Jump to content


  • Content count

  • Joined

  • Last visited

About BioBrian

  • Rank
    500+ Post Club

Profile Fields

  • Location
    Mountain River
  • Country

Recent Profile Visitors

6,161 profile views
  1. Dear Rolo, Thanks for the Rockwell article - it reinforces what I have learnt (with my own ears and speaker experiments), and adds some good perspective to the subwoofer issue. I am now confused at why you still have the idea of joining the 2 subs as a mono signal. I think you need another Dayton 250; use one amp, and one JBL woofer, per channel. With the DTQWT, I have done measurements on the woofers (not to publishing standard!), but have gained some understanding of how the speaker system works. The rear woofers have that large C-coil, as you say, and it is double-size, due to the woofers being in series (in the single-woofer-version it is only 10 mH, not 18). It is not a bad thing! It doesn't amplify the bass; it increases the impedance to higher frequencies, leaving the lower frequencies at full volume. If Troels says it cuts at 150 Hz, or whatever, that only means that the rolloff starts there. However, it is a 1st-order roll-off (theoretically), which means 6 dB per octave relative to the natural curve of this woofer. This means there is quite significant output from those rear woofers, even at 1000 Hz. It really only cuts the nasty spikes further up (the ears of the hyena). He has designed the rolloff curve to give support in the mid-range to the JA8008; this is necessary to balance the air pressures inside the horn. Another thing that showed with my measurements is the amount of bass extension given by the horn itself. Placing a mike at the horn opening (at the floor) showed around 100 dB at 20 Hz - it is similar to measuring the vent output in a BR speaker - the sum of the woofer nearfield, and the vent output, need to be considered as a whole. Good luck to anyone who thinks they can calculate anything meaningful from that. My point is that choosing a crossover frequency to match an untouched DTQWT is going to be an approximation. We have to live with a sharing of frequencies from different sources: full range from front woofer, 1st order LP and intrinsic HP from rear woofers, and the lowest frequencies from the horn at floor level. My best suggestion is that you take the L & R sources to each of the 2 Dayton 250s, let them crossover at a very low frequency, take HP to the Prima Lunas and DTQWTs, and feed the LP to your JBLs. That way you get STEREO subs (which you should place close to the DTs to avoid delays), and the high quality sound you've described from the PL/DTQWT combination. 2 subs off one amplifier will kill the stereo image - I believe you would not like this . I was going to post some graphs, but it seems the latest Windows 10 update has killed my Omnimic software - .
  2. L-pads are a continuing study. I notice big differences in SPL curves, when changing resistors before, or after, the parallel coil. Generally, I alter the tweeter's attenuation more effectively by adjusting the 1st one (before the coil). The L-pad's series and parallel resistors have important independent functions: eg particularly the parallel one can induce or correct quite a sag between 5 and 10 kHz. The reason I (it was only a feeler) questioned the role of L-pad and phase, has roots in Troels' dealing with the mighty SS Be tweeter in his ATS4: http://www.troelsgravesen.dk/ATS-4.htm The hot words being that a resistor, in parallel with the 2nd series cap (otherwise looking like a 3rd Order tweeter section), "alters phase". I've occasionally had good results from ditching one or both of the L-pad resistors. I find it hard to imagine that phase is unaffected by this. There's a lot of "discussion" (dogma really, as it is so established) out there on the phase changes caused by 2nd Order and 3rd Order (etc) filters. This is background to my observations that phase of a driver could be altered by practically any passive component, but I'm also fessing that this is very early days in my personal study!
  3. Please could we stay on topic gentlemen!? This 13 pages is already long, even though it's over 2 years of experiment. The use of a quality mid-driver is critical to my design. Matt's project is very different, and will take up a lot of cyber-ink not relevant here. There are many unsolved mysteries pertaining to this build which could be addressed.
  4. Looks like a new thread, Matt. I could make a few suggestions, but I'm sure you'll get lots anyway.
  5. Hi Matt, new project sounds good! You're asking me??! I am a bit of a sucker for datasheets, and that's how I arrived at the Scanspeak D3004/662000. The 'D' means dome, and as I mentioned way back, people, including Troels, seem to ignore this one, in favour of its "pretty sister with the pointy nose" - R3004/662000. The dome version's chart looks better to me than the ring-radiator version, (including off-axis), but I haven't heard both, and I wonder if the criticism of "one part of of the dome being in different phase from the rest" actually refers to frequencies within our hearing band. Troels has recently expressed preference for the smaller "beer can" Scanspeak tweeters, but again the Ring Radiator version: R3004/602010, which are a lot cheaper, and apparently just as good as the other 'pointy-nose'. I'm still scratching my head re the Be versions - hard domes are outside your question - although they are very expensive and look nice, I haven't actually read any evidence that they sound better. But maybe it's just an 'understood'. I don't have a lot of experience - maybe others can help you more. The SBA tweeters are still spoken of very highly, and my experience with the Neo TW29RN is very positive.
  6. Indeedy! Especially in fresh Baltic Birch: Mid is in 24 mm ply; lower mid has 12 mm extra: acoustic centres as per Troels' AT 3W Classic. I expect the larger-magnitude phase issues to be below these dudes.
  7. I've really only been concerned with phase lines 30 or (sometimes a lot) more degrees separated. The wavy woofer-axis line is not going to be heard - my head would be under the floor at listening distance and that angle, and the wiggles are probably due to diffraction around the driver edge at that weird angle. The relative distance from driver to mic changes, as the mic height changes - one driver goes relatively further away, as in a 2-way with sloping baffle. I expected to see large shifts in the phase curves, but this doesn't happen. At least, on the scale I would expect, considering the small wavelengths at mid-tweeter level.
  8. This is where my brain's stuck. Why do I think to keep the mic in the same position when measuring 2 drivers? The phase line doesn't change - is the good bit in the Impulse Response, below?
  9. The source of some bliss. It fits in the hole. All good so far.
  10. Thanks Rollo (as a friend might call you in this country)! It's called "design without the constraint of too much knowledge" . I have started cranking up my battered resources, with the beginning of new boxes. With Scanspeak and AT drivers, I might finally have to call these SATs! But I'm not sure about the tweeter yet, so maybe their shape (and Scanspeak Questionmark Audio Technology) might mean they get called "SQATs": we'll see. There's something deeply thrilling about Tasmanian Spring weather, and the smell of fresh Baltic Birch!
  11. Hi Rolando, and thanks for your kindness. I read about your experiments from Troels' website - very interesting, and in character! I am hoping that you stay stereo, and use "hifi" woofers, if anything, to boost the DTQWTs' bass. If you use a turntable, I would not expect very nice noises to be the fruit of your endeavours . I often wonder what would happen if Troels had gone ahead with the even larger DTQWTs, with twin 15" woofers and longer horn. I had the idea of adding another ~400mm to the base of mine, extending the horn (down to 24 Hz) and raising the drivers more to ear height, but it's a lot of work to find out. I found that using a 2-wheeled trolley is very easy and cheap - much better than wheels under your speakers. (Fix some felt padding, and come in from the side).
  12. Bit of excitement today: the arrival of my first pair of Audio Technology drivers. Very attractive! 23 I 52 17 06 SDKA with lighter 0.3 mm cone for 93.6 dB sensitivity; and 15 H 52 06 13 SDKAM, LR (rounded magnet parts, for a few extra Euro(!). The 5" mid is actually heavier than the 8" - a very chunky beast. I have high hopes for it. And it looks like new boxes are inevitable now, so I don't just chuck away all the good work so far. The top 3-ways, now back together with all SBA drivers, sound a notch better now, having put a C-Coil in for the woofer. Only a third of an Ohm less resistance, but it sure sounds better balanced, more weighty in the bass. So the bottoms of the new boxes will match the top width of the 15" woofer boxes, and have a re-shuffled baffle to allow room for the bigger mid, and more acreage for the tweeter. Something like: But I'm pretty sure the tweeter will get a separate box, as mentioned previously. The folds in the paper above show the width of the existing top 3-way: 30 mm less, each side.
  13. Hi Rolando, Did you miss my comments in my thread about single subs? If you like music, you need one for each channel! Using a single sub with my DTQWTs was horrible, destroying the directivity of the music. In any case, why not do what they want you to: send your 'source' to the plate amp RCA inputs, then take RCA cables from its outputs to the input of your Prima Luna? This plate amp is not large for a sub, so an even better reason to use 2. And the bigger the box(es), the less power needed. Best of luck!
  14. That must be immensely satisfying, Anthony. It certainly is for one far-away observer.
  15. Well desperation made me take the Troels Gravesen JA8008 8" drivers out and return them to my beloved DTQWT-12s. Oh what bliss - watching a movie last night with proper sound was very heartwarming. The SBA 8" woofers went back into the upper 3-ways, and of course I had to revert all the XO changes I'd made, so now they are close to Troels' SB 3W classic design, with a few of my changes. The 'wings' around the tweeters are still showing a flatter spl 'curve', so have stayed. With so many variables, it's hard to make worthwhile conclusions, but a few points raised themselves. The SB woofer lowered the sensitivity of the upper 3-ways, but I still had the system running as a 4-way, off one stereo amp. Surprisingly, they sounded really good with relatively a bit more juice going to the bottom woofers - I could easily live with that, as the presence in the upper drivers didn't seem to be significantly lacking. Actually, very satisfying. The woofer-mid relationship seems tighter, if less generous. I trotted out Omnimic again, to check for flaws in my memory/reconstruction of the XOs, and was pleased to see a close-to-ideal falling spl response (black line below). Tick 1. Next, I had a look at Mid-Tweeter phase, as last time it had been good with the Scan Speak tweeter, but not so good with SBA, using same XO. This time, it looks as though there's a better phase match (thin blue and red lines): ..nearly as good as with the SS tweeter. I'm guessing that this would be mostly due to changes to the L-pad in the tweeter section. Another discovery that's kind of been nudging me for weeks is that 'phase' doesn't change with either mic distance, or mic height. I can't quite get my head around that - I'm still stuck on the mysteries of Z-Offset I think. The following pic shows 3 very different mic positions, measuring just the mid-driver through XO. One at mid-tweeter axis, one at woofer axis, and one much closer (hence louder) at mid-tweeter axis: (The 3 thin lines are phase, the thick ones are spl). The conclusion I have at the moment is that changing phase cannot be done by moving drivers fore-and-aft, but by crossover components alone. (DEQX drivers: please bite tongues for now).