crossover mod - is my thinking right?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi (meddling tweeker alert!) - I have built a pair of speakers which are sounding good.:) I tamed down what I perceived to be some slight exaggeration round the tweeter crossover point, raising the level resistor from 3.2 to 5ohm - a definite improvement.:)


I would now like to experiment getting a bit more sparkle without raising those levels at xover freq.
is my thinking correct by adding components in pink circle?
is there another (easy) path that would have the same effect?


(reducing overall ohms similar to starting point by parallel resistor, and leaking across higher freq via the bypass (see attachment)


much appreciation if you can confirm I am heading in the right direction or not!
thanks
 

Attachments

  • P1020113.JPG
    P1020113.JPG
    968.1 KB · Views: 313
I have never treat speaker crossover as tone control. When I design a crossover, I avoid any (RC) bypass network. People usually used L-Pad or padding resistor next to the tweeter, but I prefer putting the resistor at the input of the crossover like yours, which has an opposite effect to what you actually want (increase HF response).


Are you sure that you cannot address the tone controlling duties electronically? Do you have proper impedance matching between sources and next stage amplifiers? What do you use as a preamp?
 
thanks so much for your reply. the system has mainly had 2 amps in front of it. naim integrated and nc400 with passive pre @25k. (nc400 Zin is c.100K, rega CD gives min load 10k), and also (briefly) a Pioneer A..? integrated (highlighting quality of the others!) findings similar with all. (have also checked out an old Cambridge DAC)


my thinking; given the ease and low cost of speaker tweeking, and given I might want to tweek back for a different room (I may move) I thought this might be worth a go.


perhaps my ears want/expect a 'bbc dip' which sounds better than using speaker as tone control! its really to supress that xover freq a little.
(they are 3 way btw)


I have never treat speaker crossover as tone control. When I design a crossover, I avoid any (RC) bypass network. People usually used L-Pad or padding resistor next to the tweeter, but I prefer putting the resistor at the input of the crossover like yours, which has an opposite effect to what you actually want (increase HF response).


Are you sure that you cannot address the tone controlling duties electronically? Do you have proper impedance matching between sources and next stage amplifiers? What do you use as a preamp?
 
given the ease and low cost of speaker tweeking,


I do use a crossover for the more permanent 'necessary' kind of equalisation.


One reason I don't tweak crossover is because usually the crossover has been designed very carefully. Even the input resistor value is chosen carefully. It's the heart of the system that has to have minimum evils.
 
What a tease you are! :rolleyes:

No drivers specified. No full crossover. A 3 way at least it seems, after some reading. :eek:

I don't know what to suggest. Apparently a 550Hz Fs notch on the tweeter, sort of thing you might use with some of the SEAS 27 tweeters.

SEAS-3-Way-Classic-mkII
SEAS-3-Way-Classic

Shallow second order tweeter filter somewhere around 3kHz. Seems familiar.

I think you need to tell us more for a meaningful answer. I get tired of trying to guess what the issue might be with vague problems. Sorry.
 
Here we go again wasting everybody's time. OP wants more "sparkle" (whatever that could mean to anybody) but no measurements of any kind, not even the electric signal into the driver.

While I might be tempted to post about REW and laptop mic's, you can do electrical with barely any investment of learning-curve or gear if you have a spreadsheet app and a means of measuring the signal.

B.
 
Last edited:
It'll only be a waste of time if Rob doesn't provide some more detail.

The answer usually lies in the question IF enough detail is provided.

Three ways are very finely tuned things:

685183d1528241442-linton-2-upgrade-steen-duelund-3-jpg


They can surprise you in that what you might think is a tweeter problem might, for instance, be cone breakup from the mid. Time will tell. More detail please. Some folks here have quite a good understanding of all the compromises in speakers. Is this speaker time aligned, for instance?
 
My apologies for not posting the details. attached xover. I am feeling a bit ignorant for this site, and am suitably shamed for lack of detail ! the mid is 22mm in front of the other drivers btw.
I am actually happy with the speaker, it's fine level tuning which I think could be done, might be educational, and not expensive.
R



It'll only be a waste of time if Rob doesn't provide some more detail.

The answer usually lies in the question IF enough detail is provided.

Three ways are very finely tuned things:

685183d1528241442-linton-2-upgrade-steen-duelund-3-jpg


They can surprise you in that what you might think is a tweeter problem might, for instance, be cone breakup from the mid. Time will tell. More detail please. Some folks here have quite a good understanding of all the compromises in speakers. Is this speaker time aligned, for instance?
 

Attachments

  • xover.jpg
    xover.jpg
    176.1 KB · Views: 139
My apologies for not posting the details. attached xover. I am feeling a bit ignorant for this site, and am suitably shamed for lack of detail...
Got it.

Sure, your tweak-by-ear might work. A lot of manufacturers today make hideously complex crossovers and must tweak by mic and by ear at the end. Damping factors might have no effect on a tweeter and so no audible downside to complexity* except unlikely the components will match well between the sides**.

I wonder how the first post might have been worded to signal you didn't want people to waste their time?

But if you do measurements, you and all of us out here will be immensely better able to offer more intelligent comments.

And maybe more important from "social media" ethics, WE will then be able to learn something out of the experiment.



B.
*ESLs generally have complex input networks due to transformers and other downstream elements.
** and you sure have no means to assess that
 
Last edited:
Ah, it's an Ellam-Flex-3W!
Ellam-FLEX-3W

There's some slight variation from the regular 2W Ellam Flex filter as you'd expect:
Ellam-FLEX

3 ohm DC tweeter, that SS D3004/6600. Rises a bit in impedance at the top.

And we're looking at this to take down the lower treble and restore the upper treble. Bit of tilt really. The images are too large to comfortably display, so I won't.

Hmm. I'm worn out already. I'll mull it over. :)
 

Attachments

  • Scan D3004-6600.PNG
    Scan D3004-6600.PNG
    29.5 KB · Views: 128
Hi Ben - my intention was to ask a yes/no question, which is what I thought I did - Allen B indeed answered yes (ish). However I can see that a yes/no question to me might not in fact be a yes/no question, something I will certainly consider if I dare to post anything else! Thank you for the information below, it is interesting to hear that the final stages of a complex xover are possibly by ear, which I think they must be - particularly if being tweaked for a particular room. I will certainly post findings if I do any mods.




Got it.

Sure, your tweak-by-ear might work. A lot of manufacturers today make hideously complex crossovers and must tweak by mic and by ear at the end. Damping factors might have no effect on a tweeter and so no audible downside to complexity* except unlikely the components will match well between the sides**.

I wonder how the first post might have been worded to signal you didn't want people to waste their time?

But if you do measurements, you and all of us out here will be immensely better able to offer more intelligent comments.

And maybe more important from "social media" ethics, WE will then be able to learn something out of the experiment.



B.
*ESLs generally have complex input networks due to transformers and other downstream elements.
** and you sure have no means to assess that
 
I am not in the measuring camp myself. Why bother when Troels and Scanspeak have done it for you? Maybe I'm lazy...

I am in the simming camp however. Being mathematically inclined. :cool:

Already a stunning finding.

Your mod does exactly Zilch, Zip or Zero! Sweet FA, as we say here.

The effect of the increased 5R resistor over 3.3R at the input is almost exactly cancelled by the 10R + 6.8uF RC. :D

Ah well.

TBH, and you may disagree, I really don't think much of second order filters or this lowish 2kHz crossover. I don't know why Troels is abandoning his trusted 3kHz crossovers at all. Has he gone completely mad? :eek:
 
Thanks Steve - it was just a very minimal tweak I was after, if the extra cap was down to 5uf would that not sneak in a little lift at 6khz?
if not then I can certainly live with and enjoy these - have had them for a year - way better than I could afford commercially, and blew my last speakers out of the water which was the brief...
Passing by Portsmouth Tomorrow - will doff my cap to you!
R
 
Hi diyrob,

Hopefully you've built this Troels design as he intended ( with the tweeter inset and mechanically time-aligned to the mid ).

I Xsimed the mid & tweeter portions of the network ( time-aligned & non-aligned ) and the response linearity suffers when not aligned ( there's a few db "hole" in the lower presence 4.5K range > preceded by a "bump" in the 1.8K area).

See; ( for the rough approximation )

attachment.php


:)
 

Attachments

  • Troels Flex3.jpg
    Troels Flex3.jpg
    76.1 KB · Views: 202
thanks for your time, and it is interesting what you have said. yes the speaker is built absolutely as per plan- my woodwork is considerably better than my knowledge of electronics! :)
its more of a delicate 'voicing' tweek rather than 'problem correction' due to personal taste I feel.
R


Hi diyrob,

Hopefully you've built this Troels design as he intended ( with the tweeter inset and mechanically time-aligned to the mid ).

I Xsimed the mid & tweeter portions of the network ( time-aligned & non-aligned ) and the response linearity suffers when not aligned ( there's a few db "hole" in the lower presence 4.5K range > preceded by a "bump" in the 1.8K area).

See; ( for the rough approximation )

710885d1540249411-crossover-mod-thinking-troels-flex3-jpg


:)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.