View Full Version : Tannoy Crossovers - Capacitor testing and evaluation
I've been playing now for 6 months with Tannoy crossovers, particularly the HPD crossovers, and thought it might be of interest to share and discuss findings on a few things, most especially criticality of capacitor choice.
First and foremost, had I been asked say a year ago about the differences various caps made to sound, I'd have been more than a little sceptical :mental: as correct capacitor value and type is all that matters...right? Not necessarily so!
Before I launch into capacitor choices, I'll start by saying with some degree of confidence that when it comes to the Tannoy circuit itself, Tannoy had it right, so any attempts to radically alter anything will mostly ruin the sound quality. Stick with their recommended values. You can ditch autoformers and use low DCR air core inductors (adding series resistance to make up the Tannoy equivalent values). That's perfectly fine.
Lets take the HPD crossover. To date, I've tried various MKTs and posh polyprops from various manufacturers and been a little disappointed with much of the initial outcome. I made the mistake of not thinking through in sufficient detail about the effects of various brands and types in various positions, instead looking for value and reasonable quality. Using Claritycaps for the whole thing proved to be a mistake, at least using the SA Clarity caps did. They may be 100% fine in other crossovers but not in the HPD one. There are two critical cap positions on this crossover, the 4u7 HF series cap and the 3u3 notch filter cap. The latter performs two functions. Firstly, it's part of the notch filter circuit itself and secondly it helps establish the roll-off shelf to the crossover point from somewhere just north of 3KHz down to about 1.5KHz.
What I've found, to cut a long story a bit shorter is that some polyprop caps have metallised foils which distort with applied voltage, and the degree of capacitor linearity and distortion is very important in certain locations...this is one of them. I've been working for months to lower the impact of the 3KHz resonance generated in the treble horn which gives rise to the infamous Tannoy "honk". The Fidelios have always been reasonable in this respect (as anyone who's heard them can attest) but since ditching the Claritycap SA and replacing with a Mundorf M-Cap Evo-oil, the transformation in SQ has been a bit significant. I was initially dismayed that I'd fallen for foo-ism, but looking closely at the capacitor physical and electrical characteristics yields a lot of answers. An industry cap specialist confirmed my suspicions that in certain locations, caps which have less plate deformation and greater linearity can, in the signal path, yield much better, smoother results with less (audible) distortion in certain situations (not a generalism, please note). The M-cap evo was recommended due to low micro phonic behaviour and low plate distortion.
End result is that voices are way more natural, as is piano, strings etc. This focusses attention on the mid range in a far more positive way than before as that resonant distortion is so well controlled, it is now barely audible.
There are more changes to come, principally with tests on the 4u7 capacitor choice, but for now, I have discovered a recipe of various caps used in specific locations within each crossover which seem to transform the humble HPD into something rather special, not that it was a slouch to begin with. I have also experimented with low value bypass caps on the bass shunt 6u8's but that's made little if no difference at all, so I would recommend that a good capacitor of the right value, with properties to suit its role in-circuit is more important than worrying whether to bypass or not.
Your experiences welcomed, and why you chose what you did.
realysm42
29-10-2013, 17:44
:popcorn:
Hi Joe
the 1u5 shunt you mention...which x-over? I have a treble roll-off 1u5 which isn't shunted. The only shunts are the bass 6u8 caps and the 3u3 notch filter cap. Were you referring to the 1u5 roll off circuit cap? I use an ESA in that location presently. I agree RE bypass...the effects are dubious. I generally use them only if doing like for like replacements for older speakers which use electrolytics where I use a polyprop low value to bypass the electrolytics to help linearity, but even then, only in the bass section as in every x-over I've tried, polys work best in the HF even if additional resistance has to be added to compensate for the ESR variance between electrolytics and polys.
The Black Adder
29-10-2013, 21:14
Hi Paul... lol, so sorry. I was cooking at the time.. lol. I meant bypass, not shunt. The 1.5uf 'bypasses' the resistors (either or both of the 10R, 20R resistors)
I'm afraid my experience isn't specific to Tannoy speakers. But I did find that changing the CkarityCap ESA I was using in the tweeter leg to a Mundorf Supreme was worthwhile.
Hi Paul... lol, so sorry. I was cooking at the time.. lol. I meant bypass, not shunt. The 1.5uf 'bypasses' the resistors (either or both of the 10R, 20R resistors)...that makes sense :D
I'm afraid my experience isn't specific to Tannoy speakers. But I did find that changing the CkarityCap ESA I was using in the tweeter leg to a Mundorf Supreme was worthwhile.
The ESAs seem to offer a slightly darker more shut in mid-range with less detail in the treble. Mundorfs certainly seem to be more detailed and smoother in response. The Claritycap MRs are something else though...very smooth and controlled but also very expensive. As I understand it, they were the results of extensive research by Claritycap into designing a low distortion polyprop by limiting film distortion when voltage was applied. I've found a mix of ESAs/M-cap Evo oils and the Jantzen superior Z caps in the Tannoys to be optimal (Jantzen used in the 4u7 position for the HF series cap). I've retained the ESAa in the bass shunts and bypass caps.
Well I'm toying with the idea of bypassing the ESAs in my new external Apogee x-overs with V-Caps.
Always bypass with a higher quality cap to gain much of the characteristics of the better cap on the cheap.
Did I say that? I did but only because I read it elsewhere. In my case it is a £240 experiment . It is hard to know if you have a result though. I reckon cap breakout boxes are the way to go for the serious trying to compare - instant switching between cap types for at least some chance of a realistic comparison.
Well I'm toying with the idea of bypassing the ESAs in my new external Apogee x-overs with V-Caps.
Always bypass with a higher quality cap to gain much of the characteristics of the better cap on the cheap.
Did I say that? I did but only because I read it elsewhere. In my case it is a £240 experiment . It is hard to know if you have a result though. I reckon cap breakout boxes are the way to go for the serious trying to compare - instant switching between cap types for at least some chance of a realistic comparison.
I think you have to be careful before making a decision on bypassing Justin as it seems to bring as many negative results as positive. It does help smooth out response in HF but at some expense of doing some odd things in the upper mids, so for that reason, I wouldn't use a bypass for mid/hf unless bypassing electrolytics as a cheap upgrade. Most good polys these days are pretty linear so I'd personally think twice before spending £240 on a bypass experiment. I have bypassed my ESAs in the bass as I'd not tried that but tbh, I can't say that there was any difference sonically and I certainly couldn't measure any difference using the mic and test tones.
There is merit on trying different caps in HF/mid circuits where they're in the signal path, but it is also costly. The end results can be worth it though. I use two methods of evaluation. Firstly I run test tones after setting pink noise (to 72dB) for various frequencies then measure response in room. Secondly, I use the excellent Chesky "Ultimate Test CD" and keep notes on key areas of subjective listening, including bass, bass texture, imaging, mid range, image depth, treble quality (detail & control) etc etc and compare the same values with each change of cap. Sometimes I notice little if any difference, other times, the differences can be quite marked (on the subjective test), depending on which cap was changed. I had one of those moments when the SA cap was exchanged for the M-Cap Evo Oil.
Where I used ESA's I have combined them with Mundorf Supreme. I'm not convinced it has made much difference though.
I've not found a huge difference between the supremes and the ESAs either Mark. Similar construction, similar dielectric, possibly similar plate distortion characteristics. The M-Cap Evos are different though and there is a difference there where I've mixed those with ESAs. The M-caps get my vote from the Mundorf range for use in crossovers without having to break the bank.
I'm not sure the construction is that similar as each supreme is effectively two capacitors wired in series, which isn't the case with the ClarityCaps.
That's very true Mark, but I was thinking more of the dielctric, geometry and foil used. They certainly don't seem to be much different in how they affect sound, whereas the M-cap is markedly different (better imho).
Yes, I was wondering if the dialectic difference is what you were thinking of. The Mundorfs are quite a fundamentally different construction to many other caps though - how relevant that is is another matter of course.
As you say, the Mundorf M-Cap Evo Oil are a similar price to the ClarityCap ESA and so relatively reasonably priced in the context of other 'exotic' capacitors.
Well I'm toying with the idea of bypassing the ESAs in my new external Apogee x-overs with V-Caps.
Justin , Should you decide to go Teflon I would recommend the Copper V-Cap , for similar £££ ,and my own slight preference , you might consider the Duelund VSF Copper Black
My (now) optimised versions:
http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7494_zpsf955202c.jpg
Notch filter improvements are significant using the M-Cap Evo Oils...no need to mess with resistor values (which would affect depth of notch or width of notch, so compromised effects either way). 3KHz resonance beautifully controlled to the point that its all but inaudible with most music at moderate volumes.
The Black Adder
30-10-2013, 14:26
Looking superb.. Nicely set out too and the coils are well positioned.
Which value tap did you go for on the HF coil btw?
Thanks Joe
I use the electrically flat setting for inductor equivalent which is the Green (2.9mH). This connects back to the Orange (6mH) so the 50R resistor needs dividing in value by the reflected impedance ratio, in this case about 2. The shunt is therefore 25R and not 50 (a common mistake when building fixed inductor Tannoy circuits is to forget this). The M-Cap Evo-oil is the drum-shaped white cap to the top right hand side of the PCB. Trouble is with all the experimenting, I managed to lift one of the tracks this morning but it' hardly surprising as I've used this PCB as a test bed and soldered/desoldered it many times now! From now on, I'll be creating hard wired external x-overs for anyone wanting these now that I reckon I've nailed the design/selection of components.
The Black Adder
30-10-2013, 17:18
lol.. you should have seen my test boards. A bit of a mess but all well worth it.
Listening now for a fair few hours to the Superiors in place and what was taken as greater clarity and detail is now starting to sound like too much brightness compared with the ESA caps. I'll give them 10 hours or so to bed in but tempted to go back to the ESAs for the 4u7 position as their balance was more neutral. Strange how the same capacitor values can sound so different between makes.
The Black Adder
31-10-2013, 09:52
I tried some Superiors on both the HF and notch filter and found that the sound was open in detail but seemed strained somehow. I had two test boards running that I switched. The other boards using ESA's. (I chose the ESA's as a base standard).
I really like ESA's. The thing is when you use MR's they are so much more involving but yet could also be quite unforgiving on the detail so you need to cool them down a bit so to speak and so mixing (as you have found out) it up with a PIO on key sections really does let the MR's breathe better and give the sound much more presence and energy. I've made some crossovers with ESA's and mixing them with MR's and although the strength of the sound is compromised slightly that combo also works very well.
The only thing with MR's is that they take a good month to really start to break in (using them every day for at least 3 hours). With the testing I did I set (three test board pairs) each combo a good month to bed in. It took ages, lol. I think Marco found this with his MR's too.
The Superiors did get better btw after the long break-in.
I tried some Superiors on both the HF and notch filter and found that the sound was open in detail but seemed strained somehow. I had two test boards running that I switched. The other boards using ESA's. (I chose the ESA's as a base standard).
I really like ESA's. The thing is when you use MR's they are so much more involving but yet could also be quite unforgiving on the detail so you need to cool them down a bit so to speak and so mixing (as you have found out) it up with a PIO on key sections really does let the MR's breathe better and give the sound much more presence and energy. I've made some crossovers with ESA's and mixing them with MR's and although the strength of the sound is compromised slightly that combo also works very well.
The only thing with MR's is that they take a good month to really start to break in (using them every day for at least 3 hours). With the testing I did I set (three test board pairs) each combo a good month to bed in. It took ages, lol. I think Marco found this with his MR's too.
The Superiors did get better btw after the long break-in.
Interesting. The MCap Evo oil isn't a PIO but same principle using a thick metal coated Polyprop winding in oil soperhaps similar benefits. The ESAs take some beating IME where crossovers are concerned and I've never seen much benefit in spending more. I'd rather put the cash into better "something else" like a cartridge or more vinyl! I'll give the Janntzens a little while more but it oughn't to take them that long to form properly...a few minutes/hours/days? My experience so far suggests that what you get after a day or two tends to what you get and any differences after that I think may be me getting used to the sound.
The Black Adder
31-10-2013, 11:06
Sure, it's all horses for courses.
I totally agree, ESA's are very hard to beat for the money. In my Hiarachy terms it's ESA's, MR's and Duelund Casts. The latter being just a bit bonkers in regards to price. I've only used the cast caps once and my verdict was 'very impressive' but if they were for my own personal use the outlay would've not have been a viable option. How they make those things must be some extremely expensive process.
With break-in times, personally I need to get used to a certain sound and get my head around it but that's just me being awkward.. and/or maybe a little geeky. But we all have a built in 'sounds horrible-o-meter' I suppose and if it's tripped then it's gawn.
Ali Tait
31-10-2013, 15:44
Listening now for a fair few hours to the Superiors in place and what was taken as greater clarity and detail is now starting to sound like too much brightness compared with the ESA caps. I'll give them 10 hours or so to bed in but tempted to go back to the ESAs for the 4u7 position as their balance was more neutral. Strange how the same capacitor values can sound so different between makes.
Differing ESR perhaps?
I initially presumed so Ali but in likelihood, the ESR values should be comparable as they're both similarly constructed Polyprops. The rating of the Jantzens is 800V v's 250 for the ESA caps but that shouldn't alter perceived brightness. the larger value is meant to exhibit slightly less distortion but the trade off is that big values need a little more bedding in time.
Instead of guesswork, I need to make time to measure the in-room response using a mic and some software to evaluate. I'll get round to that in a few weeks.
Ali Tait
31-10-2013, 19:04
Who's Allen?
Who's Allen? Sori Ali...long day...another 5am start leaves the brain befuddled by the end of the day! Edited ;)
Justin , Should you decide to go Teflon I would recommend the Copper V-Cap , for similar £££ ,and my own slight preference , you might consider the Duelund VSF Copper Black
Yeah my DAC has some CuTF caps in and sounds great. However, after discussing it with the chap building the x-overs he is pretty anti silly priced caps. I think I'll go all ESAs and maybe accidentally mess with V-Caps later on.
The current setup in my Apogees is Clarity PX bypassed with ESAs. The speakers sound utterly awesome to me as they are. There's a complete mass of upgrades in the new ones and I'm sure V-Cap bypass on ESAs will just be icing on the cake.
Instead of guesswork, I need to make time to measure the in-room response using a mic and some software to evaluate. I'll get round to that in a few weeks.
You actually bought one in the end then, after me telling you to NOT buy an SPL meter over and over again? LOL.
Ali Tait
31-10-2013, 21:41
Sori Ali...long day...another 5am start leaves the brain befuddled by the end of the day! Edited ;)
S'ok, just kidding.:D
You actually bought one in the end then, after me telling you to NOT buy an SPL meter over and over again? LOL.
I have the right tools for specific jobs Justin so don't sound so surprised. I don't know about "over and over again" but you did have a bee in your bonnet about SPL meters...what have they ever done to you :lol: An SPL meter is still a useful tool. Calibrated full range usb mics have come on leaps and bounds and with the right software are an invaluable aid in loudspeaker testing and tweaking. Hope you don't mind me saying but I don't tend to do what people tell me to do...I make choices based on need, and generally speaking, I understand that need better than those doing the telling ;). It's my day job.
I think you have to be careful before making a decision on bypassing Justin as it seems to bring as many negative results as positive. It does help smooth out response in HF but at some expense of doing some odd things in the upper mids, so for that reason, I wouldn't use a bypass for mid/hf unless bypassing electrolytics as a cheap upgrade. Most good polys these days are pretty linear so I'd personally think twice before spending £240 on a bypass experiment. I have bypassed my ESAs in the bass as I'd not tried that but tbh, I can't say that there was any difference sonically and I certainly couldn't measure any difference using the mic and test tones.
There is merit on trying different caps in HF/mid circuits where they're in the signal path, but it is also costly. The end results can be worth it though. I use two methods of evaluation. Firstly I run test tones after setting pink noise (to 72dB) for various frequencies then measure response in room. Secondly, I use the excellent Chesky "Ultimate Test CD" and keep notes on key areas of subjective listening, including bass, bass texture, imaging, mid range, image depth, treble quality (detail & control) etc etc and compare the same values with each change of cap. Sometimes I notice little if any difference, other times, the differences can be quite marked (on the subjective test), depending on which cap was changed. I had one of those moments when the SA cap was exchanged for the M-Cap Evo Oil.
Both bass and mid/treble sections have ESA bypasses on my current speakers which let's face it - you know quite well.
I have the right tools for specific jobs Justin so don't sound so surprised. I don't know about "over and over again" but you did have a bee in your bonnet about SPL meters...what have they ever done to you :lol: An SPL meter is still a useful tool. Calibrated full rage usb mics have come on leaps and bounds and with the right software are an invaluable aid in loudspeaker testing and tweaking. Hope you don't mind me saying but I don't tend to do what people tell me to do...I make choices based on need, and generally speaking, I understand that need better than those doing the telling ;). It's my day job.
Bollocks.;)
Both bass and mid/treble sections have ESA bypasses on my current speakers which let's face it - you know quite well.
The point being what? That bypasses work well in your system compared with no bypasses? Not sure I understand the point you're making Justin.
Just so everyone understands - that was only a light hearted boollarks issued to Paul. Nothing serious. Since we're basically mates I think people need to know that and am worried it might have been taken out of context.
The point being what? That bypasses work well in your system compared with no bypasses? Not sure I understand the point you're making Justin.
It was just for info - since you were talking about bypasses in different places and their effectiveness. All I am saying is that the ESA bypasses must be working well on both PX sections. TBH though I have no idea what real effect they have had since I've never removed them.
:lol: No offence taken....I know you too well ;)
It was just for info - since you were talking about bypasses in different places and their effectiveness. All I am saying is that the ESA bypasses must be working well on both PX sections. TBH though I have no idea what real effect they have had since I've never removed them.
The true value of a bypass as I understand it, is that when using particularly high capacitance values, it allows a more linear response, particularly at HF, than otherwise might be possible using say something less than suitable such as a cheap electrolytic (ie if you bypass an electro with a small value poly, it improves response significantly). However, there's evidence amongst many who use bypass caps that it can also do strange things with the mids, so it's not always the way to go. If trying it, always best to use a much higher quality cap for the bypass and make it between 1/10 and 1/100th the value of the cap being bypassed. I'm not really convinced that polyrops need it or indeed benefit much by it.
I have on occasion used capacitor bunching to make up large values though (ie between two and five caps in parallel to make up a large value) as in theory, this is meant to emulate a more ideal capacitor response. I don't have the test equipment to measure how effective that is, so only have the theory to go on. If I end up doing more of this work, I'll have to invest in more measuring kit (signal generator and a scope) as it's the only way to be sure.
337alant
17-11-2013, 08:37
Reffc
Excellent write up and very usefull, thanks for posting
http://www.humblehomemadehifi.com/Cap.html
This is also a very usefull link to a set of Capaciotor tests
Alan
Reffc
Excellent write up and very usefull, thanks for posting
http://www.humblehomemadehifi.com/Cap.html
This is also a very usefull link to a set of Capaciotor tests
Alan
Thanks Alan. Yes, I visited that site but would have been interested more if there'd been some measured data to complement the subjective opinions, as these can only be based in the single context of that reviewer, his speakers and hearing. Personally, I found that I preferred the ESA capacitors over the Jentzens alluded to on that test.
What WOULD be interesting is for someone to produce a list of capacitors describing the construction and measured distortion from each when subjected to various test signals. I would be interested to see if this correlated to listening tests, as the two together could form a more reliable guide. For now, the list of capacitors that I would recommend for crossover use are as follows:
Budget category
Ansar polypropylenes: well made, inexpensive and can be ordered to reasonable tolerances. Highly recommended;
Solen Polyproplynes: As above really, a complete no-brainer and can be ordered to very tight tolerances by request from suppliers such as Jerry at Falcon Acoustics.
Mid Priced:
Claritycap ESA: superb tonality when compared in like-for-like values with other similarly priced capacitors used in the signal path in mid range and HF sections of crossovers. Balanced, clean and whilst some caps lend an air of presence which can border on brightness (which is audible but odd since crossover points haven't altered), the ESA's unoversally seem to be even handed where I've tried them. I cannot recommend these highly enough.
Claritycap MR: Claritycap undertook detailed research into voltage induced plate distortion in crossover capacitors and their meaured results proved rather eye-opening. The MR was specially developed to limit this distortion to a fraction of what other polypropylene capacitors seemingly generated under measurement and has been described as the ultimate crossover capacitor. I have my reservations about this though. They are undoubtedly good, in tests against the ESAs they offered a very slight sonic advantage in clarity and also when used in the signal for mid range units, they seemed to exhibit slightly more body to the sound, but the differences seemed minor. Their biggest drawback is cost. The are almost prohibitively expensive.
Mundorf M-Cap Supreme: I like these almost as much as the Claritycap ESAs. Very similar in many respects. Slightly different plate design with Mundorf making their own claims about why this is the ultimate speaker crossover capacitor. I don't know about that but it shares a similar clarity when used in HF signal path to the Claritycap ESAs, although it is slightly more expensive.
Mundorf M-Cap Evo Oil: A strange one this. Again, a "different" physical construction to just about any other capacitor I tested. They claim a very short, almost "lossless" signal path with ultra-low ESR, combined with exceptional tolerance, but it also has very low equivalent series inductivity (ESL). Of all the capacitors trialled in my Tannoy crossovers, none made such a startling difference (and I use the term advisedly) as this one when used in the notch filter to reign in the audible effects of the 3KHz horn resonance that all MG and HPD tweeter horns suffer from. Quite how a capacitor does this, I'm not at all sure, but it was tested repeatedly with the same audible results...ie a significant improvement in clarity of signal and reduction in audible lift/distortion.
It is also a very good cap to use for the main HF signal path and best of all is that the prices are very reasonable.
One other trick I tried with various caps was to use the solens and other lower cost caps and bypass them with small value Mundorff MCap EVO oils with some good results. This can be rather hit and miss sonically but it is a way to improve linearity with increasing frequency.
What seems to matter more than the label on the capacitor though is getting the values right.
Get the values SPOT ON for your crossovers, and off the shelf budget capacitors can make a crossover that sounds every bit as good, if not better than a crossover which uses posh and expensive caps but of values which aren't making the most out of the best ranges (crossover points) to run each drive unit. That and modifying older crossovers with impedance flattening circuit additions to both low and HF sections of crossovers can have a dramatically beneficial effect over simply spending £100's on capacitors.
337alant
20-11-2013, 05:20
Thanks again Reffc I think I will order some of those Mundorf M-Cap Evo Oil caps for my Mid Tweeter Xover
Alan
Hi Paul
I finished the x-overs and they can be seen here on my site: http://www.jkwynn.co.uk/HPD_HTML/HPD_Thumbs.html
Just to supply the background for folks, I have been in touch with Paul for many months regarding the x-over design as we found we were developing along parallel lines.
I can confirm his findings regarding the ESA's, they work very sypathetically with the Tannoy HF driver.
He also convinced me to try the Mundorf EVO Oil in the HF notch filter and again I would agree that it works extreamly well in that location.
:thumbsup:
Hi Paul
I finished the x-overs and they can be seen here on my site: http://www.jkwynn.co.uk/HPD_HTML/HPD_Thumbs.html
Just to supply the background for folks, I have been in touch with Paul for many months regarding the x-over design as we found we were developing along parallel lines.
I can confirm his findings regarding the ESA's, they work very sypathetically with the Tannoy HF driver.
He also convinced me to try the Mundorf EVO Oil in the HF notch filter and again I would agree that it works extreamly well in that location.
:thumbsup:
Great result Ken ;)
Looking at your latest x-overs, they are now almost identical to my own design, the components are pretty much the same. It remains for me the optimal crossover for the HPD drivers so good to get the feedback that those components/values are working well for you too.
Powered by vBulletin® Version 4.2.3 Copyright © 2024 vBulletin Solutions, Inc. All rights reserved.