|
Nederlands Forum over Oude Radio's |
Re: 3-weg filter Luidsprekerbox AD10100-5060-0160
Geschreven door Thomas op 15. Jun 2006 01:18:27:
In reactie op: 3-weg filter Luidsprekerbox AD10100-5060-0160 geschreven door Ton Habets op 23. Okt 2005 16:58:55:
I was looking for someone to sell replacement AD 10100/W4, because after 31 years, the dome over the magnet has become brittle and actually fell out! This is how I found your comments through Google. But I didn't find any supplier yet.
I also built the kit ADK3540 when I was very young but used them with the original filter only for a couple of years. The concept of a phase-linear loudspeaker then fascinated me, after I had heard the Beovox Uniphase at an exhibition in my native Stockholm. This lead me to investigate if something similar could be designed be replacing the passive crossover with an active one. The unusual box from Bang&Olufsen used an extra "filler" element to compensate for the cancellation that occurs when connecting the elements of a 12dB/octave filter in phase. It hence had three elements for a box which is actually a two-way system! And the filler element must have a high efficiency so as not to make the impedance curve drop too much.
My finding was that the Philips elements could be used with fully electronic filters, giving linear phase at the mid-range to treble crossover. The highpass for the tweeter is left at 4500Hz (12dB/octave). The mid-range element is fed with the sum of a l810Hz lowpass (12dB/octave) and a 2820Hz bandpass with 6dB/octave; this effectively makes it a first-order lowpass with a bump at 2820Hz, but the element can operate up to 12000Hz, so this is no problem. This level of the "filler" signal i.e. the bandpass component can be independently adjusted according to taste.
However, the bass to mid-range crossover could not be made without a compromise. This had two causes. Most important, the overlap in frequency range of the elements was clearly insufficient, so neither of them could be loaded with the filler signal in addition to what is needed for traditional filters. Obviously this is because of the small enclosure for the AD 5060/SQ4 on the one hand; why the woofer falls off so sharply after a peak at 3000Hz on the other, I don't know. The other cause is the greater depth of the larger cone, as compared to the squaker when mounted on a plane baffle. And I did not want to replace the boxes. An acceptable solution was to combine a 18dB/octave bandpass with a 12dB/octave highpass. They of course have different phase lag and lead at the crossover frequency, and the acoustic delay at 610Hz just about compensates for this. Such an active filter whas complemented with three 40W output stages and monted on chassis concealed in a "podium" under the box, which let the tweeter come in the height of the ear when the podium stands on the floor.
These investigations of course started with an analysis of the passive filter that the kit came with. I found that it involved a much worse compromise apart from being passive. The components are numbered as Henk mentioned (except that the capacitor in series with the tweeter is C2 in my assembly manual; there are only to C's, so I don't quite understand Huub's contribution):
C1=50uF and C2=5uF bipolar electrolytic as labelled. The coils I have measured to be L1=1.3mH, L2=0.17mH and L3=0.25mH, because they are not labelled, except from L1 with a type number only. I cannot remember to have seen this sizing of a filter neither in a textbook nor in a practical guide. It reveals crossover frequencies of 270Hz for the bass and 760Hz for the mid-range instead of the specified 500Hz. This gives a larger phase-shift from both, so that the elements can be connected out of phase, which you would in a "pure" design according to theory not do! This explains why the response of the ADK3540 box is about 3dB lower than the average level!!! Apart from the phase response that gave localisation of soloist etc difficult (compared to simple one-element loudspeakers!), this was my major objection against this system which, as you said, was quite inexpensive in 1975. (For CDs made with modern multi-microphone technique, the linear phase response is of little joy. But when listening to old records and broadcasts, which I know have been made with two closely spaced microphones, the stereo panorama opens up. I enjoy them, how noisy they may ever be!)So far, this letter was a brief summary of what became my graduation thesis. Unfortunately, now I possess only one of the original filters, after using the other in a mono box, which was thrown away. The remaining one (and L1 and C1 from the lost one hopefully) I can send to you, if you like. They have been unused for almost 30 years. Sorry that I didn't find your thread earlier.
Now that one element is broken and the filters have become noisy in the trimming potentionmeters after four relocations back and forth to Germany, I have a good reason to buy new boxes. My colleague suggests some of the active boxes from www.me.geithain.de . They are expensive, but maybe I will enjoy them for another 30 years, if I am lucky ...
Best regards
Thomas
- Re: 3-weg filter Luidsprekerbox AD10100-5060-0160 anton tan 15.6.2006 10:51 (0)