quisk starting to work...
My son suggests "Try running Linux as a guest OS in VirtualBox (hosted on the Mac). I can¡¯t make any promises, however.¡±
73
toggle quoted message
Show quoted text
Subject: Re: [softrock40] quisk starting to work...
Date: 21 April 2014 05:14:36 CEST
Hi Sid, I have been lurking into this group for several years, and I see how people refers to you as the Quisk expert for Linux builds. I am trying to build Quisk not on Linux, but on Mac OS X. Do you know if anybody has successfully built quick on OS X? My builds failed because I do not have ALSA on a Mac, but quick is supposed to work with port audio which is available on the Mac. I posted my question on the N2ADR group as N2AR is supposedly the one who wrote quisk, but all my posts were ignored. I just need to know what to change to use portaudio, as right now every build complains about a missing alas header.
73 de
Jean-Paul AC9GH @ EN61vr99 louijp ? at ? yahoo ? dot ?com
|
|
|
?
Without looking at the circuit, I would *guess* that the
resistors were placed there to suppress any spurious
oscillations.? It's also a place to measure current flow
in some circuits by measuring the voltage across
the resistor.
?
Charlie, N0TT
?
?
Well that was disappointing.......
Instead of paralleling a second
10 ohm resister across each of R30 and R31, I went for broke and simply
bridged them out of the circuit.
As nearly as I could tell it made no
difference at all in either the received signal level nor the level of the
noise floor. And I was watching both with 1 dB resolution...... there was no
change. So obviously I do not understand all I thought I knew about the
circuit.
So what would I change to add 6-10 dB of gain to the op amps
for a test?
Warren Allgyer
9V1TD
?
|
Well that was disappointing.......
Instead of paralleling a second 10 ohm resister across each of R30 and R31, I went for broke and simply bridged them out of the circuit.
As nearly as I could tell it made no difference at all in either the received signal level nor the level of the noise floor. And I was watching both with 1 dB resolution...... there was no change. So obviously I do not understand all I thought I knew about the circuit.
So what would I change to add 6-10 dB of gain to the op amps for a test?
Warren Allgyer
9V1TD
|
Re: RXTX sparking and arcing (still having waveform issues)
Here is the image I forgot to post.
|
Re: RXTX sparking and arcing (still having waveform issues)
Not having power issues any more.
Waveform issues happen in HDSDR and other SDR programs when I try to view the Waterfall image. ?I receive the attached image constantly, no matter what frequency I tune to in my super-band.
I connect the usb to the computer, the power cable to the power strip the computer runs on, and two 3.5mm audio cables between the board and a creative xfi pro.
I just want to know how to receive any signals at this point, then I will want to work on transmit!
|
Sent: Wednesday, April 23, 2014 9:22 AM
Subject: Re: [softrock40] Re: softrock40
?
?
Original Message -----
Subject: Re: [softrock40] Re:
softrock40
>
>
> Here are some tests done by DG8SAQ with
and without those resistors at the input to the mixer.
>
http://dg8saq.darc.de/switchimpedance/index.shtml
>
> This from the
above link: ¡°Apparently, by bridging R43, R44 one can obtain a match closer to
50
> Ohms while obtaining 30% more audio
signal.¡±
>
John,
Ah! I was thinking of the other 10 ohm
resistors in the op-amp inputs.
But will the reasons for and effects of these
resistors be similar?
And that reminds me that Tony began to use
lower-noise op-amps. That DID increase sensitivity.
73 Alan G4ZFQ
?
Well I¡¯ve modified the input to my RXTX.? The
input is now antenna coax feeding a 7 pole Chebychev band pass filter followed
by a Guanella balun feeding Minicircuits broad band transformer connected
directly to the mixer input.??? No ten Ohm mixer input resistors,
no tuned home wound transformer.? Rightly or wrongly the thinking being the
Minicircuits broadband input transformer would maintain a better balance than
the original home wound tuned transformer hopefully reducing the need for the
ten ohm swamping resistors.? The primary of the broad band input
transformer connects to the Guanella balun to decouple the unbalanced input from
the now broad band mixer input transformer.?? One side of the input side of the Guanella is grounded and connects
to the seven pole BPF.??
I did these mods a couple of years ago just for the
sake of experimentation and it seems to work ok with fewer receiver images on
the screen.? At work I now have access to a VNA but at the time of doing
the mods I didn¡¯t so couldn¡¯t make comparative tests.??? I am not
really considering doing anything more with the RXTX but am instead looking at
the Hermes lite with its direct digital conversion input.? From what I have
read it looks like a low cost way of trying DDC SDR
John G3UGY
|
Re: RXTX sparking and arcing (still having waveform issues)
What did you do that made things better?
What do you mean by 'waveform issues'?
Are you using an oscilloscope to observe a waveform?
?
If you are using a line powered scope, make sure that it is plugged into the SAME power strip as the RXTX power supply.
Otherwise, you can have ground loop currents flowing and the scope will add those to the waveform you observe.
If you are talking about something displayed on your computer screen by the software, please describe exactly what you are doing and observing.? Exactly what software are you using, how is the RXTX connected to your computer, and so forth.
|
I agree with you Alan. If I were changing anything I would do the op amp input resistors rather than the ones feeding the mixer. If I am reading the schematic correctly both are involved in setting the op amp gain but the ones before the mixer may also? have some effect on swamping the switching load. After the mixer is a high impedance point and only affects gain set...... I think!
My plan when I have test time is to set a baseline level for the system and then tack a second 10 ohm resister across each of R31 and 30 and then check the system gain and MDS. I will report back the results.
If someone has already done this please speak up and save me the trouble.... :-)
Warren Allgyer
9V1TD
|
Original Message ----- Subject: Re: [softrock40] Re: softrock40
Here are some tests done by DG8SAQ with and without those resistors at the input to the mixer.
This from the above link: ¡°Apparently, by bridging R43, R44 one can obtain a match closer to 50
Ohms while obtaining 30% more audio signal.¡±
John, Ah! I was thinking of the other 10 ohm resistors in the op-amp inputs. But will the reasons for and effects of these resistors be similar? And that reminds me that Tony began to use lower-noise op-amps. That DID increase sensitivity. 73 Alan G4ZFQ
|
?
?
Here are some tests done by DG8SAQ with and without those resistors at the
input to the mixer.? ?
?
This from the above link:? ¡°Apparently, by bridging
R43, R44 one can obtain a match closer to 50 Ohms while obtaining 30% more audio
signal.¡±
?
I tried it myself a couple of years ago but didn¡¯t make a
lasting note of the results.? Be interesting to read your
findings.
?
John G3UGY
?
?
toggle quoted message
Show quoted text
Sent: Wednesday, April 23, 2014 2:54 AM
Subject: [softrock40] Re: softrock40
?
?
VERY interesting observations Roger. I like your style!
My tests so
far do not indicate to me that the op amp noise is the limiting factor. I have
found I hit the noise floor of the sound card first and, frankly, I have not
documented the noise floor of the op amps. That is why I think the op amp gain
increase would work.
On the other side, at the high end, the limiting
factor seems to be the op amp power rails. I measure very low and very
consistent IMD as I increase input signal right up until the point that the op
amp output hits a little over 4.8 V P-P. Then, as the op amps begin clipping,
the IMD goes up dramatically and unacceptably as you might expect.
It
seems to me this barrier could be breached by running the op amps at a higher
supply voltage.
So here is my plan to add 5-6 dB of gain without
affecting the top end clip point:
1) Play with the input resistors as you
suggest to add the gain.
2) Jury-rig a 9.1 V zener (because I have a bunch in
the junk box) to power the op amps at 9.1 VDC rather than 5.0. This should add
5-6 dB of headroom and bring the clip level back to where is was before the gain
was increased.
I am not sure if I will run into a headroom issue on the
sound card input ....... I would not be surprised if I did. But it is worth a
try I think.
Warren Allgyer
9V1TD
|
Original Message ----- Subject: Re: [softrock40] Re: softrock40 The 10 ohm resistors have been discussed before. From memory I think you have covered all points:- 1) Prime purpose to swamp switch variations. 2) Forms part of the op-amp gain setting. 3) Does attenuate the RF signal. I wonder how much? I see it as being in series with the op-amp input impedance. Is that likely to be very low?
Interesting re. changing gain of the op amps. If the Rx II MDS is
limited by the noise floor of the op amps, then it is not clear to me
that changing the gain will affect this in a beneficial way. The spec
sheet for the op amp suggests to me that the circuit is well designed
and the noise is limited by the op amp itself, not by the gain resistors.
Warren and I discussed this earlier, I think without any agreement. Afterwards I did a few practical tests on 10m after. MDS figures mean nothing to me, I do not know the noise level at the antenna:) Softrock on dummy load raised soundcard noise floor, maybe 10dB. (EMU Asio 24 bit although lesser cards gave similar results, just a lower noise rise) Disabling mixer and/or LO made no difference to the noise level. On 10m connecting antenna showed no increase in noise. I setup a battery powered generator at the bottom of the garden underneath the antenna producing a weak signal about 10dB above the noise.. I checked the SNR. Then agsin with a 8dB preamp. This confirmed my antenna noise was low, 6dB increase in SNR. Back to no preamp. Reducung op-amp gain reduced SNR. Increasing op-amp gain, no increase in SNR. Getting the balance of gains correct is difficult. I have been told that reducing the gain of the op-amps will increase dynamic range. On the lower bands this could be true, but it seems to me the same can be achieved by using antenna attenuators. The more I look and read discussions I feel that the basic Softrock is not easy to improve. 73 Alan G4ZFQ
|
Roger
I replied to Milt before I read you post.
Please take the measurement after pulling the USB plug to the receiver. This effectively removes the mixer and anything behind it from contributing to the noise. I don't think it will make much difference as my tests square with yours: especially with a 16 bit sound card, the card limits the noise floor.
However I come to a different conclusion as a result: by increasing the op amp gain you will push the level of S+N up through the floor of the sound card and gain exactly what you want, a noise floor determined by the external environment rather than internal receiver components.
Is that as logical as it sounds when I write it?
WA
|
Milt You may well be correct. And I think it is pretty simple to verify it. In a real use case, if the noise floor rises when the antenna is connected, then nothing in the receive chain is limiting the MDS. In the case where the noise floor does not rise, such as when connected to a signal generator, then the limiting factors would be, in system order, front end filtering, mixer loss and noise, op amp noise, and sound card noise. So far I have not found a case where mixer or op amp noise were higher than the noise floor of the sound card.? But I admit that I have lost interest once the system noise floor dropped below -135 dBm or so because then it becomes an intellectual discussion with not much practical operational impact. I don't recall any testing I have done that has seen mixer noise rise to this level. But it an interesting question and will undoubtedly lead to insomnia and an irritated wife tonight. :-) Warren Allgyer 9V1TD
---In softrock40@..., <w8nue@...> wrote :
Hi Warren,
Interesting re. changing gain of the op amps. If the Rx II MDS is
limited by the noise floor of the op amps, then it is not clear to
me that changing the gain will affect this in a beneficial way.
The spec sheet for the op amp suggests to me that the circuit is
well designed and the noise is limited by the op amp itself, not
by the gain resistors.
However, if you are interested in experimenting, I have a thought.
Those 10 ohm resistors before and after the switching mixer seem
to me to be non-essential and possibly decreasing the sensitivity.
I was thinking of bypassing them myself as an experiment. I was
thinking of this from the perspective of eliminating some thermal
noise and simple resistive loss. However, eliminating them would
also increase the gain of the op amps since the input impedance,
together with the feedback impedance, determines the gain. The
mixer itself presents, I think, a source impedance of 12 ohms, so
eliminating the 20 ohms of resistors should increase the voltage
gain by X3 or so. I would be very interested to hear what, if any,
change in MDS measurement eliminating those resistors would make.
(A caveat, I am writing this after two glasses of wine.)
73,
Roger
Hi Warren and Roger,
I believe that the noise figure of the SoftRock is determined by the
conversion loss in the "mixer" and any front end filtering.? This
will determine the MDS.? A low noise amplifier ahead of the Softrock
mixer should reduce the noise figure and improve the MDS, but could
reduce the dynamic range unless the op amp gain is reduced.? The
UHFSDR by WB6DHW has a lower MDS, but utilizes a low noise amplifier
at the front-end.
I'm also of the opinion that the 10 ohms resistors serve to swamp
out variations in the ON resistance of the mixer switches.? With out
the resistors, there could be considerable variation in the I and Q?
overall gains.? This could present some problems with balancing the
gains (and phase) of the two channels.
Just my 2 cents.
Milt
W8NUE
|
Warren & Milt, re. noise limit. I measured, with an oscilloscope, the noise level of my Ensemble II op amp output (with no antenna connected) as 0.4 mV peak to peak (with a 4.5 kHz passive audio LP filter to eliminate the 2 mV LO rf noise coming thru. A 16 bit sound card analog to digital converter referenced to 1 V rms, 2.8 V pp, can theoretically sample down to 0.04 mV pp, if I have done the math right.
If this is correct, the limiting noise floor would not be that of a properly functioning 16 bit sound card and increasing the gain (and output noise level) of the SoftRock op amps would not improve the MDS. Please correct me if I have this wrong.
I suspect that some sound cards, and/or b/c of noise pick up from inadequate shielding, might not actually reach .04 mV as this would be down in the 3 least significant bits range, so in practice increasing the op amp gain might help. I guess it is worth trying.
Roger
|
Milt, thank you for the info on the I/Q balance issue. I had not thought of that. When I said, rather hastily, that I thought the 10 ohm resistors were non-essential, I was thinking from the point of view of the op-amps, and the audio frequencies. In addition to the issue that you raised, there is the matter of the load impedance presented to the band pass filters. I believe that the current SoftRock design is well conceived to present a 50 ohm load to the filters, as the filters were presumably designed for. If I understand the circuit, the 10 ohm resistor on the input side of the mixer (plus the mixer ON resistance of 4 ohms or so, X 4, gives an aprox. 52 ohm load to the filters. (The 10 ohm resistor after the mixer, if I understand, does not count for the RF load as the op amp side of the mixer is at zero ohms to RF b/c of the capacitor there.) Thus, eliminating the first 10 ohm resistor would upset the BP filter performance. Perhaps it would be best to try just eliminating the post mixer 10 ohm resistors as an experiment.
|
Hi Warren,
Interesting re. changing gain of the op amps. If the Rx II MDS is
limited by the noise floor of the op amps, then it is not clear to
me that changing the gain will affect this in a beneficial way.
The spec sheet for the op amp suggests to me that the circuit is
well designed and the noise is limited by the op amp itself, not
by the gain resistors.
However, if you are interested in experimenting, I have a thought.
Those 10 ohm resistors before and after the switching mixer seem
to me to be non-essential and possibly decreasing the sensitivity.
I was thinking of bypassing them myself as an experiment. I was
thinking of this from the perspective of eliminating some thermal
noise and simple resistive loss. However, eliminating them would
also increase the gain of the op amps since the input impedance,
together with the feedback impedance, determines the gain. The
mixer itself presents, I think, a source impedance of 12 ohms, so
eliminating the 20 ohms of resistors should increase the voltage
gain by X3 or so. I would be very interested to hear what, if any,
change in MDS measurement eliminating those resistors would make.
(A caveat, I am writing this after two glasses of wine.)
73,
Roger
Hi Warren and Roger,
I believe that the noise figure of the SoftRock is determined by the
conversion loss in the "mixer" and any front end filtering.? This
will determine the MDS.? A low noise amplifier ahead of the Softrock
mixer should reduce the noise figure and improve the MDS, but could
reduce the dynamic range unless the op amp gain is reduced.? The
UHFSDR by WB6DHW has a lower MDS, but utilizes a low noise amplifier
at the front-end.
I'm also of the opinion that the 10 ohms resistors serve to swamp
out variations in the ON resistance of the mixer switches.? With out
the resistors, there could be considerable variation in the I and Q?
overall gains.? This could present some problems with balancing the
gains (and phase) of the two channels.
Just my 2 cents.
Milt
W8NUE
|
VERY interesting observations Roger. I like your style!
My tests so far do not indicate to me that the op amp noise is the limiting factor. I have found I hit the noise floor of the sound card first and, frankly, I have not documented the noise floor of the op amps. That is why I think the op amp gain increase would work.
On the other side, at the high end, the limiting factor seems to be the op amp power rails. I measure very low and very consistent IMD as I increase input signal right up until the point that the op amp output hits a little over 4.8 V P-P. Then, as the op amps begin clipping, the IMD goes up dramatically and unacceptably as you might expect.
It seems to me this barrier could be breached by running the op amps at a higher supply voltage.
So here is my plan to add 5-6 dB of gain without affecting the top end clip point:
1) Play with the input resistors as you suggest to add the gain.
2) Jury-rig a 9.1 V zener (because I have a bunch in the junk box) to power the op amps at 9.1 VDC rather than 5.0. This should add 5-6 dB of headroom and bring the clip level back to where is was before the gain was increased.
I am not sure if I will run into a headroom issue on the sound card input ....... I would not be surprised if I did. But it is worth a try I think.
Warren Allgyer
9V1TD
|
Hi Warren,
Interesting re. changing gain of the op amps. If the Rx II MDS is limited by the noise floor of the op amps, then it is not clear to me that changing the gain will affect this in a beneficial way. The spec sheet for the op amp suggests to me that the circuit is well designed and the noise is limited by the op amp itself, not by the gain resistors.
However, if you are interested in experimenting, I have a thought. Those 10 ohm resistors before and after the switching mixer seem to me to be non-essential and possibly decreasing the sensitivity. I was thinking of bypassing them myself as an experiment. I was thinking of this from the perspective of eliminating some thermal noise and simple resistive loss. However, eliminating them would also increase the gain of the op amps since the input impedance, together with the feedback impedance, determines the gain. The mixer itself presents, I think, a source impedance of 12 ohms, so eliminating the 20 ohms of resistors should increase the voltage gain by X3 or so. I would be very interested to hear what, if any, change in MDS measurement eliminating those resistors would make.
(A caveat, I am writing this after two glasses of wine.)
73,
Roger
|
Thanks for the confirmation Roger and for the research. It is comforting
to know that my measurements are not off in left field as they
sometimes are.
The RXII with good software should not be
discounted as a top notch radio. And I think the fact that it is a
little low on gain is more of a concession to dynamic range than it is
to the inherent performance of the passive switching mixer.
David Turnbull
has advocated simply increasing the gain of the op amps rather than
adding a preamp and I have to say I tend to agree with him. Since the
entire receive chain prior to the op amps is passive it seems that
adding gain here would have exactly the same effect as adding a preamp
and, again as pointed out by David, without adding potential for
increased IMD.
And it is only a resistor change. The weekend
approaches and it is too hot to be outside, so maybe some
experimentation is in the offing.......
Warren Allgyer
9V1TD
|
Warren Allgyer