View Full Version : Ultimate crossovers?
After many months I have now finalised and completed the external crossovers for my Yamaha NS-1000M speakers. The components are held in a laser cut acrylic cradle and hard wired on the reverse - the circuit diagram is etched on the top surface. This structure is then bolted to another acrylic layer (at ten points) that is itself screwed to the chassis. The input is via 4mm banana sockets and the output is via locking Neautriks. A great deal of care has been taken to optimise component layout, choice and positioning. This is the best I can do without spending a huge amount more. If cost were no object I would probably build each of the three legs into its own case. Not shown (will take more pics tomorrow) are the granite plinths each crossover sits on - these are coupled to the crossovers via stainless steel ball bearings.
http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/41694599-63BB-4B45-8ED4-E9DDD6B9DA7A_zpsodfpmzo0.jpg
http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/04688627-CDAC-4F8F-9F2B-C65B3951C6AF_zpsdlcsye17.jpg
http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/54702AD4-E87A-46D9-94DE-A5ABCF3B6FE3_zps64lgyio1.jpg
Electronic components used are ClarityCap ESA and Mundorf Supreme capacitors - Monacor polypropylene for the bass. The inductors are all by Jantzen (highly recommended) and the resistors are also Jantzen (10W MOX).
istari_knight
05-04-2014, 19:10
Thats a sexy looking pair of crossovers :cool:
How do you rate the ESA's ?
Nice work I am sure they work well with the drivers
Thanks guys :). The crossover is essentially the same as the standard Yamaha one and the component values are the same. The difference is that I have split it into three distinct separate elements for bass, mid and treble. I've also deleted the L-pad adjustment pots from the speakers and these are replaced by the resistors that are now in the crossover.
I'm pretty pleased with the ESAs but in both the positions they are used I combined them with Mundorf Supreme and when I swopped the tweeter cap from an ESA to a Mundorf I was surprised how much better the Mundorf was. The problem with running any other caps is a combination of cost and size - each crossover is already the size of a sheet of A4 paper and I just don't have room for capacitors that are any larger!
...I am sure they work well with the drivers
I've been pondering the above John and I have to admit that I don't understand what you mean? I haven't invented this crossover, in terms of circuit design and component value it is the same as the one Yamaha designed for the speakers (second order for all legs) - as such, one would hope that it would work with the drivers.
It will work well with your speakers
Thats a sexy looking pair of crossovers :cool:
How do you rate the ESA's ?
It's not easy to honestly say exactly how I find the ESAs - I have done some capacitor experimentation but any level of proper investigation would involve spending hundreds of pounds on capacitors just for the sake of comparison! Obviously I've read all the online capacitor comparative reviews and my conclusion is that I think you have to take them with a generous pinch of salt. The 'accepted wisdom' is that the ESAs sound a bit tonally dark - I can't say I have found that particularly. It would have been interesting to try ClarityCap MRs or all Mundorf Supreme but in the values I would need that would have added hundreds to the cost of each crossover (they are hardly budget constructions already) and meant that at least two separate cases would have been needed.
It will work well with your speakers
Umm.. Yes, indeed - it won't work with any other speakers, but then it's not supposed to :). It's a long way from the single capacitor to the tweeter that my previous speakers used though - imagine the complexity if it used higher order slopes!
It's not easy to honestly say exactly how I find the ESAs - I have done some capacitor experimentation but any level of proper investigation would involve spending hundreds of pounds on capacitors just for the sake of comparison! Obviously I've read all the online capacitor comparative reviews and my conclusion is that I think you have to take them with a generous pinch of salt. The 'accepted wisdom' is that the ESAs sound a bit tonally dark - I can't say I have found that particularly. It would have been interesting to try ClarityCap MRs or all Mundorf Supreme but in the values I would need that would have added hundreds to the cost of each crossover (they are hardly budget constructions already) and meant that at least two separate cases would have been needed.
Neatly done Mark.
RE the highlighted bit: you do have to take them with a MASSIVE pinch of salt as they are (in isolation) pretty (utterly) meaningless (the blunt truth).
I have spent £100's on evaluating capacitors and not just by ear but by comparing distortion with frequency, microphonic behaviour etc etc. There is less difference between certain brands than marketing hype and forum chatter would have you believe in reality.
What people just can't seem to grasp is that the value of a specific capacitor for a specific position in a crossover, it's tolerance rating, voltage rating, microphonic behaviour and linearity AND the speaker system that they're used in are all far more important than whose name appears on the outside. More often than not, a budget capacitor of exactly the right value will "sound" better than a very expensive capacitor of 5% difference in value from where it should be.
What determines value is very specific to each driver's impedance curve and crossover point. There's NO such thing as an identical crossover for all speakers of the same model (or at least, there ought not to be) as each will have slight differences that might require slight alterations in cap or inductor values, although QC is getting so good these days that those differences are slight.
The point is that you simply cannot generalise on what cap will sound better in speaker "A" just because a particular capacitor works well in speaker "B"
Having said that, I rate ESA's very highly because they are designed for very low distortion results, and like all polyprops are very stable and linear even with moderate temperature change, but they are also very consistent on tolerance which is one of the main reasons that I use them.
There are other caps which I use for specific locations due to being less prone to microphonics than others too, so if going down the DIY route, whilst it is fun to experiment, there can be no hard and fast recommendations other than get the values and tolerances right!
How did you go about measuring microphonic behaviour within capacitors? I know ClarityCap did an investigation into this a few years ago and concluded that capacitors were immune from the effects of external vibration. However, they ere sensitive to resonance set up within the capacitor by it's very operation. To be honest, I'm surprised if you have managed any meaningful measurement as I had rather assumed that most polypropylene caps measured very similarly.
How did you go about measuring microphonic behaviour within capacitors? I know ClarityCap did an investigation into this a few years ago and concluded that capacitors were immune from the effects of external vibration. However, they ere sensitive to resonance set up within the capacitor by it's very operation. To be honest, I'm surprised if you have managed any meaningful measurement as I had rather assumed that most polypropylene caps measured very similarly.
Actually Mark, I didn't claim to measure, but to compare (there are plenty of data sheets available) and the comparisons I made were on data sheets and listening tests for an identical set up. I must have gone through a bucket load of capacitors! There have been studies undertaken (as you point out) by Claritycap and by Mundorf; in fact the ESA caps were the direct result of Claritycap's two year research programme into microphonics undertaken with Salford University. Claritycap tested 303 capacitors and concluded that where differences in sound occurred at all, it wan't due to ESR or dielectric loss, but due to mechanical resonance under applied signals (as you rightly point out) but there's isn't the only study as it has been known for a long time that capacitors of various types can be susceptible to distortion due to the applied signal. Claritycap altered their design from the SA range to modify the film used for the metallised film caps to damp out the deformation occurring when the film resonates upon applied voltage (and with other manufacturers, they design the film and how it's wound to minimise reaction to applied voltage). So yes, point taken, microphonics isn't the right description of the problem which is more internal resonance due to plate deformation upon applied voltage. My fault...poor description.It's an interesting read by the way for anyone interested and details should still be on Claritycap's website.
The main point was though, from my own experience, that the type matters less than the value and tolerance when comparing many polyprolylene capacitors, as most caps manufactured to similar price points use their own research to develop the most effective films for the very reasons that Claritycap have given. I have found very little difference between different brands at a similar price point with just one or two exceptions, but have found that Claritycap and Mundorf are both very consistent with tight tolerances which matters a lot to me. They are also sensibly priced which is important. For non signal path applications such as shunting across positive to negative in filters, I find no significant difference using budget or posher capacitors, again providing the values are correct.
The other point though which I'm sure you appreciate is that the danger when taking colour from some of these on line comparisons is rarely do they specify tolerances, capacitor use (position) and repeatability of testing. Inevitably they tend to be based on purely subjective assessments which whilst relevant to one party for one set up, might be wholly irrelevant to another party's set up and use. That's where the danger lies.
Ah, I see - you are quite right, you didn't say that you did the measuring - my assumption - apologies. I see what you mean now by microphony - I did wonder if you meant the internal resonance. To be honest, I completely agree with everything you say about tolerances etc. and, as you say, the majority of comment is purely subjective with the reviewer using them in a specific system of unknown worth.
As an example Mark, I was called to help a friend trace a perceived problem with an AN DAC (can't remember which one). When he played (any) music files through the DAC, with the amp switched off (nothing from the speakers) you could clearly hear the music. The noise was traced to some large Mundorf Z-caps in the output stage of the DAC. The internal plates were resonating over a wide enough frequency band to cause the capacitors to act as miniature electro-static speakers! I don't know what the measurable distortion plot would have looked like, but as this is an example of signal losses where electrical energy is being lost in mechanical work and heat over a specific audio frequency, it stands to reason that changing those capacitors for something which did't behave that way would reduce system signal losses. Problem is that the DAC may have been voiced with those specific components (I would sincerely hope not though as one would hope that caps were not used to "lose" information!).
Caps, just like inductors, are not ideal components. They have additional parasitic losses which include reactive inductance and electrical series resistance. You can add to that resonant plate losses which as yet are difficult to find in any text books. Regarding inductive reactance and ESR, this makes the capacitor behave in a non ideal way but usually not in any significant manner as far as audio bandwidth is concerned. The issue with reactive inductance is that at some point, the capacitor will have a resonance point (electrical) where on paper, instead of going to zero ohms as frequency rises, the inductive reactance kicks in and the capacitor starts to act like an inductor instead so that impedance starts to climb. This is usually in the Mhz range though.
There's a lot more to a capacitor than whose name appears on the outside, so perhaps more meaningful when considering trying different caps would be for folk to look at specific capacitor construction, materials and how much attention is paid to plate reactance for hifi applications. There are two basic approaches: Some manufacturers develop metalised poly films that are strong enough to be wound very tightly indeed, so that whilst the reactance mechanism doesn't change (ie the applied voltage still applies this resonant force to the wound plates), the windings are so strong and tight that the losses become insignificant in audio terms and distortion is measurably very low. The other way is to use films that in themselves are highly damped so reducing reactive losses and distortion.
Funnily enough, I've just been reading this - now that is synchronicity!
Secondly, has anyone tried hooking a cap up to a relatively high power signal source up to 20 AC volts with a high wattage 8 ohm dummy load in place of a speaker? With some caps, at about the 10 volt level, they literally start to 'sing' if you listen closely. Other caps don't. I'm not sure if it's a microphonics effect or something else. Give it a try and see for yourself.
Check out at this link starting at post #185.
Burning in hi-pass capacitors - Page 10 - Techtalk Speaker Building, Audio, Video, and Electronics Customer Discussion Forum From Parts-Express.com
http://www.diyaudio.com/forums/multi-way/205836-do-someone-compared-these-mundorf-caps-3.html
Here are three pictures showing the complete structure of each crossover. First there is a 40mm thick piece of polished granite. On this rest four stainless steel ball bearings and these are located by acrylic discs (these are just to stop the balls from rolling away, they don't support the balls except laterally).
http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/base_zps17b1f50f.jpg
On top of this is another acrylic layer. This locates around the balls - the feet of the crossovers also locate into holes in this layer. The ball bearings protrude about 2mm and make contact with the base of the crossover.
http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/baseandframe_zps9e2bff25.jpg
http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/baseandframeandcrossover_zps2fb80144.jpg
--------
I may well build a similar support structure for my phonostage and my pre-amp.
Were the air core and ferrite cores as per the original or did you chose air core/ferrite cores for specific specs/locations? And if so, why for each?
I'm looking to build a crossover for a pair of DIY speakers made by someone else a long time ago but who later went active.. so I need to build passive xovers for them. Got the schematic but not sure how to choose..
The original inductors were all ferrite cored ones. I purchased the replacement Jantzen ones from Audio Components (both strongly recommended). With an inductor their are two important measurements - the value in mH and the resistance (DCR). If either of these values are different to the original component you will alter the crossover. Although much larger than the original inductors my replacements all have the same value and DCR.
Air cored inductors are less efficient as an inductor but more linear than ferrite cored. Because they are less efficient they need to be larger and because they use more wire the resistance increases. To lower the resistance the wire must be thicker. The result of all of this is that comparing ferrite and air cored inductors of the same value and resistance means that the air cored one is very much larger (and more expensive) than the ferrite version.
In my case I used an air cored inductor for the tweeter and the upper leg of the mid. The lower leg of the mid and the bass use ferrite cores. The reason these are not air cores is because a/ they would have been huge and b/ they would have been very expensive (a single bass inductor of 5mH would have been £90 or so). Finally, I couldn't find air cores in the correct value with low enough resistance. You will see that a lot of commercial designs do the same as I did and for much the same reasons.
337alant
08-04-2014, 13:45
Excellent job of those Xovers there Mark :stalks::eyebrows:
Alan
Thanks Alan :). They weren't originally intended to be such a major element of the Yamaha refurbishment but one thing led to another and they ended up becoming a bit of a labour of love!
Hi Mark
A very tidy and well thought out job, it is good to nice workmanship still exists.
I would like to make a suggestion should you be looking to change a few things, V-caps if you are looking for a close as its possible to 'perceive' no caps in the signal pathway, then these should be on your list.
In all my years of auditioning components and making detailed notes, nothing and I mean nothing comes close, except the Auricap Teflons (both are not inexpensive) however they are simply the dogs dangles. I have nothing to do with the company that makes them, however for close to 10 years I have installed close to 1200 of their caps.
The Telfon / copper is just so right, the only bugbear is the physical size worth trying some out at some point.
Did you know Yamaha sold out the plant that use to make the drivers for the NS series Pioneer back in the mid 80's I think, now there are the only process plant that can vapor deposit Beryllium up to a 16.5cm cone size some feat!
Vapour depositing isn't too difficult - I've done it myself at uni. Just need a good vaccuum .
On an industrial process it's another matter - one up for the DIYer I'd say!
In previous life Vapor deposition was part of the company I worked for many strings, again implementation is the key.
Beryllium is very different material to work with and it took many years of working out correct processes to enable them to deposit that most awkward of materials on the periodic table especially at that larger a diameter.
Beryllium has an amazing heat dissipation quotient as well and a very rigid internal structure coupled with a light weight makes it ideal for aerospace applications as well as loudspeaker drivers!
It is extremely difficult to produce having to go through numerous processes to generate Beryllium oxide with only three countries in the world actually producing this material, suffice to say it is not a cheap option.
Hi Mark
A very tidy and well thought out job, it is good to nice workmanship still exists.
I would like to make a suggestion should you be looking to change a few things, V-caps if you are looking for a close as its possible to 'perceive' no caps in the signal pathway, then these should be on your list.
In all my years of auditioning components and making detailed notes, nothing and I mean nothing comes close, except the Auricap Teflons (both are not inexpensive) however they are simply the dogs dangles. I have nothing to do with the company that makes them, however for close to 10 years I have installed close to 1200 of their caps.
The Telfon / copper is just so right, the only bugbear is the physical size worth trying some out at some point.
Did you know Yamaha sold out the plant that use to make the drivers for the NS series Pioneer back in the mid 80's I think, now there are the only process plant that can vapor deposit Beryllium up to a 16.5cm cone size some feat!
Yeah, I did hear about the plant sale - hence the emergence of the TAD beryllium drivers ;).
The big problem with V caps (other than their size) is their extraordinary expense! One day perhaps - that probably will end up being a three box crossover (per speaker)!
The beryllium mid dome of the NS-1000Ms is amongst the very best I have ever heard - but then I also rare the TAD speakers highly :).
I've just had a look at the V-caps and HiFi Collective only do them up to 1uF so I guess you mean to use them as bypass caps? Mind you, a 1uF cap is over £500 which seems a tad ridiculous.
I've just had a look at the V-caps and HiFi Collective only do them up to 1uF so I guess you mean to use them as bypass caps? Mind you, a 1uF cap is over £500 which seems a tad ridiculous.
It is, in the case of using as a crossover bypass cap. A cap needs to do a job. Just using one with very low distortion, (including low plate resonance), linearity and the correct value is mainly what you need from a crossover cap. Nothing more exotic. All Polys today have plenty low enough ESR and are very linear with good tolerance to temperature changes. There are perfectly good affordable caps available for the job, some of which you have already used. The driver choice itself and correct cap value within a filter circuit have more impact on SQ than using the foo-ist of foo caps.
RE inductors, my view is that there's little wrong with laminated steel or P core for higher values (say above 2mH) providing that they're over-specified for power rating and don't saturate. Cheaper and less likely to couple with neighbouring inductors. RE layout, the golden rule I use is never to place them with their axis crossing another inductor's axis to avoid interference issue. Placed like this, you can have them surprisingly closely spaced (2mH typically at 100mm from neighbouring ones as opposed to twice that distance if both placed vertically axis-wise...ie the same way). Of course, going external allows more space usually so air core, if affordable would always be my first choice.
There is quite a lot of debate relating to the use of bypass caps in crossovers, whether they make any difference at all and whether that difference, if it does exist, is a good thing. From what I've read, the most commonly used V-Cap appears to be the 0.01uF and often in valve electronics (one for you Marco).
I did look at using air cored inductors throughout but the bass inductor is 5mH and that would be a mighty large inductor (over £80 too). In addition, to get one with the right resistance (low enough) Jantzen would have had to make me one especially (which they can do).
One of the changes I made to the crossover when finalising it was to move the two mid range inductors further apart. Even though they are aligned appropriately this seems to have been beneficial (air cored inductors having the disadvantage of a larger radiated field).
One thing I did when building these crossovers was to measure all the parts I took out and all the new parts I put in. With the Yamahas I am lucky that the circuit diagram, with component values, is easy to find. Tolerances of the original components were pretty large and pair matching is one of the areas that has been significantly improved through the use of tighter tolerance components. One thing that did surprise me was how closely toleranced the Jantzen inductors were - very good indeed.
There is quite a lot of debate relating to the use of bypass caps in crossovers, whether they make any difference at all and whether that difference, if it does exist, is a good thing. From what I've read, the most commonly used V-Cap appears to be the 0.01uF and often in valve electronics (one for you Marco).
I did look at using air cored inductors throughout but the bass inductor is 5mH and that would be a mighty large inductor (over £80 too). In addition, to get one with the right resistance (low enough) Jantzen would have had to make me one especially (which they can do).
One of the changes I made to the crossover when finalising it was to move the two mid range inductors further apart. Even though they are aligned appropriately this seems to have been beneficial (air cored inductors having the disadvantage of a larger radiated field).
It's surprising just how they are affected Mark with things like proximity to conductive plates or magnetic fields. The rule of thumb is not to have the axis crossing. Troels has measured the effects of various layouts and has produced a handy crib sheet which mirrors my own findings. The link fwiw is here and it remains a very useful guide:
http://www.troelsgravesen.dk/coils.htm
Thanks for the link Paul. Actually I have seen that before and studied it during the design process - it's part of the reason an aluminium sleeve is not being used. The mid and treble inductor may not be ideal but I have offset them a bit and I felt it was the best compromise - I did look at a number of different layout options.
Thanks for the link Paul. Actually I have seen that before and studied it during the design process - it's part of the reason an aluminium sleeve is not being used. The mid and treble inductor may not be ideal but I have offset them a bit and I felt it was the best compromise - I did look at a number of different layout options.
Yup...it can be a compromise. Listening and measuring response is a must really but I can't see you'll have any problems. It's nice to see someone produce such neat work, well done.
Thanks Paul - I try to consider all angles :). The crossovers took about six months in total so I didn't exactly rush in to it. They aren't a commercial exercise though - way too much time and effort to make.
http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/95557C2B-B598-4EF8-9CAA-B194659E4F02_zpsrzib2rq6.jpg
Going back to v-caps...I have some 0.1uF Teflon/ Cu in my power amps as interstage caps. Very, very transparent - incredibly so. I have no idea though about how'd they sound or otherwise in an xover.
Umm... OK - but I don't see them being used much in speaker crossovers so it's all a bit pointlessly academic. As I said before, the values are too small to be used as anything other than bypass capacitors and the concept of bypass caps in crossovers is quite a controversial one.
Probably is indeed academic given the vcap prices for large values. Nice job BTW. Very nice.
The other problem with the V caps as far as crossover bypass is concerned is that the Vishay MKP1837 / 0,01uF MKP 100VDC is also highly regarded and they are 5 for £3.00 (or 8 matched to 1% tolerance for £11).
The V Caps claim to fame is that they use Teflon as the dialectic and, as you no doubt know, small value Russian Teflon capacitors are available on eBay for a fraction of the V Cap prices (less than 30 for 30 of them - under a £1 each).
I have various Russian teflons, the ones I have need to be uncased if to not sound over-bright. You'll know there's more a cap than simply the dialectic material but that's a big topic. The V-Caps are very expensive and the law of diminishing returns is strongly in play here. Certainly I have found it harder to identify the characteristics of capacitors in xovers than in amps. I personally would prioritise placing expensive and hopefully good sounding boutique caps in amplifiers much more so than in xovers, the typical uF values used in xovers are also a significant factor driving up cost. The usually smaller caps in amps are more affordable though still extortionate. I've tried bypass caps on my tweeters and couldn't hear any difference so I don't use them and main main speakers today don't have a tweeter, my alternative speakers do use tweeters.
Powered by vBulletin® Version 4.2.3 Copyright © 2026 vBulletin Solutions, Inc. All rights reserved.