[rick57]

Hi Hugh

What are the sound-quality optimal transformer VA values for each Aksa amp, either (yet to be decided) dual mono or stereo?

Thanks!

[andyr]

Hi Hugh

What are the sound-quality optimal transformer VA values for each Aksa amp, either (yet to be decided) dual mono or stereo?

Thanks!

Rick,

If you're bothering to build an AKSA - rather than a cheap Jaycar "special" (because you appreciate a high-quality sound) then you build them dual mono - ie. one power transformer per power supply per amp module.

Hugh's away for Easter right at the present time so you might have to wait a few days to get the definitive answer but, as I understand it, he recommends a power trannie of VA = approx 3 x the AKSA output.  IE.:
* AKSA 100 ... 300VA
* AKSA 55 ... 160VA
* AKSA 25 ... 80VA.

At least, this is what I use on my AKSAs.  If you build, like PSP did, a special-purpose AKSA 100 with lower DC rails and a beefed-up PS (which enables it to cope with a 2ohm load and deliver much more current) then you would use a 500VA transformer.

Regards,

Andy

[rick57]

Hi Andy,
 
Yes, I almost certainly will go dual mono.

Hugh's away? Thanks, I may have been puzzled at his lack of response.

Thanks for your info - the VAs you quoted, are for dual mono?
Do you know the approx difference cost of the Aksa 55/ 40/ 25?

Where the Phoenix mids in parallel are 3.5 ohms, what would be the sonic benefit of PSP’s higher current-100 watter with lower DC rails and beefed-up PS, over a standard 100 watter,  or as I was tending towards, a 55 watter?

Regards,

Rick

[andyr]

Hi Andy,
 
Yes, I almost certainly will go dual mono.

Hugh's away? Thanks, I may have been puzzled at his lack of response.

Thanks for your info - the VAs you quoted, are for dual mono?
Do you know the approx difference cost of the Aksa 55/ 40/ 25?

Where the Phoenix mids in parallel are 3.5 ohms, what would be the sonic benefit of PSP’s higher current-100 watter with lower DC rails and beefed-up PS, over a standard 100 watter,  or as I was tending towards, a 55 watter?

Regards,

Rick

Hi Rick,

As far as Hugh is concerned, there is no difference in the cost of the 55/40/25 - as the only difference between the two is a resistor or two, to adjust for the different DC rail voltages.  You may save a few $ as you go down in watts/DC rail voltage, because of the trannie prices.

Yes the VAs I quoted are for dual mono (the only way to fly IMO!!    ).

Now, you asked: "Where the Phoenix mids in parallel are 3.5 ohms, what would be the sonic benefit of PSP’s higher current-100 watter with lower DC rails and beefed-up PS, over a standard 100 watter, or as I was tending towards, a 55 watter?"

Mmmm, tricky Qu ... I think you need to ask "Tinker" who has Phoenixes, or Hugh himself.

Is that the main difference between the Phoenix and the Orion ... the Phoenix had two mids?  I personally would think a "normal" 55N+ would be fine driving 2 mids in parallel, if this is a 3.5 ohm load.  The 100 watter would possibly have too much power and might overdrive the mids?

The reason PSP chose a 100 with lowered DC rails (making it approx a 90 watter) is that this enabled the amp to cope with a 2.5 ohm load (2 Peerless woofers in parallel).

Regards,

Andy

[rick57]

Andy

> the main difference between the Phoenix and the Orion ... the Phoenix had two mids?

Yes the Phoenix has two SS 8” midwoofers (21W/ 8554), a 3.5 ohm load. The Orion has a slightly higher steeper XO and a single Seas Excel 8” midwoofer.

Yes, lower DC rails should get down to a 2.5 ohm load.

> I personally would think a "normal" 55N+ would be fine driving 2 mids in parallel, if this is. The 100 watter would possibly have too much power and might overdrive the mids?

The 21W/ 8554 is rated 100 watts, so should be ok.


It may sound a little more dynamic with 100 watts??


Regards,

[PSP]

Hi Rick,
Regarding amp selection and your choice of rail voltage, much depends on what you like to hear, your budget, and perhaps your long term audio plans.

For example if you "someday" plan to have a second system in the bedroom, you might use a 55N+ for the mids with the intention that you would eventually move the 55N+ to the bedroom system and drive the mids with a 100N+.  Since you live in Melbourne, it may be possible for you to someday audition a 100N+ in your system to see if it adds enough benefit.  

My room is large, my current speakers (GR-Research Paradox 3) are 90db efficient, and I like large orchestral works at 75-85(max) db.  With this setup and my tastes, I like the 100N+ much better than the 55N+.  I truly don't hear any difference in resolution, high end extension, etc. between the two, but the 55N+ can get a bit conjested when the music gets complex and heavy while the 100N+ is still sounding clear.  The 100N+ is also a lot more dynamic.

For the Orion (which is under construction:  amps done, next build the ASP, then the speakers), I will use a 100N+ with 43v rails (and beefed up power supply) to drive paralleled XLS woofers.  For the mid, I will try 55N+ and a 100N+ (with standard 49v rails) to see if the 100 adds dynamics, and the tweet will be driven by a 55N+.  At some point, I might try dropping the tweeter amp to 25w, but only after listening to the system for a long while.

With my current system (90db MTM floorstanders with passive crossovers in a large room) I can hear a definite advantage for the 100w amp over the 55.  With the Orion (or your Phoenix) we will have:
 25 or 55 wpc in the tweeter
 55 or 100 wpc in the mids
 100 wpc in the bass.

The drivers are connected directly to the amps, so we can probably double the effective power... so I will have roughly 4x more power availble.  By this argument, I will be surprised if I find that the 100N+ has a significant advantage over the 55N+ for the mids.  

Currently, I have the 100N+ with 43v rails in my system.  It sounds at least as good as the stock 100N+ with 49v rails.  It seems to me that there is a slight but noticiable advantage in the bass, more authority.  I have my GK-1 on the bench right now (doing a bunch of ginger's mods, etc.)  When I get the GK-1 back in the system, I will investigate the bass performance of the 43v amp more thoroughly and report back.

I wish I could be more definitive, but I have to aportion my listening time to (a) listening while doing work, (b) listening while enjoying music, and (c) listening while being an audiophile and experimenting with the system.  Makes for slow progress...

Peter

[MikeC]

Rick

My AKSA configuration, specifically for Orions is:
Tweeter - AKSA 21 N+ with 75VA laminated transformer
Midrange - AKSA 55N+ with 160VA toroid
Bass - AKSA 100N+ with 43V rails and 500VA toroid

Hugh is very willing to discuss tweaking of the design for different applications, which is originally how the low power version came in to being, for high efficiency systems. I'm not sure who first brought up the Orion bass optimised AKSA 100, me or Peter (Hugh will have to adjudicate   ), but Hugh was very willing to discuss the concept and offer advice. I also have several small tweaks to specifically optimise each amp for its operating range, some of which were suggested by Hugh. He gets personally involved with each of his customers.

Peter

I was under the impression you were up and running with your Orions - I may just beat you. My ASP is in its case and tested, all three amp sections are complete and the cases are nearly finished, and the Orions are awaiting final assembly (the veneering of my side panels was rejected 3 times   ). The only hiccup is battling to get one channel of my Welborne Labs attenuator working, otherwise my GK-1 (also modded) is also ready for testing. I can't wait, it has been way too long without a system.

Mike

[PSP]

Mike,
You will very definitely beat me... I have a spare preamp (an old Foreplay) that frees the TLP for duty in my main system while I work on the GK-1.  Also, since I built amps to support several multi-amping scenarios that I later abandoned, I have spare amps (that will all get used up in the Orion and  bedroom systems).  So... this allows me to listen to music (sinful sloth!) instead of soldering and banging on wood!  

I am eager to hear your impressions when you can finally throw the switches and hit "Play".  GK-1 + AKSA amps + Orion will be one extreme system    !

Good luck and sweet music,
Peter

[andyr]

... When I get the GK-1 back in the system, I will investigate the bass performance of the 43v amp more thoroughly and report back. ...

Please do, Peter ... I eagerly await your findings!!

The reason is my Maggies are known to lurve current!     In my 3-way active system, I use an AKSA 100 module for each bass panel but I've always wondered whether doing the "43v" mod (so the amp can deliver much more current) might make Maggie sound better?    

Regards,

Andy

[rick57]

Guys

I’m likely to in the 1st instance run my Phoenixes “hybrid” :

Tweeter - AKSA 21 N+ or 40N+, laminated transformer
Mids - paralleled, AKSA 55N+ maybe with amorphous cores  
Bass – paralleled, home theatre Rotel 200 w or Marantz 140 w

Then yes I should try other AKSAs and compare.

Is the voltage gain the same on all AKSAs, as Linkwitz says the XO allowed for, if so what is that gain?

Cheers

[AKSA]

Rick,

Gain is 38.3, or 31.6dB, on ALL AKSAs, regardless of upgrade level.

I generally go for three times the power in the trafo per channel over the amp;  so a 55W gets 160VA, a 100W gets 300VA, a 25W gets 80VA.

This is not hard and fast, and is based on the charge pulse to the filter caps on each mains cycle.  It's pretty high, so to avoid loss of rail voltage, a larger trafo is a good idea.

For similar reasons, too much filter capacitance will overstress the trick diodes because the charge pulse is much higher.  The size of this charge pulse can interact with the voice coil current return;  too much capacitance is not always a good thing.....

Cheers,

Hugh

[rick57]

Thanks Hugh

That makes things easier

Cheer   s

[bhobba]

Hi Rick,
Regarding amp selection and your choice of rail voltage, much depends on what you like to hear, your budget, and perhaps your long term audio plans.

Absolutely.  Your best course of action is to crank up your amp to the maximum you would listen at your listening position then grab an SPL meter and find what the speakers are outputting at 1M.   When I did this I found it was about 80db or each speaker doing about 77 db - at my listening position it was about 73 db - a person speaking abut 1M away is 40 - 60 db so this level is actually louder than natural for the vocal music I like.  Now we need to add headroom.  For normal music it is about 15db extra.  Orchestras can get up to 25db occasionally = (orchestral peaks can be 130 db in the pit or close up - a lot less at a normal listening position - the max is more likely 100 - 110 db there - but reverb I hear can do strange things.).  So unless you listen to orchestral music allowing 15 db on top of that is fine.  This gives you a max 77 + 15 or 92db.  Then you need to take your speakers into account - my Axis LS88's are 95db efficient  - with an 8 ohm watt.  This means with them even a 1 watt amp is good enough for me - so the 25w AKSA would do fine.  However listen to orchestral music and you need an extra 10db headroom - or 10w - so 25w is still fine for me.  However raise my maximum listening level to 80db (I would not go higher than 85db - prolonged exposure to that can cause hearing damage) and I will need the 55w even with 95 db efficient speakers.

However I have recently purchased SP Tech Time-pieces which are 85 db efficient and whacko - you need 550 W if I wanted to go up to 80 DB with orchestral music - oh dear.  Fortunately I never listen at this level - 73 db is about my max so for me with orchestral music I need 102db max which means a 55 w is fine.  In fact I rarely if ever listen to orchestral music so even a 25W would be good enough.  

What does this tell us?  If you listen at lowish (say 75 db or lower at your listening position) volume levels to anything but orchestral music then you can get by with a low powered amp such as the AKSA 25, which reputably has wonderful sonics, even on relatively inefficient speakers.  Again keeping inefficient speakers if you listen to orchestral music at low volumes then the 55w would be fine.  Listen at 80 DB or more then you will run into problems even with the 100 W Aksa.  Your only choice then is more efficient speakers or a BIG ASS AMP other than the AKSA.

It is interesting to note that while rock music can be very loud - reaching average levels of 110 db or more - it is not that dynamic - requiring a puny 10 DB headroom.  So if you wanted to reproduce that - well your out of luck - but you would be anyway because there is no way your neighbors would let you do it - and even if they did watch out for hearing damage.  As I said I would never exceed 85 db at my listening position to avoid hearing damage - so for rock music this would be 89 db from my speakers or about 100 db with head room - so a 55w would be fine with 85db efficient speakers even for rock music - if you listen at sane levels.  It is orchestral music that actually puts the most strain on a amplifier power due to its high dynamic range - and even then only occasionally.

Thanks
Bill

[andyr]

Absolutely.  Your best course of action is to crank up your amp to the maximum you would listen at your listening position then grab an SPL meter and find what the speakers are outputting at 1M.   When I did this I found it was about 80db or each speaker doing about 77 db - at my listening position it was about 73 db - a person speaking abut 1M away is 40 - 60 db so this level is actually louder than natural for the vocal music I like.  Now we need to add headroom.  For normal music it is about 15db extra.  ...

Mmmm, Bill,

I suggest all your measuring with an SPL meter relies on the transient being slow enough for the meter to respond to the transient?

Yes you can do the maths and perhaps, as you say, 15dB headroom is sufficient - or 25dB if you like orchestral music ... but I wonder?     I suspect there could be transients lasting, say, a hundred milliseconds which require much more headroom thyan that, to deliver the leading edge faithfully ... and I suspect the SPL meter won't be able to pick this up.

So you should really add another 10dB, minimum (ie. a total of 35-40dB) for the required headroom, before you can be certain your amp is not clipping.

IE. a 55w AKSA will NOT deliver you clip-less rock music with those 85dB-efficient speakers.

Of course, if you have active speakers then you can use multiple AKSAs to overcome their inefficiency - my Maggie IIIAs are specced (by Magnepan) as 85dB efficient and I run them with a 100, a 55 and a 25 ... which delivers the approximate equivalent of 750wpc (into 4 ohms), if you use the simple "doubling of power for each amp, in an active setup" rule.

I would hate to run them with a single AKSA 100 (218wpc into 4 ohms) and they would give out, on an AKSA 55!  

Regards,

Andy

[bhobba]

So you should really add another 10dB, minimum (ie. a total of 35-40dB) for the required headroom, before you can be certain your amp is not clipping.

In real life it is possible for very short lived transients to occur over 25db more than the average - even in rock music.  Although for rock music that is unlikely since it is mostly amplified electronics that is pre compressed to be played over speakers at a concert
http://en.wikipedia.org/wiki/Amplitude_compression
'This reduction of the dynamics makes the drum sound louder even though its peak may be at the same dynamic level it was before compression and make up gain were applied. It is common to compress rock music or pop music heavily to make it sound louder without introducing undesirable distortion that would be introduced by exceeding the dynamic limits of a reproduction device or medium.'
and
'Vocal performances in rock music or pop music are usually compressed in order to make them stand out from the surrounding instruments'

Regrettably over the last several years the dynamic range of Rock music has been compressed to an average of 10 db:
http://en.wikipedia.org/wiki/Loudness_war
'And while some may debate the severity of the loudness war during its humble beginnings in the 1990s, there is little doubt in the minds of the vast majority of audio enthusiasts that nearly all rock and pop CDs released on large corporate labels (and not just RIAA-owned ones; independent metal label Century Media has been stated as being a major offender, for example) in the 21st century are simply unacceptable sound-wise, with −10 dB being the standard for the past several years, very often being pushed to −9 dB (and on some occasions, even a dB or two louder!). Exceptions to today's hot standards are practically non-existent at this point, and the chances of the situation reversing are very slim due to the fact that knowledge of and concern over the loudness war is still mostly exclusive to audio enthusiasts and individuals within the audio field, demographics who make up an extremely small portion of music buyers.'

This is of course very sad from a audiophile point of view but still true.  It even has filtered down to pop music.

Over and above this for other music 18 db is the usual headroom used by most engineers:
http://en.wikipedia.org/wiki/Headroom.

Again, regrettably, it is only specialty recordings by Chesky etc that may go up to 24 db for orchestral works - although SACD and DVD audio recordings may change this in time.  Personally though I do not believe SACD and DVD audio are the future for audiophile reproduction.  I believe that is 24/196K files prepared from master tapes using lossless compression sold over the internet by specialist companies to discerning audiophiles.  But until that happens I stand by my numbers.

But being audiophiles let us assume we will eventually get recordings with 24db headroom as standard - (no one I am aware of is proposing greater headroom than this even though is real life it can occur).  So for 85db speakers at say 8-9m listening position and listening at an average of 80 db we need to have a max of 104 db at the listening position or 114 db max at 1 m from the speakers or 111 db produced by each speaker.  This equates to an amp that can provide 500 w per channel.  Listen at levels like I do a bit closer to the speakers (70 db at about 4-5M) and you get by with the 55W - actually the 25W would be good enough.  The moral - listen at low levels to get full detail, get a BIG ASS AMP, efficient speakers, or, as you correctly point out, invest in active crossovers of some sort.  My preferred solution - listen at lowish volume levels.

Thanks
Bill

[bhobba]

Over and above this for other music 18 db is the usual headroom used by most engineers:

Whoops guys that link should be:
http://en.wikipedia.org/wiki/Headroom
As Andy correctly points out real life sounds need 35 db headroom.  But this is reduced to 30db at the mic and 24 db at the mastering tape then to 18 db at the CD.  So if you listen to CD's 18db headroom in the amp is enough.

In the future with 24db available then unless one wants to go for really powerfull amps actively amped speakers would seem the way to go - as Andy correctly points out.  Either that or simply reduce the volume you listen at - 70 - 75 db at my listening postion is more than enough for me.  Speaking loudly or singing one meter away is only about 60 db.

I suspect to fully recreate dynamics we will need advances in technology such as 32 bit digital sources, super mikes and cheap powerfull audiophile quality digitial amps.

Thanks
Bill