The Anatomy of the Acapella Violon - shocking find

For quite a while now I have been having trouble with the bass on my Acapella High Violon Suboktav 2001. With the help of a friend, we have dismantled this speaker and studied the internal construction and measured the crossover points. I am hoping that my findings will be helpful to all of you Acapella owners.

The bass problem is this: bass can not keep up with the speed of the midrange and top end. On some recordings, the bottom end becomes disconnected - you can hear music from the midrange and the top, followed by the bass response a microsecond later. Furthermore, the bass is poorly controlled and flabby. From my other Audiogon threads, you can see that I have been wondering whether the damping factor of my Cary CAD-211AE amps is sufficient to control the wild bottom end, and whether a solid state amp will cure this problem.

The current iteration of the Violon is Mk. IV. I am not sure what a "Violon 2001" is, I am guessing either Mk. I or Mk. II. Acapella's own website does not reveal any secrets, all it says is that the High version of the Violon has an additional driver inside.

Anyway, this is what we found.

PLASMA TWEETER: 4th order high pass crossover (24dB/oct) at 5000Hz. Measures very flat all the way to the limit of measuring equipment. Incidentally, Acapella marks the recommended tweeter level with a pencil mark on the tweeter volume pot. At the minimum recommended range, the tweeter comes in 12dB ABOVE the reference SPL. I had to wind the tweeter almost all the way down to get a flat response.

MIDRANGE HORN: First order high pass crossover (6dB/oct) at 450Hz, with a very gentle taper between 3dB/oct - 6dB/oct from 5000Hz and up. Goes all the way up to 10,000Hz. The shallowness of the low pass section of the band pass crossover makes me suspect that it is relying driver rolloff.

This is a surprisingly wide band of frequencies (4 1/2 octaves) to ask a horn to handle. As you know, horns are tuned to work over a fairly narrow frequency range and the response drops off at either extreme of this range. Wavelengths which are too long for the horn do not couple with the horn. Wavelengths which are too short will bounce around chaotically. That is only the theory however, because the horn measures very flat between 450Hz - 5kHz.

The integration between the tweeter and midrange horn is very good.

BASS UNIT. As advertised, there are two 10" drivers in the unit, and both appear to be SEAS drivers. The external driver is run through a passive crossover from the binding posts, with a low-pass first order crossover at 450Hz.

And now, the surprise. The internal driver is run directly from the binding post with no crossover in between. In other words, it is run full range, relying on driver rolloff only.

I can think of no advantages for a setup like this, only disadvantages.

Firstly, the drivers are wired in parallel. This will drop the impedance, making it difficult to drive them with valve amps (Acapella supposedly voice the speaker with the Einstein OTL).

Secondly, a configuration like this will result in destructive interference between the two drivers, ESPECIALLY if the crossover introduces phase problems in one of the woofers. Given the other woofer is crossover-less, any difference in phase will definitely cause interference.

Thirdly, running a woofer full range will cause cone breakup at the top, which will muddy the lower midrange (exactly what I have been hearing).

Possible solutions:
- wire both drivers to the crossover, maybe in serial configuration to increase the input impedance (Zin),
- disconnect and remove the internal driver, i.e. convert the speaker from the "High Suboktav" version to normal Violon,
- remove the passive crossover entirely and use a preamp-level crossover (active crossover).

Now, I am no speaker designer. I am just an enthusiast struggling to understand these things. But this just makes no sense to me. I am hoping that someone with more experience will be able to explain why Acapella made these design choices, and what you think of the possible solutions.

Showing 9 responses by atmasphere

Amfibius, Duke's comment sounds like one to be paid attention to. So if you proceed, keep his comments in mind!

The crossover parts would change as the impedance would be different and you would have to set up for the same notch frequencies, and if Duke is right, probably a notch for the unwanted secondary impedance peak.

I'm glad I make amps and not speakers!! :)
Amfibius, I have a simple idea that you might try, since it appears that you have access to people that are speaker savvy. See if you can arrange to put the woofers in series and readjust the crossover to support the same notch frequencies.

Tube amplifiers, unless endowed with a lot of feedback (which makes them sound more solid state) are not be happy with an 8 ohm midrange and tweeter, along with a 4 ohm bass unit. The bass will be loose and flabby sounding, owing to the increased distortion of the amplifier (NOT as popular mythology has it, lessened 'damping factor').

Setting the woofers at 16 ohms will make the speaker much easier to drive with almost any tube amp. The result will be improved detail and a greater sense of authority; this will be a load a tube amp can handle. It will also take care of the rear woofer not being crossed over.

BTW they may voice the speaker with the Einstein, but if they had done that with the woofers in series, the result would have been even better. Sixteen ohms is an easy way to get smoother sound and more detail out of any amplifier.

Rlawry, The speaker you have has only a single woofer and is an 8 ohm load. The way an amplifier handles that is quite different from how it reacts when the woofer load is 4 and the rest of the speaker is 8, especially if the amp is question is a tube amplifier. Your speaker is an easier load for tubes (and transistors for that matter) so your results will be a bit different!
Zombie thread come to eat your brains!!
Amfibus, you have 10 meter speaker cables??! OK- in order to get the most out of the speakers, you will want to shorten those up- over 5 feet and most speaker cables chomp into the impact and definition in a big way! If I had known that earlier, I would have recommended getting that sorted out *before* even taking the speaker apart.

Four ohm speakers tend to be very critical of speaker cables (16 ohms OTOH *almost* don't care). IOW you won't get away with a 10 meter run, regardless of your amplifier.

So I would try an experiment wherein you temporarily use a short run of cable that is only a meter or meter and a half (so you will have to move your preamp and front end to do this), do this with your Cary 211s running full range on the four ohm tap, and see how it sounds then (and report back of course).

I understand that you may need to have your preamp and the rest of the system much further away, but there are preamps that are in fact intended for that (and that are also all tube and zero feedback, so you won't corrupt the sound of the Violins). We'll solve *that* problem *after* you establish the effect of the cable!
Amfibius, 5 meters is still *really* long. Many 4 ohm speakers will simply not allow lengths like that- Magnaplanar (otherwise a fairly easy load) is a good example- with them its beneficial to turn the amps around backwards and run 6-inch speaker 'jumpers' from the amp to the speaker (the biggest improvement by doing this is in the bass BTW, although the mids and highs get better too. This is one of the arguments for monoblock amplifiers). Seriously, 3 feet or so is about the maximum for many 4 ohm speakers.
Oops! never mind... :)
Hi David, When you are dealing with 4 ohm speakers and you are running a long cable, first off the resistance of the cable actually becomes part of the total impedance driving the speaker.

The second thing to understand is that all speaker cables have a property known as 'Characteristic Impedance'. This is not the resistance or impedance of the cable, rather it is the property of the cable such that it will have its best performance when terminated by a specific impedance. For example, a cable with a characteristic impedance of 8 ohms will perform best when terminated by an 8 ohm load. Its actual DC resistance will likely be quite low as will its AC impedance, however.

Even if we were talking about a cable with a 4 ohm characteristic impedance, there are no speakers that are exactly 4 ohms over their entire range, so there would be termination errors. Most cables have a characteristic impedance that is much higher than 4 or even 8 ohms so there is a greater error that results.

The error is in the form of reflected energy which smears the waveform. This increases as the cable length is increased; a simple way to reduce the error is to shorten the cable- hence a good argument for monoblock amplifiers.
Amfibius, based on the results that you posted, it looks like the woofers are in parallel, and so are 4 ohms while the rest of the system is 8 ohms.

If this is so there will not be a tube amplifier made that will play the bass at the right level. Transistors will do OK, but then the highs will be dry as you experienced.

I would consider putting the woofers in series if you want to use tubes. This will get the bass working right, so long as the crossover for the two drivers is the same. If not, it might be a long road ahead for you to get this sorted. Good Luck!