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

[xrk971]

Some sound clips of this speaker recorded with my phone and powered by my SuSyLu 100w Class A amp:

Test of SuSyLu 100W Class A Amp - YouTube

Example of speaker in dark “espresso brown” stain and satin lacquer with Duratex coated baffle.

Example of beautiful build by as8912:

Another Example of a floor standing FAST TL by Plott:

Example of a floor standing TL version of this speaker by Moutik:

Example of a smaller sealed version with PS95-8 full range drivers by Jimk04, here.

I wanted a speaker design worthy of the exceptionally high performing ScanSpeak 10F/8424 Discovery 3.5in full range driver. I decided that a short sealed Dagger TL coupled with a sealed bass unit would give me the best overall transient response and clarity. I decided to mate the 10F/8424 with the superb Dayton RS225-8 aluminum cone woofer for bass duties. The design ended up as a 24 liter volume for the woofer and a 1.1 liter 3-sided pyramid Dagger for the 10F/8424. The important external dimensions are: 10in wide baffle and 24 liter internal volume (sealed). You can make the height and depth to suit your individual case, but it will need to be deep enough to house the Dagger. The internal dimensions I settled on were 8in wide x 12in deep x 16in tall. The short sealed TL is made of three 6in wide x 12.5in long triangles. Construction will use 1in thick pink XPS foam and regular foam core for the short TL. The front baffle will also use a thin sub-floor plywood in order to provide adequate support for the heavy 8in driver. The design will have the option of a bass reflex vent that can be sealed for use with a Linkwitz transform, or opened up for some additional bass when group delay is not an issue. The system will use miniDSP for XO and EQ duties and bi-amped. Here are predicted results in 4pi space with baffle step and diffraction effects accounted for. XO will be at 500Hz with 4th order LR. The RS225-8 is an exceptionally flat response driver and choosing a slightly higher XO frequency will keep the distortion on the 10F very low while giving more weight to the mid bass as it comes from an 8in woofer. Calculations show that with Linkwitz transform, bass extension should be about 42Hz (f3) with max SPL around 101dB.

SPL vs freq for sealed case at 24v (no Linkwitz transform) - the hump at 2.5kHz is the baffle diffraction:

Impulse response for sealed case:

Group delay for sealed case is 2.5ms at 100Hz and 5ms at 50Hz:

For bass reflex option, here is SPL vs freq at 24v:

Here is group delay for bass reflex (2.5in dia x 11in long vent - rear firing) case, not too bad actually (about 10ms at 50Hz)

Here is the Sureply thin baffle with driver holes and drivers to show basic look:

Here is construction on XPS enclosure beginning:

Here is the enclosure with Dagger TL internally mounted and bracing, wiring, and terminal cups installed:

Currently waiting for liquid nails glue to dry...

Edit April 27, 2015: New LR2 XO @350Hz and new sound clips of many different genres. Have a listen, this speaker sounds fantastic!

10F/8424 & RS225-8 FAST Ref Monitor - Page 2 - diyAudio

Here is the new LR2 XO @350Hz:

Update April 30, 2015: New Butterworth 1st order linear phase XO (BW1 XO):

Phase for BW1 XO:

Step Response for BW1 XO:

200Hz Square Wave with BW1 XO:

1.6kHz Square Wave with BW1 XO:

New sound clip with 1st order BW1 XO: http://www.diyaudio.com/forums/atta...or-10f-8424-rs225-8-fast-bw1-350hz-clip-b.asc

Edit May 14, 2015: Stereo Pair completed!

Edit Dec 16, 2015: forgot to post final lock-down configuration with Harsch XO:

IR:

Step response:

More info on final setup here: 10F/8424 & RS225-8 FAST Ref Monitor

Distortion:

Group delay about 5ms at 50Hz:

Edit Jan 5, 2018: using passive 1st order transient perfect XO for past 2 years.

Recent implementation with film caps and air core:

frequency response and phase using passive XO:

Impulse and step response:

Latest tweaks to the passive 1st order XO with simulation files:
10F/8424 & RS225-8 FAST Ref Monitor - Page 95 - diyAudio

Edit Dec 3, 2018: BOM for XO is here.

Edit: Dec 27, 2018 - Wxn just built a pair using SB23NRXS45-8 and TG9FD-4 with DSP from a DCX2496 here. The results look great and he did a great job on the cabinet:

Nice waterfall:

Edit: Dec. 27, 2018 - I should have mentioned that Aatto built a nice set using the SB23NRXS45-8 with 10F/8424 as well but using a bass reflex box and miniDSP and also with passive XO, but can't remember what he is currently using here:

Also, forgot to mention the nice set with RS225/10F that I built in espresso brown stain - these are working out real well as the WAF is excellent as evidenced by the fact that they are upstairs next to the TV :

And of course, here is my current setup. These have been my main speakers for over 2 years now and, I keep coming back them because they sound so nice and balanced. The sharp percussion and wonderful stereo imaging also bring me back. Here they are in my speakerlab, now re-done post basement flooding (back in July 2018) with a new engineered wood flooring, so the echos are a mess and sound treatment is definitely needed now that carpet is missing. Looks nice and clean though - the clutter is building up again (mostly foam core speakers in storage around the corners).

Edit Dec 16, 2019: Thanks to JPS64 for making a super layout for the XO PCB. This will make building this speaker that much simpler.

Edit Dec 21, 2019: new TL floor stander for RS225:
10F/8424 & RS225-8 FAST / WAW Ref Monitor

Edit Dec 22, 2019: Just wanted to remind folks that there was a design made for this speaker XO with an RS100-8 and RS225-4 here:
10F/8424 & RS225-8 FAST / WAW Ref Monitor

Edit Mar. 20, 2020: Revised plans for the TL in all 3/4in BB plywood.

Photo of the TLs in action next to the sealed version:

Edit March 30, 2020: Vistaton B80 Variant with modified XO

Edit May 19, 2020: this is the XO for the 4ohm RS225-4. You can use your favorite 8ohm 3.5in full range on top. Just make sure it’s 8ohms and adjust R1 to taste.

Edit July 24, 2023: Keantoken developed a crossover specific to the TC9FD-8 to be used in this speaker with the RS225-8. The response looks very smooth and has a wonderful step response. More info here:

https://www.diyaudio.com/community/threads/10f-8424-rs225-8-fast-waw-ref-monitor.273524/page-196

Schematic:

Measured frequency response appears to have 1st order XO near 1500Hz.

 

[zman01]

Good stuff. Look forward to updates.

 

[RajkoM]

Very nice

 

[xrk971]

Speaker Completed with Measurements

The glue is almost dry - could use a few days to really cure, but I was impatient and took some data anyway.

Here is the completed speaker - looks like a nice standmount studio monitor.

The measurements were made with the speaker on a stand with 10F/8424 at 43in above the floor and mic at 0.5m away. This setup allows use of 4ms gate to look at above 300Hz frequency response without room reflections.

Here is the measurement showing the LR4 XO at 500Hz and the effect of applying the Linkwitz transform (a rather aggressive one to get 30Hz):

Here is the harmonic distortion for no LT (very low HD over the entire range):

Harmonic distortion with LT (about 1% THD at 50Hz at 82dB):

Here is the Impulse response - very sharp response with no ringing:

This is the high resolution detail of the frequency response with 4ms gate and 1/48th octave smoothing (this speaker has a response that is flat within about +/- 2dB. There is no EQ applied to the 10F/8424 just about -3.8dB level matching to compensate for baffle step loss of the woofer:

Here is the measured group delay with the LT turned on (about 12ms at 50Hz - probably due to the very aggressive LT to hit 30Hz):

How does it sound? Well you can imagine from looking at the graphs... Very very nice!

This speaker really has turned out a lot better than I anticipated given the XPS foam construction. The box resonance seems to be minimal and not contributing to the HD in substantial way. It makes the speaker extremely lightweight. I have not installed the vent yet - maybe later, but so far, the bass is quite nice and I am not missing it. The sealed alignment has very tight timing due to minimal group delay.

Here is a first sound clip - change .asc to .mp3 to listen.

 

[5th element]

You cannot go wrong with the RS225 it is an excellent driver regardless of price, the fact it's so cost effective just makes it even better.

I too have used it in a similar configuration, that being a small sealed box with an LT. It sounds wonderful like such and quite surprised me with how much bass it could put out in such a 'limited' configuration, one could say.

 

[xrk971]

You cannot go wrong with the RS225 it is an excellent driver regardless of price, the fact it's so cost effective just makes it even better.

I too have used it in a similar configuration, that being a small sealed box with an LT. It sounds wonderful like such and quite surprised me with how much bass it could put out in such a 'limited' configuration, one could say.

You are right, it is a special driver regardless of price. I really like it. Can't find anything else of similar performance for even 3x the price.

 

[freddi]

Great project ! - - do you think a lower priced version would greatly suffer? - say Silver Flute woofer and PA130? (or trusty Vifa TC9) - how rigid will be that lamination when dry?

 

[xrk971]

Silver Flute and TC9FD can work well too - not as refined as this setup though. I don't think I have ever had a speaker that was this flat in response without EQ.

 

[BYRTT]

.....I don't think I have ever had a speaker that was this flat in response without EQ.

Beautiful data so far with what seems as excellent drivers. Sound clip sounds great but guess with little more time to tweak last bit you release new ones. Very inspiring thanks share build ...

 

[BYRTT]

Think can be educating and interesting see graphical how 2. order acoustic high pass filter for box and a Linkwitz transform performs therefor attach some data below.
With a wire i loop back my sound card and in REW set "JRiver MC 20 virtual sound card device" as output device, this way JRiver DSP engine sets inside the loop and in REW settings in DSP filters will be measured real time. Data used for filters shown down below plot area in REW and in the screen copy of DSP engine.

 

[xrk971]

I tried 2nd order symmetric and it leaves a big hole at the XO point. Should I move xo points around so asymmetric? I tried moving LR4 to 350Hz and it actually seems smoother - so will have a listen there for a while. MiniDSP makea all this very easy to try.

 

[5th element]

I tried 2nd order symmetric and it leaves a big hole at the XO point. Should I move xo points around so asymmetric? I tried moving LR4 to 350Hz and it actually seems smoother - so will have a listen there for a while. MiniDSP makea all this very easy to try.

That's normal for a 2nd order filter you have to invert the polarity of one of the drivers for them to sum correctly.

 

[xrk971]

That's normal for a 2nd order filter you have to invert the polarity of one of the drivers for them to sum correctly.

Thanks for the tip - I will give it a try.

 

[mor2bz]

http://www.pispeakers.com/Speaker_Crossover.pdf

see pg. 42

 

[wesayso]

Nice work! Could you show the step response? If I were a rich man I'd order 50x 10F for obvious reasons .

 

[xrk971]

LR2 XO @350Hz and new sound clips

Ok, I implemented the Linkwitz-Riley -12dB/oct (LR2) XO at 350Hz, flipped the polarity of the 10F as suggested by 5th Element and others. The sound is even better now I think. Very cohesive and engaging sounding. One of the best speakers I have heard (up there with the Trynergy and my reference all horn loaded system). The quality of the bass is really surprisingly good for such a simple sealed alignment. I can imagine it would sound even better in a wooden box. The clarity and sharpness of the highs is also really remarkable. Listen to the guitar in the SRV clip E.

Here are the new measurements...

SPL vs Freq for the LR2 XO @350Hz:

Harmonic Distortion:

Freq response (4ms gate and 1/48th oct smoothing) in detail (2dB/major div) with phase:

Impulse response:

Group delay (still greater than I expected with the new LR2):

I did not get Wesayso's request for Step function until after I was done gathering data and left the lab for the day.

Sound clips include two from the Subjective Blind Comparo thread, and the old Spanish Harlem standard. Have a listen, I hope that at 320kbps mp3, you still appreciate the quality of the sound. I am listening to the raw 96kHz .wav file and it sounds superb - some of the best I have heard.

 

[BYRTT]

..........Group delay (still greater than I expected with the new LR2):

..........

Seems not so bad at all when added all system filters upon each other group delay ends up pretty close to the measured (see at 40Hz point both show 15mS).

Filters used for test was:
1. Acoustic high pass filter at 52Hz 2.order BW to model sealed box.
2. Electric linkwitz transform at 52Hz BW down to BW at 30Hz.
3. Electric LR 2.order at 350Hz XO filter.
4. Acoustic hole system band pass in a high pass 4.order at 27Hz and a low pass 4.order at 19kHz.

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

 

[xrk971]

Thanks for the excellent analysis Byrtt! So what do I need to do to get GD as low as possible? What if I get rid of system high pass filter that is supposed to prevent woofer over excursion? My amp doesn't have enough volts to cause that anyway (19v laptop brick probably barely hits xmax). Go to 1st order XO? Less agressive LT to only 38Hz vs 30Hz? But truth be told, the 15ms doesn't sound so bad to my ears - I have not optimized for time alignment yet - maybe there is something to be gained there?

 

[irribeo]

Considering it is a recorded playback I think E sound very good. There are some differences listening here with headphones, funniest one is clearer treble detail for recorded playback versus original when listening to drum at start of Eclip , bass is a little tighter original Maybe my original is different (mine is 9:14 lenght StevieRay/Reed on DifferentWaves sample CD for Beosystem2500)

 

[xrk971]

Irrebeo,
Also, don't forget I am recording in stereo the mono right channel (only) of a stereo recording. So the content from left speaker is missing. Interesting that treble in recording was higher - the response seems to fall off above 16kHz - and is measurably flat. My bass is off by 15ms at 40Hz as Byrtt showed and my measurement confirms. Looks like I need to get rid of woofer protection HPF, maybe try 1st order XO even? Get LT to be not as aggressive? The LT as it is, gives solid 35Hz punch so really makes a difference on some recordings.

Thanks for the feedback.