Help with Crossover

edited September 2015 in DIY
Hi!

How are y'all? I haven't been around much (real life...), though I think I shall be again once I build a set of Colin's SECA amps. But I have an issue which I hope someone can help with:

My 'speakers (Triangle Antal ESW) are very lovely things, yet I have heard they (and all Triangles of this vintage) are easily improved. The trick is to:
  1. Remove the varistors from the crossover boards (which make up a protective element re overload)
  2. Replace the ceramic encapsulated wire wound resistors with something better of equivalent values


As I'm rained off today, I thought I'd tackle number one. I just need to be absolutely sure I'm right before I proceed (I have little electrical knowledge).

image
Treble and midrange board.

Looking at the crossover board (mid & top), I think the varistor is the small tan coloured disc - am I right? (The ceramic resistors are clearly visible in this picture too).

image
The bass board, same varistor?

image
Close up (Bass board).

As an aside, I am pleasantly surprised at the quality of the manufacture, the boards are thick with loads of copper, the soldering and internal wiring is frankly excellent. It's good to see some air cored inductors on mid and top too, though I suspect they (like everything on the boards) is made to budget. I seem to remember that placing them on edge (90 deg) from the board would be good practice though I may be imagining that, and it could be fiddly. Maybe some nice polyprop capacitors (not silly expensive) will also be beneficial.

I think the quality is explained by the fact Triangle manufacture everything in house including drivers (the tweeter is a horn loaded titanium compression driver). The compromises by the fact these retailed at under a grand in the mid-late nineties.

Comments

  • A little more googling gave me the confidence to identify the Varistors and whip them out. Well, it did stop raining. It was either commit or go out and do something useful! I took them out and replaced them with single run links of TQ black.

    So, sounding better? Don't know yet. It sounded wonderful before and it sounds wonderful now. I might put the serious amps on in a bit (if there is time).
  • edited September 2015
    OK, two glaring errors in my OP:

    Only the tweeter coils are air core (scope for improvement perhaps?)
    These were on sale in the mid to late 'noughties, not the nineties!

    So, as to sound. Removing these varistors has most certainly made a valuable difference to me, the level of presence and impact has jumped (these were hardly shy loudspeakers in the first place). This is real front row stuff! There is a new level of texture and timbre, in that musical cues I was aware of previously are now obvious. The separation of events within the soundstage that give clues to the ambience of a recorded space is vastly improved, listening to choral works is stunning! All this takes place in an even wider, taller soundstage - in fact the sense of scale is vast. This means all that is going on is spread about the room, thus avoiding any fatigue. I guess I'm saying over all frequencies, the sound is more extended (by which I don't mean they go deeper/higher).

    All in all, a nice little improvement which comes from simply taking something out. Can't get cheaper than that!
  • Hi Alan, how are you doing mate, do you still have the Old  Magisters?
    I know little about covers, you need PAC ( Paul) to pitch up here i Guess, 
  • edited September 2015
    Hi James,

    Im good thanks. Great to see you haven't electrocuted yourself (yet)! :P I bet your new amplifryer sounds great though.

    The Maggies were completely refurbed, and sold to a very happy camper in Somersert. I believe he is still on cloud nine!
  • image

    So, thinking about those nasty ceramic cased Wire Wound resistors shown above... The Picture shows 5W 4.7ohm with the 'J' denoting 5% tolerance (the other is 2.2ohm). Am I right?

    As I understand it (Ha!) wire wound resistors have some inductance, as they are basically coils inside. On the face of it replacing with non wire wounds could be a good idea, but what if the inductance of the wire wound resistor was intentionally a part of the speaker voicing process?

    So perhaps it's best to replace like for like with better wire wound resistors, like Jantzen or similar. Maybe some nice 10W jobbies with a 1% tolerance? There are also Mills non-inductive wire wounds as an option.

    However part of being better quality means the inductance is kept to a
    minimum, which reintroduces the prospect of unintentionally altering
    the character of the speaker (bad).

    I'd welcome some observations here, I don't want to go overboard but I need to understand a little about what I'm doing.

  • A|an, Just a thought Varistor are normally open circuit until the voltage reaches the value, so shorting it out is wrong remove yes. Now if it short then it is a semiconductor fuse or dead.
    Next the Nice big resistors may not be used in the protection arrangement and maybe there to match the sensitivity of the drivers in the box to the filter. So do not short, these guys are not fools and have more than likely got the design right.

    Be good now OK

    Col
  • Good to see you Col, I trust you had good hols? I'll be in touch soon.

    I took the Varistors out and replaced with wire links. Was that wrong? Should I cut the wires?

    The white 'coffin' resistors are properly part of the crossover I'm sure, not just for protection. I understand that they can be replaced with better components which do the job better, the internet told me so! I don't want to alter the crossover values at all or change the design, just maximise what has been designed.

    I want to know if I should put in equivalent value wire wound resistors with tighter tolerances and higher power handling? These will also have lower inductance - would that matter?
  • Alan said          "I took the Varistors out and replaced with wire links. Was that wrong? Should I cut the wires?"

     yes please it shorts the amp out via the resistors to protect tweeter.
    Resistor change yes use Non-Inductive Welwyn type I try and find them for you.
  • I appreciate that Col, I really do. Would that be better than the Jantzen I linked to four posts up?

    I just took the Xovers out and snipped the wire and guess what?....

    No sound. Zilch. Nada. The inky blackness was impressive though! I resoldered the link and all my 'Tree-ongly' [/french accent] goodness is restored. So you have my apologies Col, for I plainly have mis-described something or not understood sufficiently, for I plainly have misled you. Accidentally. Sorry. #-o
  • PACPAC
    edited September 2015
    Alan

    I'm guessing that the varistors were in line with the circuit, and not parallel, in which case removing them will cut any power to the circuit!  Replace with straight wire if in series.

    You wont really notice a huge difference in resistor types, trust me.  The more important thing by far is tolerance!  Use decent 1 to 3% tolerance resistors.  Welwyn are ok.  There's no such thing as a truly inductive wire-wound, even the cheapest ones measure so small in inductance as to be utterly meaningless in context with a passive crossover circuit;  We're taking fractions of a hundredth of a MilliHenry!!!  For noise, again, being a much higher  voltage circuit than say a preamp, there's no real advantage to so called low noise resistors, because the noise threshold with the applied voltage and current is appreciably well below audible threshold.

    Just use appropriately rated METOX or wirewounds.  I recommend 10W for bass duties and 5W for HF and that's probably overkill.

    The cheaper resistors are just as well suited for use as more expensive ones BUT are usually not as close on tolerance and tightening up here is what brings improvements to sound quality and pair matching because most modern drivers are very closely matched to begin with.  This means that to get what used to be called "reference standard" in matching for drive units is possible straight out of the factory these days, so close pair matching and tight tolerances on crossover components is important.

    Another set of improvements would be to replace those very cheap (and probably very nasty and wide tolerance) electrolytics with tighter tolerance versions (minimum 3 to 5% if possible) from the likes of Mundorf (very cheap).  Not all electrolytics sound the same as there's quite a few considerations (tolerance is only one consideration) in performance.


    I'd leave the inductors alone.  You wont get a huge improvement by changing those at all.  Cored inductors are best for LF duties provided that they are adequately rated.  Do not be tempted to switch to air core for the LF section because
    1) it will increase likelihood of cross-coupling with mid/hf inductors due to higher operating currents and wider field coverage of electromagnetic field and;

    2) it will lower speaker bass unit sensitivity due to a higher DCR.  This could also slightly affect crossover slope.  If you must spend money here, do a like for like with a FERON cored inductor such as the low/"Zero Ohm" versions from Mundorf which have negligible hysteresis but personally, I wouldn't think in this case you'll notice much of an audible improvement.


    The biggest improvement imho would be to remove the crossover boards , work out the circuit and then mount the components, hard wired to a ply board, with decent spacing for the inductors and in doing so provide a better damped mounting arrangement, hard wiring and removal from the influences of the speaker magnets on the inductors plus reduction in resonance applied to the crossovers.

    Hope this helps!


  • ^What he said.^
  • Good points Paul, and thank you. I have no reason to doubt anything you say. Regarding capacitors being a worthy target for replacement, my previous experience would agree with that.

    I will do the resistors simply because I have heard several anecdotes particularly relevant to my speakers, where it is said to have made a worthwhile difference. Just as removing the varistors has (same sources). Also it is easy and will cost little. Caps to follow. Thank you for allaying my concerns over the wire wounds and inductance.

    I have no real plans to swap the inductors, nor rebuild the crossovers externally, though I can certainly see the value in it.
  • edited September 2015
    Oh dear.


    Coupled Inductors = It the value and the current not the core or the air gap.

    Noise on Resistors = value below 10K are very low noise above then this is not true, https://en.wikipedia.org/wiki/Johnson%E2%80%93Nyquist_noise

    Inductor in large wire wound resistors = http://www.resistorguide.com/inductance/

    PCB or Plywood no PCB every time just get the tracks right I thought PAC used PCB in his design.

  • PACPAC
    edited September 2015
    Hi Col

    air cores tend to have a wider coupling range (I have tested both in various orientations using a calibrated meter with an applied current feeding them, and there are measurable differences or changes to inductance values depending on both spacing and proximity to things like conductive casings or loudspeaker magnetic fields).  In cored inductors the core limits the coupling field more so they can be spaced closer, and yes, the higher values with greater current couple more as you say (and as I hinted at already!).

    Noise is a non issue really for crossovers (passive) as most values will be well below 10K, (most being between 1 and 50 Ohm depending on position).

    I have measured inductance of all types of resistors used in crossovers  using very accurate meters and none exhibit inductance likely to influence the crossover design even by a measurable fraction.  Non inductive types sold as such are usually just a marketing gimmick to steer people to buy those makes imho.


    I use ply mainly with hard wired crossovers.  Nothing at all wrong with a well made, hefty PCB which I have used, but I don't like the thinner budget PCBs with narrow tracks and closely spaced components for passive crossovers.

    Not looking for a debate, just offering my own preferences based on experience having built and tested hundreds of the things
    ;)

    Happy to have the conversation in private with you if you'd like to discuss further but have offered what I have offered above for Alan's benefit.
  • No probs, thick track is good.
  • OK, new wirewounds are in, those ceramics are gone.

    Sounds even more...Wow! Gorgeous upgrade this, like going a speaker or two up the range! I know it's all economics, but it's hard to understand how a loudspeaker is so handicapped by component choices, and the Triangles are (I think) better than most to start with!

    I haven't got around to the BiPolar caps yet (they're having a good day).

    Differences are to do with texture and imaging (the imaging!), and a taming of some hitherto unnoticed harshness perhaps?


  • BiPolar are they on ‎CLi2O3 or diazepam ? 8-X
  • Nah...just means that the speakers have to be in the mood to play music  :))
  • And not ugly music as one Sam says what do you think Dave.
  • Royds lurve ugly music, even if Sam doesn't. 
  • AntiCrap said:
    BiPolar are they on ‎CLi2O3 or diazepam ? 8-X
    Col, I finally stuck the capacitors you recommended in the other day. The system has been in fairly heavy TV use for a few days now and I have waited to have my first 'proper' listen with the big amps. I was not initially impressed but what I heard when I first plugged these in the other day. More bass, for sure, but where had the glorious mid-range disappeared to?

    Fast forward to today, and... :D Nice one, Colin! Everything is better so far as I can tell. It seems very easy to hear subtle details of texture and tone which I wasn't aware of before. The Triangles have become almost head-phone like in the way they present these tiny nuances, and this is in no way lost at high volume either. And it was really high at one point... with no edginess or shoutiness on the part of the speakers. Just a really well sorted wall of sound between them.

    By sorted, I'm referring to tone, which remains just on the warm side, timbre and integration of the drivers. Really well integrated, seemingly, er seamless!? The soundstage remains hugely wide and the 3d depth of image has almost recovered to its previous holographic nature following crossover surgery. It's certainly heading in the right direction. The Triangles continue to shine a light into the intentions of each performer in an almost telepathic manner, just as before.

    Also, I noticed while giving them a work out, the bass is more extended. Useful bass seems to have gained at a semitone extra, perhaps a little more. I noticed this on Jame's Blake's 'Limit to your love' and several Massive Attack and Younger Brother tracks.

    All in, a nice little reward for a bit of time spent on a rainy afternoon. I can't say I enjoyed making the mods, but I'm now glad I did. I'll post some more if they change any further.
Sign In or Register to comment.