In article , Arny
Krueger
wrote:
"Jim Lesurf" wrote in message


If you don't employ oversampling or resampling then you don't risk
problems with some intersample peaks being above 0dBFS. Oversampling
and resampling systems need to allow for values as much as 3 or 4 dB
above any individal values in the source stream to avoid overflows.
Using a non-oversampling DAC followed by analogue filters with enough
headroom means the digital system doesn't need this extra 'bit' or so
above 0dBFS.

Intersample peaks above 0 dBFS can easily happen whether or not
oversampling or resampling is used. It is a natural consequence of the
reconstructed data passing through certain data points.

Indeed. However if the DAC does not generate intersample values it doesn't
have to cope with them. That is left for the following analogue stages.

The good news is that they are rare in natural data. Another good reason
not to normalize to 0 dB FS.

Indeed. Alas it does happen. Indeed, reading a computer mag some months ago
I was alarmed to see an 'expert' blandly explaining how to use sotfware to
compress sound levels up to 0dBFS. No sign he had a clue about the problems
this could cause.

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 , Peter Chant
wrote:

QUAD used for the PSU in their 34 preamp might avoid switchon
thumps/clicks. That uses an op-amp to center the ground about
the
polarities of a single psu potential difference. Means 'both
rails come on
together' as soon as that op amp wakes up. :-)

If I follow you correctly then the ground, phono screens on input
output, are connected to 0V which is as op-amp output?

If I get a chance I'll dig out my copy of the diagram and put a scan of the
relevant bit on the web so the details will be clear. But the trick they
employed was to use a nominally floating single pd from the PSU. Then had
an op-amp generate a mid-way voltage which is then used to define and hold
a '0V' line for the bulk of the circuitry. This then is connected to the
screens on the phono sockets, etc, so becomes 'ground'. They also chose to
make the mid-point *not* the half-way point and produced asymetric rail
voltages for other reasons. Quite an interesting design.

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 , Peter Chant
wrote:

QUAD used for the PSU in their 34 preamp might avoid switchon
thumps/clicks. That uses an op-amp to center the ground about
the
polarities of a single psu potential difference. Means 'both
rails come on
together' as soon as that op amp wakes up. :-)

If I follow you correctly then the ground, phono screens on input
output, are connected to 0V which is as op-amp output?

If I get a chance I'll dig out my copy of the diagram and put a scan of
the
relevant bit on the web so the details will be clear. But the trick they
employed was to use a nominally floating single pd from the PSU. Then had
an op-amp generate a mid-way voltage which is then used to define and hold
a '0V' line for the bulk of the circuitry. This then is connected to the
screens on the phono sockets, etc, so becomes 'ground'. They also chose to
make the mid-point *not* the half-way point and produced asymetric rail
voltages for other reasons. Quite an interesting design.


Using an op-amp to create a mid-point ground wasn't new in the Quad 34. I
first met it in a Weir bench PSU in the early 1970s (albeit with a discrete
component amplifier). It's a convenient way of creating a variable, balanced
dual rail supply as there is no need to try to make two regulators track.
Simply use one variable regulator to generate the rail-to-rail voltage and
rely on the amp to create the mid point.

I used an op-amp to create a mid-point ground in an active audio
balanced-unbalanced converter that I built in the late '70s and have used
the same technique in various other similar boxes since. All those had
equal voltages either side of the ground, but I noticed that a PPM card made
by Audix that I acquired from eBay recently used an op-amp to generate a
ground that wasn't mid-way between the rails.

David.

Hi,

Just to let people know that I've added some more info to

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

This now gives some values for the distortion under various conditions.
(Finally got around to doing reliable measurements on this!)

Also a little more info on the battery PSU connection.

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

On Tue, 03 Aug 2010 16:29:26 +0100, Jim Lesurf
wrote:

Hi,

Just to let people know that I've added some more info to

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

This now gives some values for the distortion under various conditions.
(Finally got around to doing reliable measurements on this!)

Also a little more info on the battery PSU connection.

Slainte,

Jim

The battery part was interesting. Could this DAC be the makings of a
decent amplitude modulator?

d

On Tue, 03 Aug 2010 16:29:26 +0100, Jim Lesurf
wrote:

Hi,

Just to let people know that I've added some more info to

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

This now gives some values for the distortion under various conditions.
(Finally got around to doing reliable measurements on this!)

Also a little more info on the battery PSU connection.

Slainte,

Jim

A question occurs. Is the output of the DAC bipolar? If not the net
half scale DC could be compromising the transformers a bit. Maybe a
big cap in series?

d

In article , Don Pearce
wrote:
On Tue, 03 Aug 2010 16:29:26 +0100, Jim Lesurf
wrote:

Hi,

Just to let people know that I've added some more info to

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

This now gives some values for the distortion under various conditions.
(Finally got around to doing reliable measurements on this!)

Also a little more info on the battery PSU connection.

The battery part was interesting. Could this DAC be the makings of a
decent amplitude modulator?

I have wondered if - over a limited range - you could adjust the volume by
changing the PSU voltage. However I have no idea if it would be OK to
*quickly* vary the rail voltage. And so far as I can tell thus far you are
limited to the range from around 2 - 5.5 V and only get about 5-6dB change
in level. I assume there is some internal smoothing since ripple isn't
evident when I use the (fairly small) wall-wart PSU.


On 03 Aug in uk.rec.audio, Don Pearce wrote:


A question occurs. Is the output of the DAC bipolar? If not the net half
scale DC could be compromising the transformers a bit. Maybe a big cap
in series?

No sign of any 'half rail' dc offset on the output of the DAC. Just checked
it quickly with a DVM to confirm. Around 0.2 mV dc. This is with nothing
except a (high impedance) DVM connected and using the battery supply I
made. (Currently providing about 4.95Vdc.)

The AU-D3 is designed to be used as a normal DAC in HiFi and AV systems,
but meant for people in the 'installation' trade so a sensible price, etc.
I don't think 'audiophiles' were on their sales radar or they might have
added a zero. :-) TBH if there were 2.5Vdc on the output I doubt the
transformers would be working anything like as well as they are! I also
doubt people would be happy to use it in installations.

And when used with no o/p load the transformer saturation is symmetric at
LF. Which I'd not expect if there were significant dc feeding into the
transformer.

So far as I can tell thus far, it is a decent little DAC. Bargain for 35
quid. Maybe a reviewer sometime will tell us if it is better than a 1000
pound DAC. 8-]

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:

....
I've just been keeping a mental note to see if it 'lasts OK' or I find they
need recharging annoyingly often. If I get a chance I'll simply measure the
current drain once or twice and extrapolate an estimate. Then update the
webpage with this and any other extra info I've collected.

I'm sure that is fine, or it will be one of those things about it that people can
tweak to death.

-Mikkel

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. :-)

It doesn't use any fancy items made of unobtainium, etc. I
deliberately
bought all the bits from CPC so they'd be easy to source.
But if anyone
wanted, it might make a nice basis for experiments and no
doubt could be
'improved' or tweaked in all kinds of ways. That said, the
unit described
works nicely here and I'm enjoying listening using it.

A cap at the input to limit bass might be a good idea. If
you allow the transformer to define the bass limit, it'll
push the core into its non-linear region and cause
distortion everywhere.

That would then highlight the fact that you can't get clean
bass at an adequate volume with your transformers. A little
maths would suggest that it can't be done even with optimum
transformers.

Your point that the winding resistances are helpful could be
misleading. If the winding ratio was optimal, then winding
resistance would not be a benefit at all. In your case
winding resistance is so high that you are half way to a
resistive divider. If you compare the output from the
optimum divider to that from the optimum transformer with
negligible resistance, then clearly the transformer would
result in much higher output. You have sacrificed much of
that advantage.

If you start by specifying the optimum transformer, then you
could see if such a thing is available. After all, line
level signals and headphones have been around a long time,
so if transformer coupling was a good idea, you'd think
they'd be commonplace.

As a starting estimate based on the common 10:1 impedance
ratio rule-of-thumb and your estimate of 200 ohm output from
your DAC, you'd want a 2k:32 audio transformer with winding
resistances no more than 200 and 3.2 ohms. Primary
inductance needs to be such that it's reactance is equal to
its source resistance of 400 ohms in that case, more or
less, at the -3dB frequency. Then your DAC output would be
happy with 2k, your 'phones would see about 6.5 ohms which
ain't
perfect still but that's transformers for you...that's why
valve amps need global feedback. It's commonly assumed that
headphones are so near a resistive load that source
impedance doesn't matter much, but I believe that's a
contentious issue.

The reason audio transformers are specified in terms of two
resistances, rather than a winding ratio, is because it's
only when used for that particular purpose that you should
expect full audio bandwidth. You seem to be a couple of
octaves short. You need much more primary inductance.

That would make for a fairly chunky transformer with a
winding ratio of 8:1, or going on 6:1 if you cheated and
went for 1k:32. That'd give you a quarter or a third of a
volt output from your DAC's full whack, which might be
enough for you but most ppl would consider it inadequate.
Basically, what you are trying to do is impossible. Look at
it this way, 2V into 2k is a maximum of 2 mW, or 4 mW if you
go for 6:1. Take from that losses in the inevitable
resistances, and what you have left can't be enough.

Once you realise you need MORE POWER, and therefore a power
supply, there are umpteen simple designs available. A dual
opamp, a couple of caps, and a few resistors isn't hard to
do. I know you just wanted to check out the transformer
idea, but you could have done that in your head, I would
have thought, and then done something more sensible instead.

Ian

In article 8L_6o.74338$2%2.70629@hurricane, Ian Iveson
wrote:

[snip interesting comments]

Once you realise you need MORE POWER, and therefore a power supply,
there are umpteen simple designs available.

Well, I found that I *didn't* need 'more power' - even in lower case! :-)
The result sounds loud enough for me, Indeed, I'm using it with the volume
wound down by about 6dB most of the time. So in my case, experiment
disagreed with your theory. :-)

However I appreciate that *others* might like 'more power'. FWIW I am
looking at alternatives, and others can do so if they wish. I'm not trying
to sell the design or tell anyone that the circuit I made is perfect and
would suit everyone. Just that it is an example that is *very easy to make
for people new to electronics* and can give useful results.

A dual opamp, a couple of caps, and a few resistors isn't hard to do.

Yes, I could easily have done that. Although of course your list isn't
complete. e.g. need for a board/support and psu. And the maker being
confident they could put that together without problems - particularly if
they have no scope, test kit, experience, etc, etc.

But the design idea was to make the circuit easy for *others* to build who
*haven't* already built many bits of electronics and are deterred by what
they usually see. Something they could assemble easily with no real
support, kit, or experience. I also liked the idea of seeing how 'minimal'
I could get in that respect.

A problem here is that those of us who have over many years become familiar
with practical electronics building can easily forget how confusing or
daunting it can be to someone with no experience starting with no equipment
or components. Particularly if they have no local friend or relative who
can show them how it is done.

I know you just wanted to check out the transformer idea, but you could
have done that in your head,

But then I would not have known the actual performance or been able since
to enjoy using it. I find I enjoy actually listening to music more than
just reading a score. ;- Nor would I have established that I didn't
personally need 'more power'. Indeed, so far the rise in LF distortion
isn't a problem in my case, either. Music isn't usual a 20Hz 0dBFS tone or
similar - at least the type I listen to. Although when I tried the HeadDAC
with a test recording of a synth rock concert on TV with a fair bit of LF
it sounded OK to me... at least to the extent that "The Pet Shop Boys at
Glastonbury" would sound OK to me, anyway! 8-]

I would have thought, and then done something more sensible instead.

I'm not stopping you (or anyone else) from doing a 'better' design and
posting it on the web. if you do and it looks good to me, Indeed, I'd
welcome/encourage you doing so. I'd also consider telling people about it
as an alternative if you put up the results. So do what you think is
'sensible' and let us know!

That said, you might find it worth building the HeadDAC design and trying
it yourself in practice with some phones like the EH350s. Then say if -
with phones of that kind of sensitivity and impedance - it seems
unacceptably quiet to you.

BTW I am also myself thinking of simple active designs. As a possible
alternative for more confident/experienced people. *And* of perhaps getting
more capable and appropriate transformers. Alas, CPC and the other 'usual
suspects' for home construction electronics didn't seem to have these. At
least I couldn't find any to fit my 'minimal fuss' brief for the design
aimed at drawing newcomers into DIY/modding. But I'd like to have
'improved' versions to point people to if they wanted something 'better'
(for them) and were willing to go for something more complex or expensive.
That said, I'd imagine that many such are already available. Indeed, I did
buy a small 'kit' of one as part of my experiments but haven't reported on
it.

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