Hi All,
I'm itching to do my first speaker project and internet research over the past few weeks has made me fascinated with full range loudspeakers, specifically the pensil family using alpair drivers.
In this family of designs I'm not overly concerned about cost or complexity but I am going in circles on which size to build. My place a little small right now but I plan on moving in a year or two so I'd like to build bigger and have something nice to bring along.
My listening space is 11' square with 8' ceilings though the left hand side of the room opens up to different room. A little small I think.
My musical tastes are all over the place but chest bruising bass is not a goal. I like a little thump but prefer an overall flatter response. I keep the eq flat.
Part of me wants to go big and make a pair of super pensils with 12p drivers, but I can't imagine my space justifies it (about 2m ear to cone). Am i more likely to get better performance with a smaller driver/box?
Thanks in advance for any advice,
Brian
Sent from my SM-N900V using Tapatalk
I'm itching to do my first speaker project and internet research over the past few weeks has made me fascinated with full range loudspeakers, specifically the pensil family using alpair drivers.
In this family of designs I'm not overly concerned about cost or complexity but I am going in circles on which size to build. My place a little small right now but I plan on moving in a year or two so I'd like to build bigger and have something nice to bring along.
My listening space is 11' square with 8' ceilings though the left hand side of the room opens up to different room. A little small I think.
My musical tastes are all over the place but chest bruising bass is not a goal. I like a little thump but prefer an overall flatter response. I keep the eq flat.
Part of me wants to go big and make a pair of super pensils with 12p drivers, but I can't imagine my space justifies it (about 2m ear to cone). Am i more likely to get better performance with a smaller driver/box?
Thanks in advance for any advice,
Brian
Sent from my SM-N900V using Tapatalk
Hello Brian!
I would recommend a 3-5" driver. Personally I like 4" drivers. 3" drivers currently only work down to 100hz or so, while you can eek 80hz out of a 4" and crossover the lower frequencies to a dedicated subwoofer. The larger your speaker driver is, the more it will beam at higher frequencies. The human ear is most sensitive to frequencies from ~3k to ~5k, so you want a good amount of consistent directivity up to this point in my opinion. Otherwise sound reflected off your walls/floor (with a beaming driver) won't be similar as the direct sound from speaker to ear, which will result in poorer imaging. You can find a table of speaker driver size can beaming freq here.
If price isn't an issue, you can't go better than the Scanspeak 10F/8414G-10 which is readily available from Madisound. It was a top 2 in recent blind testing here (only beaten by 1 vote) and measures dead flat 30 degrees off axis (speakers facing straight out, not pointing to ears).
If you want to play with something cheaper, the Dayton Audio PS 95 also did well in these blind tests.
If you want to try something new and untested, Mark audio just came out with two new drivers, the Alpair 5 gen 2 and the Pluvia.
Happy building! Be sure to share any future builds with us here.
I would recommend a 3-5" driver. Personally I like 4" drivers. 3" drivers currently only work down to 100hz or so, while you can eek 80hz out of a 4" and crossover the lower frequencies to a dedicated subwoofer. The larger your speaker driver is, the more it will beam at higher frequencies. The human ear is most sensitive to frequencies from ~3k to ~5k, so you want a good amount of consistent directivity up to this point in my opinion. Otherwise sound reflected off your walls/floor (with a beaming driver) won't be similar as the direct sound from speaker to ear, which will result in poorer imaging. You can find a table of speaker driver size can beaming freq here.
If price isn't an issue, you can't go better than the Scanspeak 10F/8414G-10 which is readily available from Madisound. It was a top 2 in recent blind testing here (only beaten by 1 vote) and measures dead flat 30 degrees off axis (speakers facing straight out, not pointing to ears).
If you want to play with something cheaper, the Dayton Audio PS 95 also did well in these blind tests.
If you want to try something new and untested, Mark audio just came out with two new drivers, the Alpair 5 gen 2 and the Pluvia.
Happy building! Be sure to share any future builds with us here.
Echoing the above welcome -good to see someone new taking the plunge into speaker design / building. 
OK, wading into this. I would be careful about jumping straight in with something like the Super Pensil. The slightly smaller standard Pensil 12[P] has what I consider to be the superior balance of properties. It gives up a little on the bottom end, but that is compensated for with a touch of extra gain & slightly tighter control over the driver. YMMV on that front. I designed the pensils, so to an extent they reflect my preferences in terms of alignment etc.
Room size per se isn't much of an issue, in that (short of very large horns) it's difficult to have 'too large' a speaker for a given space. If you have too much of something, it's easy to get rid of it; increasing what doesn't exist in the first place is somewhat harder. With that said, if you're new to this, and might be moving at some point in the medium term, I'd probably test the waters first by building something smaller rather than going to the immediate expense of a large speaker. See what you think based on experience & then look at moving up in size (and probably cost although that's not a given).
Regarding driver size, beaming, FR etc., I don't entirely agree with all the conclusions, as there are a number of caveats to keep in mind:
-The table linked to regarding beaming is predicated on the assumption of pistonic behaviour and does not automatically apply to wide-band drivers as-is. In some cases (many) it does, but this is not invariable and the frequencies can vary somewhat. It depends on the driver and how it was designed. Those with moderate - deeper cone profiles will tend to become more directional at high frequencies compared to other drivers of the same cone diameter with a flatter profile, e.g. Jordan, MA & some TB units, which are designed with an eye toward maximising the resonant properties of the main cone at the expense of some nominal response linearity. You can get similar variation with drivers that have additional sub-cones (whizzers) etc., where behaviour varies depending on the mechanical XO frequency and the size & profile of the main & sub-cones. So it's not quite as simple an issue as is often thought, or not when dealing with wideband drivers at any rate.
-Related to the above, I would suggest being a little careful of assuming a ruler-flat on-axis response is necessarily a good guide to what might be called 'practical' performance. It's very tempting to use that as the baseline, and it does have value -I doubt many would argue on that front. However, many wideband drivers have a response trend that is rising at higher frequencies. This is not because their designers are fools who lack the wisdom of us forum-dwellers who obviously think we can do a better job than they can, but because they are trying to compensate for the naturally greater directionality of cone drivers at higher frequencies. Basically, they're putting a lift in the on-axis response to give a wider useable listening axis, which is not necessarily a bad thing, as the majority of people, especially if you don't have a dedicated listening room, do not tend to listen purely on the driver's axis. As a trade-off, it's not a bad one, although it does put them at a disadvantage if you then use the on-axis response as your comparative basis vis-à-vis drivers that are intended to be ruler-flat (or as close as possible) under those conditions.
So much for driver design / conditions. In terms of size, it's possible to get < 100Hz LF out of 3in drivers, although in outright terms that comes at the price of increased harmonic distortion and limited output / dynamic range, so whether that's worth it depends on your specific requirements & situation. Some of the little 3in units are very nice though; the PS95 and Alpair 5.2 are mentioned; Fostex's FF85wk is another. Quite a few decent TB drivers available too, while the Vifa TC9 and Faital models have developed a following. As far as cabinets go, the TABAQ is probably the best known example of forcing small drivers low & is very popular. Frugel-Horn Lite doesn't go so deep, but is a little more efficient in loading terms as a horn variant. Lots of other options around, so have a look and see if anything appeals.
4in units will get you plenty of other options in terms of drivers & enclosures. The 5in size will get you more LF, at the price of a little less refinement at the top (that's purely a rule of thumb and doesn't automatically hold good -depends on the drivers). As you go up in size, so you tend to get greater sensitivity, more LF output capacity and dynamic range.
If you like the pensils, then if you haven't tried a wideband driver before, I would be inclined to one of the 4in or 5in model sizes initially, which will give you an entry point and some flexibility in terms of adjusting the alignment to your room. Later on, if / when you move, you may want to revise and / or go for something different, be it larger, smaller or whatever. In general, it's good policy to factor the room into your speaker choice or design, since rooms are probably the most variable part of the chain. Either way, welcome & don't hesitate to ask for advice. That's why we're here: to share thoughts etc.
OK, wading into this. I would be careful about jumping straight in with something like the Super Pensil. The slightly smaller standard Pensil 12[P] has what I consider to be the superior balance of properties. It gives up a little on the bottom end, but that is compensated for with a touch of extra gain & slightly tighter control over the driver. YMMV on that front. I designed the pensils, so to an extent they reflect my preferences in terms of alignment etc.
Room size per se isn't much of an issue, in that (short of very large horns) it's difficult to have 'too large' a speaker for a given space. If you have too much of something, it's easy to get rid of it; increasing what doesn't exist in the first place is somewhat harder.
With that said, if you're new to this, and might be moving at some point in the medium term, I'd probably test the waters first by building something smaller rather than going to the immediate expense of a large speaker. See what you think based on experience & then look at moving up in size (and probably cost although that's not a given). Regarding driver size, beaming, FR etc., I don't entirely agree with all the conclusions, as there are a number of caveats to keep in mind:
-The table linked to regarding beaming is predicated on the assumption of pistonic behaviour and does not automatically apply to wide-band drivers as-is. In some cases (many) it does, but this is not invariable and the frequencies can vary somewhat. It depends on the driver and how it was designed. Those with moderate - deeper cone profiles will tend to become more directional at high frequencies compared to other drivers of the same cone diameter with a flatter profile, e.g. Jordan, MA & some TB units, which are designed with an eye toward maximising the resonant properties of the main cone at the expense of some nominal response linearity. You can get similar variation with drivers that have additional sub-cones (whizzers) etc., where behaviour varies depending on the mechanical XO frequency and the size & profile of the main & sub-cones. So it's not quite as simple an issue as is often thought, or not when dealing with wideband drivers at any rate.
-Related to the above, I would suggest being a little careful of assuming a ruler-flat on-axis response is necessarily a good guide to what might be called 'practical' performance. It's very tempting to use that as the baseline, and it does have value -I doubt many would argue on that front. However, many wideband drivers have a response trend that is rising at higher frequencies. This is not because their designers are fools who lack the wisdom of us forum-dwellers who obviously think we can do a better job than they can, but because they are trying to compensate for the naturally greater directionality of cone drivers at higher frequencies. Basically, they're putting a lift in the on-axis response to give a wider useable listening axis, which is not necessarily a bad thing, as the majority of people, especially if you don't have a dedicated listening room, do not tend to listen purely on the driver's axis. As a trade-off, it's not a bad one, although it does put them at a disadvantage if you then use the on-axis response as your comparative basis vis-à-vis drivers that are intended to be ruler-flat (or as close as possible) under those conditions.
So much for driver design / conditions.
In terms of size, it's possible to get < 100Hz LF out of 3in drivers, although in outright terms that comes at the price of increased harmonic distortion and limited output / dynamic range, so whether that's worth it depends on your specific requirements & situation. Some of the little 3in units are very nice though; the PS95 and Alpair 5.2 are mentioned; Fostex's FF85wk is another. Quite a few decent TB drivers available too, while the Vifa TC9 and Faital models have developed a following. As far as cabinets go, the TABAQ is probably the best known example of forcing small drivers low & is very popular. Frugel-Horn Lite doesn't go so deep, but is a little more efficient in loading terms as a horn variant. Lots of other options around, so have a look and see if anything appeals. 4in units will get you plenty of other options in terms of drivers & enclosures. The 5in size will get you more LF, at the price of a little less refinement at the top (that's purely a rule of thumb and doesn't automatically hold good -depends on the drivers). As you go up in size, so you tend to get greater sensitivity, more LF output capacity and dynamic range.
If you like the pensils, then if you haven't tried a wideband driver before, I would be inclined to one of the 4in or 5in model sizes initially, which will give you an entry point and some flexibility in terms of adjusting the alignment to your room. Later on, if / when you move, you may want to revise and / or go for something different, be it larger, smaller or whatever. In general, it's good policy to factor the room into your speaker choice or design, since rooms are probably the most variable part of the chain. Either way, welcome & don't hesitate to ask for advice. That's why we're here: to share thoughts etc.
Last edited:
Planet10, Xaborus, Scottmoose,
Thank you for the insights. I should have posted sooner as these responses have helped limit the options and therefore makes diving in seem more feasible. Relative to this project, here are my takeaways so far:
Since navigating the pro/cons for using a different sized driver or box would require a perspective I don't yet have I will take the simple/reasonable path - which at this point sounds like the Alpair 10 in a "normal" pencil enclosure.
Looking at the datasheets for the 10p and 10m, it looks like the p has more of what is important to me. It goes a little lower, has a hair more sensitivity, and looks a little less erratic at 15/30 degrees off axis. The m looks hotter at the upper end of spectrum but both are spec'd to 25khz and I doubt I'll be hearing anything up there anyway.
AM's website didn't have a Pensil plan for the 10p, is the second link in planet10's first post here still the best one to work from?
Thanks again,
Brian
Thank you for the insights. I should have posted sooner as these responses have helped limit the options and therefore makes diving in seem more feasible. Relative to this project, here are my takeaways so far:
- I don't need to be worried about over-building, though I may be unwise to do so, it wouldn't sabotage my work
- Similar to what I've been reading in other posts 4"[100mm] seems to be the best starting point in the full range world
- The reduced performance in terms of spl and bass response of an Alpair 10 vs 12 is a good trade off for other reasons
- The added size of a "super" pencil does extend low end but compromises balance
Since navigating the pro/cons for using a different sized driver or box would require a perspective I don't yet have I will take the simple/reasonable path - which at this point sounds like the Alpair 10 in a "normal" pencil enclosure.
Looking at the datasheets for the 10p and 10m, it looks like the p has more of what is important to me. It goes a little lower, has a hair more sensitivity, and looks a little less erratic at 15/30 degrees off axis. The m looks hotter at the upper end of spectrum but both are spec'd to 25khz and I doubt I'll be hearing anything up there anyway.
AM's website didn't have a Pensil plan for the 10p, is the second link in planet10's first post here still the best one to work from?
Thanks again,
Brian
Consider a mass loaded transmission line (MLTL) enclosure. It can be implemented as a small tower enclosure and designs exist for the Mark Audio Alpair 10.2, 10.3, 10P, etc. I've posted several designs on this forum (search 'Jim Griffin MLTL') that are straight (non-tapered) MLTLs that are easy to build (akin to conventional vented box construction).
Bob Brines at:
Brines Acoustics
also posts on this forum has several designs and various price points from just plans, flat packs, to finished speakers.
A petite MLTL would be perfect for your room and music yet it would have enough bass to please you.
Bob Brines at:
Brines Acoustics
also posts on this forum has several designs and various price points from just plans, flat packs, to finished speakers.
A petite MLTL would be perfect for your room and music yet it would have enough bass to please you.
Right. They're just a specific alignment thereof. One of Jim's (or Bob's) designs will get lower; I've tuned a little higher to get more 'slam' through the 60Hz - 120Hz range, and used TL levels of damping to provide a near-unreactive impedance load. Name your poison https://www.youtube.com/watch?v=EZ9se8i4ujs -they're all good.
Whoops... I was thinking an A10 was 10cm for some reason and fit along with 3-5" recommendations above. It's clearly a 6" driver. Still it looks to be a popular unit though so probably a good starting point.
I know I said my musical taste was all over but there is a heavier end to the spectrum... not a lot of testimonials out there praising these for heavier guitar tracks. Think I'll be disappointed? I'm hoping for good all-round performance, not just lighter music. I know no one can predict my preferences but does any see my expectations as being way off? I do have sub with a high pass filter I can dust off for the low end.
I'm still trying to get my around the relationship of watts to spl by way of efficiency... I mean I get it in principle but not intuitively yet. An 89db/W speaker with a 30W handling should be able to deafen me, correct (not a goal, a frame of reference)? And even 2 watts would get me into the low 90s for spl... that's plenty for smallish rooms right? For the record I'll probably break the speakers in with a headphone amp while I build something with more guts: either a 10W kt88 se or a little 25W chip amp.
Thanks again,
Brian
Sent from my SM-N900V using Tapatalk
I know I said my musical taste was all over but there is a heavier end to the spectrum... not a lot of testimonials out there praising these for heavier guitar tracks. Think I'll be disappointed? I'm hoping for good all-round performance, not just lighter music. I know no one can predict my preferences but does any see my expectations as being way off? I do have sub with a high pass filter I can dust off for the low end.
I'm still trying to get my around the relationship of watts to spl by way of efficiency... I mean I get it in principle but not intuitively yet. An 89db/W speaker with a 30W handling should be able to deafen me, correct (not a goal, a frame of reference)? And even 2 watts would get me into the low 90s for spl... that's plenty for smallish rooms right? For the record I'll probably break the speakers in with a headphone amp while I build something with more guts: either a 10W kt88 se or a little 25W chip amp.
Thanks again,
Brian
Sent from my SM-N900V using Tapatalk
.
The 10 in Alpair 10 does signify 10 cm… actual cone size thou, it is a nominal 5 ¼" speaker.
dave
Hey Brian!
You are correct, it only takes a few watts.
Professionals in a studio set the volume to 86dB at the listening position. Any higher and they would risk NIHL (Noise Induced Hearing Loss), sense they listen all day. It's also good to note that at higher volume levels, your ear tries to protect its self from NIHL, and tenses up a muscle to lower/block some sound. So higher volume not only risks NIHL, but also you can imagine that sound quality isn't the greatest when your ears are trying to STOP the sound. Here is a guideline to safe listening levels.
Physically, sound intensity doubles every 3dB, but we perceive a 10dB increase as double the loudness (90dB twice as loud as 80dB).
An externally hosted image should be here but it was not working when we last tested it.
You can get a cheap TPA3116 amp (very good class D) for $35 from parts express. I would strongly recommend getting a working amp before you DIY one. That way if your DIY speakers sound "wrong", you know it's the speakers and not your DIY amp. If you want to DIY an amp later on, I strongly strongly suggest a gainclone LM3875 kit from audiosector. Gainclones measure/ show all the acoustic characteristics of tube amps (A.k.a. prodomanently even order distortion that reduces as the harmonic multiple increases) with all the reliability of solid state, and the huuuuuggeee power savings of class AB vs class A tube amps. Audio sector has tried virtually every possible configuration of the gainclone, and only sells the best sounding configuration he's tried (of course you have to trust his subjective taste).
Anyway, pretty much any fullrange speaker I would only recommend down to 80hz and crossover with a dedicated subwoofer. Trying to get even 40hz-20khz will result in major compromises and overly complex speaker cabinets that are impossible to mathematically predict. I agree with going with a MLTL build. They are extremely easy to build with little math. Just build a skinny bass reflex, and you have a MLTL. Many people report that they find "skinny" speakers image better than fatter ones, so you have nothing to lose.
A little more in depth to MLTL design, is that the CSA (cross-sectional area) of the speaker cabinet (WxD) should be 2-4x your speaker drivers cone surface area (Sd), and never less than 1x Sd. So yeah, build a skinny bass reflex with CSA 2-4x Sd, and you have a MLTL. At worst it will just work as a regular bass reflex.
If you have ANY other questions, absolutely feel free to ask.
Last edited:
Just a comment on the volume.
When I designed TABAQ I set the internal cross section to 4 times Sd. This was done to get as much output from the opening as possible. The higher volume, the higher SPL from the opening.
To control the bass, to get a gentle roll of, the TABAQ is rather heavy stuffed in the upper 2/3 of the line.
The "best" cross section / volime is of course depending on tha actual driver to be used.
Designing a quarter wave has several parameters to use: That is why it is so fun:
Driver.
Tuning and shape of the cabinet (defining the tuning based of driver Fs and Qts)
Cross section (volume).
Where to place the driver (how close to the closed end).
Density of stuffing.
Distribution of the stuffing (typical in upper 2/3).
Baffle step compensation (if you find the need for this, when you listen to the speaker).
Have fun!
Bjørn
When I designed TABAQ I set the internal cross section to 4 times Sd. This was done to get as much output from the opening as possible. The higher volume, the higher SPL from the opening.
To control the bass, to get a gentle roll of, the TABAQ is rather heavy stuffed in the upper 2/3 of the line.
The "best" cross section / volime is of course depending on tha actual driver to be used.
Designing a quarter wave has several parameters to use: That is why it is so fun:
Driver.
Tuning and shape of the cabinet (defining the tuning based of driver Fs and Qts)
Cross section (volume).
Where to place the driver (how close to the closed end).
Density of stuffing.
Distribution of the stuffing (typical in upper 2/3).
Baffle step compensation (if you find the need for this, when you listen to the speaker).
Have fun!
Bjørn
You know what? I think I have some PCB kits for LM3875 (or was it LM3886) in storage along wit a PS board and a transformer I pulled from a broken Event Studio Precision 8 studio monitor. That should be my first stop when I DIY an amp. I do also have a 2x75W Carvin amp that I can use that sounds decent, though it currently exists as part of my guitar rig. Either way, its good advice: Test the speakers with a known performing amp before building something.
I plan to build the Pencil design so the internal volume and cross-sectional area are determined. Its an easy build too when compare to some other more complex options.
And indeed, I have pulled the trigger on a pair of 10p from Madisound, along with some binding posts. I'm thinking of doing the back of the pensil in 2 pieces: A larger lower section that is glued in place for more rigidity and small upper section that is removable for wiring and stuffing... or maybe I'll just build it to print.
I see, that makes sense. I guess it now comes down to which a given person is referring too when they spec a size range. Cone vs OD.
I see, that makes sense. I guess it now comes down to which a given person is referring too when they spec a size range. Cone vs OD.
BJosephs,
Chamfering the inside of the driver cutout is generally a good practice for wide-band drivers, and recommended for Mark Audio drivers.
Since the frame is pretty thick for the Alpair 10P, you might have to add and 1/4" layer of material assuming that you will use 3/4" material for your build.
Chamfering the inside of the driver cutout is generally a good practice for wide-band drivers, and recommended for Mark Audio drivers.
Since the frame is pretty thick for the Alpair 10P, you might have to add and 1/4" layer of material assuming that you will use 3/4" material for your build.
If you are working with 18mm ply, and if you make a 10mm rebate to accommodate the driver frame (without frame cover), you are left with around 8mm behind the driver. Some say that's enough with good quality ply to support the driver.
The points where the screw will go in - they do not need to be chamfered - the areas around them, yes.
The points where the screw will go in - they do not need to be chamfered - the areas around them, yes.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.