basshead Posted October 21, 2023 Posted October 21, 2023 I started thread here over 18 years ago and the only thing the same since that thread was created is two old 12inch G&S Designs GE7012 subs powered by a Teac Reference A-H500 (2x50W) amp. Now the HT is virtually complete (Paul Carmody Amiga towers (made by someone else), off-the-shelf in-ceiling speakers in an enclosure with baffle angled towards the rear of the room, and (DIY) Paul Carmody's Sunflower centre channel), and rear speakers (DIY) using the same Dayton Reference series drivers as the Amiga and Sunflower), it's time to build a better enclosure for these subs as they sound terrible (and seem to be much lower output in the HT space). From playing with a DATS and WinISD, it seems these subs just might need to be retired. I reconed (surround only) them perhaps a decade ago (as the surrounds had perished) and they had also been reconed by Tony Doran perhaps 25 years ago (and cones were reinforced with layers of PVA glue), so I was sceptical the TS parameters would be even close to factory specs. Hence the DATS, and the measurements I got were quite abit off the data preloaded with WinISD (assuming they're the same model drivers). For example, the DATS measured (using added mass to cone) Qes 1.1, Qms 6.636, Fs 33.55, Vas 67.1L & Qts 0.948. When I load a new project in WinISD the sealed enclosure is almost 770L with Fsc of 35Hz and Qtc 0.994! A bit of googling indicates such a high Qtc is better suited to infinite baffle. That probably isn't going to work in a 4 x 4.5m room (and with a low ceiling). After playing around with various enclosure sizes and also modelling the existing enclosure in WinISD, I seem to be getting some unusual curves. I even considered 4th order bandpass but apparently they are very inefficient and I'm not keen on buying another amp to replace the Teac. I've attached a screenshot showing the curves. I'll post more details later...
basshead Posted October 22, 2023 Author Posted October 22, 2023 Explanation of the curves from the plots in the above post... "Ported" is the light green curve with near-flat response between 25-45Hz at 104db. This is with isobarik ticked in the driver tab. If I switch it back to "standard" then the curve changes to a much higher output with curve from about 104ddb at 20Hz - 110db at 45Hz. 150L encl tuned to 25Hz. "Ported (by tutorial)" with dark green curve is 80L tuned at 20Hz Red curve is one I threw in just to see what a random different driver would look like. 120L tuned at 23Hz "Existing box common enclosure" is essentially what is describes; the box minus the partition in the middle (as there are 2 drivers). I switched this to iso-barik for this go get a peak at just over 40Hz (dark blue curve). If I switch it back to standard, the curve becomes a similar shape to "existing" box but far higher level; 25Hz at 100db - 50Hz at >113db. 100L tuned at 35Hz "existing box" 50L tuned at 35Hz. Dark green curve shows peak at 60Hz. No wonder it sounds terrible. I vaguely recall using WinISD to design this box over a decade ago but possibly used factory TS parameters as I only got the DATS recently.. Finally the "sealed" grey curve is using the DATS measured data without any changes; 767L enclosure with Fsc 35Hz. I'm probably leaning towards trying out the ported (light green curve) then adding the 2nd driver inverters (for iso-barik) to see if I can get the flatter response and remove the 40-60Hz peak. Available space (pending wife approval) for a new enclosure is around 670L. That's only 100L below the Vas of one G&S driver. I've heard pre-built HT subwoofers sound much better and go much lower and louder than these two G&S 12's, and in much smaller enclosures. Hence why I'm keen to build a new enclosure for them. But if I'm going to build a box that large, I'd be very tempted to buy a couple of 18's... and then I'd need a bigger amp than the Teac... Any thoughts welcome!
davewantsmoore Posted October 26, 2023 Posted October 26, 2023 On 21/10/2023 at 2:53 PM, basshead said: 770L I suspect you have either measured the parameters incorrectly, or loaded them into winisd wrong. Your "ported tutorial" should have produced a result with a flat response.
Recommended Posts