[kyrill]

Hi Hugh

Lost my email where you specified what high grade capacitor i need to connect in line with the tweeter to protect it from the current peak when you shutdown the amp
it is the Aurum Cantus G3 ribbon tweeter ( http://estore.websitepros.com/1736754/Detail.bok?no=216) 8 ohm

It is directly connected to an Aksa 32 watt version ( Aksa 55 watt but with smaller transformer) in an active crossover setup (DEQX)

: )

[AKSA]

Hi Kyrill,

Nice to hear from you!  I hope you and your family are well, and that the AKSA continues to give good service.

These caps are definitely a good idea for a ribbon tweeter, direct connected.

My tweakers tell me the most appropriate is a Mundorf Supreme, or a Sonicap Gen II, or possibly a trusty SCR bypassed with a much smaller one of the above.

In this situation you are looking for the best possible sound, and this series cap is important.

Cheers,

Hugh

[kyrill]

He he

Good morning Hugh

what value and voltage?

 

[aurelius]

Hi Kyrill,

What is the crossover frequency you've set the DEQX at?

Simplifying the situation a little, a capacitor in front of a speaker acts like a simple first order high pass filter with a crossover frequency of fc = 1/(2*pi*R*C). 

I would want the capacitor-imposed crossover point to be a couple of octaves below your DEQX crossover frequency, so that it has minimum impact on standard operation.

If you plug your DEQX crossover frequency as f into the following formula:

C = 1,000,000/(12.6*f)

and then round up to the next nearest value in uF, you will end up with a suitable value. 

RE: voltage, anything greater than 20V will be fine.

[andyr]

Hi Hugh

Lost my email where you specified what high grade capacitor i need to connect in line with the tweeter to protect it from the current peak when you shutdown the amp
it is the Aurum Cantus G3 ribbon tweeter ( http://estore.websitepros.com/1736754/Detail.bok?no=216) 8 ohm

It is directly connected to an Aksa 32 watt version ( Aksa 55 watt but with smaller transformer) in an active crossover setup (DEQX)

: )

Hi kyrill,

Why do you think you must have a series cap to "protect" the ribbon tweeter from "the current peak when you shutdown the AKSA 32"  ?

I run a 3-way active system with an AKSA 25 (shortly to be a LF!!  : ) ) directly connected to my Magnepan ribbons.  This I venture to suggest would not be any more "sturdy" than your Cantus G3 ribbons and I have had abolsutely no issue in terms of "power-off thumps".  That's the beauty of an AKSA, IMO!

Regards,

Andy

[AKSA]

Hi Kyrill,

My apologies;  I was in a screaming hurry to get out the door to listen to LM's new Lifeforce (that's Lynton M., who lives in an adjacent suburb!) and shortshrifted your question.

I will work an example according to Aurelius' (Mark C's) equations, based around 3dB corner frequency drops.  I will assume the ribbon is 8R;  but in truth if its 4R, double the capacitance, and if 2R, double it again.

Assuming 1000Hz DEQX crossover, we have 1000000/12.6*1000 which is 79.36uF, npv would be 82uF.

The precise impedance at 1000Hz is 159000/1000*79.36, which is 2R, gives us a 3dB point of 8R at four times lower frequency, viz two octaves, that is 250Hz as Mark has suggested.

If the crossover is 500Hz OR the ribbon impedance is nominally 4R, double this capacitance to 158.7uF, npv 2 x 82uF, OR 100uF and 68uF.

Voltage rating need not be too high;  63V would be fine for a low power AKSA, but you may find it necessary to buy 100V or 250V ratings because these are all that is available.

Finally, you might like to mix 'n match, since this will save money.  A good policy is around 80-85% cheap cap, like a Solen, with 20-15% an expensive cap, like the Mundorf or Sonicap.

In this instance, for 79.36uF (and the exact value is not at all critical), I would suggest 68uF of Solen, and 12uF in the quality cap.

Hope this helps!

Andy,

Ben (Tinker) had this issue as well, and after much discussion we decided that since the infrasonic switch off spike was considerable and lasted about half a second, it might be best to cap couple his sensitive tweeters.  I agreed;  he has settled for 39uF of SCR, and without bypassing it seems fine with little or no degradation of sonics on his Phoenix.

Cheers,

Hugh

[andyr]

Andy,

Ben (Tinker) had this issue as well, and after much discussion we decided that since the infrasonic switch off spike was considerable and lasted about half a second, it might be best to cap couple his sensitive tweeters.  I agreed;  he has settled for 39uF of SCR, and without bypassing it seems fine with little or no degradation of sonics on his Phoenix.

Cheers,

Hugh

Hugh,

Did Ben actually try not using a cap ... or was he too afraid of frying his tweeter that he put one in "just for sure" and so has never listened to his Phoenixes "naked"? 

Because if he hasn't, he doesn't know how they might sound like?    So it's hard to say whether there's any sonic degradation or not!   

And I can't believe that Maggie ribbon tweeters (which are supposedly thinner than a butterfly wing!  ) are more robust than other types of tweeter - particularly the cone drivers used in the Linkwitz speakers!     And I've never had a problem.

Regards,

Andy

[Russell Dawkins]

For the finicky, wouldn't it be nicer to have the option of a high quality switch in parallel with the cap?

The switch would be open only for amp power on/off (which ideally would not be each time the system is used, the amps being left on and the idling consumption tolerated).

This would allow the use of a cheap cap - the switch should sound better than the best cap in the world.

That's what I would do, anyway.

[andyr]

For the finicky, wouldn't it be nicer to have the option of a high quality switch in parallel with the cap?

The switch would be open only for amp power on/off (which ideally would not be each time the system is used, the amps being left on and the idling consumption tolerated).

This would allow the use of a cheap cap - the switch should sound better than the best cap in the world.

That's what I would do, anyway.

Absoloootely no way, Russell!!     IMO, putting a switch in the way of the output signal will seriously compromise the sonics.

If you were using a cheap (read crappy) "receiver" then you wouldn't be able to hear the difference but when the amp is an AKSA - even an N+, not necessarily a LF - you certainly will!   

Regards,

Andy

Regards,

Andy

[Russell Dawkins]

So, I take it you think a capacitor permanently in series with the tweeter will sound better than a switch contact?

 

I did say high quality switch!

Am I to presume your speaker wires are soldered to their respective terminals - or even better, wire wrapped? If not - if you are using banana plugs, for example - then a good switch contact will represent a similarly good connection.

It is my turn to be surprised!

I can't imagine that even an Audio Note silver-foil-in-oil cap will approach the sound of a set of quality switch contacts in a SPST switch.

I'd be interested in hearing Hugh's opinion.

[AKSA]

Russell,

Thank you for your post.  Interesting question, but one I try to avoid for a couple of reasons.  The first is that I'm not an engineer, and thus am vulnerable if I give an opinion on something so divisive, and second, this divisiveness extends across the whole spectrum and can produce enormous friction and name calling.

Have a look at the 'Audibility of Output Coils' thread in the SS section of DIYaudio.com to get a handle on just how acrimonious these things can be......  it ain't pretty.....

Capacitors color the sound.  The best of them, in my experience, have been the teflons and the polystyrenes, which are difficult to detect.

OTOH, some caps color the sound objectionably, like some metallised polyester caps, and some color it euphonically, like the Jensen paper in oils and the RTX hybrids, ps and pp foil, and the Sprague orange drops widely used in guitar amplifiers.  It is thus OK to say 'depends on the cap'.

Musical instruments play to a certain tone, and the electronics and source instrument are carefully tailored to that tone.  A Fender Stratocaster and Tweed amp has a certain sound, so does a Gibson with a Vox, which uses EL84s in Class A.  Traditionally, and even this has caveats with tube amps, the hifi amp strives to be faithful to the source, so we want our caps to be without color.  Yet there is evidence even in hifi, and a good example is my AKSA which is slightly colored, that people like a bit of color there too.

The argument is vehemently put in objective engineering circles that there should be absolutely no measured distortion in a high fidelity amplifier.  Yet even in direct coupled designs, there will be a blocking cap, and this will add color, not to mention the many caps in the passive crossovers of almost every commercial speaker in the marketplace.  As a general rule, less busy music like jazz quartets and simple vocals, sound very good on somewhat distortive systems, such as single ended tube amps.  But busy music, like large orchestral pieces and rock concerts sound better on very low distortion solid state amps.  And so, while the focus is correct and often ideologically driven the reality is something different.  And besides, people's tastes do vary enormously.

And so, my answer is not definitive.  Objectively, the cap in series with the tweeter will increase distortion, but if the cap is well chosen, the distortion will be limited to H2 and a smidgin of H3, and may sound very nice indeed.  Subjectively, the issue is clouded by psycho-acoustic preferences.

Simply put, whatever floats your boat!!

As you can see, I'm an artful dodger.  I'm sorry about this, but no one answer is strictly valid.

Cheers,

Hugh

[andyr]

So, I take it you think a capacitor permanently in series with the tweeter will sound better than a switch contact?

Not exackerly, Russell.  IMO there is no need for a cap ... so no need for a switch either.  

And as any cap introduces a sonic signature, if it's not absolutely needed ... why bother to introduce it?

I did say high quality switch!

Am I to presume your speaker wires are soldered to their respective terminals - or even better, wire wrapped? If not - if you are using banana plugs, for example - then a good switch contact will represent a similarly good connection.

No, my speaker wires are not soldered to their respective terminals.  I believe I would certainly get better sound if I did ... but speakers and amplifiers need to be moved occasionally, thus the "real world" intrudes into the "optimal world"!    Hence, I don't want to introduce yet another contact in the signal chain (with a switch).

And yes, bare wire wrapped around the binding posts would sound even better - for a short time until the copper surface oxidised.  I'm not the sort of person who likes to undo the wires each week to clean them, so bare wires are out for me!  

BTW, I do believe there are banana plugs ... and banana plugs.  And - unlike that scene in "Croc Dundee" - more metal is not better!    So I use Multi-Contact bananas made of gold-plated copper in a kind of a sprung cylinder format.  These make good contact all the way along the barrel and are lightweight (I subscribe to Mr Eichmann's theories about minimal metal in the signal path, so when I rebuild my AKSAs into LFs, I'll use Eichmann Cable Pods instead of the Vampire BPs I currently have).

Regards,

Andy

[Jens]

Hi Hugh,

Could you quantify the voltage swing and duration of the "thud" for each of your amps and list it?

I think this would give us something to go from. In reality, most quality tweeters and ribbons have a far higher tolerance vis-a-vis short-term effects like the turn-off thuds of the AKSA amps than most of us believe.

For example, I know for a fact that the Scan-Speak tweeters that I've been using (previously the 9900 Revelator and now the R2904 ring radiator) are very sturdy and can withstand a lot of abuse - even unfiltered. A DIY shop here in Copenhagen used to demo sturdiness of the 9900 by simply wiring the terminals of the tweeter directly to the terminals of an 80 watt amp and then turn up the volume. Nothing ever happened to the tweeter .... and they certainly did not stop after half a second!

However, with exact quantification of the "problem", each of us would be able to go to manufacturers (or knowledgable dealers) to obtain information about just how much would be too much for our respective tweeters and ribbons - and then decide for ourselves whether or not some measures of protection should be taken.

If protection is needed, I believe a switch arrangement that only disconnects the "live" wire to have the least impact on sound (provided, as Russell notes, that a high-quality switch is used).

Introducing either a switch or a cap introduces two solder joints en route to the driver. However, I believe that even the best of caps cannot compete with a simple switch point for "direct routing".

I wonder, though, if a simpler solution to protection would be simply to short output from the amp with a suitably sized resistor at turnoff (e.g. using a relay) - taking into consideration not to damage the output stage, of course. Hugh, what do you think? Also, to how great an extent would such an arrangement diminish the problem?

[AKSA]

HI Jens,

Good suggestion, but consider this one:  I'm moving away from the AKSA, to the Lifeforce, which does not suffer this problem to any significant degree, and eventually hope to build amps for retail sale, plug 'n play!!

I've measured the problem on the 100W AKSA, and it's a 12V spike last 600mS at switch off.  To put this into perspective, a 12V DC voltage on a nominal 8R speaker driver with 6.8R DC resistance amounts to 21.2 watts.  However, this is only for 0.6 seconds, and in any event dissipation is averaged over the integral of the spike's amplitude, so it's actually closer to about two thirds that, viz 14 watts.

It happens that a voice coil is rated to a particular power, usually given in music watts, and typically 100W.  This is not the same as DC watts, around 2*root2 more, so a voice coil rated to 100W music would actually be able to dissipate only 35W DC continous.  Of course, this would push the cone to the extreme of its travel, severely stressing the suspension, but it would be electrically able to withstand 15.42 volts indefinitely.  For a period of 600mS, this power could probably be upped three or four times to close on 100W or more, which of course corresponds to spikes of 26 volts.

In a passive speaker, no DC can reach the midrange or the tweeter as they are DC blocked with a cap in the crossover.  The problem only arises with direct coupled drivers, such as the woofer and the ribbons of which we speak.  Woofers are much more robust mechanically than tweeters, because they are designed to move air and often have huge voice coils which can sink a lot of heat.  Tweeters are electrically almost as strong - the XT25G voice coil can sink up to 950 watts for 10mS and that's pretty impressive, probably around 100 watts for a second - but their suspensions are more delicate and this is, of course, the primary issue here.

The Lifeforce has a much diminished thump, tiny by comparison, because the input diff pair is operational almost from switch on as it's powered from a current source.

Jen, I hope this helps!!

Cheers,

Hugh

[andyr]

Hi Hugh,

Could you quantify the voltage swing and duration of the "thud" for each of your amps and list it?

I think this would give us something to go from. In reality, most quality tweeters and ribbons have a far higher tolerance vis-a-vis short-term effects like the turn-off thuds of the AKSA amps than most of us believe.

I wonder, though, if a simpler solution to protection would be simply to short output from the amp with a suitably sized resistor at turnoff (e.g. using a relay) - taking into consideration not to damage the output stage, of course. Hugh, what do you think? Also, to how great an extent would such an arrangement diminish the problem?

Hi Jens,

With respect, I think that even if Hugh published specs, some people would still either not believe him ... or not understand what they mean, vis-a-vis their perticklar tweeter!    So why should he bother?   

I repeat, I don't believe there is a more fragile tweeter than the 1.6m long Maggie ribbon ... Magnepan even give you a thin steel strip to put over the back of the ribbon "cage" for when you move the speakers around - as moving them suddenly has been known to break the ribbons!   

Yet my ribbons have abolutely no problem with the AKSA turnoff "thump".  AKSAs are much more benign than other amps I have encountered ... well, that is unless the LF is much worse in this respect ... Hugh??

Not only that, I have on one or two occasions, accidentally turned off the active XO without turning off the AKSAs first (too much red wine with dinner!!    ).  This is a horrendous thump - but they have still survived!     So, as you say, they are more tolerant of abuse than we give them credit for!

However, for those lily-livered AKSA owners, I think your relay concept is brilliant ... I wonder if it can be effected?

Regards,

Andy

[kyrill]

hi  dear people

What nice to see the well known names again from years ago : ) ( Jens, Andy..)
Is this your chat and dating site?  aa

I had a serious Hi Fi listening depression when a big fire cracker ruined  practically the ability of hearing the  low tones in my left ear.. I thought why border listening to the best when you are crippled anyway? So I jumped to video in a dedicated Home THeater and projector and missed more and more the emotions music can give you. Until i saw Leonard Cohen on "I am your man" and the different singers like Anthony. The music put tears in my eyes back again, so my brain didn't mind the sub par performance of my left ear.

I undusted my Aksa's and oh boy.
anyway when you get older you loose high freq  hearing anyway

I ruined a ribbon already but have a tweeter that you can easily replace the ribbon. Yes the magnepan ribbon material is much more flexible. The aluminium leave of 1/10 of a mm thickness ( my guess) made a swing of a cm every time when the amp was shut down and then after a while ( the other channel is still correct)it lost its ladder pressed form (structure)

I use the 8 ohm ( 8 R?) formula of Hugh, cross over is at 2 khz, but the peak current is seen by the cap as one hz?

IS it theoretically possible to use a coil parallel tot the tweeter to shortcut very low hz to "ground" wire of the speaker?

[andyr]

hi  dear people

What nice to see the well known names again from years ago : ) ( Jens, Andy..)
Is this your chat and dating site?  aa

I ruined a ribbon already but have a tweeter that you can easily replace the ribbon. Yes the magnepan ribbon material is much more flexible. The aluminium leave of 1/10 of a mm thickness ( my guess) made a swing of a cm every time when the amp was shut down and then after a while ( the other channel is still correct) it lost its ladder-pressed form (structure).

Hehe .. well of course, kyrill.  I'm an "audiophile", right (tho' I can't speak for Jens!  ) ... so I don't have any sort of social life - I just listen to my AKSAs and Maggies by myself at home!!     And spend a lot of time on the Net!!

Aaah, so you already ruined a ribbon from your Aurum Cantus G3s?     That's terrible - obviously it really is more delicate than the Maggie true-ribbon? 

Then you need go for Russell's cap in parallel with a SPST, high quality switch!     Unless Jens' idea of an output relay can be engineered.

Regards,

Andy

[Jens]

Hugh,

Thanks for your input. I was actually thinking of both the "traditional" AKSA amps and the Lifeforce amps. Sorry, I should have been more specific.

From what you're saying most tweeters will probably be able to withstand the AKSA "bump", whereas - as is evident from Kyrill's post - some ribbons may not be able to tolerate the "swing".

However, since there is much improvement on this point with the Lifeforce amps, the best suggestion is probably to go for a Lifeforce, if you want to drive a tweeter/ribbon directly

I'm sure that is to your liking, Sir! 

Kyrill:

Nice to have you back - and that you have found the "lost" music again The human brain is a wondrous thing, which can easily compensate for left-right ear differences. Sorry to hear you ruined a ribbon. However, I believe the Aurum Cantus ribbons are easy to get hold of, as they are available from many DIY speaker outlets. I can't remember the actual name of them, but I'm sure a little googling will pay off 

Andy:

I'm an audiophile,too, but I do have a social life besides audio 

And yes, it would be nice if a non-invasive solution for the "thump" problem could be found. Huuuugh?

[kyrill]

ha  the Lifeforce


Can i find the dynamics back to solder and assemble a lifeforce? Will it fit the same housing made for the Aksa's ? And will the lifeforce have a less "thumpy" "last breath just before it dies to be born again for a next time? because: " The Lifeforce has a much diminished thump, tiny by comparison"  There seems to be an antagonism in the sentence a tiny difference with the Aksa or tiny thump compared to others?

and yes i replaced the ribbon already, doable in 5 minutes

ah well in due time, in search of the Holy Grail, i will go back to the Lifeforce to have a more pure sensory organs perceived reality experience.

[kyrill]

Hugh

Assuming 1000Hz DEQX crossover, we have 1000000/12.6*1000 which is 79.36uF, npv would be 82uF.

IS this high value an electrolytic cap? On the net i mostly see low (<20uF) Mundorf capacitance unless it is a Mundorf Elco
Is there a web address who sells non Elko Mundorf caps with such a high capacitance?

is this a right cap? Condensateur Mundorf MCAP 400V. 47µF if cross over freq is not 1000hz but 2000hz?
http://www.audiophonics.be/condensateur-mundorf-mcap-400v-47uf-p-960.html?cPath=51_193_92
i see now how expensive it is this is the very moderate price series silver in oil is much more expensive

Hugh
n this instance, for 79.36uF (and the exact value is not at all critical), I would suggest 68uF of Solen, and 12uF in the quality cap.

I suppose i solder them parallel to each other? So parallel values adds for caps 

but what is the bypassing ? 90 degrees turn of the route back to the "ground" wire?  Wikipedia tells me only a by pass story in decoupling networks in power supplies for instance