Peerless HDS Tweeter too harsh, (Nomex 164)

Jump to Latest
#161
system7 said:
Quite easy to follow really. The SC10N soft dome has a ghastly second resonance around 6kHz, represented by an S-shaped kink in the phase curve. Ideally a phase curve will be more or less a straight line. You can see that a 6" woofer has a few issues too. Another way to look at the kinks in the phase is the slope of the bends correspond to group delay. Above 6kHz, it has lost all linearity.

The non-ferrofluid XT25 ring radiator does much better, with no significant resonances or energy storage whatsoever. Almost perfect straight line. In fact a ring radiator has much in common with a cone tweeter, even down to needing a third order filter to work well. Significantly, a ring radiator pins the troublesome floppy centre of the dome rigidly, thus curing most of the resonance problems.

Nobody has mentioned one other rather obvious thing about the D2608 tweeter, is that it is a largish 30mm dome, which means the usual problems kicked in earlier and required some rolloff.

A flat frequency response says nothing about whether a speaker will sound good. What is interesting to me, is that some drivers can convey MUSICAL INFORMATION very well. Unfortunately, a soft dome does very little above 6kHz in that respect whether you roll it off with damping, or let it spit and sizzle in the usual way.

Truly one of my more interesting posts, IMO. I don't tell people their speakers are rubbish to make myself popular! 😀
Click to expand...

Hy System7, I know it's a bit old stuff, but I'd be curious to the crossover used in that Visaton W170S/SC10N build, with 2kHz cross-over frequency, whose phase plot you had attached to post #124 from this thread. can you please share the design?
 
#162
Clausen said:
Update:

Well, one day several weeks ago, I decided to do a side by side test with my old ACI's. As I said the Nomex 164 had some noticeable advantages I noticed right away after building them, but the tweeter was not one of them, which became more and more apparent.

After doing the side by side test, it was hands down the tweeter IMO. So I decided to try something of the wall. I pulled the SS D2905/9000 out of the ACI's and replaced the HDS tweeters. Then after listening added a Zobel to them.

Well I'll I can say is WOW! I know I said wow before, but since doing this, I haven't tried anything else. I'm happy now. They sound so much more musical and pleasing.

From the specs I have found I think these 9000 are 4 omh, so I'm thinking I'll try the 9500's when I have a little more room in my budget since those are 6 ohm. But other then that I don't think I'll be doing anything else until I'm ready to build new speakers!

Thanks for all the help. Some time in the near future those HDS will be for sale. Right now I put them in the ACI's since I use those in the bedroom. If anyone wants them let me know!
Click to expand...

Old thread..

I suppose the Nomex 164 version 2 has these Scanspeak tweeters.. and the crossover point is lower too - closer to 2KHz

Would it sound less hotter and more pleasing but with details at the same time ?

Am looking to build my first floorstander speaker ( current speakers are Merlin TSM BME ... its 6 years now )

regards
 
#165
#166
Hi,

I'm building Troels design and had similar problems.

Harsh-in-you-face-vocals
Unbearable for any music at loud volumes
Highs that drill into your ears
Rock & Heavy metal is bearable at low volumes

Using REW and DATS, here's what worked for me. Tried four different notch filters - 2K, 4K and 7K. The Notch mentioned in post here did not help.

Don't remember how many measurements were taken focused on the wrong area. At least for me, its been the baffle step on the midrange all along.

Here's what worked.

Baffle Step for 9.375" front width:

L - 2.25 mH
R - 6.6R​

Zobel on the midrange:


R - 8R
C - 20uF​

Rewired the original LPAD (R1051_1R / R061_18R) with a variable unit and decided to test.

New values for HF LPAD:

R1051 - 6.4R
R1061 - 18.8R​

Hope this helps.
 
Last edited:
#167
The Nomex 164 returns! PEERLESS-NOMEX-164

I don't have that one set up in a modelling program anymore, but I was interested that Joe Rasmussen's Elsinore Mk3 and 4 have the same drivers. The Peerless HDS 810921 is a Scan D2608/9130 these days.

Elsinore Loudspeaker

The Mk1 with the Vifa XT25G30 is simpler. He found the paper 830875 and the polycone 830874 almost interchangeable.

Elsinore Mk1

I just did the conversion on crossover to single drivers. Quite straightforward. Halve coils and resistors, double capacitors in the mid bass filters. Might be worth a look. A 5kHz LCR notch works better than a tank notch to reduce breakup IMO. Single drivers have the same loudness as two in series. And that Nomex driver has a heap of breakup.
 
#168
This tweeter is, overly exciting, that would be word. I think it just doesn't match the Nomex cone (ugly 4-5k break ups). And looking at the crossovers, it should be lower. Whatever. What I don't know is it the upper top that is the problem or something in the presence area around x-o point. If it's the top, small 2-3uF cap in parallel with tweeter will lower it. A small 0.05mH in series might also do the trick, but might boost low highs, depending on x-o. I would lower the tweeter level in the first place to make it less in your face. I don't know what the mk2 x-o is, hopefully more listenable. If the problem is the presence around x-o then maybe more shallow filter either tw or upper midwoof would do the trick, making sort of BBC dip and moving soundstage back.
 
#169
I often get criticized for my admiration of Troels Gravesen. But IMO, he always does good work. And explains it well.

But clearly, this speaker was a bit of a stinker. We've all had disasters. 😱

If I had invested hugely in this baby, I would be looking at cone breakup from the 830875 6" driver. That and the tendency of all 2.5 ways to overwhelm a small room with bass.

Next, on a narrow baffle is dispersion and diffraction artefacts, leading to brightness or sibilance.

Thinking about Joe Rasmussen's very successful Elsinore, I notice he uses a waveguide on the tweeter, Along with an LCR for the awkward cone breakup.

Problems abound for a cheap fix. But a waveguide tweeter looks like a plan here.

549637d1463451462-devore-orangutan-clone-devore_orangutan-jpg


The Morel CAT378 is a 94mm faceplate tweeter, which works well enough on 4.7uF with a 1mH/47uF/10R LCR. Ought to sound better, IMO. Less diffraction so smoother.
 
You must log in or register to reply here.
Top Bottom