Dejan, we're definitely on the same page, the main disagreement I think is that I believe that all recordings can be rescued - this is not theory, but personal experience; and a very satisfying goal. As regards others' systems I have zero interest in mouthing platitudes about how it sounds, these days if I hear a problem I do mention it in as diplomatic a way as I can, because my real interest, always, is to improve upon the now, the buzz for me is always achieving a higher quality than what currently is.
The type of noise is everything, I believe. Noise correlated to the event, which then is called distortion, is the real nasty in my experience - it "annoys" the brain, causes fatigue. The more random, or unconnected to the musical event is the unwanted sound component the more relaxed the brain is in digesting what it hears ... this has been my experience, and others around who have heard the same, "better" quality have confirmed that it is the preferred.
Ah, Enya! My soft spot.
I expect you now understand better why I, when I am approaching the point of explosion, put on Enya to get me back to sanity.
Sometimes, even Nana Mouskuri does the trick. Or Loreena McKennitt.
What's odd about that is that in general I do not listen to many female vocals, not do I like female vocals. But the above ladies do the trick for me, no idea how or why, nor do I particularly care. I just enjoy it.
Oh yes, I forgot about that, that's how important it is to me.
Even so, we don't have a fundamental disagreement, rather we differ in degree. You say rescued, I say a good system will make a hitherto poor sounding recording listenable, serviceable if you like. But it cannot make a poor recording suddenly sound good, in my view, just listenable.
On the other hand, a good system will turn what was thought to be at best a so-so recording into an actually good one, it may present details not heard before on lesser systems.
And on what were good recording beofore, well, now they may turn out to be not just good, but excellent.
But remember, Frank, all these definitions and descriptions are still subject to our perception of sound, and are therefore in part subjective. As noted before, love the recorded material well enough and many possible deficiencies disappear anyway for the benefit of the whole and the essence.
This is, I believe, where most audiophiles fail; they want what they percieve as sound perfection, and if they don't get that they discard the whole. Which to me means that they are in fact listening to the system, not the music.
I was fortunate to be free from that defect right off the bat. It was always music I was after, not some imaginary perfection in playing that music. And I was lucky on several occasions.
Best example - my old Dual CS 604 TT. That was a wildcard both for Dual and me, bad for them, great for me. Within a month of its introduction, the German audio public discovered that it was a few percent below the their top model(CS 701 and CS 704), but at around half the price. So, a waiting list appeared, its sales rocketed, but the sales of the bigger model dropped significantly. Dual concluded that they had just shot themselves in the foot, and quickly (within some 8-9 months, never so short a period before or after) they replaced it with model CS 606, which returned things into their natural order. I jumped on board just in time to get it.
My comprehension was the second part of the jigsaw puzzle. I realized that this model was very probably made in some form of collaboration with Ortofon, who was pushing low mass and high compliance in full force, both by their cartridges (the Concorde series, OM/LM 10/20/30, their AS212 tonearm started turning up on Telefunken TT models, etc), so I quickly replaced their factory installed Dual (actually some model by Audio Technica) cartridge with an Ortofon LM 20. A match made in Heaven. I have never heard any other TT of the day play so well with that cartridge, nor have I heard that cartridge used to such good effect with anything else, except the Telefunken models with Ortofon's own tonearm. Both the Dual and the LM20 were squarely mid priced items at the time, yet together they delivered more than the simple sum. Which is why I stuck with that combo to this day, despite the escalating prices of replacement stylii, currently at €260 from Ortofon (which puts it in the upper middle class these days).
This caused a friend, who owns a Thorens 126 Mk.2 base with an SME 3009 Mk.3 tonearm and an expensive Audio Technica cartrige, and listens to it via the same preamp-amp combo from Marantz as I use that he cannot hear any differences worth mentioning, at an overall price difference of twice the price of my setup.
Listening to West Side Story on a charity shop £1 LP. Scratches etc . What a sound ! The mods to the Quad 33 works. It almost oversteps the line. Open and sweet, deep, tight, fast, real, addictive.
When I sold Quad people would not buy it as 6 to 8 volume out of 10 was typical. The Quad volume controls were not great below 5. My instinct is this was done to protect the ESL57 speakers. Even the 303 is a bit too much for them. People felt they must be trashing the amp to use it this way. I use the Quad at 8 out of 10 always into a pure 4 ohms load. It gets hot and the very soft current limiting can be heard. Like a slightly wornout KT88 valve. KT88 is not as nice to my ears. 405 is OK but not my cup of tea. Sounds much better as a monobloc ( ask ).
The other thing about the 33 is the H setting of the tape board is just about a pure signal path ( CD or PVR + Internet when me ). I use a Quad tuner so happy to use the Quad input level. The radio inputs are next to useless for most other things. They need modifying for best results. If you have no use for phono the S1 position of the disc board can be populated for 300 mV use. It can be returned to phone use in seconds if wanted.
Denon DL110 LP12 ( glued chassis ) Ekos, 33/303 SMGa sounds stunning. The weird thing is pushing the gain up the curve has tiightened it. I can slightly hear compression, it is just on the right side.
If anyone wants to know this mod just ask. It is a 10 minute job. The DL 110 is a remarkable device. 90% will never know as it falls between typical MC and MM levels. It is a genuine MC. I paid £135 for mine new. It is a minute to midnight to do this. Also alcohol and DL110 are possible at that price.
Quad by being scientific excluded many. At best people said " not for me, I am not that person ". This dam amplifer is the real deal. My system will play any music and does. Glad it has no more watts. I would use them and it would end in tears.
When I sold Quad people would not buy it as 6 to 8 volume out of 10 was typical. The Quad volume controls were not great below 5. My instinct is this was done to protect the ESL57 speakers. Even the 303 is a bit too much for them. People felt they must be trashing the amp to use it this way. I use the Quad at 8 out of 10 always into a pure 4 ohms load. It gets hot and the very soft current limiting can be heard. Like a slightly wornout KT88 valve. KT88 is not as nice to my ears. 405 is OK but not my cup of tea. Sounds much better as a monobloc ( ask ).
The other thing about the 33 is the H setting of the tape board is just about a pure signal path ( CD or PVR + Internet when me ). I use a Quad tuner so happy to use the Quad input level. The radio inputs are next to useless for most other things. They need modifying for best results. If you have no use for phono the S1 position of the disc board can be populated for 300 mV use. It can be returned to phone use in seconds if wanted.
Denon DL110 LP12 ( glued chassis ) Ekos, 33/303 SMGa sounds stunning. The weird thing is pushing the gain up the curve has tiightened it. I can slightly hear compression, it is just on the right side.
If anyone wants to know this mod just ask. It is a 10 minute job. The DL 110 is a remarkable device. 90% will never know as it falls between typical MC and MM levels. It is a genuine MC. I paid £135 for mine new. It is a minute to midnight to do this. Also alcohol and DL110 are possible at that price.
Quad by being scientific excluded many. At best people said " not for me, I am not that person ". This dam amplifer is the real deal. My system will play any music and does. Glad it has no more watts. I would use them and it would end in tears.
I am not sure about this. As i said earlier, every musical instrument is heavily distorted, but no one calls it distortion. I think the impression has somehow to do with "contextual / thematic meaning" . Nasty distortion is "out of context"
Read all the reference about the transistors in the posts before and have a beginner question please :
How to deal between PSRR and Noise issue when it comes to choose the voltage reference in a "stable" voltage shunt (I mean stable because : without correction with an oap or similar).
I read some old threads which says for a 15V shunt reg I need a zener should be better than a TL431 (can it be from On Semi or best supplier...)
If a part of the answer is in the question as i don't need a bandgap TL431 very precise reference but more a non non noisy ref : i will choose a 16 V Zener BZX something ref. But how to deal with PSRR before. A simple CLC with just a snuber before the diodes bridge ? Need something like max 120 mA peaks include the shunt needs.
Something like a R-Core ->fast rectifier->snubber- 1K uF LL (105° to more 120 C) ->smd 47uH (common chocke?) -> 1 K uF very low ESR (this time as the first cap is a simple LL normal ESR to avoid too much pulsed current as there is after a reservoir cap after the chock).
Finally if swapping a good complementary pair like the TIP 142/147 to a discrete darlington with the high band width BS139/140 which is the best driver for it : a BC317 low noise advised above instead a BC550 ?
Does the PSRR could be better with a common chocke before the secondary or better a 1:1 isolation transforer or some already made Network filtering like a Schafner FN9675 e.g. ?
Here I illustrate with a concrete need to try to understand what is said above about a "good enough" Reg. My simple understanding for a pre, amp, analog outputstage is we don't need always a very stable voltage if trade ffs must be choose but more a good ratio between PSSR and Noise issue (two similar things of a same problzem ?).
A simple LM317 "à la Nazar" or a simple zener after the Avalanche region if voltage is known to feed a darlington without feedback correction allowed by some superregs ?
How to deal between PSRR and Noise issue when it comes to choose the voltage reference in a "stable" voltage shunt (I mean stable because : without correction with an oap or similar).
I read some old threads which says for a 15V shunt reg I need a zener should be better than a TL431 (can it be from On Semi or best supplier...)
If a part of the answer is in the question as i don't need a bandgap TL431 very precise reference but more a non non noisy ref : i will choose a 16 V Zener BZX something ref. But how to deal with PSRR before. A simple CLC with just a snuber before the diodes bridge ? Need something like max 120 mA peaks include the shunt needs.
Something like a R-Core ->fast rectifier->snubber- 1K uF LL (105° to more 120 C) ->smd 47uH (common chocke?) -> 1 K uF very low ESR (this time as the first cap is a simple LL normal ESR to avoid too much pulsed current as there is after a reservoir cap after the chock).
Finally if swapping a good complementary pair like the TIP 142/147 to a discrete darlington with the high band width BS139/140 which is the best driver for it : a BC317 low noise advised above instead a BC550 ?
Does the PSRR could be better with a common chocke before the secondary or better a 1:1 isolation transforer or some already made Network filtering like a Schafner FN9675 e.g. ?
Here I illustrate with a concrete need to try to understand what is said above about a "good enough" Reg. My simple understanding for a pre, amp, analog outputstage is we don't need always a very stable voltage if trade ffs must be choose but more a good ratio between PSSR and Noise issue (two similar things of a same problzem ?).
A simple LM317 "à la Nazar" or a simple zener after the Avalanche region if voltage is known to feed a darlington without feedback correction allowed by some superregs ?
Yes, all instruments have "distortion" associated with their physical construction and how they are played - a good example is the classical guitar, where the sound of the fingers moving on the frets is always part of what one hears - and may irritate intensely someone who merely wants to hear the pure note resulting from the strings. But that "distortion" is linked 100% with the the actual procedure of the musician creating the music - and is fully "convincing" to someone who is knowledgeable about the instrument. A "scratchiness" in the replay, OTOH, which varies very much depending upon precisely what is going on, and in this case is caused by the electronics misbehaving, just sounds completely "wrong", the flavour of that sound is totally at odds with the musical event.
Dejan, you said this before:
That's what I want from every recording. Every recording. I want to understand why the musicians did what they did, why they bothered playing that music in the first place - at the emotional level you describe. If I can't do that because I am aware of too many technical glitches in the sound that I'm listening to, then the replay system has failed to do its job. Not the recording, but the playback mechanism is not good enough - I'm after an emotional connection to the music being played, I'm not interested in being a marker looking at someone's exam paper.
And because I've found that goal to always be achievable if I go to enough effort, I don't "judge" recordings - rather, I'm interested as to whether the system is good enough to make the playback "work" at the emotional level, every time ...
We did a test similar to the link below, while I was at the university. Though being before Internet became popular, results are likely kept in professor and university records.
But results were similar. We gathered about 20 (IIRC) subjects from musicians to hifi experts, and we asked them at which point they would notice something off with the music, while we were injecting noise at varying frequencies and amplitudes.
The result was that at low frequencies it takes quite a bit of distortion before it becomes audible, while at higher frequencies distortion becomes more audible.
Experimental Study : Distortion - Axiom Audio
Actually, Frank, I consider that Belafonte LP to be my prize LP. No other LP has the same subjective value as that one to me.
Not even "Bridge Over Troubled Water", which I also love dearly, all the more so since by an odd twist of fate I had about three weeks before it was released, thanks to a friend's dad who worked at the pressing plant in UK and nicked one from the initial trial batch especially for me. These trial run LPs used to be called "alpha series" and are extremely valuable these days, since everything was brand new when they were stamped and is said to sound the best. I don't know if that's true or not, but I do know that it really has a sound quaity few can even approach. And I am a certified Paul Simon fan.
But in essence, it simpkly plays music with emotion. That's the gist of it.
Not even "Bridge Over Troubled Water", which I also love dearly, all the more so since by an odd twist of fate I had about three weeks before it was released, thanks to a friend's dad who worked at the pressing plant in UK and nicked one from the initial trial batch especially for me. These trial run LPs used to be called "alpha series" and are extremely valuable these days, since everything was brand new when they were stamped and is said to sound the best. I don't know if that's true or not, but I do know that it really has a sound quaity few can even approach. And I am a certified Paul Simon fan.
But in essence, it simpkly plays music with emotion. That's the gist of it.
My experiance is LM317 or LD1084 are your best options. You will need serrious measuring equipement to do better. I have just finished a project that looked at all possibilities. Sorry to say most oppinions given generally are highly imaginative. One that is never addressed is the ability to sink current. A friend says they use a power amp and capacitor to send in sine waves to start and see what happens. This mimics a digital use. With analogue he sees no problems in the data sheet versions. LD1084 seems no worse than LM317 and seems to get rid of heat a little better. It also can work at a slightly lower voltage before drop out. TL431 is a strange device. Just when it seems to win you find it is a runner in full body protection. You have a choice, safe or fast. If you choose fast you must be sure it's safe. It has one good thing about it. Mostly binary if it is unsafe. I have never seen a device go instantly into osciallation like it does. The higher the voltage the better it is. If using it as 2.5V 1 nF is already the upper limit or I suspect 100 uF. Nothing between. All of this is in graph form with On semi doing very good devices.
My favourite on my tests was the amplified zener. The Zener needs a bit of current to be stable. That suits very well. I used a feedback pair NPN with PNP pass device. This is so much better than a LM317 with lets say TIP2955 as shown in data sheets. Reason being the pass device is only in it's own feedback loop and isn't global. The gain will be very high which mean the zener can be sweet spotted.
The big problem I see is the voltage reference is the noise source. This mostly is the problem and is like seeing stars from Earth.We need a lot of engineering to overcome the problem we have ( when stars a vacuum ). LM317 adds complexity to make it idiot proof. As an idiot I like that. I have heard a LM317 regulate a Maranzt Model 9. It sounded great. That's is it's party piece to work at 400 V if the dynamic range can be kept to lets say 30 V. Although I didn't try it my hunch is the ideal volatge reference is a resistor and CCS. I see no reason to think it wouldn't work. If so a capitance multiplier as a pre-regulator to make up for lost ripple rejection. LM317 has better ripple rejection than many because the reference voltage can be bootstrapped to ground. This also reduces noise. LM317 can be 3 dB from vacuum status.
are you describing a CFP sitting above a Zener diode.
Output = Zener voltage minus the Vdrop of the CFP devices?
Which is NPN and PNP for a positive supply?
I couldn't agree more. For that reason, I am baffled that active speakers are not very popular in the high end segment of the market. We still live in a world where oth amps are optimized on one island, and speakers on another. Except of course in pro audio, where the systems approach is the rule.
That's right. Mostly I wanted to put forward a starting point of a low noise concept with minimal problems. The Cfbp is a good alternative to a Darlington and is less prone to trouble in this application as it's stability is not linked to some type of op amp.
My friend made a very good point. If wanting a better regulator a power amp with high slew rate looks ideal as it can sink current ( ? ) . That's what he does. I also did this using the 1957 HC Lin amp as a starting point. This was to prove it needed very few extra lumps of plastic. Form all that I have read no one seems very concerned about this. Is this the shunt advantage ? If so why don't they say so? Most seem to say shunt is good because it is not in the series path. Look again people, it is tenuous to say so. A L165 ( TDA2030/40/50 ) is not slower than a LM317 I suspect. An LED on one pin and gain on the other would be OK.
https://mail.google.com/mail/?tab=wm#inbox/14a8978e052857c7?projector=1
Has anyone got the circuit of Amstrad IC2000? I never saw inside one although many were sold. That's like a garage never seeing a a certain popular car. It might be the most awful junk, no worries I would still be interested. Mr Sugar made things the industry would not be proud of. He is often put on TV as the man who knows best. DIY Audio has a lot of brains here, none of us will be remembered as this man regardless of what any will say against him. In a way we are the fools and him the wise man.
I have to use some electrolytics in the signal path. Out of common types what should I looks at? Unfortuneatly they are polarised. Panasonic FC ? I am not interested in tweaking parts or tantalums. 100 and 22 uF. They will see no more than 6 V. I fancy FC at 50 V will be OK. Tan theta seems better on the higher voltage types. I dare say some ceramics are better than I would imagine. There isn't space for polyester or whatever.
Has anyone got the circuit of Amstrad IC2000? I never saw inside one although many were sold. That's like a garage never seeing a a certain popular car. It might be the most awful junk, no worries I would still be interested. Mr Sugar made things the industry would not be proud of. He is often put on TV as the man who knows best. DIY Audio has a lot of brains here, none of us will be remembered as this man regardless of what any will say against him. In a way we are the fools and him the wise man.
I have to use some electrolytics in the signal path. Out of common types what should I looks at? Unfortuneatly they are polarised. Panasonic FC ? I am not interested in tweaking parts or tantalums. 100 and 22 uF. They will see no more than 6 V. I fancy FC at 50 V will be OK. Tan theta seems better on the higher voltage types. I dare say some ceramics are better than I would imagine. There isn't space for polyester or whatever.
Point taken. If people go back a page or two I show with TL431. NPN 100R and PNP to follow. I suggested use a zener in place of the TL431 and loose the feedback loop. Tricky choice that one. Ears and not scope as both look fine on the scope as long as TL431 behaves. The 100 R was a lazy choice. I suspect it is not the best choice, try 47R?. I think BD139 and high speed TO3P/247 devices the best choice. The advantage of Cfbp is LDO and it should be faster although very marginally.
I wonder if a simple complimentary feedback output stage ( like a power amp ) could be used ( no VAS ) ? I doubt it even needs biasing? Then it might sink current. The bottom half doing not much usually. 20 mA bias might be worthwhile.
- Status
- Not open for further replies.