FR58EX and AC130F1 micro-FAST / WAW

Founder of XSA-Labs
Joined 2012
Paid Member
I have an old pair of Boston Acoustics HD5's with rotted foam woofer surrounds lying around for past couple of years. Always been meaning to replace the surrounds... I got some new Aurum Cantus AC130F1 drivers in the other day (same ones used in Jeff Bagby's Continuum, the advanced tech redux of the LS3/5A) and I needed a box to test them in. The HD5's were a perfect fit. They actually sounded almost passable when hooked up to the stock XO and paired with the 3/4in mylar dome tweeter. So I got to thinking, what if I swapped that tweeter for a nice 2in full range driver, and active bi-amped it for a very cute little FAST desktop monitor. Cute because the boxes are 6.25in W x 9.5in H x 7.0in D.

Edit: typo the title says FR56EX but actually FR58EX.
Fountek FR58EX 2" Neodymium Full Range Speaker Driver

I have some Fountek FR58EX's (thanks to Byrtt!) lying around from the recent data collection for the Round 5 of the Subjective Blind test thread. They sound like aluminum dome tweeters but can be crossed very low as they have an fs around 150Hz. One of the nicer aluminum drivers that don't ring. The AC130F1, is one of the nicest mid bass drivers I have tested - it has a super flat response, not unlike a ScanSpeak 10F, but only goes up to 10kHz and has no cone break up peaks.

I made a small 4-side Dagger for the FR58EX (3in base x 5.75in high foam core panels x qnty 4), stuffed tightly with polyfill. I went with 4 sides to make the volume larger and to accommodate the square cutout for the tweeter. I removed the stock terminal cup/XO board and replaced with a plain terminal cup. I added a 5-way banana post binding terminals for the woofer. Stuffed the box (circ 5.7 liters) with stock polyfill and added some more for good measure.

Here is a photo showing the detail of the Dagger as mounted with the FR58Ex:

506251d1443345174-fr56ex-ac130f1-micro-fast-fr56ex-ac130f1-fast-03.png


Here is a photo of the completed speaker front:

506253d1443345174-fr56ex-ac130f1-micro-fast-fr56ex-ac130f1-fast-01.png


Back showing the terminal cup and binding posts for bi-amping:

506252d1443345174-fr56ex-ac130f1-micro-fast-fr56ex-ac130f1-fast-02.png


Edit - update Sept 28, 2015: added diffraction control Scotchbrite pads - it really works as can be seen in measured response below:

506447d1443446626-fr58ex-ac130f1-micro-fast-diffraction-control-square-scotchbrite.jpg


Effect of diffraction control (brillo pads):

506448d1443446626-fr58ex-ac130f1-micro-fast-diffraction-control.png


I used miniDSP and a TPA3116D2 amp to do the active bi-amping. Since both drivers are essentially full range with clean flat response, I decided to go with a first order crossover at about 1kHz so that the filters have lots of overlap beyond the XO point. The full range driver goes down to 250Hz easily (2 octaves), similarly, the AC130F1 goes up to 4kHz easily (2 octaves). Why go with a first order XO, aren't those kind of archaic and no one in their right mind would use it given all the nice 2nd and 4th order Linkwitz Riley filters that are so prevalent? Well, the first order XO, if done right will give me a speaker that is flat phase and transient perfect. Kind of of like a single driver full range speaker, but in this case, a 2-way that is flat phase and transient perfect. What is so great about transient perfect? Super realistic percussion and plucked instruments if recorded live with minimal post production. Small jazz ensembles sound real and like they are there in your listening room.

Here is the measured acoustical crossover plot. The relative sensitivities of the the FR58EX an AC130F1 matched perfectly, no gain adjustment was needed. There is some mild EQ to get the responses sort of close to a 1st order and to flatten the response a bit.

506254d1443345174-fr56ex-ac130f1-micro-fast-fr56ex-ac130f1-fast-xo.png


Here is the resulting impulse response and step response. IR is very clean with little overshoot and just a little bit of fuzz which may be reflections from the driver mount and cabinet lip. The SR is indicative of a transient perfect speaker with a right triangle shape. Dare I say it, can a Continuum do this trick? Don't think so with LR2 slopes - the leading edge is probably negative going:

506255d1443345174-fr56ex-ac130f1-micro-fast-fr56ex-ac130f1-fast-ir-sr.png


Here is the measured frequency response and phase, pretty linear within a 30 deg band:

506256d1443345174-fr56ex-ac130f1-micro-fast-fr56ex-ac130f1-fast-phase.png


And here is the harmonic distortion:

506257d1443345174-fr56ex-ac130f1-micro-fast-fr56ex-ac130f1-fast-hd.png


I listened to a few songs and it sounds really good. It renders percussion like it was meant to be heard. Below are some sound clips to demonstrate the time accuracy and also vocals so you get the sense of the balance in the voicing. Change .asc to .mp3 to listen.

Square waves at 1700Hz XO frequency:
506464d1443456488-fr58ex-ac130f1-micro-fast-1700hz-square-wave.png


If you have some old cabinets lying around, consider converting them to a new FAST.

Update Nov 28, 2015: Stereo Pair completed.
517017d1448742934-fr58ex-ac130f1-micro-fast-fr58ex-ac1301f1-fast-stereo.jpg


Update Dec. 14, 2019 - a passive 1kHz acoustic 2nd order XO design for this speaker based on measurements by hdspeakerman:
802027d1576350987-fr58ex-ac130f1-micro-fast-waw-fr58ex-ac130f1-xo-schematic-v002-jpg


802028d1576350987-fr58ex-ac130f1-micro-fast-waw-fr58ex-ac130f1-xo-freq-phase-v002-jpg
 

Attachments

  • FR56EX-AC130F1-FAST-03.png
    FR56EX-AC130F1-FAST-03.png
    422.7 KB · Views: 4,939
  • fr56ex-ac130f1-fast-hd.png
    fr56ex-ac130f1-fast-hd.png
    149.4 KB · Views: 4,926
  • fr56ex-ac130f1-fast-phase.png
    fr56ex-ac130f1-fast-phase.png
    90.4 KB · Views: 6,732
  • fr56ex-ac130f1-fast-ir-sr.png
    fr56ex-ac130f1-fast-ir-sr.png
    60.9 KB · Views: 10,891
  • fr56ex-ac130f1-fast-xo.png
    fr56ex-ac130f1-fast-xo.png
    78.2 KB · Views: 9,090
  • FR56EX-AC130F1-FAST-01.png
    FR56EX-AC130F1-FAST-01.png
    361.5 KB · Views: 5,836
  • FR56EX-AC130F1-FAST-02.png
    FR56EX-AC130F1-FAST-02.png
    360.1 KB · Views: 4,330
  • FR56EX-AC130F1-FAST-Clip-1.asc
    FR56EX-AC130F1-FAST-Clip-1.asc
    1.7 MB · Views: 186
  • FR56EX-AC130F1-FAST-Clip-2.asc
    FR56EX-AC130F1-FAST-Clip-2.asc
    1.7 MB · Views: 137
Last edited:
Nice recycle. The step response might be your best yet? I guess the only down side with this speaker (in context with its intended use) is that someone might blow the tweeter if they get carried away and turn it up, but since it's intended to be a work desk speaker high SPL isn't much of a concern. How loud can you drive the tweeter in that crossover config before distortion or IMD says "turn it down"?
 
wow that looks great! I have also been wondering about trying some first order crossovers to get flat phase and transient perfect. This is a great application for it with two near full-range drivers.
I still can't get over how good all the plots look: crossover/FR, impulse/step, flat phase, and good distortion.
 
That's the nice thing about using a full range driver as a tweeter, compared to a dome tweeter, it has gobs of xmax (2.9mm) and xdamage is even more. I think you can move this ~5mm before xdamage so with a -12dB cut at 250Hz we can probably put full thermally (15w rms and 30w Max) limited power to the driver. This driver is capable of 97dB at peak power so that's probably the upper limit when you start hearing distortion - the woofer had a higher rating and is probably limited to around 102dB at xmax. With 5dB baffle step loss they are well matched up to the max of the Fountek.
 
Last edited:
wow that looks great! I have also been wondering about trying some first order crossovers to get flat phase and transient perfect. This is a great application for it with two near full-range drivers.
I still can't get over how good all the plots look: crossover/FR, impulse/step, flat phase, and good distortion.

Thanks, Jeshi. A first order slope like this can only work because both drivers are essentially full range and flat response. I have tried this with less than a full range woofer and it doesn't work because the breakup peaks of the woofer intrude. That's when resorting to a quasi transient perfect Harsch XO is a good option as its 4th order on the woofer and 2nd order on the tweeter.

I have never heard a first order XO sound this clean before. Listen to the drum rim shots in clip 1. On a non transient perfect speaker it doesn't sound real.
 
You could do a passive xo with a minimum of 1 coil, 1 cap, 2 resistors. But in reality there is some shaping and shelving to do to flatten response so maybe add one more coil and resistor for BSC. But you can't do it with a flat baffle - will need about 0.16ms delay on tweeter. That is same as a 2.51inch stepped baffle or a waveguide of same depth.
 
Why go with a first order XO, aren't those kind of archaic and no one in their right mind would use it given all the nice 2nd and 4th order Linkwitz Riley filters that are so prevalent? Well, the first order XO, if done right will give me a speaker that is flat phase and transient perfect. Kind of of like a single driver full range speaker, but in this case, a 2-way that is flat phase and transient perfect. What is so great about transient perfect? Super realistic percussion and plucked instruments if recorded live with minimal post production. Small jazz ensembles sound real and like they are there in your listening room.

Exactly! To me high fidelity means it sounds like the real thing. 'Advanced' crossovers can sound smooth and fine, but it seems to me that most peoples reference is recorded music and/or some sort of theoretical ideal of a smooth sounding stereo. I spent many years playing in bands. IMO a better reference is the sound of real instruments and voices.

Anyhow, nice job on that. 😀And I'm glad you finally tried the old fashioned archaic first order crossover. Really, the only way to do it properly is by utilizing full range drivers.

There's a few ways it can be done. The small FR crossed at 1khz like this is probably simplest, as you're far enough away from driver Fs to not run into any problems with a simple parallel crossover. I also like using a slightly larger - say 5" FR in a quasi 3 way. Crossover points about 5-700 and 4 or 5k. (suspiciously like a 70's monkey coffin 3 way😉) Usually an aperiodic loading for the full range is required.
 
Wow nice build/idea including beautiful measurements and sound clips.

This builds sound clip 1 when doing quick compare to "comparison round 4's tracks 2" the sound signature seems close to B/D (TG/TC9FD).

Phase trace for AC130F1 looks like textbook in the 90º lag point is around low 90Hz area just as a simulation program predicts.

Phase trace in 1-2kHz area takes a little dip if one virtual smooth it by eyes, think that could indicate some fine tuning could be worth looking into, its probably the spl difference between LF and HF but could be slopes precision or actual XO point too. Electric XO point set at 1kHz but looks more like 1500-1800Hz acoustic, attach below how XO area vertical lobes look like with unsymmetrical lobe that BW1 always 90º apart phase perform for 750/1000/1250/1500/1750/2000Hz, guessed ctc distance to 110mm by looking at picture.

Second attachment is vertical lobe at 1kHz if drivers in phase at XO point as with even order filters and guess Harsch would be as that.
 

Attachments

  • Lobeing_BW1.PNG
    Lobeing_BW1.PNG
    115.8 KB · Views: 196
  • Lobeing_even_order.PNG
    Lobeing_even_order.PNG
    20.3 KB · Views: 198
Last edited:
Thanks Byrtt. It was a quick and dirty effort on my part to get the xo last night at 3am. With fresh eyes I can go in a record raw responses and import as FRD files into PCD and try to tweak and shape the responses to match textbook perfect 1st order Butterworth and see if that improves things any. To my ear this sounds very good right now but I can see that there is some non ideal behavior. I guess your point is that of the slopes are not perfect you get vertical lobing that is frequency dependent and that can ruin spatial image coherence?

Really, this woofer is too nice to use in a box like this. It was an experiment although if it sounds nice enough I may keep it. As it is, the cost of the drivers alone is more than the retail cost of the original speaker! So it better sound good. 🙂 indeed it sounds better than the HD5 could have ever have hoped to sound. Just look at the OEM woofer compared to this one - cast frame vs stamped, magnet is 3x the size, carbon fiber inpreganted polycone, and the tweeter is now aluminum dome high excursion with Nd motor and Dagger rear chamber, XO is advanced all DSP with time delay. Only thing original is the box. 😀

But a good wood box is always the stumbling point for me given no tools or wood shop.
 
Very much liking this little prototype.

Not totally convinced the 1st order thing, even with these drivers (but that's just my opinion)

As for the wooden box and having no workshop...

Well that's DIY - I bought and got my birch ply cut at a common DIY retailer locally. I jigged driver holes, sanded and finished my current speakers in my one bedroom apartment.

With respect, you have done some outstanding prototypes with foam core (and EVA along with Jeshi), and I guess if I had the chance to play with so many drivers id also do many tests - if only for my benefit. It reminds me of my early teens wanting to try loads of different drivers (mostly Audax) and not being able to afford the really cool seeming drivers.

With the thought you have put in to foamcore, I have no doubt you can pick some drivers and make the cabinet more permanent - dare I say it? Settle on a design or configuration and tweak until its to your taste.

I fancy a good plywood/silicone/EVA/silicone/wool or cotton felt composite (going from external to.internal lams)

Or even a Scan 10F/FR58x/other Blanda bowl build? 😀
Using the 8" bowls for FR and the larger ones for the AC130/other
 
Last edited:
might the jazz reference clip 1 be posted to do evaluations? - the first 18 seconds is a nice uncluttered passage which wasn't in round 4 clip 2

here you go...

It's sourced from mono (R) because I recorded clip one as right channel only. You need to take clip one and delete one channel and recombine to both ears for true comparison.

Or let me do that for you here, clip 1 again from speaker but presented in mono.

Do note that these speakers start rolling off below 90Hz so the bass will be lacking.
 

Attachments

With the thought you have put in to foamcore, I have no doubt you can pick some drivers and make the cabinet more permanent - dare I say it? Settle on a design or configuration and tweak until its to your taste.

The problem with DIY, is that the process of designing and building, then finally listening to it and hearing the WOW or AHA moment is kind of addictive. Like gambling or fishing but this is much cleaner and cheaper. For me, it's the process that I enjoy. Although I have come across several "keepers" that are still in my lab that I listen to once in a while, or all the time. You know what gets the most air time in x's lab? The mini Karlsonator, the 10F/RS225, and my full range synergy horn with bass horn for sub. For easy playability though, the K'nator gets the most use. It just has gobs of full warm bass and clean impulse response. Despite a few aberrant frequency response dips here and there, it is quite the special speaker almost all the time.

Now, this little FR58EX/AC130F1 is kind of growing on me. I can tell it's good if it makes me want to build a matching stereo pair. It's a little speaker and that makes it very attractive for desk use and not having much bass below 90Hz is a good thing when you don't want bass bothering the rest of the family late at night.
 
thanks xrk971 - that'll allow me to record some of my toys. If a Dagger were made of wood - would it be worse in performance than foam? - thin plywood can resonate at high Q - in contrast, foam seems somewhat acoustically "translucent"

hey GregB - any driver picks for the quasi 3-way which would sound "real" and allow pretty high peak spl?
 
Last edited:
Scotchbrite makes a difference

Here is a simple (and inexpensive) little add on that I have found generally makes the response smoother in the highs, and improves off axis uniformity. I have not measured the difference yet but believe I hear improved highs off axis at 45 deg. One scouring pad fits perfectly.

506347d1443388239-fr58ex-ac130f1-micro-fast-scotchbrite.jpg
 

Attachments

  • scotchbrite.jpg
    scotchbrite.jpg
    62.3 KB · Views: 1,572
Here is a simple (and inexpensive) little add on that I have found generally makes the response smoother in the highs, and improves off axis uniformity. I have not measured the difference yet but believe I hear improved highs off axis at 45 deg. One scouring pad fits perfectly.

506347d1443388239-fr58ex-ac130f1-micro-fast-scotchbrite.jpg
These are looking more and more like an LS3/5a now.
20120615_2429453.jpg


Those aurum cantos AC130f1 do look like excellent drivers.
 
Thanks Byrtt. It was a quick and dirty effort on my part to get the xo last night at 3am. With fresh eyes I can go in a record raw responses and import as FRD files into PCD and try to tweak and shape the responses to match textbook perfect 1st order Butterworth and see if that improves things any. To my ear this sounds very good right now but I can see that there is some non ideal behavior. I guess your point is that of the slopes are not perfect you get vertical lobing that is frequency dependent and that can ruin spatial image coherence?

Really, this woofer is too nice to use in a box like this. It was an experiment although if it sounds nice enough I may keep it. As it is, the cost of the drivers alone is more than the retail cost of the original speaker! So it better sound good. 🙂 indeed it sounds better than the HD5 could have ever have hoped to sound. Just look at the OEM woofer compared to this one - cast frame vs stamped, magnet is 3x the size, carbon fiber inpreganted polycone, and the tweeter is now aluminum dome high excursion with Nd motor and Dagger rear chamber, XO is advanced all DSP with time delay. Only thing original is the box. 😀

But a good wood box is always the stumbling point for me given no tools or wood shop.


Respect it was a quick and dirty effort in middle of the night and as said it is a nice build including data and sound clips, comment was not to bite at all but based our previous investigations into those BW1 show they so sensitive for small errors in slope and correct spl mix from LF and HP part plus need precision acoustic offset executed. Probably because they cooperate over so wide a bandwidth they need much more care for precion in setup than steeper filters where cooperation bandwidth is narrow.


.....I guess your point is that of the slopes are not perfect you get vertical lobing that is frequency dependent and that can ruin spatial image coherence?.....


Hope have this right as i see it : ) vertical lobing is inherent part of uneven filter slope, that's not just regarding errors in your settings for BW1 filter but because of soft slopes most of speakers bandwidth is in a XO region. LP verse HP is never in phase as even order filters is, the always 90º phase apart between LP/HP means unsymmetrical vertical lobe because in acoustic domain we sum with a ctc distance spacing and this means listening hight is relative very sensitive. When summing two BW1 filters in electric domain gives a perfect summing as first picture below, no ringing at all as with higher order but also there not the ctc spacing parameter which makes the difference for BW1 electric verse acoustic. Below picture 2 is BW1 1kHz XO point vertical lobe with ctc distance set to zero (coxial speaker or electric sum) and with 110mm as with this build, picture 3 is even order LR4 1kHz XO point vertical lobe with ctc distance set to zero and with 110mm. So reason showing the lobe plots was to remind we get tilt or sensitive listening hight especially because XO area is so wide when there is ctc spacing verse a electric device where BW1 sum is invincible and perfect in measurements.

Picture 3-6 is quick and dirty made BW1 verse Harsch in electric DSP loop, system bandwidth is set to 92Hz-20kHz
 

Attachments

  • IR-SR.png
    IR-SR.png
    26.3 KB · Views: 679
  • SR.png
    SR.png
    49.1 KB · Views: 121
  • IR.png
    IR.png
    46.6 KB · Views: 128
  • Phase.png
    Phase.png
    53.8 KB · Views: 134
  • FR.png
    FR.png
    84.3 KB · Views: 145
  • Lobeing_even_order.PNG
    Lobeing_even_order.PNG
    50.9 KB · Views: 127
  • Lobeing_BW1.PNG
    Lobeing_BW1.PNG
    36.7 KB · Views: 366
Byrtt,
Your electrical sims are right one with what I see. I always have that little rise before the sharp leading edge with my Harsch XO. Only with BW1 symmetric am I able to get the steep sudden cliff like rise on the leading edge. I have never gotten a Harsch SR to look like the one shown by Harsch in his paper.
 
Think its the plots X and Y scaling that makes the difference also the one from the Harsch paper have no real world system bandpass cutoff filters.

Picture 1 is copy the Harsch XO paper at picture 2 i set X Y scale close to that paper and it now looks lot more alike without the revealing rise before the sharp leading edge, and had this build not had high pass f3 at 92Hz the approx. 1,8mSec decay had been longer.

Picture 3-4 is IR/SR woofer section where sharper rise less ringing for BW1 is clear.

Picture 5-6 is IR/SR tweeter section where less ringing for BW1 is clear.
 

Attachments

  • Harsch.PNG
    Harsch.PNG
    7.4 KB · Views: 126
  • SR.png
    SR.png
    41.7 KB · Views: 108
  • IR_LP.png
    IR_LP.png
    54.7 KB · Views: 103
  • SR_LP.png
    SR_LP.png
    52.9 KB · Views: 93
  • IR_HP.png
    IR_HP.png
    44.4 KB · Views: 98
  • SR_HP.png
    SR_HP.png
    45.9 KB · Views: 100
Last edited: