The HiFi chat, build and modification thread!

[PurpleFrog]

The EU has said that we're all going deaf due to too loud iPods and such like. So the volume output is limited by quitea lot.

There are two cures for this:

1. Change the region in player settings to USA or whatever

2. Install Rockbox (google it - it's alternative firmware you can install for most popular players; its free and plays every codec I've heard of)

I used an older set of late 1980s headphones and they were much louder than my current ones too so maybe there is some limiting going on there too.

Nope. It's just that some headphones provide more decibels per amp than others. This has always been the case and always will be. It's a headphones characteristics thing.

The audio from the bottom of an iPod is much better than the headphone socket and is fixed.

The audio out from iPod varies with the model. On the whole, it is actually damn good - this is not subjective, because you can actually test signal accuracy with standard electronic hardware. The best accuracy of all tends to come from Sansas, weirdly - they are waaaay past the point where anyone with normal hearing and using anything like standard headphones will be able to hear better quality:

http://nwavguy.blogspot.co.uk/2011/02/s ... sured.html

To get a substantial improvement over a $40 Sansa driving $150 HD25-ii phones (this is a model that Sennheiser designed for TV news soundpeople, but which has become standard for DJs too - they have very accurate sound and are extremely hard to destroy) you really have to go to electrostatic headphones and an amp powerful to drive them.

Otoh, you can a big boost to audio quality for free if you tune your headphones using pink noise: most people have a couple of dips in their frequency response, and if you compensate for them using an equalizer you'll have a much more accurate sound.

 

[apache]

headphone amp! portable!

Was talking to a mate down the pub, and we think it has to be Class A as you can have Class A for headphones without the huge heat load when you need speaker driving power.

People think class A means "Good, better than class B" but they're actually just the simplest and least efficient amp design:

http://en.wikipedia.org/wiki/Amplifier#Class_A

What is the reason behind portable headphone amps?

Mostly marketing and superstition. Modern portable audio amps are extremely good. Except when driving very power hungry headphones - say HD600s - and extra amp will actually be harmful. But typical hifi nuts... well, these are people who spend $3000 on "high end" cd players that no one can tell from a $100 one or even a PS1 in blind tests.

Wikipedia, really?

Class A because it negates the most prevalent form of distortion in an O/P stage. The input stage and the VAS can be made essentially distortionless. The output stage is the fly in the ointment. Yes it's inefficient, but on mains power, who cares?

Yeah, I still don't really 'get' portable headphone amps though, unless you're deaf and need loads of oomph - or have decided to use low impedance phones and need a bit more current. However, a few portable jobs don't have any more current drive than your iPod. One I've just built could probably set fire to low impedance headphones - yes it's overkill, but hey, I wanted to see what I could do in a couple of inches of board space

 

[greenstiles]

Very happy with the tone controls very.

 

[apache]

Great! Happy to help. Now go out and tell the evangelists you CAN have transparent tone controls

 

[PurpleFrog]

Wikipedia, really?

I can direct you to articles by professional amp designers, but everyone knows what wikipedia is.

Class A because it negates the most prevalent form of distortion in an O/P stage. The input stage and the VAS can be made essentially distortionless.

Take a look at the link to the test of the Sansa Clip: it's amp is built into a chip along with a DAC and a 4-band equalizer. Bought in bulk, the chip costs about $2. And it's output is "essentially distortionless"! Building a class A to get distortion free sound with a reasonable component cost stopped being necessary at least 20 years ago.

 

[apache]

I can direct you to articles by professional amp designers, but everyone knows what wikipedia is.

Take a look at the link to the test of the Sansa Clip: it's amp is built into a chip along with a DAC and a 4-band equalizer. Bought in bulk, the chip costs about $2. And it's output is "essentially distortionless"! Building a class A to get distortion free sound with a reasonable component cost stopped being necessary at least 20 years ago.

That's quite OK. It's 20-odd years since I used to share a lab with 'professional amp designers', but we still keep in touch. Everyone knows what Linkedin is.

Without wanting to get into a flame war over something as unimportant as hifi, as only a keen amateur these days, I do not wish to design with an all in one chip (assuming one could buy the chip from CPC or RS) with an amp, a DAC and a 4 band EQ in there as anyone can do that given the data sheet. Where's the fun?

Also, a 5 minute google doesn't elicit it's noise or distortion specs other than those by NWAVguy - audible hiss? Jesus... I absolutely guarantee you, for example, it's basic measured amplifier noise spec will not meet what I could come up with in 10 minutes - the Clip will be hobbled by inherent noise of passive smd parts for starters, a switched mode power supply for seconds, and a switched mode amplifier for thirds. Being able to measure THD+n down to around 0.05% isn't wildly impressive either - and it may or may not be audible, but why not aim higher? Most current op-amps suitable for audio can achieve a couple of orders of magnitude better than that. <edit> Interesting to note that the Sansa's distortion is primarily odd order harmonics and the iPod is even order. Some might say the iPod is better then for this reason alone. It's a bit simplistic to state that low impedance headphones are more efficient than high too. Depends on the characteristics of the amp feeding them, the output voltage swing, current drive etc, though 1 ohm isn't too bad for a portable. I'm guessing the Sansa has been designed like this as most portable phones are low impedance? I don't know, not too familiar with portable type phones.

Paralleling NE5532's as OP devices will drop output impedances far far lower - and reduce noise into the bargain. You could just use conventional OP devices though and end up with as close to zero as sensible. It wouldn't fit with portable sensibilities, but that's not really what I'm interested in.<end edit>

Coming back to class A, I didn't realise this Clip was a cheap portable suffering poor battery life as it is. Obviously, class A isn't suitable for portable (and component cost certainly isn't 'reasonable'), though you may get away with it in an enthusiast headphone amp but is it worth it? Who knows, not tried it, but the *measured* distortion (assuming you've not made a mess of the input stage and VAS) WILL be lower distortion than the 'Clip' - which will definitely measure worse than most domestic amps as, as previously mentioned its amplifier will be so called class T (really class D but with a hint of marketing) type in order to make decent battery life from a tiny Lithium cell. Class D, T or whatever it's called has it's own set of problems. For home audio (connected to the mains) there is certainly no reason to use that topology - other than price (or if you're one of those odd sort who think that a £10 eBay 'T-amp' sounds better than a well designed conventional amp). From memory, because of the inductive filters on the output to take out the couple of hundred kHz switching frequency, they only really work properly with a very narrow loading range - but this wont matter for 99% of the MP3 using public. I believe that some computer speaker amps etc don't bother with a filter at all and just use the fact that the speaker wont reproduce the switching frequency. That's properly inelegant though...

There's lots of design issue references to Class D out there, so important it is to the portable market. Though it's not necessary to dig too deep to realise that it's not a 'hifi' topography. Having said that, there's a lot of web space devoted to saying it's incredible. That's the web for you. For sensible views, there's several JAES papers out there. I have references to the JAES papers buried deep in books but it's too sunny outside to spend time looking for them - if you're seriously interested, I can have a look this evening for you?

Don't get me wrong, current portable technology is mind blowing. The storage capability, the battery life, the user friendliness etc, but it's been built for a particular demand. Cheap and cheerful audio on the go for a market that values being able to store a million albums, and enough eBooks to fill the British library more than it values absolute quality. I don't believe (and nor does industry) that the ultimate quality rests with an iPod just yet. It doesn't need to.

I'm not sure now where your original point was going. As far as sound quality goes, any competent designer could come up with an amplifier that measures better (and therefore potentially sounds better) than that in a mass market portable device. On the flip side, a portable device as smart and cheap as a Clip is way beyond the realms of DIY and will always be so. The proper hifi, home based market is a different ball game - very few physical or economical restraints.

Look at that! The 'hifi chat' group really is!

 

[greenstiles]

Just like to add, how ' useable ' the new tone controls are !....yes it's great for tweeking fairly good recordings to taste......but i listen to a fair bit of 60's american acid rock, psych, and 80's alternative scene etc, and some of the 60's compilation boxes (5/6) cd's are sometimes so out of kilter, even though they say they painstakely remastered all the singles to sound the same kinda quality, this thing is a life safer...........things like, ' pebbles ' collections are fairly well remastered, but stuff like the 'trash box' is so, all over the place, i've had the tone controls to very very different settings indeed to bring some of this stuff back into line, and it really does make them listenable to again, BUT without having to dull them too much and take the life out of them etc etc.......very very ' useable ' ...a simple word but one that means a lot in hi -fi land.

 

[apache]

Good news. Do you reckon the turnover frequencies are about right then?

 

[greenstiles]

Treble/high........spot on..........at first I thought the bass/low could have been a little different, but after playing a good cross section of music, i realized that any initial doubt about the bass was due to some really bad recordings I was playing at fist, but as mentioned before, this thing is so useable, that I have found myself using a lot of it's range to add or take away (mainly in the bass) also with the treble/high you can reduce it a lot without making the sound dull. So for my music collecting, yes perfect focus and range.

 

[apache]

If I recall correctly, I made the peak boost / cut for LF around 30Hz (turnover around 400Hz rather than the typical 1kHz). I couldn't really take the boost much lower without introducing a steep high pass filter around 10hz to prevent problems when using it with vinyl. It would be fine with a digital source as there's no LF crap like there is with vinyl.

Good info though, and things I may think about for MKII.

Started on pre-amp design yesterday. Probably going to go the modular route centred around a standalone line level input buffer / selector / volume control module and PSU module which would form a basic no frills pre. This could be expanded then with RIAA module, tone module, headphone amp module. This opens up a simple and lower cost upgrade path such as L/R PSUs etc.

The whole lot will be designed as dual mono for minimal crosstalk. I've started looking at input buffer, and am looking along the lines of multipath voltage followers which will offer noise floor (so the input performance becomes independent of source impedance) and distortion below measurement limits, think 0.00x type distortion. Yes, I appreciate there is conjecture that higher distortion is inaudible, but if you can have much lower at minimal extra cost, it would be remiss not to aim for the best possible performance.

Will keep you posted on progress.