10F/8424 & RS225-8 FAST / WAW Ref Monitor

To bring GD as low as possible you right the fast one is get rid of the electric 27Hz HP filter in miniDSP and should give approx. 6,5mS at 40Hz, and if recordings is free of low frequency rumble it should be no problem. Will probably end up around having 8-10mS delay at 40 Hz.

Present setup is minimum phase domain all over and the tools for GD at low frequency seems:

Electric linkwitz transform, the more low we transform the lower or less GD we get, but trade off is the higher gain needed the lower we like to transform, the actual one from 52Hz down to 30Hz takes out 9,5dB SPL in BW Q config.

Electric a fair high value capasitor can be used for woofer and bring some small better GD timing but don't think these excellent drivers and setup deserve a cap that most often has own signature and not being 100% transparent.

Electric one giving only a small number is go from the 350Hz 2.order XO to 1.order 500-700Hz and as example at 600Hz XO will gain 0,630mS less GD at 40Hz.

Acoustic volume tricks is also possible but think alignment is probably carefully set at a right point where Xmax load is in control.

To get better measured GD timing than about 10mS at 40Hz think next step is FIR filters but they will cost delay between real time at the track played and the sound into room.
 
Last edited:
Welcome xrk971,
For info here is FR and GD for a linkwitz transform from BW Q at 52Hz to BW Q at 30HZ. The fun thing is from 500Hz down to 39Hz timing is faster than real time :rolleyes:.
 

Attachments

  • FR.png
    FR.png
    37.3 KB · Views: 2,577
  • GD.png
    GD.png
    56.2 KB · Views: 2,568
Founder of XSA-Labs
Joined 2012
Paid Member
Welcome xrk971,
For info here is FR and GD for a linkwitz transform from BW Q at 52Hz to BW Q at 30HZ. The fun thing is from 500Hz down to 39Hz timing is faster than real time :rolleyes:.

Funny thing, the shape of the curve reminds me of the paradox of quantum tunneling and superluminal speed of particle penetration through a quantum barrier. Apparently, this is not any violation of physics... :)

Alternate View Column AV-75

So we hear the music before it happens :D
 
If I were in your shoes I'd be building the second enclosure by now ;).
The 10F might surprise you with 6 dB/Oct filters, at least if I remember the distortion graphs right. Just try it and measure... The roll off from the enclosure the 10F is in will help to control it I guess.
FIR filters would be fun to try too! But first get that other cabinet ready!
 
Founder of XSA-Labs
Joined 2012
Paid Member
0:24s and 0:25s in clip is remarkable clearer/"easier to follow" than on my cd. Mono could be reason too, but it had me listening a couple of times:)

I wonder why that might be? I am playing from a CD made from 96kHz of a .wav file of high res recording. Still, playing from a CD through analog output which makes it defacto 44.1kHz 16-bit material.

I am listening to the 24-25 second spot and hear what you are referring to as some low level detail of a rapid drum rattle sound. Cool.
 
Very kool new project x. :D The 10F seems to be a superb driver, and a FAST like this is the ideal implementation for it.

Since someone is going to ask eventually, allow me... How would you compare the 10F to your fave Vifa TC9FD WRT realism and detail? The sample sounded quite good, but near the limits what will be masked by the signal chain electronics and mp3 compression.

RE the crossover: if you're not getting a deep null with a normal phase connection and 2nd order, it's not going to have perfect phase matching with typical reverse connection either. 3rd order Bessel is worth a try. 1st order acoustic is my favorite if feasible. It just sounds real. The downsides are lobing, SPL limitations from excursion limits, and the wide bandwidth needed from both drivers, not to mention any corrections needed to fix driver issues beyond the xover knee. You'd probably need to cross at ~500+, and ideally have flat response for two octaves on both sides of that. IOW the woofer to 2k and the 10f to 125.
 
Founder of XSA-Labs
Joined 2012
Paid Member
If I were in your shoes I'd be building the second enclosure by now ;).
The 10F might surprise you with 6 dB/Oct filters, at least if I remember the distortion graphs right. Just try it and measure... The roll off from the enclosure the 10F is in will help to control it I guess.
FIR filters would be fun to try too! But first get that other cabinet ready!

You are right about the second box! No second thoughts here - just need to find time. Sureply baffle is cutout already. Need to buy one more 2ftx2ft pink foam panel as my initial estimates of materials needed was off. Btw, I use lots of pink Fiberglas stuffing - how apropos. :)

I can't wait to hear this in stereo, and also maybe to get a more powerful amp than 50 watts/ch to really experience the kick. Btw, this is one of the few speakers I have messed with (aside from my big bass horn) that is able to play the extreme bass line of Suzanne Vega's "Headshots" without distorting.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Very kool new project x. :D The 10F seems to be a superb driver, and a FAST like this is the ideal implementation for it.

Since someone is going to ask eventually, allow me... How would you compare the 10F to your fave Vifa TC9FD WRT realism and detail? The sample sounded quite good, but near the limits what will be masked by the signal chain electronics and mp3 compression.

RE the crossover: if you're not getting a deep null with a normal phase connection and 2nd order, it's not going to have perfect phase matching with typical reverse connection either. 3rd order Bessel is worth a try. 1st order acoustic is my favorite if feasible. It just sounds real. The downsides are lobing, SPL limitations from excursion limits, and the wide bandwidth needed from both drivers, not to mention any corrections needed to fix driver issues beyond the xover knee. You'd probably need to cross at ~500+, and ideally have flat response for two octaves on both sides of that. IOW the woofer to 2k and the 10f to 125.

GregB,

Thanks. :D

Fortunately, the RS225-8 is one of the flattest woofers in existence usable up to at least 2k. The 10F in the current Dagger starts falling off at 150Hz I think. So I can try the 500Hz 1st order and see what happens.

I did get a pretty deep null in 2nd order which went away when I flipped polarity on one of the drivers.

I think the TC9FD and TG9FD are hole-for-hole matches with the expensive 10F sibling. I could just plop it in and listen. However, the weaker motor of the TG9/TC9 require a larger volume than 1.1 liter to get a sufficient deep extension. XO may have to be set a bit higher back to 400Hz or so...

Although, as you say mp3 may be the limit so a lossless FLAC file may have to suffice, albeit very short FLAC clip.
 
The 10F might suit with 6dB/octave electrical filters, but the RS225 certainly does not. It has a massive peak in the third order distortion at 1200Hz that will not be properly suppressed if one switches to a first order electrical filter. You should stick with a 2nd order.
 
Founder of XSA-Labs
Joined 2012
Paid Member
The 10F might suit with 6dB/octave electrical filters, but the RS225 certainly does not. It has a massive peak in the third order distortion at 1200Hz that will not be properly suppressed if one switches to a first order electrical filter. You should stick with a 2nd order.

Thanks for the heads up on the 3rd order HD peak. I haven't used it long enough to realize this limitation. It appears to be very flat response wise through 1200Hz.
 
What i said about GD prediction in 8-10mS at 40Hz was some guessing looking at the system bandpass. Did a new run and see it could end up better but it depends following is true, that actual driver have TS data as printed in datasheet and actual box volume is tuned to exactly BW Q then at 40Hz says 5,9mS. Plot below blue is new prediction and olive with the old HP filter. Could be curious to hear data where real measurement ended up without the electric 27Hz HP filter.

Filters used for test was:
1. Acoustic high pass filter at 52Hz 2.order BW for model of sealed box.
2. Electric linkwitz transform at 52Hz BW down to BW at 30Hz.
3. Electric LR 2.order at 350Hz XO filter.

DSP setup for filters testing is same as explained #10.
 

Attachments

  • GD.png
    GD.png
    56.2 KB · Views: 2,130
Founder of XSA-Labs
Joined 2012
Paid Member
Byrtt,
Again, thank you for the great analysis. My internal box volume truly is 24liters but then there is some bracing and RS225 self volume so I am guessing closer to 22 liters when all is said and done. However, doesn't the pink fiberglass stuffing change the box Q as well as some wall flex of the XPS walls? I can measure the box impedance with the wonderful DATS tool and see what the impedance sweep provides. The RS225-8 measured parameters were close but Qts a bit higher than spec. IIRC I measured Qts of 0.43 vs spec of 0.38. Fs was close as well as Vas.
 
Nice one.

If I was needing near field studio monitor this design would be an obvious choice. Am looking forward to hearing them someday... However, I'm not a fan of off center tweeters (even though the benefits are understood). How much damage would a trapezoid profile do to the HF? The crossover point appears low, so it seems baffle surface area is helpful for the tweeter no?
 
Founder of XSA-Labs
Joined 2012
Paid Member
Nice one.

If I was needing near field studio monitor this design would be an obvious choice. Am looking forward to hearing them someday... However, I'm not a fan of off center tweeters (even though the benefits are understood). How much damage would a trapezoid profile do to the HF? The crossover point appears low, so it seems baffle surface area is helpful for the tweeter no?

My original plan, believe it not was a trapezoid baffled Dagger much like the one seen earlier in the Hypercube FAST thread. Impatient and wanting a quick build - and also never having built a traditional prismatic off center tweeter box before, I decided to give it a go to avoid having the complications of a trapezoid upper driver. The smaller baffle size would indeed have some more baffle step loss. I think the larger area is helping to keeping flatter.

They do sound very nice and the flat response is indeed useful as a studio monitor. Hope you get to hear them someday. If you are curious what the trapezoid baffle does, the response posted in the TC9FD Most Bang for the Buck thread would give you and idea. That baffle is 7in at the top and 12in on the bottom and 13 or 14in tall IIRC.

Here is the trapezoidal baffle:

469721d1425569701-vifa-tc9fd18-08-best-bang-buck-10f-8424-min-phase.png


Compared to offset rectangular baffle:

479717d1430079914-10f-8424-rs225-8-fast-ref-monitor-10f-8424-rs225-8-fast-lt-eq-4ms-gate-fr.png[
 
Last edited: