I reran the sim with a 8 ohm W170S and the very similar 90dB SC10N tweeter, Philip. Interesting.
That speaker is definitely peaky at crossover and probably over bright on the tweeter level.
I found that 4.7R works nicely on the bass shunt. A 15R across the tweeter will take level down 3dB if it is bright. I modelled a 6kHz 15R/0.33uF notch on the bass too, which aligns phase and reduces breakup too. It looks as though that all creates a hole in the response at crossover, but that is mostly tweeter diffraction effects I think, so should not be a problem. It will sound much smoother.
If it is a problem, you can reduce the 8.2uF to about 4.7uF. The 4.7R should be a 10W type. The 15R resistors can be a 3W or 5W wirewound or ceramics. The little capacitor should be a MKP polypropylene, but a cheaper MKT would probably work too. Happy hunting! This is worth trying and shouldn't cost much.
Best Regards, Steve.
That speaker is definitely peaky at crossover and probably over bright on the tweeter level.
I found that 4.7R works nicely on the bass shunt. A 15R across the tweeter will take level down 3dB if it is bright. I modelled a 6kHz 15R/0.33uF notch on the bass too, which aligns phase and reduces breakup too. It looks as though that all creates a hole in the response at crossover, but that is mostly tweeter diffraction effects I think, so should not be a problem. It will sound much smoother.
If it is a problem, you can reduce the 8.2uF to about 4.7uF. The 4.7R should be a 10W type. The 15R resistors can be a 3W or 5W wirewound or ceramics. The little capacitor should be a MKP polypropylene, but a cheaper MKT would probably work too. Happy hunting! This is worth trying and shouldn't cost much.
Best Regards, Steve.
Attachments
Many thanks Steve for all the work you put in to diagnose where the crossover is deficient, prescribe mods and re-run simulations.
I did notice in one of your past posts you remarked there is a tendency for many posts on forum to fizzle out before reaching a conclusion. Well I think you have brought this one nicely to a conclusion in that any owner of EB2S speakers will now know ‘what to do’ to improve them. If the owner is to be me then I’ll need some serious skills development before being let loose with a soldering iron.
One thing on my mind is that the Monacor DT-99 used in the EBS2 speakers is more efficient than SC 10 N (93dB v. 90dB) and going by the FR plots from manufacturer the SC 10 N is flatter above 1.5kHz than DT-99. Would you say any capacitor/resistor component values need changing because of this difference, particularly the 15R across the tweeter?
I'd say those two tweeters are rather similar.
All this stuff is brought to you courtesy of the Hilbert Transform:
https://en.wikipedia.org/wiki/Hilbert_transform
All this stuff is brought to you courtesy of the Hilbert Transform:
https://en.wikipedia.org/wiki/Hilbert_transform
Attachments
About the tweeter slope, Philip, I haven't changed it from the original design, just given you an option to take the level down a little, since it might be on the bright side.
Tweeters fall off in level off-axis, so it's all a bit subjective anyway:
http://www.seas.no/index.php?option...fc&catid=45:seas-prestige-tweeters&Itemid=462
All the other advice stands. It's not a huge project. The 40W soldering iron and unleaded solder will be the most expensive purchase. But useful tools to have IMO.
Best regards, Steve.
Tweeters fall off in level off-axis, so it's all a bit subjective anyway:
http://www.seas.no/index.php?option...fc&catid=45:seas-prestige-tweeters&Itemid=462
All the other advice stands. It's not a huge project. The 40W soldering iron and unleaded solder will be the most expensive purchase. But useful tools to have IMO.
Best regards, Steve.
Thanks for the 'heads up' with the Hilbert Transform - as you say, it's built into Boxsim so that does the hard work for us.
No, it's not a huge project but as a beginner I'm puzzled by how to fit in the additional components given that pcb was designed so it was 'just right' for the crossover as specified by Arcaydis. I suppose the 3 resistors aren't that big and assume same goes for the extra capacitor on the notch filter.
Another thing that puzzled me is the FR response for the tweeter. In this simulation (with notch filter and added resistor on bass shunt?) the tweeter level rises about 2dB from 2kHz to 3 kHz, then 6dB from 3kHz to 4kHz and then about 1dB up to 5kHz.
Whereas in the simulation with tweeter offset the level rises 6dB from 2kHz to about 2.8kHz and then another 4dB from there to 5kHz.
If I didn't know which graph was which I'd pick the first one as showing simulation with 15R across the tweeter to take level down 3dB. But I'd be wrong! Why is that?
