YABADAS - Yet Another Boring Awful DIY Audio Speaker

Folks,

As we all seem to be in speaker building mood (woodworking in the garden in open air is incredibly therapeutic), here my other project....

What I wanted was a pair of speakers that offer a good representation of the "average" Audiophile speaker (well, not quite THAT average).

I almost bought a pair of Dynaudio BM15 Passive Monitors (they are pretty good for the money), but I figured I could make something more suited to my taste myself.



Having excellent chinese connections and having noted previously that HiVi make a reasonable set of improved of Dynaudio cone drivers (eg not stamped metal frames but Esotec style cast frames and cones sourced supposedly from the same source as Dynaudio) so I had a look.

Incidentally they also make Germany's favourite ribbon tweeter, found with it's original Horn in the £ 30,000 Ascendo System M and with different front panels in Burmesters top of the range speakers and the most recent incarnation of the Qudral Titan (and a few others).



This treble unit is actually quite something, pretty constant directivity behaviour at 70 Degrees radiation angle horizontal, reasonable vertical dispersion, 95db/2.83V with 7 Ohm resistive impedance and decent power handling and around ten times the radiating surface of a 1" dome with a fraction of the weight.

Similar drivers where measured by some people on the net as having very high distortion, as the German K+T discovered the problem is that screws of the back plate are often not tight, allowing leakage. Tighten them and the tweeter shows above 2KHz around 0.2% THD at 95db.

I always liked the bigger Dynaudio's and I figured that this hornloaded ribbon could likely keep up with one of the 8" or 10" Woofers.

HiVi's early efforts in dynaudio copy woofers where not so ideal, their more recent modification/copy of the Dynaudio 24W100, namely the D8.8+ with an esotec style cast chassis appeared good on paper.



It has 15mm linear throw peak/peak (28mm nonlinear) and a 100mm Voice coil, so it should have a lot less thermal compression than most common "HiFi" Speakers. Distortion is not as good as some more recent drivers, but with mostly < 1% THD at 95db/1m above 200Hz it's tolerable. The original Dynaudio 24W100 does no better, well, this driver design is getting on being nearly 20 Years old.

Independent measurements from K+T and Hobby HiFi in Germany confirmed that the measurements published by HiVi are reasonably honest and realistic. So I ordered a set of drivers up from the factory.

I selected the smalles box that can still fit the drivers and the Tangband 8" X 12" Racetrack Subwoofer Drivers I had around and decided would make a swell passive radiators....

This gives around 25 Liter net with an LF tuning of around 30Hz, giving 86db/2.83V/1m and a -10db (simulated in halve space) of around 25Hz (-3db simulated in halve space would be around 45Hz). Given the usual room gain, this somewhat overdamped LF alignment should allow a resonable flat response in room down to the -10db limit.

Powerhandling is fine, the driver should be able to remain within in the linear excursion at rated (150W RMS) power down to below 30Hz, this is matched by the excursion ability of the passive radiator, so a good deal of maximum SPL should be availbe at low frequencies if the Amplifier can keep up, making up for the abysimal efficiency (85db/1W/1m).

Boxes where build quite solidly and professionally in 18mm Birch Ply, sadly I could not get them veneered and the professional, while routing in excellent joins etc. had made a bit of a hash out of the woods surface. He had filled in all the problem areas and sanded them flat, but colours did not match and all, so I could not keep the original surface and just oil it. My solution was a fairly thick plastic laminate, not ideal, but good enough the speakers are allowed to stay in the living room....

FRD Driver models from either manufacturer data or from the independent measurements appeared fairly close, so I I decided to run with the "average". A nice little bonus with the drivers chosen is that the inset ribbon tweeter behind it's horn is pretty much timealigned with acoustic center of the (very shallow cone) woofer.

Starting points for the crossover in modeling using the various FRD tools was actually the Dynaudio Xennon 2 crossover (24W100 & D28 Tweeter, 1st order crossover) but ended up changing a lot (a control setup based on the 24W100 & D28 Datasheets also showed significant problems with the original crossover).

I ended up with a crossover at 2150Hz and with acoustic 4th order linkwitz riley slopes. This requires of course electrical slopes that look nothing like 4th order LR. Both slopes are best called "transtitional" though the final rollout is 1st order electrical on the woofer and 2nd order electrical on the tweeter

Crossover components are of reasonable quality with a hefty Air Core inductor on the woofer and a copper foil inductor for the tweeter highpass, capacitors are a mixture of Mundorf Tinfoil, the new Mundorf RFX Series and some MIT Multicap on the Tweeter and MKT & MKP to make up the right value on the woofer.

Just using the Behringer DEQ and pink noise in room shows a reasonably flat response, I need to measure the response using some MLS based pseudo anechonic stuff and maybe tweak the X-Over a little.

I also need to fine tune the added weight for the passive radiators, but for now the sonics are sufficient that they can play music in the main system and quite demanding stuff at that.

In general terms the speakers sound a little more boring than what I'm used to, but they still manage good low and high level dynamics (one of the reasons to choose the 4" Voice Coil 10" Driver), reasonable coherence and all that.

In terms of some of what "audiophiles" and "High End Phreaks" listen for they probably do slightly better than my open baffles, when it comes to sheer involvement and emotional communicativness they loose out, but not enough to turn me off.

So, just the ticket to tune the sound of electronics that are meant to be used with "generic high end speakers".

Ciao T

 

You just couldn't bring yourself to get a pair of B&W's then?

They look pretty much identical to the SLS! Very nice. Pity the woodworker messed up the surface of the wood though.

I didn't realise the D8.8+ had a 4" voice coil! My my...

How much do those tweeters cost then? I have seen them before but they didn't exactly jump out at me as being very high quality.

Out of interest, what do you think of the Monitor Audio GR10's? They area bit old now but I still think they are one of the best sub £1k standmounts about. They are simply enjoyable IMO. I would have thought such a thing would serve well as a reference for common hi-fi.

 

Hi,

The guy who now designs ribbons for SLS used to work for HiVi.....

Yeah, he kinda did these as favour and rush-job, so to complain would be too curlish....

Yeah, just like the good old DA's. Big voicecoils, big magnets, magnets inside the voicecoil and fairly small cones. Not a bad recipy, IMHO.

Let us say that at oem sample pricing they are very affordable.

Not much. I don't "get" MA speakers....

Ciao T

 

I also have a pair of passive BM15's that I am using in the study.

I have not been that impressed. I have found the bass a bit boomy and the midrange has a nasal quality to human voice that I don't like. I get the impression that bass driver wants to be in a bigger cabinet as part of a 3-way system?

 

Those look terrible.

You really need to get some woodworking skills behind you. You do seem to be getting a little better though so keep trying

 

Yeah I heard the distortion was low.

 

I think his woodwork skills looks ok - however, I think the ribbons looks a bit big foir that speaker, and those white stopper things at the top and bottom look out of place tho...

 

Hi,

What for? They are temporary fixtures with a limited shelf life. I was merely interrested in getting them together quickly and use them for the reasons they where made for.

So, I'm sorry, but I'm not even trying...

Ciao T

 

Was just a little bit of trolling. It seems to be quite popular these days.

 

Hi,

Well, it extended to buying pre-cut Laminate and really nasty high solvent content contact cement and gluing the sheets on, then simply finding a style for the edges and corners that did not look too terrible as I lacked the tools to mitre the Laminate.... ;-)

They look okay in the living room.

ER indoors inspected them closely and only complained about the sloppy way in which some spare tonearm counterweights has been "super steel" epoxy glued to the Tang Band woofers used as passive radiators. Being advised that this was the backside she was suitably mollified.

First the "stoppers", those are actually chinese chess pieces which are employed in methodes analogue to the so called "Shun Mook" disks. I tend to get a new Chinese Chess set every time a chinese audio friend visist and I use them for all sorts of jobs and in all sorts of positions.

In common with Shun Mook disks the chess pieces feature mysterious oriental symbols on top (well, to us gweilos they are mysterious anyway) and having many chess pieces in your living/listening room seriously enhances you intelligence (disclaimer, they only enhance intelligence if actually used to play chess!!!).

For the ribbon size, the woofer is a 9.5" (lets split the difference and call it 10") which means even with the Dynaudio style equal surface for center "dome" dustcap and outer cone "ring" they only go up so far.

Again, being derived from Dynaudio Technology they peak a little (but well damed, even on a waterfall) and then roll quite steeply, in fact the initial acoustic rolloff is steeper than a 4th order crossover!!!! even though it eventually settles at a final 2nd order rollout.

The result is that woofer forces a crossover point somewhere between 2 & 3KHz and requires a first order lowpass with some response peaking dialed in via the drivers impedance compensation network. I found that a near perfect 4th order acoustic rolloff happens for the woofer at 2150Hz as -6db point.

What that means is that a ribbon is needed that can be used quite low. The specific one I selected has a 5" long ribbon and a 1400Hz resonance. So it is just right to be crossed over with a 2KHz -6db point. By the time we hit the region below the resonance where distortion rockets sky high we have already attenuated the drive signal to the ribbon 18db, so it receives very little drive and therefor cannot distort much.

The final reason for selection was to match as closely as possible the horizontal radiation pattern in the crossover region (and to have an even but attenuated frequency response at least 30 and better 60 degrees degrees off axis).

There are not many options for that kind of Job....

Ciao T

 

Hi,

The BM15 is tuned to high, lengthen the port (the old bunch of drinking straws trick works fine), that will reduce the boom.

The rest, I suspect the 1" Dome and 24W100 have way too different radiation pattern in the crossover region and the crossover is too high.

You might want to see if what you hearing is the off axis response flare around 3KHz where the dome comes in. You could try some felt or foam "waveguides" to even this out....

I basically cannot abide dome tweeters that are not horn loaded to make them constant directivity, which automatically tends to boost the lower band end enough to allow a fairly low crossover point.

Same with my ribbons, with horn they are flat down to 1.6KHz at 95db/2.83V/1m without Horn getting them down to 2KHz is a chore and they are only 92db/2.83V/1m at 2KHz.

I suspect the active BM15's are a better solution as the crossover is steeper and lower in frequency.

What turned me off the BM15 was when the guy at KMR said he thought the PMC AML1s where drastically better.

Ciao T

 

Not following you re. the drinking straws. I should be able to extend the port using some old PVC tube so may give it a try. The manual says cabinet is tuned to 43Hz in the standard BM15 design. The crossover freq. of 2.7kHz (and 6dB slope) does seem high. The driver integration between woofer and tweeter is pretty poor.

I agree that the active version must be better (haven't heard it) as the passive one doesn't deserve the praise I have read.

 

Get a load of drinking straws and fill the port with them, you can leave the straws sticking out of the port a bit to lengthen it. It is also supposed to reduce port turbulence or something I think.

 

Gotcha. I kinda guessed that's what he meant. Does any particular brand of drinking straw sound better?




Kidding.

 

Russ Andrews ones.

 

Hi,

As Simon says, get a load of large diameter drinking straws, fill the port completely with them, letting the "look out" either inside the box or outside.

As the straw can be longer than the port you effectively tune the port lower, in addition this whole arrangement makes the port more lossy reducing port Q, thus making the overall port output a little lower in level but overall broader.

Yes, you can also do that.

There is definitly a problem there, see here:

http://measurements.hi-fi.ru/loudspeakers/bm15p.html

However the problem is much worse towards one side of (mirror imaged) speaker. You can also see the bass hump from too high port tuning there.

That said, the time domain, waterfall and FR measurements really do not look that bad....

I heard the actives a few times and liked them.

Ciao T

 

You could also vary the lengths of the straws, broadening the ports tuning. This is supposed to reduce port turbulence even more!