The most important thing that one needs to do with filters like that is to design around the acoustic response and not the electrical. As far as I can tell xrk all you've done is dial in the respective filter order in the miniDSP and little else. To make this work correctly you need to make sure that the acoustic result of applying the electrical filters is what you want. For example applying a 1st order electrical filter rarely results in a decent 1st order roll off, this is why your previous attempt at a first order at 350Hz was it? actually looked like it was more at something like 600Hz or whatever.
The same is true here. Both the roll offs for the 2nd and 4th order LWR filters should be 6dB down at the crossover frequency if the drivers are integrated and filtered properly. This is clearly not the case. The end acoustic roll off is a combination of the electrical roll off of the filter + the drivers natural roll off and until you've designed the filters to also correct for the drivers innate responses, so that the end acoustic roll off actually looks like the desired target, there is little point in doing any comparisons.
The same is true here. Both the roll offs for the 2nd and 4th order LWR filters should be 6dB down at the crossover frequency if the drivers are integrated and filtered properly. This is clearly not the case. The end acoustic roll off is a combination of the electrical roll off of the filter + the drivers natural roll off and until you've designed the filters to also correct for the drivers innate responses, so that the end acoustic roll off actually looks like the desired target, there is little point in doing any comparisons.
5th Element,
Thanks for your constructive criticism. I know that overall acoustic XO seems to be at 500Hz so need to take a good look at this. I think the baffle step falloff must be what is pushing it higher than where I set it as both drivers are about as flat in response as one can get. A lot to learn with this XO business and miniDSP lets you set it arbitrarily anywhere you want, but as usual, there is a lot of consideration as to where it should go.
Thanks for your constructive criticism. I know that overall acoustic XO seems to be at 500Hz so need to take a good look at this. I think the baffle step falloff must be what is pushing it higher than where I set it as both drivers are about as flat in response as one can get. A lot to learn with this XO business and miniDSP lets you set it arbitrarily anywhere you want, but as usual, there is a lot of consideration as to where it should go.
I did listen yesterday to the 3 samples. I shouldn't have because I had light developing headache. Easiest to listen to was B then. Today I find myself trying to prove to myself why, maybe C but especially A seem to have little louder treble. Also around 30 seconds A might be little off, both C and B sound cleaner/clearer in that transition. Won't be listening longer today because of yesterdays light headache, life is easier without
So Monday also for the 2nd run poll. It is fun, thanks for the clips.
So Monday also for the 2nd run poll. It is fun, thanks for the clips.
I did listen yesterday to the 3 samples. I shouldn't have because I had light developing headache. Easiest to listen to was B then. Today I find myself trying to prove to myself why, maybe C but especially A seem to have little louder treble. Also around 30 seconds A might be little off, both C and B sound cleaner/clearer in that transition. Won't be listening longer today because of yesterdays light headache, life is easier without
Thanks for the feedback on the sound of the different XO's. It's good to know that you can hear a difference at all.
Construction of the stereo mate has begun. I have the main box with 5 panels built and the liquid nails has had 3 days to cure. Will move to the final installation of bracing, Dagger, and front baffle reinforcement Sureply soon. Looking forward to listening in stereo. I have a possible choice of using an old class A/B (LM38xx-based) for the bass section and keep the TPA3116D2 for mids and treble duty. Still need to make the oscon mod and bootstrap snubber mod on the new TPA3116D2. Or should I buy a new higher power class D. Amp for the bass? The sims show that I need 250watts if I really want to hit peak SPL. Sustained SPL is 80watts. Above 100watts class D amp power supply prices really skyrocket - hey coat more than the amp.
A funny thing happened, I had the XO sound clips on my computer all laid out and renamed as A, B, C and MS, in their wisdom, forced an OS update and rebooted my PC overnight without my permission. Of course, I lost the files and now can't remember the answer key. I could probably muddle through it and retrace my steps but it will take a few days. Sorry...
The second unit for stereo is awaiting for glue (Liquid Nails, non volatile version) to dry on final capping operation on main box (Dagger, braces, terminal cups, fiberglass stuffing, and wiring installed), then the wood baffle needs to be glued on. Should be going in stereo in maybe 2 days.
The second unit for stereo is awaiting for glue (Liquid Nails, non volatile version) to dry on final capping operation on main box (Dagger, braces, terminal cups, fiberglass stuffing, and wiring installed), then the wood baffle needs to be glued on. Should be going in stereo in maybe 2 days.
Last edited:
A funny thing happened, I had the XO sound clips on my computer all laid out and renamed as A, B, C and MS, in their wisdom, forced an OS update and rebooted my PC overnight without my permission. Of course, I lost the files and now can't remember the answer key. I could probably muddle through it and retrace my steps but it will take a few days. Sorry...
On the contrary what some people believes, it was quite easy to hear the difference of B file to the others. I'm maybe wrong, but I believe the A and C files had too much overlap between drivers. I got the same result when running my naked ob basses too high.
Regarding differences between A and C, it was impossible to hear with my Grado headphones in my laptop.
Have you back tracked your files yet? Would be interesting to see what the difference is.
Peter
Perhaps you forgot to charge the settings in the Windows update dialogue to "download but do not reboot "?
Bob
OS/2 was better. too late now.
I fully understand the limitations of this sort of thing, but I do plan on listening to the samples - just haven't gotten to it yet.
I've always been interested in trying a B&O filler type crossover, but have never gotten to that either.
For listening gear I have HD600 cans and a sound devices usb interface. Pretty pro stuff, but it's still resolution limited, mostly by the cans.
too late now.I fully understand the limitations of this sort of thing, but I do plan on listening to the samples - just haven't gotten to it yet.
I've always been interested in trying a B&O filler type crossover, but have never gotten to that either.
For listening gear I have HD600 cans and a sound devices usb interface. Pretty pro stuff, but it's still resolution limited, mostly by the cans.
Stereo Finally!
The glue is still a bit wet but I took some preliminary data and listened a bit. Externally it looks dry but the internal surfaces and braces still have moist Liquid Nails and this makes a difference in the harmonic distortion as I found out.
I first tested the impedance to ensure that the boxes matched. The Dagger is a pretty close match, the sealed RS225 is a little off with the unit 2 having a slightly higher Qts, 0.65 vs 0.61 for unit 1. I may have to adjust the fiberglass stuffing to balance them.
10F/8424 Dagger Impedance (purple is bare driver and blue and green are in Dagger):
RS225-8 Impedance:
Here is a photo of the pair on stools for me to listen to:
Here is the comparison of unit 2 (Left) compared to unit 1 (Right) response for LF and HF drivers and overall system response. Unit 2 is dark blue, Unit 1 is light blue - pretty close match.
This is harmonic distortion with wet glue, resembles earlier HD measurements that I made of unit 1:
This is the HD of unit 1 now that the glue has dried and the internal braces are securely affixed, significant reduction in HD! So I will have to wait a week or so to make sure the glue really dries before making more measurements:
Again, these are HD measurements at 0.5m and 0.71 volts just so that I can see things are working or not. In this case it shows that glue is still wet.
Here is the response of unit 2 with a 4ms gate applied and some very small tweaks to the PEQ to get the response and absolute phase very flat:
I am still waiting for a new TPA3116D2 amp to arrive, or maybe a TDA7492 to power the bass drivers. In the meantime, I am using my car audio amplifier. This class AB amp has a different time delay/phase than the TPA3116 so I will have to redo the phase angle setting.
So how do they sound? I set it up on a couple of stools as stand mounts and listened to my usual test tracks. Sounds excellent and the imaging is very good. The full range and woofer integrate very well for a convincing sound stage. On my "Live at Small's" Jazz CD, they have a very good ambient recording inside a little club and I can place all the instruments. Blues Travelers sounded exceptionally good with crisp and clear guitar attacks and haunting sounds of the harmonica augmented by superb bass guitar lines and kick drums. All instrument phantom images could be located on the stage. These speakers image exceedingly well and sound very balanced so you can listen forever and not get fatigued or bored.
I took the RS225 off to let the box get ventilation so that the glue dries faster. The sound will indeed improve with age as the glue gets drier. I won't be listening again until maybe another week. Hopefully the new amp arrives by then.
The glue is still a bit wet but I took some preliminary data and listened a bit. Externally it looks dry but the internal surfaces and braces still have moist Liquid Nails and this makes a difference in the harmonic distortion as I found out.
I first tested the impedance to ensure that the boxes matched. The Dagger is a pretty close match, the sealed RS225 is a little off with the unit 2 having a slightly higher Qts, 0.65 vs 0.61 for unit 1. I may have to adjust the fiberglass stuffing to balance them.
10F/8424 Dagger Impedance (purple is bare driver and blue and green are in Dagger):
RS225-8 Impedance:
Here is a photo of the pair on stools for me to listen to:
Here is the comparison of unit 2 (Left) compared to unit 1 (Right) response for LF and HF drivers and overall system response. Unit 2 is dark blue, Unit 1 is light blue - pretty close match.
This is harmonic distortion with wet glue, resembles earlier HD measurements that I made of unit 1:
This is the HD of unit 1 now that the glue has dried and the internal braces are securely affixed, significant reduction in HD! So I will have to wait a week or so to make sure the glue really dries before making more measurements:
Again, these are HD measurements at 0.5m and 0.71 volts just so that I can see things are working or not. In this case it shows that glue is still wet.
Here is the response of unit 2 with a 4ms gate applied and some very small tweaks to the PEQ to get the response and absolute phase very flat:
I am still waiting for a new TPA3116D2 amp to arrive, or maybe a TDA7492 to power the bass drivers. In the meantime, I am using my car audio amplifier. This class AB amp has a different time delay/phase than the TPA3116 so I will have to redo the phase angle setting.
So how do they sound? I set it up on a couple of stools as stand mounts and listened to my usual test tracks. Sounds excellent and the imaging is very good. The full range and woofer integrate very well for a convincing sound stage. On my "Live at Small's" Jazz CD, they have a very good ambient recording inside a little club and I can place all the instruments. Blues Travelers sounded exceptionally good with crisp and clear guitar attacks and haunting sounds of the harmonica augmented by superb bass guitar lines and kick drums. All instrument phantom images could be located on the stage. These speakers image exceedingly well and sound very balanced so you can listen forever and not get fatigued or bored.
I took the RS225 off to let the box get ventilation so that the glue dries faster. The sound will indeed improve with age as the glue gets drier. I won't be listening again until maybe another week. Hopefully the new amp arrives by then.
Attachments
Last edited:
Liquid nails (non volatile) resembles latex caulking loaded with some sand I think. It dries slowly and is semi flexible and adds some mass to help damping as I use a thick layer between two panels as constrained layer damping. I don't have a problem with gaps - I glop it on heavily and use my fingers to make clean fillets at joints. Probably main reason to use it is low cost. $2.40 per tube at HD. I find once dry, the bond with both XPS foam, or foam core, or wood is very good.
Does anyone have a good link to how to build an analog op-amp Linkwitz transform circuit? There are a bunch of custom PCB makers in China that will make 10 2inx2in 2-layer boards for $10. An all analog XO and LT pre-power-amp may be an interesting option for me to think about and possibly try.
input buffer and fan-out to HP and LP is 1 op amp, 1st order active Butterworth is 2nd op amp stage, Linkwitz transform is 3rd stage, time delay is 4th stage, buffer/driver is 5th stage. 10 op amps total for stereo - so can be done with 5 dual op amps all in tiny SMT.
Maybe TI FilterPro and verified with TINA is what I need to fire up now...
Avoid all the big caps and coils of a passive, and in qnty, much cheaper than DSP.
A little digging has turned up an article and a spreadsheet for calculating component values.
https://www.trueaudio.com/st_lkxfm.htm
John L Murphy, the author of the above website and spreadsheet is also the author of Dayton's TrueRTA and DATS, all great stuff so I trust his calcs even more now.
input buffer and fan-out to HP and LP is 1 op amp, 1st order active Butterworth is 2nd op amp stage, Linkwitz transform is 3rd stage, time delay is 4th stage, buffer/driver is 5th stage. 10 op amps total for stereo - so can be done with 5 dual op amps all in tiny SMT.
Maybe TI FilterPro and verified with TINA is what I need to fire up now...
Avoid all the big caps and coils of a passive, and in qnty, much cheaper than DSP.
A little digging has turned up an article and a spreadsheet for calculating component values.
https://www.trueaudio.com/st_lkxfm.htm
John L Murphy, the author of the above website and spreadsheet is also the author of Dayton's TrueRTA and DATS, all great stuff so I trust his calcs even more now.
Last edited: