Sound Quality Vs. Measurements

Status
Not open for further replies.
I just did the experiment I mentioned before . My OB speaker and a budget priced super amp . Rotel RA 931 ( fully EQ'ed ) . Superficially as good as the amp I am using . Valiant attempts at Rotel to be state of the art . It is awful .

So I used the Magneplanars + Rotel . Not bad , almost OK . Very boring .

Not content I tried my Dynaco A25's test speakers . What a surprise . Very good . The bass is the one note type , it tries to be articulate and almost gets there . By far the best sound of the group . Even treble detail excellent . This is from two dull sounding devices . The Dynaco's usually don't sound this good especially when high grade amps .

What this proves is low distortion and a good reputation mean nothing if it isn't harmonious . One paper my OB speakers don't need better than the Rotel . Also on paper the Rotel is better than a Quad 33 / 303 . In reality the Quad is leagues better . I did not expect this . The Quad is the one amp I never sold . Mostly it was my kitchen amp with Mission 761 speakers , sublime combo . I didn't give it a second thought . I hit hard times and was forced to use it as main amp . Gradually I learned how superior it is . It took me 22 years . The story of the amp is my brother repaired it for an old gentleman called Mr Haynes . My brother was a very careful man . He went through everything and did new for old caps and the best 2N3055's he could get . Staff at work wanted to buy it . I came to work and was asked " have your heard that amp your brother did " ? Eventually Mr Haynes died . His grief stricken wife insisted we buy it . I took it home as I couldn't sell it , it was grief and you don't sell stuff like that . Now even more so as my brother aged 48 died from the flu 3 years ago . I am only so sorry it took me so long to find out the truth . I build and design my own amps . Mostly what I build is better . However not so much that I give a dam .

For the gentleman who said about measurements . The Quad and the Rotel is not mysterious . You find the mismatch by trying music then find the answer . If you ask me the Quads regulated PSU is what I think I am hearing . A very low colouartion PSU . If you remove the regulator the amps becomes rather bouncy and full of zest ( careful because the voltage is a tad high ) . The regulated version sounds cool and honest . The latter helps the OB speakers . Theory will make you think the regulator is irreverent . Many caps is brute force regulation .

Good has come of this . I have to do a party for a 90 year old lady who was a Land Army Girl . Rotel and Dynaco's and a lap top . The Rotel is a repair so a good workout . It was to be the Maggie's and the Quad .
 
One small thing . The OB's are rather positioning friendly . I didn't think it likely . I think that the PA origins is helping . My old Pro Audio friend Nigel Copin said damping factors is wrong route to success . Speakers should be stiff suspension designs like in olden days . We thought this eccentric and fun . I think he has a point .

Nige C once said he connected a 9V battery to the feedback loop and it was the magic answer to our quest . I never did get to see what he did . His best friend a PHD in electrical engineering and I tried to work out what he had done . We never saw him again .

http://www.soleberry.net/P1010262_1.jpg
 
Last edited:
Yes when setup wrong , when setup correctly no, they add to the ambient sound-field creating a much better , more realistic soundstage and imaging possible from a point source speaker.

You just happen to have heard ,



1. A bad ESL speakers (plenty of those) or
2. A bad setup ( even more of these).


I'm not a big ESL fan of those using wide panel to reproduce HF, segmented like this is what i would do if doing an ESL ...

King Sound Electrostatic Loudspeakers - YouTube




:)

I've heard good and bad of both ESL's and setup.
Still couldn't get away from the room sound interfering and the 'headphone' position dependence...this is not unique to panel/OB's of course, .

LOL, That link you give has no sound ??? ...other YouTube videos are ok on my PC speakers.


Dan.
 
Last edited:
If you ask me the Quads regulated PSU is what I think I am hearing . A very low colouartion PSU . If you remove the regulator the amps becomes rather bouncy and full of zest ( careful because the voltage is a tad high ) . The regulated version sounds cool and honest . The latter helps the OB speakers . Theory will make you think the regulator is irreverent . Many caps is brute force regulation
Yes! The "rather bouncy and full of zest" will mean that difficult recordings will turn into a mess, become unlistenable to; the "cool and honest" should be capable of revealing all the inner detail of heavily layered recordings ...

Gee, isn't it great that conventional THD measurements and the like tell you all about this !! :rolleyes:, :p
 
Point is many learned books tend to suggest well designed amps do not need fancy power supplies . D Self almost insists that is so . I suspect this is because at idle no ripple is heard . Until clipping it is well hidden . Or is it ?

Build a class A amp . Suddenly you have problems . What Dvv and others are doing is fit a class A PSU to a class AB amp . Brute force regulation . As we head to 3 amps current my guess is 5000 uF ( 2 x 10 000 in series with centre 0 V ) will sound dirty .

Julian Vereker gave one word of warning with active regulators . The PSU transistor should be the fastest in the amp .

One could consider fitting a BDY 56 or whatever to the Quad 303 regulator . Unlike most amps one is enough . The amp is single rail 67VDC regulated . My experiments with output caps makes me think they are much maligned components . They offer superb protection against DC faults or signals that tend to DC . I doubt if the finest valve output transformer is superior . I have measurements of that and I can say that is definitely true . On the Quad I use a separate film cap to tweeters . That's a very cheap upgrade which you nearly always need regardless .

I am sure the various experts do not think a regulated PSU is a waste of time . It is something they gloss over . One simple thing to do is regulate the VAS to input . I usually have my outputs a few volts higher which is the opposite of text books . If so simple RC filtering can be enough from the raw DC of the dumpers .

The example of the Quad is exaggerated . As the amp is designed with a regulator it is no surprise it is not brute force before the series transistor . Even so no obvious hum and I suspect nothing jumping out in terms of measured difference . Chalk and cheese sound . Both are better than each other depending on what you want . The Quad version long term is nicer . I repaired for a chap called Rufus his Quad 303 . It had gone short in the series transistor ( CE ) . About 85 V from memory instead of 67 V . He was very upset as to him it sounded worse when repaired . The other fault condition gives about 23 V . Can't remember what causes that ( zener ? ) . Saw it a few times , perhaps 20 years ago now .
 
I recon the listening room and what's in it can make more of a difference than most people realise. I like a slightly lively acoustic but when i have guests round the system sounds flat. If I do listening comparisons the human content of the room definitely has an affect on sound quality. Most concert halls have reflections too which can add sparkle or make music sound hellish depending on many factors including how full it is. How would you go about measuring that never mind trying to reproduce it at home. I'm just glad I don't have that job.
 
Nige, a regulatred PSU is a double edged not sword, but a mammoth razor blade.

It can stabilize the amp and make it sound way better than without it, but it can also ruin it completely. The problem is making a stabilizer which will be up to the task.

The single transistor per line stabilizer I use for the IPS and VAS is just fine for just that purpose, where the currents usually end by 40 mA or so. But regulating Amperes, and many of them, is a whole new ball game.

In my limited experience, the regulator must actually be another power amp, with its own equivalents of the IPS, VAS, driver and output stage. And I agree it must be faster than the amp it is serving, which is not very easy when your amp slews at 200 V/uS.

I haven't tried it, but I suspect MOSFETS would do well here.
 
Last edited:
PSU colouration .

As the time machine goes backwards we get amplifiers that need brute force power supplies to over come their weakness of ripple hum rejection . Sometimes a choke was used in CLC Pi filter . The Quad 303 a series regulator .

The series transistors makes sense . It should be doing less work than the outputs . If a bridged output one can even dispense with the output cap if a crowbar is used to protect the speakers . This allows one transistor to serve one complete amp . Naturally that isn't set in stone that it must be only one . If two there might be slight separation advantages . Perhaps the measurements would look like monoblocs with two .

B41456B9229M - EPCOS - CAP, ALU ELEC, 22000UF, 100V, | Farnell UK

I came to the conclusion all amps should start from a cap like this . A pair for the usual modern amp . This allows the better 0V connection if + 0 - . As soon as multiple caps are used the risk is 0 V is goodness knows where . 17 amps should be OK for a big amp . If we have the series regulator we can build in a current limiter . Contrary to popular opinion they make an amp sound better . How come ? They tend to sound like class A amps coming to the limit . The usual current limiting seen on amps doesn't seem to do anything except ruin the sound . I can always brake amps with that . The simple way to do a current limiter is a series resistor before the series transistor . What looks like a constant current source also is excellent . I was told it isn't called that if not asked for full current .

If preferred raw DC can be fed to the dumpers . In theory that should be fine . If the voltage higher than necessary the dumpers act like a series regulator until clipping . If over volts that should not happen . Also if VAS clips first that looks very nice on the scope .

So the proposal is 90 V into 22 000 uF . One carefully chosen series transistor ( DMOS ? ) . Current limiter . Output voltage might be 80 V . Amp power 75 W rms 8 R . In bridge 300 W rms 8 R . Low R will be fine , it will softly limit if asking too much .

Listening to my set up last night I felt I have it right and all other hi fi is daftly wrong . Putting speakers in boxes is bloody daft . Example , never in my life did I hear so many bass notes . Most music has none except a thump . No worries a thump will do nicely . Organ music is unbelievable . Every dam pipe added adds , you hear each and the blending . They rasp at LF ! Asked if better than my Maggie's I said yes . I wouldn't dare run them so hard . I have a feeling it requires a better man than me to brakes these . Quad 63's also . I was offered a pair for $0 . They wouldn't last long and not convinced they are better even where they should be . I would say what I have has better voice quality . That is perhaps as material is mixed on speakers more like mine .

Have listened to many efficient speakers . 105 dB / watt Klipsche . 115 dB / watt Western Electrics ( circa 1940 , pair ) . All left me perplexed . In both cases a big amp tried and the usual 2A3 amps . I was completely underwhelmed . The Spruce Goose of speakers .

Now I know why . The box stopped any magic . In my case I am using 1 watt max at 1 kHz . I would defy anyone here not to be concerned about ear damage if wanting that to be the preferred volume . At 50 Hz I will be using 6.3 W . I peak the EQ at 30 Hz + 16 dB and roll at 6 dB below 20 Hz . What I have is massive headroom on an amp even Quad knew to be puny . BTW valve fans it's just as good . Would you believe the distortion is about 0.03 % just before clipping including preamp . Preamp - 3dB 200 kHz . Distortion is in exponential decay starting at - 80 dB second harmonic at 5 watts . There is crossover distortion . It simply masks detail . One has to know very good amps to know that . The circuit looks like a less sophisticated 1965 transistor radio . Devices struggle to cost 5 cents . The 33 will clip at the drop of a hat . You must avoid that .

EQed speakers usually sound rubbish . Now I know why . Long before the phase shifts resolve the box spoke out . It is like a game of tennis where someone is at the net ruining the game with an extra racket . If you like they are at net level catching the odd ball or bass note .

I listened to the Dynaco's carefully . I bought them recently for $40 and they are pristine . I had these in 1974 and thought them the only speaker that could play Sibelius . Moody speakers . Well the aperiodic loading does work . They hint OB . Some base just isn't there and then a mighty low note comes out . I would describe them as having a lazy man at the tennis net . They force sitting centrally to image and sound spacious . They are not a good friend of the Quad , they like the minnow Rotel . That is bonkers as both usually sound muffled via things they don't suit . Together they are not .

The proposal is 2 x aperiodic sub woofers . 15 inch basses with Qts of 1.2 . The aperiodic bit a big port full of drinking straws . The bass will work 15 to 50 Hz .

I solved using polystyrene baffle . Minimal sizes plywood square mini bafflers front and back . Hole cut with hot wire . The front and rear clamped in compression . Ply champhered to work better and look nice . Edges of the PS in box frame of maple . 5 degree back tilt .

Now to the crunch . I visit many houses . The ubiquitous Krell is shown to me by man with a jealous look on his face that I should have ( jealous of his own good fortune ) . Sounds is like a strangled NAD 3020 . I have heard the same at Jimmy Hugues house sound stunning ( Level 17 , Crushed by the wheels of industry on 12 inch EP ) . How do you hi fi dealers live with yourselves to allow yourselves to take the money and run ? It is not customer choice as not concert they went to sounded like that . It would be you went to a restaurant and they gave you tooth paste to eat .

How did we get so far off track ? Simple . We never had that place down the garden where we could make music and stuff anyone else . 2 x 4 feet x 2 . Is it so difficult to think we all should have that ? I am divorced yet she won't leave the house . It isn't the usual . I am the one who still sorts her life out and she needs that . Her face when she saw the speakers . Not a leg to stand on . I am always helping out with the finances . All I say is these might be money one day . Like hell they will , we all know that .
 
Nige, a regulatred PSU is a double edged not sword, but a mammoth razor blade.

It can stabilize the amp and make it sound way better than without it, but it can also ruin it completely. The problem is making a stabilizer which will be up to the task.

The single transistor per line stabilizer I use for the IPS and VAS is just fine for just that purpose, where the currents usually end by 40 mA or so. But regulating Amperes, and many of them, is a whole new ball game.

In my limited experience, the regulator must actually be another power amp, with its own equivalents of the IPS, VAS, driver and output stage. And I agree it must be faster than the amp it is serving, which is not very easy when your amp slews at 200 V/uS.

I haven't tried it, but I suspect MOSFETS would do well here.

The thing to grasp is this must work . There are bigger problems down stream . It is a leap of faith to say what the dumpers didn't do a series transistor will . The big advantage is the hum can be miniscule . I bet you a substantial amount of money most people especially me will not realize these advantages . A skillful person will as he gets 0 V nailed . I sometimes see exquisite star earth systems . I am just too lazy to do that . I play about until my analyzer gives me the nicer rubbish then give up .

Something I feel to be true . Negative feedback is mostly a thing to control speakers . Bung in some 10R emitter resistors and see how it compensates if you doubt this . Thus if current is enough in the PSU it's going to work .

Your oversize rail compensation caps seem very OK . For you a regulated VAS with under voltage option might be a free lunch . A shunt regulator if you must if that . I am concerned that no one challenges shunt regulators . With a bit of care series seems to work better . Fast transistors are not rare especially when cheap MOS . Second harmonic distortion does subtly change when a series device used . It is usually for the better . MOS in series usually doesn't oscillate , why should it .

You might ask why I use the Quad . That might be like asking why someone uses a classic Mercedes ? It reminds him of how he might progress in life to be a real world professional . BBC asked Quad how the amps were made so cheaply ( the retail price 1/4 the realistic price said BBC ) ? Quad answered 25 000 units must be sold before a profit is made . Quad made the mistake of saying it to be the worlds best amp as no one on the planet can truly tell if something is better . It cost them many friends and yet was mostly true . If one doesn't have the speakers to prove that it is not Quads fault . Quad didn't if asking in the ESL 57 .

Hi fi is like cars without aerodynamics . Just keep designing bigger engines . Trouble is the car is wrong in other ways then . If down force that's acceptable
 
Last edited:
Sound Quality Vs Measurements....ok, it's too late to do any measurements.
250px-HomeEarthRodAustralia1.jpg
Last night, I went outside with a torch and a bucket of detergent solution, found my house MEN earth stake and flooded the area with the detergent solution, and then ran another 100L or so of tap water to soak the detergent into the pure sand ground.

My stereo now sounds nicely cleaner and clearer.


Dan.
 
Last edited:
...

I came to the conclusion all amps should start from a cap like this . A pair for the usual modern amp . This allows the better 0V connection if + 0 - . As soon as multiple caps are used the risk is 0 V is goodness knows where . 17 amps should be OK for a big amp . If we have the series regulator we can build in a current limiter . Contrary to popular opinion they make an amp sound better . How come ? They tend to sound like class A amps coming to the limit . The usual current limiting seen on amps doesn't seem to do anything except ruin the sound . I can always brake amps with that . The simple way to do a current limiter is a series resistor before the series transistor . What looks like a constant current source also is excellent . I was told it isn't called that if not asked for full current .

...

I beg to differ.

Large caps do have smaller output impedances, but then, two halves of the big one will have an even smaller output impedance and larger available current reserve.

But that is secondary, as the benfits are relatively small. What IS the issue is that large caps do not filter well the upper mid and high ranges. Because of this, we need paralleled caps of smaller values, in my case 100 uF // 3.3 uF // 0.1 uF to take care of the higher range. I have tried it so many times that there is no doubt left at all in my mind that this is the way to go.

Nige, getting this right I think is on a case by case basis, I don't think there's any niversal formula one can apply as is. You just have to try and listen, over and over again until you get it just right.
 
Sound Quality Vs Measurements....ok, it's too late to do any measurements.
View attachment 423437
Last night, I went outside with a torch and a bucket of detergent solution, found my house MEN earth stake and flooded the area with the detergent solution, and then ran another 100L or so of tap water to soak the detergent into the pure sand ground.

My stereo now sounds nicely cleaner and clearer.


Dan.

You surely did clean up your sound, Dan. :D :D :D
 
Dvv , Point is your 4700 uF decouplers more than do that . Add a 0.1 uF PTFE ?
rose bush next to ground spike is good . Rose bush dies = bad sound .

Never mix PME with TT grounding if you live in a street . Under fault condition may be seeing > 200 A . If allowing for that it will be a very good ground .
 
Last edited:
The good news is no negatives when using successful OB designs to compare . 18 dB is fine as there is tons of headroom from amp and driver . I think because I have abandoned getting better than Quad 57's I stand a chance .

The subs will be double and be aperiodic ( not port in function , not IB , not OB. Port - IB compromise ) . I will square wave tune them . Again if needing weird EQ they will get it . Time alinement is a problem . Even with my cavalier attitude looks matter . Under the OB would be ideal . Would firing up behind the OB stand a chance ( seems wrong ) ? 15 to 50 Hz is about what they will be . If they need 400 watts they will get 400 watts .

Gilbert Briggs said 2 tons is the mass of a cabinet with OB performance . As simply as he could say- 60 dB cab re-radiation . That's even ignoring bounce back through the cones .

I get the feeling a 10 inch unit would be better if to the same recipe . Listening to Cliff Richards Living Doll for the 90 year old Land Army Girl I was stunned . I am not his No 1 fan , well maybe that has changed . Also Bill Hayley When the saints . It was like he came to my old youth club and performed live . Little imperfection that suggested an unpolished version . Jazz on Radio 3 form Cheltenham . Wow . Oh how I wish I still had Charlie Hayden " For a free Portugal " on A&M Horizon .

What is stupid is I dreamed this up 30 years ago and just talked about it to anyone who would listen . What a fool , I was right and the maths do work . Robin Marshall told me the bones of it . He said I mistook the Quad sound for being electrostatic .

http://www.hifizine.com/2011/09/prototyping-dipole-bass-system/
http://www.usspeaker.com/BETA12LT-1.htm
 
Last edited:
Dvv , Point is your 4700 uF decouplers more than do that . Add a 0.1 uF PTFE ?
rose bush next to ground spike is good . Rose bush dies = bad sound .

Never mix PME with TT grounding if you live in a street . Under fault condition may be seeing > 200 A . If allowing for that it will be a very good ground .

What's with the abbreviations?

Nige, I have no idea what you are talking about - what's "PTFE", "PME", etc?
 
Status
Not open for further replies.