Crossover nightmare!!!!!!!

[alan-1-b]

Hello Lucas - look in your "Private Messages" file on this web-site.

*************************************************

For 72uF/100 volt bipolar electrolytic caps, go to:-

AudioLoudspeakers <--- Click on this.

and to their "Components" page, then to "Capacitors", then look under "Alcap" 100v.

Scroll down to 72uF @1.62,
then scroll down to the bottom of that page to see exactly how that cap has to be ordered.

This company has 4uF and 6uF in 400volt "Solen" metalized polypropylene at low prices,
but a little better can be bought from:-

Save money by building your own high performance loudspeakers or those who want to install their own home cinema, surround-sound or multi-room home entertainment system <--- Click on this.

go to their "Crossover Components" page, then to "Capacitors", then to "Cross Caps" -{the black ones}.
These are 400volt rated, and use a Zinc-Aluminium alloy for their metalization, and that may be better than the plain Aluminium of most other MKP types.

On the first page of is 33uF @7.52 including UK Postage !

On the second page is 6.2uF @3.26 including UK Postage,
and 3.9uF @2.13 including UK Postage.

Tolerance is +/- 5%.
3.9uF is so close to 4uF that the difference will not be heard,
and similarly for 6.2uF versus 6uF,
and those as slightly lower and slightly higher values in Series in the 44 cross-over will part compensate for each other,
AND they are in the correct proportions for their relative capacitance values in a 3rd Order cross-over, which this one is.

I would buy these instead of "Solen".

Wilmslow Audio "Supersound" are 400 volt "Solen" re-branded.

For those who can buy from USA, all the above cap values are available from Madisound in ClarityCap PX series at good prices, and from Sonic Craft in "Sonicap", both of which are said by various DIY users to be better sounding audio caps.
I like the ClarityCaps, and soon I will be trying the Sonicaps.

UK "Soniqs" branded are ClarityCaps, but unfortunately only available in a limited range of values.

************************************************** *****

Resistors will have to be installed in electrical Series with the 33uF and 6uF caps in the midrange filter to simulate the ESR of the original electrolytic caps there OR otherwise the crossover points will change AND the slopes of the filters will change.

Add 2.7ohm/5watt wirewound in Series between the 6uF and Ground.

Add at least 1ohm, preferably 1.5ohm, and a bit larger if you don't want a too-forward mids' sound, in 5watt wirewound in Series with the 33uF cap.

Time has run out.
I have to go now.
I continue this next time.

[alan-1-b]

Originally Posted by alan-1-b

(1) - These are 400volt rated, and use a Zinc-Aluminium alloy for their metalization, and that may be better than the plain Aluminium of most other MKP types.

************************************************** *******

(2) - Add at least 1ohm, preferably 1.5ohm, and a bit larger if you don't want a too-forward mids' sound, in 5watt wirewound in Series with the 33uF cap.

(1) - After what-ever was the original development of film capacitors,
Tin was next used for the metal inside, and that is said to be better than Aluminium.
It seems that Aluminium has been used instead of Tin because it is cheaper and/or easier to use.
I have forgotten the reasons that aluminium is not optimum for use in plastic film caps, but perhaps it may be that aluminium metalized plastic is harder to damp securely and thus resonates under applied voltage stress which causes additional internal voltage generation via the piezo-electric effect.
I do not know for sure, however I suspect a Zinc-Aluminium alloy may be less resonant in combination with the plastic film it is applied to as metalization.

Let's listen and decide !

******************

(2) - the ESR of an electrolytic cap is not constant - it varies across the frequency spectrum
-{but it is still a Resistance and Not a Capacitive Reactance}-
thus is different in the 500Hz region to what it is in the 5kHz region.
I estimated the ESR in the 5kHz region for the 6uF cap in the mids' filter, because that is primarily where it is acting,
however the 33uF cap is passing a wide band from below 500Hz to above 5kHz,
thus any fixed resistor added to simulate ESR will have to be a compromise.

If you prefer strong lower mids and non-prominent upper mids, then use a larger value resistor, perhaps as high as 2.7ohms, though I suggest starting a bit lower.

If you prefer prominent upper mids, and don't mind the lower mids being a bit recessed, then use 1ohm.

I'd start with 1.5 ohms, and listen, then decide what to do next.

The caps in the tweeter circuit do not need ESR simulating resistors, because the original caps were fairly low ESR,
BUT, the treble sound will be distinctly different with new polypropylene caps to what it was with the old caps,
and it MAY be a bit too much different in character to the sounds from the midrange driver.
To integrate the sounds from them both better I recommend adding about 0.5ohm in electrical Series between the 6uF cap and the tweeter.
This resistance is too low to noticeably attenuate the output level - it will only slightly modify the tone, to blend better with the character of the mids.

I'll Post about suitable types of 0.5ohm resistors next time.
I have to go now.

[alan-1-b]

For the Tweeter series' connected resistor, ideally a non-inductive type should be used, because any Series connected Inductance will attenuate the upper treble level, even if only very slightly with the very small inductance present in a 0.5ohm, medium-low power, wirewound resistor.

Different Manufacturers list low-ohm resistors in different ways, thus you may see:-
0.47ohm; 0.51ohm; 470mohm; 500mohm; 510mohm; ... and even a 560mohm will be OK.

USA residents and Mail-Order buyers can obtain Mills MRA-5 , 5watt non-inductive wirewound in one of the above values.

Only 12watt Mills are available in the UK, and those are too large in size for this sensitive application in the treble, and no-where near that amount of power dissipation is required here.

The current at 5kHz and above is a fairly small part of music signals,
and thus through 0.5ohm there will only be a very small voltage drop, thus low power.
A good quality 2 watt resistor will suffice, but a metal element type is preferable because the connection of metal wire leads to metal resistive element is electrically more secure for small sensitive signals than what may be the case for mass-produced cheap resitors of other resistive element types, except for the special applications types made by "Vishay" and "Caddock".

In the UK a Caddock MP930 in 500mohm/2.25watt/non-inductive can be bought from "Farnell" for 3.55.
This is a small size Radial format resistor.
Solder carefully to its connection tabs, and those are the legs to support the resistor on.
Keep the body of the resistor in air - not touching anything - so that its entire surface area can remain at the same temperature.
Its 30watt rating is ONLY for when it is attached to a Heat Sink, and that is not practical, nor necessary, inside a loudspeaker enclosure.
It will remain quite cool inside during normal music playing in this application, even to quite high volume levels.

If you don't want to pay 3.55, then for low prices Farnell also sell:-
"Welwyn" good quality UK manufactured wirewounds.

Welwyn type WP2S is 2.5watt@25*C and is available in 470mohm and 560mohm for about 30p.
The Inductance will be very low for 500mohm in this 2.5watt size.

The WP*S series are an enhanced pulse handling capacity type, thus will be assembled very well, much better than for standard wirewounds, hence suitable for sensitive audio signals.

For the ESR simulating resistors in the midrange filter,
and in the bass filter for application if one has bought polypropylene caps to replace the electros there,
Welwyn WP4S is 4watts@25*C and will be sufficient power dissipation because only low ohm values are being used in those locations.
WP4S is available in 1ohm; 1.2; 1.5; 1.8; 2.2; 2.7; 3.3 and 3.9ohms.

If you can't buy from Farnell, then the basic Welwyn W** series wirewounds are available from various sellers.
Buy from their W22 series, and not from W21 series -{except for the 500mohm}- as those are a bit too low in power rating for these applications.
{For 500mohm a W21 is sufficient dissipation, but its not the ideal resistor to use in series with a tweeter,
but if you can't find better ... there are a lot worse !}.

One wants the resistors to stay cool inside the loudspeaker enclosure.
Hot resistors can work OK in Power Supply applications, but not in audio signal applications, especially if their temperature fluctuates with signal level because that causes voltage generation at the junctions between the wire leads and the resistive element.

For ALL resistors do not bend their leads at the junction with their bodies.
Leave at least 1/8" and preferably 1/4" of wire extending straight out, and bend it there.
Support the resistors body by standing it on its bent leads, and with the body not touching any surfaces - especially not touching the surfaces of capacitors or inductors.

***********************************************

I'll discuss winding one's own Inductors next time.

[PaleRider]

When using resistors in low-freq; use several in paralell to increase powerhandlig.

[wintermute]

Hi Tinitus I'd suggest looking up Dielectric Absorbtion (DA) as a parameter of capactitors, if you don't think that different dialectrics can make a difference to the sound. Basically it is a measure of how much lag there is in a capacitor giving up it's energy. Ideally a capacitor should give back all of its stored energy instantly on demand.... however depending on the dialectric used it will range from almost instant to having quite a lag. It is the ones with the lag that cause the most problems... The issue is that they give up some of the energy when asked, and some of it a bit later... I think you should be able to see what the problem is with the energy that gets given up after it was required...

I've only read about the effects with respect to signals that are amplified, obviously any distortions introduced by a capacitor will be greatly increased if there is a gain stage following.... I don't know how much of an effect the DA will have when used in a crossover, but I guess the voltages and currents in a speaker can be quite high so my reasoning would tend to lead me to the conclusion that the effect should be audible.

It seems once again I'm posting this A link to a section of Walt Jungs web site have a read of his picking capacitors articles. Yes they were written nearly 30 years ago (which is all the more reason for posting them as the myth that there is no difference in capacitor types seems to perpetuate) but they are well worth a read IMO

Tony.

[LucasAdamson]

Thanks Palerider - good advise, but as I'm being a cheapskate and using bipolar electros in the bass (ie replacing them) the ESR should be about the same, so no resistors required.

[tinitus]

Hi Tony

Sure it matters, everything matters
But like with everything else its mainly the construction that matters
If a component is poorly designed and manufactored no fancy material will save it

MKT caps are quite adequate
Those yellow ones sold from english company linked to ealier looks much like the flat/oval silver coulered ones sold through Monacor, and they are fine, but have become a bit costly
What I say is, even the worst dialectrical plastic thing still have better dialectrics than paper stuff
We also know that the best are wound with foil/paper
So, no I dont buy it

I also said I would avoid bipolar electrolyte caps at any cost
Or at least the cheap ones
I know there are better ones (smooth), but they also cost serious money, and available only in smaller values
The bigger ones are the roughed ones, and suppposed to be less good fore xo work
I admit I have only heard a few
But what I do know is that I can do a very high end design with the cheap ones
If any design depend on any exotic meterial, Im off

Im also fed up with all this non inductive stuff going on fore years now
The few film resistors I have tried all sounded significantly worse than the good old wirewound cement, no matter where its used

In my youth the most exotic caps were polycarbonate
Then out of the sky came polypropylene
People went mad and couldnt get enough
To me its just seemed to sound worse and worse the more of this nasty stuff being used

So no, any salesman would have a real hard time to convince me of anything
And the more they talk the worse it gets
Its all bogus to me
I keep my money in my pocket and spend them on more important things

[PaleRider]

Originally Posted by LucasAdamson

Thanks Palerider - good advise, but as I'm being a cheapskate and using bipolar electros in the bass (ie replacing them) the ESR should be about the same, so no resistors required.

So you don`t use any impedancecorrection?

[tinitus]

Oh, I forgot to say I know, or think I know, why those "crappy" oval/flat MKT sounds so good
Because they are "flat" pressured
When they pressure them after "rolling" they get a really tight construction
Mr Duelund also discouvered this through his many visits to the danish part of Jensen
It was the first step to the now famous DCA
His request fore a different and better leads connection was more difficult to fullfil
But he made it happen, in a small messy apartment in the middle of Copenhagen
He wholehartedly hated all this new fancy plastic stuff
He is truly missed

[LucasAdamson]

Originally Posted by PaleRider

So you don`t use any impedance correction?

The original design didn't use any, and in the bass circuit there are no changes being made. I am just replacing the 35 year old bipolar electrolytic 72uF caps with near identical modern bipolar electrolytic 72uF caps, so I don't think it's required, is it?