[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[linrad] Linrad & slew-rates; Re: [linrad] Homebrew I/Q-chains and Op-amps, compiling Linrad




Hi Roger!

It seems that the hash table is no show-stopper at present, according
to Leif's recent observations. Lasse/OH5XP advised me to use additional
commands like  xhost +localhost  as Knoppix does not allow direct root
logins like in the old pre-version4 days. In between I ran 'alsaconf'
for running sound. I get some xlinrad starts, but the following few
error messages I get consistently:

When trying to save the parameters (W):

[1029]routine:thread_main  file: menu.c
Failed to open/create the file par_user int

When trying to set sound configuration (U):

[1016]routine:set_analog_io  file:lsetad.c
----------------------------

I get stuck on these two errors; no Linrad operation yet.
I did not mix up the linrad/xlinrad options, but I wanted
to keep the 01-37 version with svgalib as a backup in case
that the 02-14 does not work in X. Well, both are still "0".


Noise and slew-rate (op-amps/W3SZ)

Regarding the selection of low-noise operational amplifiers, your
question is very valid. Even the simplest circuit deals with many
(often independent) parameters, and it is very hard to oversee the
full situation, and all the requirements. There is much more than
the noise-figure, and some parasitic effects don't show up in the
modeling, but finally only during use and measurement. Leif knows
the requirements much better regarding the Delta 44, but as I have
understood, the voltage swing at the input is quite high for full
scale resolution of the A/D-converters in the low-gain position.

Now if you go for 192 kHz sampling, and you want to make a clean
sine wave at near 100 kHz frequency, you need to kick the voltage
up to 10 volts in much less than quarter of the 100 kHz period,
let's say in 2 us... This converts to a slew rate of about 5V/us,
so you see that LT1128 does not make it, as it makes more a linear
ramp than a sine wave. The same problem is encountered with a RC4136.
Therefore for full cranking at the upper frequency limit with 192 kHz
sampling it is better to take something in the order of 20V/us, like
the AD797 or the much cheaper NE5532P. There is plenty of choice and
alternatives for the selection of this output operational amplifier.

The problem with me is that when I get in the circle of comparing
all those specs, I never get to the phase of building the circuit!
It is best to make the prototype, test its limits, and then improve
by replacing or redesigning the weakest links in the schematic.
However, very often we succeed in the first (sluggish) run.

                                   73,  "Zaba"  OH1ZAA/2


At 16:05 6.6.2006, W3SZ wrote:

Hi, Zaba, and all,

I am glad to see that you are making progress!  Building small receivers
is fun!

I wish I could help you with the hash table stuff, but I have no idea what
the problem is.  I have only the most basic idea of what a hash table is
or does.  For those like me who have only a vague idea of what a hash
table is, here is an interesting reference.  It won't help your problem
though, as its way to basic and general.

http://en.wikipedia.org/wiki/Hash_table

Since you mentioned possible op amps to use for the final stage of your
receiver I have a question for you and the group.

Around 2000 or so when I built my simple EME receiver to use with Linrad,
it looked like the AD797s had the best noise specs and so this is what I
used.

Now it seems that some of the cognoscenti tout the LT1128A and don't even
mention the AD797.

The typical noise at 1 kHz for the AD797 is 0.9 nV/sqrt[Hz] and that for
the LT1128 is 0.85, and max figures are 1.2 for the AD797 and 1.1 for the
LT1128, so the LT1128 is ever so slightly better in this respect.

The gain bandwidth of the AD797 is 110 MHz and that of the LT1128 is 20 I
think;  I guess that doesn't matter for our baseband amps.

The slew rate of the AD797 is 20 V/usec and of the LT1128 is 5V/usec.

So my question is:

For use as the 'last stage' of amplification before the soundcard for a
system sampling at up to 192 kHz, are these two chips pretty much
equivalent given the [to me] tiny differences in noise, or is the LT1128
clearly superior, and if so, why.  If the LT1128 would have advantanges
under some conditions, for some projects, what would those conditions or
projects be?

If there is not a big difference in performance for this purpose, why is
the LT1128 currently the 'ultra low noise op amp of choice'?  I don't
think it is cost, as from Digikey they both cost 8-9 USD in single
quantities.

I apologize in advance if this is a stupid question.  An electronics
wizard I am certainly NOT.  Hence the attempt to learn.

PS I got a really pleasant surprise yesterday when I tried to compile and
run Linrad 02.14 on an old P3 with a pre-Fedora version of Redhat Linux
installed and everything compiled and ran perfectly, including XLinrad!!

Thanks and have a great day, all, and

73,

Roger
W3SZ

On Tue, 06 Jun 2006 06:00:24 -0400, 'Zaba' OH1ZAA <zaba@xxxxxx> wrote:

Sir Roger & all!

Though I have the complete WSE-line here, I am still missing
the proper power supply (-15V/+15V) amperage to fire them up.

On the other side it is fun to build something small from time
to time, so I constructed the low noise I/Q-amplifier according
to SM5BSZ/Leif's Figure 2 in "Linux PC-radio. Optimised direct
conversion receiver for 144 MHz using standard Schottky diode
mixers" [http://www.sm5bsz.com/linuxdsp/hware/optiq.htm].

As it turned out the 2N2222A is an excellent choice for the
input-transistor. I had some 1980's stock, but ON Semiconductor
[http://www.onsemi.com/PowerSolutions/product.do?id=P2N2222A] is
listing this P2N2222A model with plots of noise figures below 1 dB
over a wide range of impedances and bias. I used the RC4136 for
the op-amps, but if you need to steer a Delta 44 to full range
with the 96 kHz setting, then  it may be better to take a NE5532
or a similar type for the output stages, with a higher slew-rate
than the RC4136.

--
Roger Rehr
W3SZ
http://www.nitehawk.com/w3sz/

#############################################################
This message is sent to you because you are subscribed to
 the mailing list <linrad@xxxxxxxxxxxxxxxxxxxxx>.
To unsubscribe, E-mail to: <linrad-off@xxxxxxxxxxxxxxxxxxxxx>
To switch to the DIGEST mode, E-mail to <linrad-digest@xxxxxxxxxxxxxxxxxxxxx>
To switch to the INDEX mode, E-mail to <linrad-index@xxxxxxxxxxxxxxxxxxxxx>
Send administrative queries to  <linrad-request@xxxxxxxxxxxxxxxxxxxxx>


#############################################################
This message is sent to you because you are subscribed to
 the mailing list <linrad@xxxxxxxxxxxxxxxxxxxxx>.
To unsubscribe, E-mail to: <linrad-off@xxxxxxxxxxxxxxxxxxxxx>
To switch to the DIGEST mode, E-mail to <linrad-digest@xxxxxxxxxxxxxxxxxxxxx>
To switch to the INDEX mode, E-mail to <linrad-index@xxxxxxxxxxxxxxxxxxxxx>
Send administrative queries to  <linrad-request@xxxxxxxxxxxxxxxxxxxxx>

LINRADDARNIL