But since a Zobel needs a resistance anyways, it may just as well be taken care of (partly) by the ESR of the cap. It's more about tolerances and degradation in this particular example why Elco's may not be preferred.
There is a lot of thing who work against good localization. May i list some of them ?
- Speakers used above their non fractioning limit (different peaks of high Q in the response curve, rarely the same between the two speakers): Avoid large band speakers.
- Filter precision and their phase behavior (avoid multiways enclosures).
- For the same reason, avoid high slope filters.
- Different directivity curves at the crossover between 2 adjacent ways in the same enclosure (avoid medium-bass/ treble of too much different size).
- Bad phase alignment between the speakers in the same enclosure.
- Speakers with too large directivity witch emulate too much the resonances of the room and reflexions on the walls
As you see a lot of them are contradictory.
I addressed those problems in this way with less compromise as possible :
Only two ways enclosure. But, if you don't want to overpass the fractionating limit for your bass or low medium driver, you need a large band medium-treble able to go that down in frequencies up to the highest limit.
Only horns can answer this requisite without fractionating, with very few available drivers able to up enough in frequencies.
And only horns can have an emitting surface equal to the one of the bass driver.
A controlled directivity horn was a requisite.
I we want to have our cones/membranes not fractioning, even after their Xover frequency when it is still audible, we need hight slope filers. Need to be precisely paired both in phase and level curves.
Last edited:
Even when large values are needed, a better part is polyprop which comes in high values (100mfd) and high voltage. Even large values made for PF correction are better and cheap (when that matters). But, in many cases it just doesnt matter all that much compared to other larger issues.
Me, I would not use (bipolar) electrolytics - unreliable, wide tolerances, variable to too many things and shorter life to mention a few other reasons. but they are cost effective. I go with active cross-over which cost a lot more than any combination of caps or type becuase I get Over-All better results than passive cross-overs. -RNM
Me, I would not use (bipolar) electrolytics - unreliable, wide tolerances, variable to too many things and shorter life to mention a few other reasons. but they are cost effective. I go with active cross-over which cost a lot more than any combination of caps or type becuase I get Over-All better results than passive cross-overs. -RNM
Last edited:
50 of them 100V (you know the size and price) for each enclosure and just a zobel ? And NO NO NO audible benefit ?
Did-you hate electrolitic caps that much ?
Your active crossover will not solve the impedance peak of the speakers near their resonances. This network will remain in my active filter project.
i measured the 10 years old electrolitic caps, and they are still ok.
I believe you can accept to change them all 5 years, as a normal service.
Last edited:
Then I wouldnt do it. Maybe elsewhere. Reliability is high on my list, however. I know someone will chime-in and tell be they never had an issue with them. But, there are others that have. All these exceptions dont make a rule. -RNM
Again, it is just a Zobel. Paralleled with ~~0.1 ohm of internal resistance of the filtering coils + the 0.1ohms of internal amp impedance+cable resistance, on one side. and paralleled too with the speaker.
They do not participate to the crossover frequency in any way.
You can even remove the network. It is just a little better that way, amp side. and better than to leave a 150 ohm peak resonance impedance here. NO ISSUE.
This network SOLVE the impedance peak of impedance at resonnance, amp side: the amp will not have to deal with DV/DC disparities frequency dependant. It is all about.
It is time to rehabilitate those poor elctrochemical caps. They are not that bad.
Last edited:
That be me.
My tigers (swtpc 250 from '78) seem to still power up.
My dynaco 200 by 2 (date unknown) still powers up.
Alas, my SOTA computer (Systron Donner, 1960) had a problem...
Even though I tried (in vain) to reform them (3 days, 5% increment per hour), the supply caps decided they were no longer of this world...they passed on to the other side...silently but not without smoke or smell.
Seems they kinda need to be not dry. Go figure. Guess they'll never meet 10 power 8 MTBF rate, eh? Sigh..
jn
If you never tried to zobel a loudspeaker for totally flat impedance curve, and listened (and measured) the difference, you don't know the benefit of it.
I do not like to loose money for nothing. Even for audiophile cables.
Last edited:
jn,
This is crazy. I think you missed my simple point. Once again, please understand that I was only making a comparison of time delays - nothing having to do with settling times. I do not disagree with you that settling time might be an issue with 500 feet of mis-terminated cable.
My simple point was that effective differential time delay introduced by crossover component tolerances can be orders of magnitude larger than the differential time delay that might be created by speaker cables. This was all in the context of imaging.
BTW, every couple of years some nut-case comes along and argues that speaker cables should have a characteristic impedance of 8 ohms so that they are properly terminated by an 8-ohm loudspeaker, even if the run is only 10 feet. This is completely silly.
At the same time, we are wise to be mindful of the impedance presented to the amplifier by the combination of the speaker cable and the loudspeaker out to perhaps 10 MHz - impedance that in some cases can get nasty as a result of reflections at high frequencies that might de-stabilize some amplifiers. Please go to Figure 18.4 on page 376 of my book to see what I am talking about. A 120-ohm Zobel at the loudspeaker end assures at least some kind of termination at high frequencies, since many loudspeakers begin to look inductive and Hi-Z as the frequency approaches the MHz region.
Cheers,
Bob
You both agree on that from the early beginning, it seems. In fact it seems you agree on near everything, just you forgot to tune your violins the same key. Reason why i both punished-you, naughty boys.
Funny you mention the behavior of coils up to MHz, i was talking about that for zobels.
Last edited:
Ah, good, that helps.
.
Now calculate the settling time, say to 80% of final value, for a step from a source to load, 3 meters, dielectric coefficient of 4, line Z 100, load Z of 4.
Use either the ideal eq, or include R and G. Or, simply go to the belden site, they actually graph some of this. Hint, the results are worse.
You attempted to use simple prop delay into a matched non reflective load as a response to a calculation of mismatched termination settling time.
I'm not that easy. And, I expect better.
I know precisely what you were saying, what you were doing.
I cannot argue the nut-case aspect. I certainly meet that description.
However, diss me at your own risk. If you haven't figured it out by now, hopefully you soon will..I kinda know what I speak of. I tend not to speak when I am not sure.
I am very happy being quiet and allowing the experts speak their stuff on the forum.. When I am not quiet, you have to figure out why..
cheers, jn
OK optical source to amp, if you have bits anywhere in the path no need for any new complaints.
Neutron, I think you have said you have conducted extensive research on aural perception.
Can you tell us when & where?
I'm interested cos I have some small experience of this but am a bit rusty cos a beach bum for more than a decade.
______________________
Bob, I think its unlikely any of your designs need such elaborations.Bob Cordell said:
However, it is highly likely (certain??) that certain $zillion Golden Pinnae designs would be embarrassed if such exotic precautions are not taken.
Last edited:
Oh, please, not you, PMA. It is a LCR serial circuit.
You can use a very high resistance value for the coil (little wire diameter), you just have to subtract this value from the resistance value. As you just have to substract the inductance of the cap from the inductance value (low under the tolerance of the coil).
Last edited:
- Status
- Not open for further replies.