I'm now satisfied with the tonal balance of my Usher 8945P & Peerless 810921 2-way. I regard this revision as the final crossover for this design. The design page has been updated:
http://www.geocities.com/woove99/Spkrbldg/Usher_2way/Usher_2way.htm
The difficulty in voicing this design was due to the HDS tweeter's rising response in treble. We can notice this in various measurements of the tweeter:
Joe R's - http://www.customanalogue.com/elsinore/elsinore_images/diyaudio/HDS_FR.gif
Zaph's - http://www.zaphaudio.com/tweetermishmash/810921-FR.gif
Mark K's - http://www.markk.claub.net/Testing/SS6600_peerlessHDS/graphs/Peerless_HDS_tweeter_FR.GIF
augerpro's - http://www.htguide.com/forum/showpost.php4?p=354368&postcount=47
If you look at off-axis measurements in some of the above graphs, the response peak above 10 kHz is flattened when listening off axis. But note that if we used other tweeters that did not have this property, their off-axis responses in the range would be much more reduced. In my experience of voicing the design, the 810921's upper treble is audible no matter what the listening axis is, unless the axis is extremely off-axis (e.g., more than 20 degrees).
I found that, without addressing this issue, the tweeter sounds thin and harsh especially with orchestral music with lots of violin. So, to enjoy this tweeter's excellent clarity in lower treble (and some of upper mids), which is in fact the defining characteristic of any tweeters, its upper treble and some upper end of its mid treble must be suppressed using a contour circuit or a notch filter.
In my previous version, I mildly suppressed it using a narrow, shallow notch filter, but it turned out not enough. So in my final crossover, I widen and deepen it a little, and it then gave me good tonal balance in treble.
http://www.geocities.com/woove99/Spkrbldg/Usher_2way/Usher_2way.htm
The difficulty in voicing this design was due to the HDS tweeter's rising response in treble. We can notice this in various measurements of the tweeter:
Joe R's - http://www.customanalogue.com/elsinore/elsinore_images/diyaudio/HDS_FR.gif
Zaph's - http://www.zaphaudio.com/tweetermishmash/810921-FR.gif
Mark K's - http://www.markk.claub.net/Testing/SS6600_peerlessHDS/graphs/Peerless_HDS_tweeter_FR.GIF
augerpro's - http://www.htguide.com/forum/showpost.php4?p=354368&postcount=47
If you look at off-axis measurements in some of the above graphs, the response peak above 10 kHz is flattened when listening off axis. But note that if we used other tweeters that did not have this property, their off-axis responses in the range would be much more reduced. In my experience of voicing the design, the 810921's upper treble is audible no matter what the listening axis is, unless the axis is extremely off-axis (e.g., more than 20 degrees).
I found that, without addressing this issue, the tweeter sounds thin and harsh especially with orchestral music with lots of violin. So, to enjoy this tweeter's excellent clarity in lower treble (and some of upper mids), which is in fact the defining characteristic of any tweeters, its upper treble and some upper end of its mid treble must be suppressed using a contour circuit or a notch filter.
In my previous version, I mildly suppressed it using a narrow, shallow notch filter, but it turned out not enough. So in my final crossover, I widen and deepen it a little, and it then gave me good tonal balance in treble.
I agree this tweeter sounds like it has more HF energy than others, even more than the response would indicate. In a couple bookshelves I built the SS 9800 response actually dropped much less than 810921 off axis, which you would think would contribute to a somewhat similar overall power response to the 810921, with it's rising response on axis, but very steep drop off axis above 10k. But still the SS 9800 is the "quieter" tweeter by far.
I'm really not sure what all the factors are contributing to this difference. Just using external EQ dropping the response in the 10k-20k is helpful, but not the magic bullet. I found dropping the response in the crossover region (1750 hz in this case) was most helpful, but with the woofer contributing here too it's not clear to me which was affecting the sound more. I haven't had the chance to move the XO point much yet, Jay what did you find when adjusting the XO point? Have you tried an external EQ for quick changes? I found that very helpful.
After messing with the EQ a lot I'd have to say this tweeter just has the characteristic of sounding somewhat "bright" or high energy. The SS 9800 is the more "refined" sounding and probably the definitive "audiophile" tweeter, but in fast A/B testing the 810921 is hands down more open, more natural, and just more lifelike to my ears. The SS 9800 is super easy to get good sound from, just keep teh response flat, adjust the level to taste and it will sound great. The 810921 requires a lot more effort to tune to the individual's taste, but in the end more rewarding if you put that effort in.
Here is the revised XO for that Peerless HDS bookshelf speaker that I've developed but haven't yet built:
More experimenting to do...
I'm really not sure what all the factors are contributing to this difference. Just using external EQ dropping the response in the 10k-20k is helpful, but not the magic bullet. I found dropping the response in the crossover region (1750 hz in this case) was most helpful, but with the woofer contributing here too it's not clear to me which was affecting the sound more. I haven't had the chance to move the XO point much yet, Jay what did you find when adjusting the XO point? Have you tried an external EQ for quick changes? I found that very helpful.
After messing with the EQ a lot I'd have to say this tweeter just has the characteristic of sounding somewhat "bright" or high energy. The SS 9800 is the more "refined" sounding and probably the definitive "audiophile" tweeter, but in fast A/B testing the 810921 is hands down more open, more natural, and just more lifelike to my ears. The SS 9800 is super easy to get good sound from, just keep teh response flat, adjust the level to taste and it will sound great. The 810921 requires a lot more effort to tune to the individual's taste, but in the end more rewarding if you put that effort in.
Here is the revised XO for that Peerless HDS bookshelf speaker that I've developed but haven't yet built:
More experimenting to do...
Hi Jay,
First of all, I have to say I enjoy your site, particularly the fact you are proposing SW files of your projects! Thanks a lot for our diy community!
I'm planing to use these drivers for a friend of mine; playing with your SW file I ended up with a second order (+RLC) filter on the woofer (sort of challenging your comment about getting a good phase tracking, lol). This is the chart graph of your last best sounding filter:
and this is mine:
Please, can you comment about how you guess it could sound regarding your experience with your Usher 2-way?
EDIT: obviously comments from other members are welcome.
Thank you very much.
First of all, I have to say I enjoy your site, particularly the fact you are proposing SW files of your projects! Thanks a lot for our diy community!
I'm planing to use these drivers for a friend of mine; playing with your SW file I ended up with a second order (+RLC) filter on the woofer (sort of challenging your comment about getting a good phase tracking, lol). This is the chart graph of your last best sounding filter:
An externally hosted image should be here but it was not working when we last tested it.
and this is mine:
An externally hosted image should be here but it was not working when we last tested it.
Please, can you comment about how you guess it could sound regarding your experience with your Usher 2-way?
EDIT: obviously comments from other members are welcome.
Thank you very much.
augerpro said:
Glad to hear about your experience, Brandon. I agree that the 810921 is not easy to get right sound from. I don't know what's going on, either. I think it has its own tonal characteristic---somewhat "light" and "thin." These are the best words I can come up with. In fact, through my several xover modification attempts, it kept this tonality. With my final xover, it's not harsh at all but very sweet and crisp. Sounds pleasing to my ear.
Using the EQ should be a very good way to find out what frequency range most affects a spaker's certain tonal balance. You're right in that a tweeter's low end gives its most defining sonic characteristic. In the xover voicing of my Usher 2-way (and also my RS180 2-way), I found that if I reduce the contribution of the tweeter around 1.5 kHz (1.2 k to 2 kHz) too much, violins in orchestral music don't sound right. I tried to achieve similar voicing between my two designs, but I concluded that the different tonal characteristics of the Peerless and the Seas 27TDFC couldn't be made equal by xover tweaking.
As for your new xover, I'd move the notch a little down to 1.2 k - 1.4 kHz. I think we need to suppress SPL from 8 kHz to 15 kHz. Above this range, our ear is not very sensitive.
crazyhub said:
Your version will definitely sound different from mine. But in what ways? It's difficult to say without actual listnening. As you can see from my discussion with augerpro, the Peerless tweeter sounds more sensitive than its spec indicates. What you need to do is to actually build and voice the speakers by ear. If you don't want to bother, I recommend to build my version and tweak it according to my suggestion. My version was strongly influenced by my experience of building and voicing the Usher design. So I believe it's a safer bet.
According to my voicing experience, I predict that your version will sound "thin" and "dry" due to the wide dip in the 500 Hz to 2.5 kHz range compared to mine. That's all I can say at this point.
Ah, I was wondering what you were talking about! The dip should have been centered at 15k, I think it drifted when I input standardized values. But I have to build up the cap and unwind the inductor anyway so I'll shift it back.
The dip from 1k-3k worked best when paired with the Nomex woofer. Both seemed to be voiced by Peerless to work with each other and thus have sort of a light or free sound. When I paired it with the RS150 the results were similar to yours, the dip was really unneeded and more a personal taste thing. Of course the RS150 was crossed 100 Hz lower and much steeper after the initial roll off, so it wasn't contributing near as much energy in that area as the Nomex.
Definitely an interesting tweeter.
The dip from 1k-3k worked best when paired with the Nomex woofer. Both seemed to be voiced by Peerless to work with each other and thus have sort of a light or free sound. When I paired it with the RS150 the results were similar to yours, the dip was really unneeded and more a personal taste thing. Of course the RS150 was crossed 100 Hz lower and much steeper after the initial roll off, so it wasn't contributing near as much energy in that area as the Nomex.
Definitely an interesting tweeter.
yes Yay
thans for yr thoughts on Xovers and free access to them
I have also choosen the 810921 for my main OB speakers but i "tame"them with my DEQX digital preamp (www.deqx.com)
I hear a tremendous amount of details and echo and many voice overs and the treble sound is very sweet and subtle when it needs to be
however i am building 2ways for my HT with a sub
and plan to use yr speaker designs for it
thans for yr thoughts on Xovers and free access to them
I have also choosen the 810921 for my main OB speakers but i "tame"them with my DEQX digital preamp (www.deqx.com)
I hear a tremendous amount of details and echo and many voice overs and the treble sound is very sweet and subtle when it needs to be
however i am building 2ways for my HT with a sub
and plan to use yr speaker designs for it
this very fine tweeter is out of production - go to
Tweeters | Tymphany
Who knows, which of the showed parts are the successor ?
The D2608 from Scan Speak looks similar - unfortunately there is ferrofluid in the air gap in use - go to
http://www.scan-speak.dk/datasheet/pdf/d2608-913000.pdf
Thank you for advices.
Tweeters | Tymphany
Who knows, which of the showed parts are the successor ?
The D2608 from Scan Speak looks similar - unfortunately there is ferrofluid in the air gap in use - go to
http://www.scan-speak.dk/datasheet/pdf/d2608-913000.pdf
Thank you for advices.
Last edited:
This has been debated several times here. AFAIK, the Peerless HDS tweeter has always been manufactured by Scan Speak in DK, and the D2608-9130 Discovery tweeter is the same as the Peerless. If you compare the SS datasheet with the Tymphany one (you can find one at Europe-Audio), you'll notice that the FR graphs and the VC data are exactly the same.
Ralf
Ralf
you are right.
Peerless HDS Typ 810921 (old datasheet)
http://www.lautsprechershop.de/pdf/peerless/peerless_hds_dt26.pdf
scan-speak d2608-913000
http://www.scan-speak.dk/datasheet/pdf/d2608-913000.pdf
Are there a version from the same tweeter without Ferrofluid ??
good advice - thank you. Also this models looks good:
SB Acoustics :: SB29RDNC-C000-4
SB Acoustics :: SB29RDC-C000-4
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.