ~85dB, while allowing the top octave through.
With a 67cm tall box, and "typical" other dimensions (20-30cm wide, 20-30cm deep), I could have as much as 50L of internal volume to play with, which is triple what the Peerless needs for a typical BR. This is why I'm calling for suggestions for atypical enclosures, such as a 4-fold TL or BVR.
*I noticed, when picking my poison, that all the FF range have an on-axis spike (it looks like most Fostex drivers do). The Fostex graphs often underplay the peaks, I've seen them measured elsewhere at +13dB. The FF85WK has its peak at ~12 kHz, and very low excursion, which is what decided me on the 105, as my most acceptable set of compromises.
**although I do have some other drivers I could try if the Fostex / Peerless combo fails miserably.
I feel like perhaps I didn't explain my intentions properly.
My plan is to optimise the Fostex for off-axis listening. This is to avoid its on-axis spike at 8 kHz*
Off axis, the FF105WK rolls off >10 kHz. I intend to use simple eq to pad it down 3dB (or more) below 10kHz.
Therefore the Fostex driver's midband efficiency in my application will be well below the Peerless' midband efficiency. Moving up to an 8" or 10" bass driver is tempting for extra LF extension, but not really needed for efficiency matching (and also bad for my budget)**
Details will be something like this:
Peerless woofer in vented box, lowpassed 300-500Hz (baffle step frequency).
--> starting point will be 5.6mH coil (because I have some already), 30uF cap, with expected passband efficiency ~85dB
Fostex in sealed box. No large cap to in the 300-500Hz range: the driver / enclosure will be rolling off anyway, plus the sealed box will limit its excursion. I think this is what Greg was thinking of when he referred to problems with a highpass cap.
As I stated in my first post, my starting minimal crossover will be a 8ohm resistor, paralleled with a small cap (maybe 1.5uF) to bring the Fostex's response down to ~85dB, while allowing the top octave through.
With a 67cm tall box, and "typical" other dimensions (20-30cm wide, 20-30cm deep), I could have as much as 50L of internal volume to play with, which is triple what the Peerless needs for a typical BR. This is why I'm calling for suggestions for atypical enclosures, such as a 4-fold TL or BVR.
*I noticed, when picking my poison, that all the FF range have an on-axis spike (it looks like most Fostex drivers do). The Fostex graphs often underplay the peaks, I've seen them measured elsewhere at +13dB. The FF85WK has its peak at ~12 kHz, and very low excursion, which is what decided me on the 105, as my most acceptable set of compromises.
**although I do have some other drivers I could try if the Fostex / Peerless combo fails miserably.
2.5 litre net will give a butterworth at about 120-130 Hz. Too low for a corner on your XO. And you'll need to deal with theimpedance peak.
I always like an aperiodic midTL as they help flatten the impedance peak making it easier to get a passive XO to work.
Normalized for your nominal impedance, we found that 6mH/30 uF series worked pretty well with previous gen FF85 and EL166.
It should also be noted that to "XO at the baffle step" somewhere between 0.707 and 1 x the nominal BS is the range to explore.
dave
I always like an aperiodic midTL as they help flatten the impedance peak making it easier to get a passive XO to work.
Normalized for your nominal impedance, we found that 6mH/30 uF series worked pretty well with previous gen FF85 and EL166.
It should also be noted that to "XO at the baffle step" somewhere between 0.707 and 1 x the nominal BS is the range to explore.
dave
Attachments
Planet10,
Any tips on sizing a stuffed TL for either FF105WK or FF85WK? I might actually be doing this with FF85WK, as I have two 6.5" midbass drivers forcing me to make a 180° on their intended use. I'm thinking a 4" ABS pipe. I would prefer for the cabinet external depth not to be over ~8", is that too short? Polyfill any good? I assume a fairly high density, progressive? Please do spill all your trade secrets. 😀 😉
Naturally I'll be measuring if get down to doing this. 🙂
Hollowboy - not trying to hijack your thread, the info will be relevant to you as well. 😉
Any tips on sizing a stuffed TL for either FF105WK or FF85WK? I might actually be doing this with FF85WK, as I have two 6.5" midbass drivers forcing me to make a 180° on their intended use. I'm thinking a 4" ABS pipe. I would prefer for the cabinet external depth not to be over ~8", is that too short? Polyfill any good? I assume a fairly high density, progressive? Please do spill all your trade secrets. 😀 😉
Naturally I'll be measuring if get down to doing this. 🙂
Hollowboy - not trying to hijack your thread, the info will be relevant to you as well. 😉
Doesn't that describe the Woden labyrinth series for which "we" built protos a couple of months back? Not exactly a highly stuffed "aperiodic" TL such as in the midrange of the current SilverFlute/ FF85WK, MTMs, or the floorstanding EL166/FF85K of a couple of years ago, but still quite effective .
I think IG is asking what the minimum TL size requirements are for midrange/HF use only. Good question, and I wish I knew the answer.
It would seem that you would want the 1/4w stub frequency to be at least as low as your crossover, but I guess it could be fudged shorter. OTOH, a 300hz pipe is only 11.25" long (depending on your altitude... 😀)
I suspect stuffing near the exit of a pipe really does the most in flattening impedance - which is of course precisely where you don't want stuffing in a full range TL. It would be worth some experimentation.
It would seem that you would want the 1/4w stub frequency to be at least as low as your crossover, but I guess it could be fudged shorter. OTOH, a 300hz pipe is only 11.25" long (depending on your altitude... 😀)
I suspect stuffing near the exit of a pipe really does the most in flattening impedance - which is of course precisely where you don't want stuffing in a full range TL. It would be worth some experimentation.
Our midTLs so far have not really been designed, but fitted into the space available.
The 3 FF85 (2 x FF85k, 1 x FF85wk) have been end-loaded ~ 10:1 tapered TLs as long as the box is deep, which has been on the order of 7-8". (Tysen, EL166 MTM ML-TL, Silver Flute W14 stand-mount MTM.
The A7.3 midTL in the elipsa is a bit more complex as the available shape was a slice of the long truncared elliptical cross-section of the box with the driver nominally at the 50% mark. The end near the terminus has a couple partitions ro make that side's length effectively longer & narrower.
Key is the stuffing. It gets denser towards the terminus. If you want to be thorough you can measure the impedance looking for the lowest resonance peaks. I did this with a 4" PVC tube and Alpair 5 before deciding that taper was a good idea.
More can be seen in the Tysen FAST thread.
dave
Please, hijack away, it is all of interest.
A random driver I pulled out of a TV measures flatter and sounds a little easier on the ear than the Fostex. For a moderate Qts driver (marketed as BR optimised), it has a very tilted up / lumpy response.
http://quickbrownfish.weebly.com/fostex-experiment.html
A random driver I pulled out of a TV measures flatter and sounds a little easier on the ear than the Fostex. For a moderate Qts driver (marketed as BR optimised), it has a very tilted up / lumpy response.
http://quickbrownfish.weebly.com/fostex-experiment.html
In our cases, just to fit the midTL in, they ended up quite large, probably larger than needed.
dave
Another boombox speaker vs the Fostex (blue). Close mic, cardboard test baffle, minimal smoothing, lots of background noise. I tested a bunch of drivers, and most had some notch around 1kHz, so I think that was from some flaw in the setup - but the measurement was good enough to compare drivers against each other.
It is a 4ohm driver, so sensitivity was the same but efficiency -3dB.
It was a Sharp brand driver, that looks exactly like this:
Soundlabs Group R 10 S - 4 Ohm
Most salvage drivers are not great, but some surprisingly naff looking drivers do pretty well 🙂
It is a 4ohm driver, so sensitivity was the same but efficiency -3dB.
It was a Sharp brand driver, that looks exactly like this:
Soundlabs Group R 10 S - 4 Ohm
Most salvage drivers are not great, but some surprisingly naff looking drivers do pretty well 🙂
Attachments
- Status
- Not open for further replies.