In article , Peter Chant
wrote:
Jim Lesurf wrote:

This is just to let people know that I've now put up a webpage
describing the design of the simple headphone DAC I've been
experimenting with during the last few weeks. The page is at

http://www.audiomisc.co.uk/HFN/Headp...C/HeadDAC.html

The aim was to produce something inexpensive and easy to build so that
anyone willing to use a soldering iron and drill some holes in a metal
box could make one if they wished. :-)

Any reason you did not build an emitter follower instead as a buffer?
Or is that just a bit too complex for your aims above?

Chose this approach my aim was to make it as 'simple as possible' for
people who had not previously built anything. So a 'minimal' design. Must
confess that having thought of using transformers I became curious to see
how well the cheap 'CPC' ones might work. :-)

Of course, there is no reason I or someone else can't now suggest
'improved' versions - like better transformers or using buffers. Similarly,
adding crosstalk bleed networks which some may prefer for headphone
listening. :-)

And I'd expect that others might prefer a physically 'tidier' construction
that I am using. Never been much of a 'mechanical engineer'. ;-

But having listened to the unit with Sennheiser EH350 phones my main
reaction is that the results sound more enjoyable than I'd expected. The
only tweak I plan at present is to to add a limited-range balance control.

FWIW later tests and all listening have been with 4 x NiMH AA (1700mAh)
cells as the power supply. These have so far lasted a couple of hours but
I've no idea how much longer they will manage before needing a recharge.

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

"Arny Krueger" wrote in message
...
"Eiron" wrote in message

On 27/07/2010 10:36, Fed Up Lurker wrote:

This TDA1541 machine that is itching for you to NONOS it
and

Maybe you could explain what you mean by NONOS for those
of us who haven't kept up with these things.

NOS usually means New Old Stock. Maybe people should read the added "NO"
as "no, don't waste your time with it.

Building new gear out of obsolete parts from a fast-moving technology like
DAC is something that people with way too much time on their hands seem to
want to obsess over.


Hi Arny
NONOS is "non-oversampling", which isn't a truly full or accurate acronym.
It is the disabling/by-passing the digital filter pre conversion (as opposed
to the post conversion analog filter ).
There is plenty available about this simple mod online.
The controversy that surrounds this DIY* mod is that the traditionally
accepted to be required brickwall filtering is defeated. This of course
results in quantization is not pushed (oversampled, 2x, 4x, 8x etc) out
of the audio band. To an engineer as yourself that would be alarming
and indeed it measures painfully, but the result can be magical.

*It isn't just a DIY modification, a number of hardware manufacturers
have issued NONOS models, particularly Audionote, T+G and Project.
(Some maybe termed "Filterless")
Also the *early* Pioneer Legato Link Link and Wadia Digimaster were a
a sort of stepped roll off between brickwall and slow sloped.
There was of course a number of variable coefficient filter players
in the early days, though none could be termed full non-oversampling.
I've had success with a saa7350, but a convoluted process to get there.
Many have NOS'd PD2028's and apparently the results with a NOS'd
PCM1702 can be spectacular (so thats my next project).
The reason the ancient Philips multibit chip TDA1541 is the focus for
so much NONOS'ing is it is the easiest to NOS, just 3 pins by-passed.
And TDA1541 players abound on ebay, car boot sales, garage sales etc.
As an engineer you Arny should investigate NONOS.

In a search use following terms:
non-oversampling dac
NON O/S dac
Filterless dac
You will be swamped with info, but start he
http://www.lampizator.eu/Nonoversampling/NOS.html
http://www.lampizator.eu/LAMPIZATOR/...r/TDA1541.html

And of course my latest uploaded clip of one of my NOS'd cd players
but again YT got a bit urgent re copyright infringement.
http://www.youtube.com/watch?v=gFo3Ywc6OMA

I only resurfaced in this group in the hope of getting a tech journalist
to have a try, and I would've supplied the TDA1541 CD player!
I just wanted once in the mag brigade that one of them would pen
an accurate article on the subject, but semi-retirement and not
rocking the boat seems to be the order of the day in the publishing
world (how much do I have to bait you Jim?).
So Arny, I have no doubt you know your way around a PCB and know
how to read a schematic, and by-passing 3 pins would be breeze
for you, so if you want to try NONOS Arny, I'll happily guide you.

Jim Lesurf said...

32 ohm headphones are very common as are 16 ohm headphones.

Yes. IIRC I mention one of the phones I used were about 18 Ohms, but the
three others I tried all came in around 32-35 Ohms.

I used several pairs of Sennheiser 414's before they went out of
production. They had a nominal impedance of 2,000 ohms and could be
operated from speaker outputs without problems. My first pair came
fitted with DIN ls plugs.

Theoretical question, Jim. If you were to redesign your DAC
specifically for HD414's would the transformers still be necessary?


--
Ken O'Meara
http://www.btinternet.com/~unsteadyken/

"Fed Up Lurker" wrote

NONOS is "non-oversampling", which isn't a truly full or accurate acronym.

True. The letters "non" should not be capitalised as they are not the
initial letters of words.

It is the disabling/by-passing the digital filter pre conversion (as
opposed
to the post conversion analog filter ).
There is plenty available about this simple mod online.
The controversy that surrounds this DIY* mod is that the traditionally
accepted to be required brickwall filtering is defeated. This of course
results in quantization is not pushed (oversampled, 2x, 4x, 8x etc) out
of the audio band. To an engineer as yourself that would be alarming
and indeed it measures painfully, but the result can be magical.


Really?, it measures crap but sounds "magical"? This is the same sort of
nonsense that we get from the vinyl brigade. Early CD players were
non-oversampling, oversampling was introduced later because it sounded
better. Had those early CD players sounded "magical", oversampling, and
later single-bit converters, would never had taken hold.

David.

In article ,
UnsteadyKen
wrote:
Jim Lesurf said...

32 ohm headphones are very common as are 16 ohm headphones.

Yes. IIRC I mention one of the phones I used were about 18 Ohms, but
the three others I tried all came in around 32-35 Ohms.

I used several pairs of Sennheiser 414's before they went out of
production. They had a nominal impedance of 2,000 ohms and could be
operated from speaker outputs without problems. My first pair came
fitted with DIN ls plugs.

I also used to use HD414s for many years. But my old ones have
disintegrated and AFAIK you can't now get them, or the replacement bits. I
liked the sound overall for general test use, but found them uncomfortable
if used continuously for a long time. The new EH350s I have seem more
comfy. I suspect that have more extended bass as well, but can't be sure as
I am relying on comparing with memory. :-)

Theoretical question, Jim. If you were to redesign your DAC
specifically for HD414's would the transformers still be necessary?

Not sure, but it seems plausible that they wouldn't. Certainly the AU-D3
happily drives 1k. So a pair of 414s with a 1k or 2k2 volume pot would be
likely to be fine, I guess. That said I can't recall the sensitivity, so a
transformer might be useful. I'd need 414s to decide, and mine are long
past use.

FWIW One of the things that made me think 'transformer' was thinking "If I
still had HD414s their high impedance would mean I wouldn't need
buffers...".

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

In article , Fed Up Lurker
wrote:


I only resurfaced in this group in the hope of getting a tech journalist
to have a try, and I would've supplied the TDA1541 CD player! I just
wanted once in the mag brigade that one of them would pen an accurate
article on the subject, but semi-retirement and not rocking the boat
seems to be the order of the day in the publishing world (how much do I
have to bait you Jim?).

You can try baiting me as you wish. :-) However I just do what interests
me and write about what my mind chooses. Too old, slow, and crusty to be
led. If I felt the need to 'conform', I'd be using Windows... ;-

One advantage I have is that I don't write for a living, and don't mind if
I seem unfashionable or cloth-eared.

So far, they've happily published what I write, sometimes complete with my
spelling errors. 8-] My interest is really just in giving readers some
'heads up' starting points and directions they can investigate if they so
choose. I do like to vary the topics from month to month and this did fit
nicely into one column. FWIW I find some other 'columns' (not in HFN) a bit
predictable and monotonic as they keep on about the writer's pet
hobby-horse or favourite area.

However when what I've written on the topic appears, you can try writing in
and adding 2p-worth. If others do similarly... editors love 'trends'.
Either way, I do intend to come back to it in future since it does interest
me. Just can't say when or how until I do. Particulary now The Proms have
started *and* the grass needs cutting. :-)

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

In article , David Looser
wrote:
"Fed Up Lurker" wrote

NONOS is "non-oversampling", which isn't a truly full or accurate
acronym.

True. The letters "non" should not be capitalised as they are not the
initial letters of words.

It is the disabling/by-passing the digital filter pre conversion (as
opposed to the post conversion analog filter ). There is plenty
available about this simple mod online. The controversy that surrounds
this DIY* mod is that the traditionally accepted to be required
brickwall filtering is defeated. This of course results in
quantization is not pushed (oversampled, 2x, 4x, 8x etc) out of the
audio band. To an engineer as yourself that would be alarming and
indeed it measures painfully, but the result can be magical.


Really?, it measures crap but sounds "magical"? This is the same sort of
nonsense that we get from the vinyl brigade. Early CD players were
non-oversampling, oversampling was introduced later

Erm. Even the 1st generation Philips players used x4 oversampling. So that
wasn't introduced 'later' IIRC.

because it sounded better.

IIRC they were primarily concerned by two factors. One being that
oversampling made post reconstruction analogue filtering easier/cheaper.
The other was that they weren't confident at the time that they could make
16 bit dacs at consumer prices/quantities. So my impression was that the
initial decision was based on what they thought they could make more
cheaply and reliably for a mass market.

Above based on my recollections of reading the special issue of 'Philips
Tech Rev' on the launch of CD Audio.

That said, companies like Philips do have 'form' when it comes to promoting
new ideas that "sound better" when they really mean "extend our revenue
stream". Think SACD/DSD. ;-

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

"Jim Lesurf" wrote in message
...
In article , Fed Up Lurker
wrote:


You can try baiting me as you wish. :-) However I just do what interests
me and write about what my mind chooses. Too old, slow, and crusty to be
led. If I felt the need to 'conform', I'd be using Windows... ;-

I don't want you to conform! I chose you because you don't conform.


One advantage I have is that I don't write for a living, and don't mind if
I seem unfashionable or cloth-eared.

Agreed, there is indeed a fashion following trend amongst some
journalists...


a bit predictable and monotonic as they keep on about the writer's pet
hobby-horse or favourite area.

Yep...

However when what I've written on the topic appears, you can try writing
in
and adding 2p-worth. If others do similarly... editors love 'trends'.

Do they?

Either way, I do intend to come back to it in future since it does
interest
me.

I've got you hooked.

Just can't say when or how until I do. Particulary now The Proms have
started *and* the grass needs cutting. :-)

My maid does that for me.

"Jim Lesurf" wrote in message
...
In article , David Looser
wrote:
"Fed Up Lurker" wrote

NONOS is "non-oversampling", which isn't a truly full or accurate
acronym.

True. The letters "non" should not be capitalised as they are not the
initial letters of words.


What a pillock you are Dave, I can't stop larfing,
Now back to Jim's post:



Erm. Even the 1st generation Philips players used x4 oversampling. So that
wasn't introduced 'later' IIRC.

because it sounded better.

IIRC they were primarily concerned by two factors. One being that
oversampling made post reconstruction analogue filtering easier/cheaper.
The other was that they weren't confident at the time that they could make
16 bit dacs at consumer prices/quantities. So my impression was that the
initial decision was based on what they thought they could make more
cheaply and reliably for a mass market.

Agreed. But a whole heap of other factors too that would only baffle Dave.
A key one being how early "16 binary digit" recordings were made.


Above based on my recollections of reading the special issue of 'Philips
Tech Rev' on the launch of CD Audio.

I was just about to mention that.


That said, companies like Philips do have 'form' when it comes to
promoting
new ideas that "sound better" when they really mean "extend our revenue
stream". Think SACD/DSD. ;-


But it is a cosy two way relationship. The pen pushers need an evolving
hardware industry and steady advertising revenue, not that I have a
problem with that, but healthy cynicism is required by we the punters
whose dosh is being steered along the latest fashion path.

Slainte,

Why thank you.

"Jim Lesurf" wrote

Erm. Even the 1st generation Philips players used x4 oversampling. So that
wasn't introduced 'later' IIRC.

Digital audio pre-dates the commercial introduction of CD players by a fair
few years, and those early professional ADCs and DACs all operated at
Nyquist rate (ie. they were non-oversampling). And most first-generation CD
players used non-oversampling DACs, Philips were the exception.

because it sounded better.

IIRC they were primarily concerned by two factors. One being that
oversampling made post reconstruction analogue filtering easier/cheaper.
The other was that they weren't confident at the time that they could make
16 bit dacs at consumer prices/quantities. So my impression was that the
initial decision was based on what they thought they could make more
cheaply and reliably for a mass market.

True enough, making monolithic converters with true 16-bit linearity at
consumer prices was a challenge, and analogue filters with the necessary
performance are expensive. So yes, one of the drivers for going to
oversampling was cost. But superior audio quality was also part of it. At
the time that they introduced their 4x oversampling players Philips' claim
was that 14-bit converters with 4x oversampling sounded as good as 16-bit
without. I'm not aware that this claim was seriously challenged at the time.

Between 1985 and 1995 I built a number ADCs and DACs for a hard-disk based
recording system. The early ones used expensive 16-bit professional
converter modules (not chips) with in-house designed analogue filters. Later
we tried 4x oversampling based on the Philips chip-set, and finally units
based on delta-sigma chips from Crystal Semiconductor. These later were
definitely the best sounding of the lot.

So I have a fair bit of experience of building, testing and evaluating units
using a variety of different types of DAC modules and chips and I can say
with confidence that the idea that non-oversampling DACs sound "magical"
compared to oversampling and delta-sigma designs is sheer poppycock.

But I'm somewhat confused as to what "fed-up lurker" is doing. If I
understand his post correctly he's removing the digital filter from the
signal path without changing the analogue filter at all. So the analogue
filter is now totally wrong for the spectrum it's being asked to handle. Is
it the presence of alias components in the output audio that he thinks
"magical"?


David.