I have used my own DSP active OBs for several years & am part way through building my third set - a 3-way pair based loosely on the PAP concept.
Previously I used 2-way for midrange/tweeter, crossed to a pair of subs at 100hz. The 3-ways are designed to go down to 40-50hz with the subs below that & will use 4 of the 6 stereo channels available from my two DEQX units (HDP-5 master - HDP-3 slave. The slave will control the two subs).
Amps are Dynavector (bass), Quad (mid) & Graham Slee (Treble). Drivers are Mark Audio Alpair 7, Alpair 12 with twin Beyma 15" per speaker, subs M&K & B&W.
Once finished & the drivers complete burn-in in my music room, I will disassemble and take some anechoic measurements outdoors so the DEQX processors can create correction filters (time & phase alignment etc), then reassemble in the room & finish the setup, including slight room eq & time alignment to the subs. The subs are already time and phase aligned in the music room & I have confirmed that the RS232 & digi links between the two processors work correctly.
Progress so far:
Baffles cut, routed & veneered. Baffle sides & the woofer cutouts have a 10mm radius. Midrange/treble will be a flush fit.
Frame construction with tweeter baffles at the correct angle for seating position head height. The wooden side runners may prove to be rather flimsy when I try to move the finished speakers - I will have to be careful. All hardware used is stainless steel including nyloc nuts (at the moment I’m using temporary wing nuts & square nuts until I’m satisfied everything fits).
Baffle mounting screws are held in place by small rubber O-rings so that panels can be fitted from the front. Decoupling spacers between baffles & frame are shown (I intend to leave the bottom baffle/driver in place when moving these & maybe the top one as well if it’s manageable but the others will come off, the Beyma's are heavy monsters).
Soundcare SuperSpikes used as mounting feet, fitted to M6 threaded inserts. The feet are internally decoupled which also allows easy tilt adjustment. I've taken the precaution of snugging up several nylon cable ties around the timber at the base on the fronts as I suspect this may be a weak point. I will trim these later.
Rear uprights with speaker terminals mounted on wooden plates. This original design used rectangular wood section but I am unable to bend the aluminium side braces sufficiently without stress cracks to get the correct angle. These have changed to a triangulated section (see below).
Clamping angled top sections for tweeter baffles in place, then drilled, glued & screwed:
Aluminium side braces cut, angled & drilled at the front end only so far. I found that bending these to 45deg weakened the aluminium too much so I’m unable to fold round behind the rear support (35deg is strong & stress free), These will be bolted through from the sides using threaded inserts angled into the wood. Trimming & drilling the rear ends will take place after trial fitting to the side runners & lined up with the rear support :
Initial assembly of the side braces:
Rather than try to bolt the side braces with a gap to the rectangular rear support as shown in the speaker terminal picture above, I used my saw table to strip a larger piece back to the correct 35deg angle for both rear supports. These are the next thing to work on, trimming to length, painting & re-mounting the speaker terminal blocks.
An approximate image of how a final speaker will look (bottom braces resting on a wood offcut). The rear support will be cut down to finish flush with the top side braces & in this picture it is a piece of the original rectangular section. Now these are triangulated, the speaker terminals will face forward as in the picture.
As the main load bearing braces are double thickness, I next need to remove these from the frames and araldite/clamp together before trimming to length/drilling the rear holes. As I only have 2 large clamps and there are sixteen of these halves, this may take a few days.
To be continued....
Andrew
Previously I used 2-way for midrange/tweeter, crossed to a pair of subs at 100hz. The 3-ways are designed to go down to 40-50hz with the subs below that & will use 4 of the 6 stereo channels available from my two DEQX units (HDP-5 master - HDP-3 slave. The slave will control the two subs).
Amps are Dynavector (bass), Quad (mid) & Graham Slee (Treble). Drivers are Mark Audio Alpair 7, Alpair 12 with twin Beyma 15" per speaker, subs M&K & B&W.
Once finished & the drivers complete burn-in in my music room, I will disassemble and take some anechoic measurements outdoors so the DEQX processors can create correction filters (time & phase alignment etc), then reassemble in the room & finish the setup, including slight room eq & time alignment to the subs. The subs are already time and phase aligned in the music room & I have confirmed that the RS232 & digi links between the two processors work correctly.
Progress so far:
Baffles cut, routed & veneered. Baffle sides & the woofer cutouts have a 10mm radius. Midrange/treble will be a flush fit.
Frame construction with tweeter baffles at the correct angle for seating position head height. The wooden side runners may prove to be rather flimsy when I try to move the finished speakers - I will have to be careful. All hardware used is stainless steel including nyloc nuts (at the moment I’m using temporary wing nuts & square nuts until I’m satisfied everything fits).
Baffle mounting screws are held in place by small rubber O-rings so that panels can be fitted from the front. Decoupling spacers between baffles & frame are shown (I intend to leave the bottom baffle/driver in place when moving these & maybe the top one as well if it’s manageable but the others will come off, the Beyma's are heavy monsters).
Soundcare SuperSpikes used as mounting feet, fitted to M6 threaded inserts. The feet are internally decoupled which also allows easy tilt adjustment. I've taken the precaution of snugging up several nylon cable ties around the timber at the base on the fronts as I suspect this may be a weak point. I will trim these later.
Rear uprights with speaker terminals mounted on wooden plates. This original design used rectangular wood section but I am unable to bend the aluminium side braces sufficiently without stress cracks to get the correct angle. These have changed to a triangulated section (see below).
Clamping angled top sections for tweeter baffles in place, then drilled, glued & screwed:
Aluminium side braces cut, angled & drilled at the front end only so far. I found that bending these to 45deg weakened the aluminium too much so I’m unable to fold round behind the rear support (35deg is strong & stress free), These will be bolted through from the sides using threaded inserts angled into the wood. Trimming & drilling the rear ends will take place after trial fitting to the side runners & lined up with the rear support :
Initial assembly of the side braces:
Rather than try to bolt the side braces with a gap to the rectangular rear support as shown in the speaker terminal picture above, I used my saw table to strip a larger piece back to the correct 35deg angle for both rear supports. These are the next thing to work on, trimming to length, painting & re-mounting the speaker terminal blocks.
An approximate image of how a final speaker will look (bottom braces resting on a wood offcut). The rear support will be cut down to finish flush with the top side braces & in this picture it is a piece of the original rectangular section. Now these are triangulated, the speaker terminals will face forward as in the picture.
As the main load bearing braces are double thickness, I next need to remove these from the frames and araldite/clamp together before trimming to length/drilling the rear holes. As I only have 2 large clamps and there are sixteen of these halves, this may take a few days.
To be continued....
Andrew
Last edited:
Edit 2 : and a really solid base ! As well as the top ( put counterweights ???)
These will be heavy enough without counterweights - on each speaker the bottom pair of Beyma 15LX60V2s weigh over 20kg alone & the main alu side braces are 9.5mm thick.
Finished the build and have done an initial indoor measurement & calibration.
Completed rear frames: Beyma 15" drivers were mounted to their baffles prior to fitting as they are so heavy, Alpair drivers were added to the fronts after this picture was taken. All the frame mounting holes & threaded inserts lined up correctly & the weight of the Beymas did not pull things out of alignment.
Cabling connected:
Side view of a completed speaker:
Graham Slee amps powering the Alpair 7s @ frequencies above 2khz, these are very sweet & transparent sounding:
Initial readings taken in the centre of the room (each speaker moved to this position with some basic sound absorbtion as shown). The mic (calibrated Earthworks M23) was at 600mm midway between midrange & treble drivers and separate readings were also taken at 1m with the 600mm plots giving the most accurate data. Crossovers are at 200hz 16th order and 2.5khz 20th order, both linear phase (it has been recommended I experiment with lower order slopes and different Xover points which I will do although DEQX algorithms automatically correct most issues - I can detect no pre-ringing or beaming with this setup so far but I will probably become more picky as the initial love affair fades a bit).
Raw data for the two speakers, measurements in-room as shown above, not perfect but OK to get me started.
LH:
RH:
Impulse response as shown below (first image) was reasonably clean until 10.7ms with a small floor reflection affecting the reading. This means that the data is fairly accurate down to around 120hz so I have only corrected from 150hz to 6k for now.
For comparison I have shown an outdoor measurement of my older OBs below this which is 'anechoic' to 31ms at 1m (corrupted only by tiny ripples after that - reflections from the mic/stand). I plan to do a similar outdoor measurement of the new speakers sometime in March or April, weather permitting. They will be on a 1m high platform above open lawn & nothing will create significant reflections but the drivers need to be fully burned in before I do this.
Here's frequency response after group delay, time & phase alignment 150hz - 6khz. The uncorrected excursions from the upper response of the tweeters are caused by predominantly ceiling reflections as these are very tall speakers - outdoor measurements will resolve that later. Although bass extends fairly flat to around 50hz, the room reflection at 10.7ms would affect a calibration & I know from previous experience that this would add a nasal quality which is completely absent from a typical OB speaker:
So, I guess the real question is, how do they sound so far?
Very good indeed - not too dissimilar to my other OB speakers but with faster and more open bass, well integrated with the transient dynamics of the midrange which is what I was aiming for. These drivers also have quite a long way to go before they have completely loosened up. I have accurately time/phase aligned both existing subs with 4th order LR crossovers & it's noticeable if I switch them out.
Overall this setup sounds absolutely huge in both height and width & even though I am not tampering much with the treble response at this stage I must say it is already smooth sounding (Boris Blank Anthology FLAC album was used for setup which is very clean, dynamic and well mastered & I've listened to a lot of other stuff since, vinyl & digital). The soundstage has great depth whenever the original mastering includes it & I notice differences in the acoustics of the recording space more than I did with any other speakers in this room.
These have a way to go but already I'm getting probably 80-90% of the result I believe I can achieve later with more accurate measurements and calibration.
....the two DEQX processors are indicated in this image, HDP-5 black, HDP-3 silver.
(note: The subs are already corrected & time aligned to the OBs by the second DEQX unit using previous outdoor measurements - their response is almost flat down to 16hz in-room & not 'correcting' the OBs below 150hz doesn't mean those frequencies are not being reproduced by the speakers, it's just that they use the raw driver response as shown in the measurements whereas DEQX adjusts to a time-group delay-phase-frequency aligned flat response wherever you 'window' the measurements, in this case so far - just between 150hz-6khz. Everything above & below that is as shown in the LH & RH raw data measurement images)
Completed rear frames: Beyma 15" drivers were mounted to their baffles prior to fitting as they are so heavy, Alpair drivers were added to the fronts after this picture was taken. All the frame mounting holes & threaded inserts lined up correctly & the weight of the Beymas did not pull things out of alignment.
Cabling connected:
Side view of a completed speaker:
Graham Slee amps powering the Alpair 7s @ frequencies above 2khz, these are very sweet & transparent sounding:
Initial readings taken in the centre of the room (each speaker moved to this position with some basic sound absorbtion as shown). The mic (calibrated Earthworks M23) was at 600mm midway between midrange & treble drivers and separate readings were also taken at 1m with the 600mm plots giving the most accurate data. Crossovers are at 200hz 16th order and 2.5khz 20th order, both linear phase (it has been recommended I experiment with lower order slopes and different Xover points which I will do although DEQX algorithms automatically correct most issues - I can detect no pre-ringing or beaming with this setup so far but I will probably become more picky as the initial love affair fades a bit).
Raw data for the two speakers, measurements in-room as shown above, not perfect but OK to get me started.
LH:
RH:
Impulse response as shown below (first image) was reasonably clean until 10.7ms with a small floor reflection affecting the reading. This means that the data is fairly accurate down to around 120hz so I have only corrected from 150hz to 6k for now.
For comparison I have shown an outdoor measurement of my older OBs below this which is 'anechoic' to 31ms at 1m (corrupted only by tiny ripples after that - reflections from the mic/stand). I plan to do a similar outdoor measurement of the new speakers sometime in March or April, weather permitting. They will be on a 1m high platform above open lawn & nothing will create significant reflections but the drivers need to be fully burned in before I do this.
Here's frequency response after group delay, time & phase alignment 150hz - 6khz. The uncorrected excursions from the upper response of the tweeters are caused by predominantly ceiling reflections as these are very tall speakers - outdoor measurements will resolve that later. Although bass extends fairly flat to around 50hz, the room reflection at 10.7ms would affect a calibration & I know from previous experience that this would add a nasal quality which is completely absent from a typical OB speaker:
So, I guess the real question is, how do they sound so far?
Very good indeed - not too dissimilar to my other OB speakers but with faster and more open bass, well integrated with the transient dynamics of the midrange which is what I was aiming for. These drivers also have quite a long way to go before they have completely loosened up. I have accurately time/phase aligned both existing subs with 4th order LR crossovers & it's noticeable if I switch them out.
Overall this setup sounds absolutely huge in both height and width & even though I am not tampering much with the treble response at this stage I must say it is already smooth sounding (Boris Blank Anthology FLAC album was used for setup which is very clean, dynamic and well mastered & I've listened to a lot of other stuff since, vinyl & digital). The soundstage has great depth whenever the original mastering includes it & I notice differences in the acoustics of the recording space more than I did with any other speakers in this room.
These have a way to go but already I'm getting probably 80-90% of the result I believe I can achieve later with more accurate measurements and calibration.
....the two DEQX processors are indicated in this image, HDP-5 black, HDP-3 silver.
(note: The subs are already corrected & time aligned to the OBs by the second DEQX unit using previous outdoor measurements - their response is almost flat down to 16hz in-room & not 'correcting' the OBs below 150hz doesn't mean those frequencies are not being reproduced by the speakers, it's just that they use the raw driver response as shown in the measurements whereas DEQX adjusts to a time-group delay-phase-frequency aligned flat response wherever you 'window' the measurements, in this case so far - just between 150hz-6khz. Everything above & below that is as shown in the LH & RH raw data measurement images)
Last edited:
Interesting set-up. I also have the same double-DEQX for development and started with Alpair 12P. Latter took 600 hours' gentle run-in before measurements. With 2 x 18" woofers/side, produced the most accurate vocals ever...only bass "needed" more work. Subsequent improvements have delivered enjoyment beyond $100K+ speakers. Am more than happy using DEQX, as are Kyron. Hopefully, I'll produce more compact 2- and 3-way models soon, after several further breakthroughs over the past year - but i started with 12p and DEQX! Good Luck.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.