Using PC as Terrestrial/Satellite recorder - obsolete?

Jim Lesurf wrote:

choose a system that doesn't have any fans. The box I have in mind
simply has heatpipes to a substantial metal case that works as a decent
heatsink. Use SSDs and you should get a machine with no mechanical noises.

Even without fans and discs, you can still hear a few "singing" noises
that vary with the CPU workload (not sure where this comes from -
voltage regulators? memory chips?)

In article , Andy
Burns
wrote:
Jim Lesurf wrote:

choose a system that doesn't have any fans. The box I have in mind
simply has heatpipes to a substantial metal case that works as a decent
heatsink. Use SSDs and you should get a machine with no mechanical noises.

Even without fans and discs, you can still hear a few "singing" noises
that vary with the CPU workload (not sure where this comes from -
voltage regulators? memory chips?)

Yes, I've noticed something like this with my main 'workhorse' machine
(which does have a fan that also comes on at times). I'm also not sure what
does it but I can tell when a CPU-demanding process starts and stops by the
change in pitch. My best guess is a capacitor in a SMPS singing, but its
just a guess.

FWIW That machine is close to me as I work, and in the common 'shuttle' tin
box with holes.

Conversely, a low-powered shuttle I use with the hif system has a fan I
disconnected. That doesn't do it. But it uses a large 'external
soap-on-rope' PSU with a closed box.

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 Mon, 02 Feb 2015 10:34:46 +0000, Andy Burns
wrote:

Jim Lesurf wrote:

choose a system that doesn't have any fans. The box I have in mind
simply has heatpipes to a substantial metal case that works as a decent
heatsink. Use SSDs and you should get a machine with no mechanical noises.

Even without fans and discs, you can still hear a few "singing" noises
that vary with the CPU workload (not sure where this comes from -
voltage regulators? memory chips?)

If this is with the penultimate B+ model (I've just seen the
announcement about the R-Pi 2 that's just gone on sale in the US of A
for $35.00) which upgraded the analogue voltage regulator to a
switchmode module, this could be an unfortunate resonance in its
inductor (magnetostriction of the core[1], possibly the winding[2])
otherwise I can't think of anything else that could account for this
noise (memory chips don't (normally) make 'singing noises').

[1] In the absence of standing DC, magnetostriction of ferrites
normally results in frequencies emitted at twice the fundamental.
Adding a DC component (which I'd expect would be the case with a
switching regulator) will allow excitation of resonance at the
fundamental frequency itself, along with a reduced or suppressed
second harmonic response.

Normally, the switching frequencies in such compact regulators tend
to be in the dozens to hundreds of KHz, if not right up to MHz in some
cases, so aren't usually audible. I'd expect the small size of the
ferrite component in this case to preclude emission of audible
frequencies (bats might complain, if they could but I'd expect even
young puppies would fail to notice these particular 'ultrasonics').

[2] Since the inductor winding is subjected to the same laws that make
the electric motor possible, this too can be a source of sonic
emission. Since it's most likely that this switching regulator will be
running at more than 50KHz, you could be hearing the intermodulation
between the noise currents generated by CPU activity induced onto the
internal 3v3 supply rail and the switching frequency pulses.

The usual cure for this sort of problem is to apply a blob of hot
glue to the inductor to act as a resonance damper (normally, such
blobs of damping material would be applied directly to the windings
and core of a naked inductor. Without taking another close look at the
R-Pi B+, I suspect the inductor is an smd packaged component where
the blob of hot glue will have less effect on damping resonances
within the smd package and will act more as a muffler than a damper
(although it will still work as a damper on any packaging resonances).
--
J B Good

Johny B Good wrote:

Andy Burns wrote:

Even without fans and discs, you can still hear a few "singing" noises
that vary with the CPU workload (not sure where this comes from -
voltage regulators? memory chips?)

If this is with the penultimate B+ model

No, bog standard PC mobo, not tried stuffing my head in the case to try
and locate where it's coming from ...

switchmode module, this could be an unfortunate resonance in its
inductor (magnetostriction of the core[1], possibly the winding[2])
otherwise I can't think of anything else that could account for this
noise (memory chips don't (normally) make 'singing noises').

I think the normal DRAM refresh would be well into the MHz and even I
don't claim to have ears *that* good, it's probably in the same
neighbourhood as TV line frequency, but shifts around depending on
activity, maybe particularly graphics activity, but the GPU is within
the CPU anyway.

[snip good stuff]

On Sun, 01 Feb 2015 22:11:40 +0000, Johny B Good
wrote:

On Sun, 01 Feb 2015 09:56:23 +0000 (GMT), Jim Lesurf
wrote:

In article , Roderick Stewart
wrote:


Use the largest fans you can fit, and choose the type that are
automatically speed controlled from the motherboard (as most CPU fans
are anyway now).

Or choose a system that doesn't have any fans. The box I have in mind
simply has heatpipes to a substantial metal case that works as a decent
heatsink. Use SSDs and you should get a machine with no mechanical noises.



I've been wondering how good those little fanless micro-PC units really
are. They're sold for use as media centres, so presumably they can play
high-def video, but I can't help thinking that if they really could do
everything a full size PC can do, then that's what all PCs would be
like. I feel sure there must be a catch. Maybe that's why "smart" TVs
are so slow, as the extra electronics in those must be small and
fanless, so presumably similar.

I suspect the problem in hardware terms isn't simply the 'fanless' but the
absence of much efficient heatsinking. Indeed, on my TV there is no
external sign of any heatsink arrangements of slots for efficient airflow.


All the internet connected TVs and PVRs
I've seen can play the actual video perfectly well, but are
infuriatingly sluggish to respond to anything else.

Again, I suspect in part this is because they use chip archtechures
optimised for TV reception, and then the makers bodge together other
aspects on the "box ticked" basis.

FWIW given the way people are using RPi's as media devices I'm also
wondering about a newer series of ARM based boards that are more capable.
All being well, I should get one eventually for running RISC OS on, but
once that's done I'll also look at Linux on them.


In which case, Jim, I think you'll find the ODROID-C1 worthy of your
attention in this regard. It's not intended to replace the Raspberry
Pi in it's envisaged role as a teaching aid, just as a better specced
SBC using very similar features to the R-Pi.


HOLD THE FRONT PAGE!

I've just spotted the announcement of the Raspberry - Pi 2 in csr-p
and taken a look at
http://www.raspberrypi.org/forums/viewtopic.php?f=48&t=61654

which details the improvements made over and above the B+ model.

This puts it more on a par with the ODROID-C1 I mentioned above
(except I haven't seen any mention of Gigabit ethernet). Even if this
first revision of the Pi 2 Model B is still cursed by a "Fast
Ethernet"(tm) lan port, the more powerful processor should be able to
extract more performance out of the USB connected ethernet port (we
might even see as much as 70Mbps out of the theoretical 100Mbps which
would be more than ample for simultaneously streaming 3 HD/4K movie
streams).

I have a sneaking suspicion that if it isn't already Gbit enabled,
the next revision will be - I can't find any mention of this anywhere,
not even when I followed the link to RS's Raspberry Pi pages where I
did, at least, discover its price in Pounds Sterling (21.88 of them, a
figure that suggests this is the pre-VAT price, implying a memorable
retail price of £26.26 - personally speaking, I'd prefer the more
memorable £20.20 figure but at least it isn't an eye watering
£50.50 :-)

It seems I may be investing in a Raspberry Pi after all. Whilst the
ODROID-C1 still has the edge over this latest R-Pi (for the moment),
the R-Pi has a wealth of projects including a few media streaming ones
to choose from which, for me at least, is an extremely important
consideration (the existance of an active news group and ready made
media streaming packages - no faffing about with web fora, I _hate_
web fora!).

HTH & HAND :-)
--
J B Good

On Sun, 01 Feb 2015 22:11:40 +0000, Johny B Good
wrote:

On Sun, 01 Feb 2015 09:56:23 +0000 (GMT), Jim Lesurf
wrote:

In article , Roderick Stewart
wrote:


Use the largest fans you can fit, and choose the type that are
automatically speed controlled from the motherboard (as most CPU fans
are anyway now).

Or choose a system that doesn't have any fans. The box I have in mind
simply has heatpipes to a substantial metal case that works as a decent
heatsink. Use SSDs and you should get a machine with no mechanical noises.



I've been wondering how good those little fanless micro-PC units really
are. They're sold for use as media centres, so presumably they can play
high-def video, but I can't help thinking that if they really could do
everything a full size PC can do, then that's what all PCs would be
like. I feel sure there must be a catch. Maybe that's why "smart" TVs
are so slow, as the extra electronics in those must be small and
fanless, so presumably similar.

I suspect the problem in hardware terms isn't simply the 'fanless' but the
absence of much efficient heatsinking. Indeed, on my TV there is no
external sign of any heatsink arrangements of slots for efficient airflow.


All the internet connected TVs and PVRs
I've seen can play the actual video perfectly well, but are
infuriatingly sluggish to respond to anything else.

Again, I suspect in part this is because they use chip archtechures
optimised for TV reception, and then the makers bodge together other
aspects on the "box ticked" basis.

FWIW given the way people are using RPi's as media devices I'm also
wondering about a newer series of ARM based boards that are more capable.
All being well, I should get one eventually for running RISC OS on, but
once that's done I'll also look at Linux on them.


In which case, Jim, I think you'll find the ODROID-C1 worthy of your

STOP PRESS!

Further to my heads up on the latest Rasberry Pi, I've just taken a
look at CPC's pricing on the Raspberry Pi 2 and its product
information.

Two points of note;

One, it's still stuck with a Fast Ethernet port and two, they're
retailing it for £3.67 more than RS's retail price. Since I can order
it via my business and RS do free postage for business customers on
orders valued at £20.00 or over (pre-VAT, of course!) that's my
cheapest option.

I don't know what RS charge ordinary customers for postage on a small
item like this but I suspect it will still be cheaper to order from
CPC due to orders valued above a fiver +VAT[1] are free postage to all
of their customers.

I'm almost tempted to bang an order into RS right now but I've waited
two or more years for a better value SBC option to arise so another
month or two is neither here nor there in the larger scheme of
things. :-)

[1] It seems they've raised their minimum from a fiver plus VAT to a
tenner (VAT inclusive price - ie £8.33 plus VAT) which is good news
for anyone ordering a Raspberry Pi 2 model B.
--
J B Good

On Mon, 02 Feb 2015 09:49:02 +0000 (GMT), Jim Lesurf
wrote:

In article , Johny B Good
wrote:
On Mon, 02 Feb 2015 01:27:47 GMT, Paul Ratcliffe
wrote:

On Sun, 01 Feb 2015 21:45:01 +0000, Johny B Good
wrote:



Properly written software by 'Real Programmers' using a less
abstracted compiled language such as C or, best of all, assembler,
will have all the necessary responsiveness you'd have every right to
expect out of such hardware resources.

With C you're always having to manage the language and worrying about
buffer overruns.

Can't say that has been one of my main worries during using 'C' for some
decades now.


With Assembler, you won't live long enough to write
anything of any size and debug it, and frankly there is no need to for
99.99% of code. C++ has to be the language of choice for native
compiled apps. these days.

Thanks for the 'heads up' on the other versions of C.

Not quite sure from the context what you mean here. But I wouldn't have
said 'C++' was a 'version' of 'C'. Its not like 'C#' or other wannabes.


C++ it is then which probably means my virtually unread C (or C+)
Programming guide that I acquired nearly two decades back will not be
much use to me then.

Actually, you can use quite a lot of 'C' in 'C++'. Indeed, most of the
'C++' programs I've written have been mostly the same as I'd have written
in 'C'. Just that I'd needed some of the features of C++. :-)

But then all my programs are really FORTRAN anyway! 8-]


I wonder if that's a BASIC reference (ISTR that BASIC was
derived/inspired from/by FORTRAN). I think I know what you mean. When
I was learning RPG my programming style was influenced by my Z80
assembler experience (you know what they say; "When the only tool you
have is a hammer, every screw looks like a nail.").
--
J B Good

"Johny B Good" wrote in message
...
my programming style was influenced by my Z80
assembler experience (you know what they say; "When the only tool you
have is a hammer, every screw looks like a nail.").

Somewhere I've got a Z80 program that I wrote for doing a bubble-sort on an
array of bytes. At the time, the BASIC interpreter on my CP/M PC was
painfully slow so I wrote a Z80 function that was passed the address of the
array (filled in the BASIC program) and the number of entries, and
bubble-sorted it; the BASIC program then used that sorted array (printed it
or whatever it did). And it did the sort considerably faster than the same
algorithm in BASIC.

The computer has long ago gone to the tip when it stopped even *trying* to
boot off the floppy, but somewhere I have a paper listing of the program. I
wonder if I could still write a bubble-sort in Z80: basically it involves
loading consecutive bytes from the array into two registers, comparing them
and swapping them if the first is greater than the second, then writing the
registers back to the array before moving on to the next.

On 29/01/2015 21:00, NY wrote:
"Paul Ratcliffe" wrote in message
...
On Thu, 29 Jan 2015 14:22:02 -0000, NY wrote:

At least TS (unlike MPG and MP4) supports
multiple streams so they preserve the subtitle stream.

Rubbish. If an MPG didn't support multiple streams, how do you think
you'd get pictures and sound?
Of course they are capable of subtitles, audio description etc.

I've not yet encountered an MPG or MP4 which has more streams that video
and (one or more) sound. Recording software that records to MPG always
gives just these, even if a subtitle stream is present in the incoming
transport from the broadcast.

In contrast, recording to WTV or TS preserves the additional subtitle
stream.

Conversion software such as VideoRedo ditches the subtitles when
converting to MPG/MP4 whereas it preserves it for TS.

I don't doubt that the MPG/MP4 standard allows extra streams, but in
practice software that I've encountered discards it when saving to these
formats.

Maybe what I should have said in my original comment is "At least
software that saves to TS (unlike MPG and MP4) preserves the subtitle
stream, whereas software that saves to MPG/MP4 often doesn't". That's a
bit less dogmatic!

I don't know for sure if MPG and MP4 are supposed to be able to contain
subtitles but MKV definitely can, and multiple audio too.

--

Brian Gregory (in the UK).
To email me please remove all the letter vee from my email address.

On Mon, 02 Feb 2015 09:49:02 +0000 (GMT), Jim Lesurf
wrote:

With C you're always having to manage the language and worrying about
buffer overruns.

Can't say that has been one of my main worries during using 'C' for some
decades now.

That would seem to say quite a lot about your programming skill or
attention to detail (or lack thereof).
Would you care to show me some sample code for concatenating two strings
of unknown length then, without the risk of buffer overruns?

Not quite sure from the context what you mean here. But I wouldn't have
said 'C++' was a 'version' of 'C'.

It's a superset of C. Everything you can do in C can be done in C++.

But then all my programs are really FORTRAN anyway! 8-]

OMG.