Re: Follow-up Antuino question
Jerry, Contrary to a popular opinion, a free running LC oscillator can be the quietest there is. Any attempt that holding it to a frequency (phase locked loop) starts adding to the phase?or amplitude noise. Consider the phase correction being applied like a modulation (which it is), hence it will spread the signal from being a single solid carrier. That's the phase noise! The circuit that I used is pretty simple. Similar to the one at? (except that I used a tap on the inductor instead of a secondary?winding).? The section on VFOs in the EMRFD book is a thorough?text on building VFOs. To put it in a nutshell, VFO design is not as important as the building of it is. You must use polystyrene?capacitors if possible, or at least NP0 capacitors. Both of these are freely available?from Mouser. I have ordered a bad full of 100 pfs and I just keep paralleling them until I hit the values I need. I built it over a ground plane, I kept the regulator away to prevent its heat from heating up the rest of the components. Most importantly, there are no hanging wires at all. All components are directly soldered to each other and the entire VFO is mounted on the back of the tuning capacitor.? When the time permits, I will write up about the transceiver. It is a full-break in CW transceiver for 40 meters with two selectable filters for SSB and CW. It has 7 watts of CW power from an IRF510. It is very similar to the W7EL's transceiver. That article is one of those that pack so many things in that each time you read those few pages, you learn something new. It is available on?
- f
Farhan,
What do you think is the major factor in the relatively poor performance of the Si5351?
I suspect it is jitter, not the fact that it is a square wave.
The Si5351 is their cheapest part at $1, there are other Silicon Labs parts
with far better jitter performance at around $10 to $15.
Would be interesting to compare results using one of those.
Make sure the other Si5351 outputs are shut down to avoid crosstalk.
I've heard it reported in this forum that the Si5351 has less phase noise (one way of
thinking about jitter) than most analog VFO's.? So I could be wrong.? Or they could
be wrong.? Or there might be good kinds of jitter and bad kinds of jitter.
This might be a case where you want to run the Si5351 with integer divide
on the output multsynth dividers, setting the frequency by adjusting the PLL multisynth.
That will give significantly less jitter than using fractional divides on the outputs.
It's is with all the stuff from QRP Labs does, but since the Si5351 has only two
PLL's/VCO's, this approach gives only two outputs.
On the uBitx, we use a fixed PLL VCO frequency of 875mhz, and divide it down
to each output using fractional output divides so we can get three independent outputs
from one chip.? Perhaps the uBitx should use?VCOa for the VFO at CLK1, VCOb for
the the mostly fixed LO into the second mixer at CLK2, and a fractional output divide
of VCOb for the relatively low frequency (and thus low jitter) CLK0 that is used for the BFO.? ?
That would burn more Nano flash for the firmware, but might be worth exploring.
Jerry, KE7ER
On Sat, Nov 14, 2020 at 08:31 AM, Ashhar Farhan wrote:
Barry,
I too built a direct conversion transceiver a few months ago. I chose a fully analog design without any ICs.
A carefully designed free running vfo is used. It is so stable that I can continuously monitor the FT8 spot!
On the other hand, when I substitute it with an Si5351, the difference is immediately obvious. The CW is noiser and the SSB audio has less fidelity. I can no longer copy the signals that were barely above the noise floor.
Given a choice, I would use a free running VFO any day. Except that it is very expensive to build them with a good slow motion and implementing multiband design is always a?challenge.
- f
|
|
|
Gordon,
Any assembly can be cut into pieces. Advantage, one can change/ improve a module at its convenience.
The best system (in my opinion) is a board with main functions and a bus with several connectors (Like a pc)
Then you will be free to modify (for example, there was a lot of discussion on the AGC) --> It's will be and addon near the audio module.
Provide in this case an interconnexion by jumper on the mother board.
There¡¯s a lot to think about interconnexions...and others, ex:1 or 2 free slots also for personal fabrications, etc.
In short, this is no small matter.
But if you want to cut out a Ubitx, is that allowed??Only the author can tell you.
cdt
|
Re: Follow-up Antuino question
Farhan,
What do you think is the major factor in the relatively poor performance of the Si5351?
I suspect it is jitter, not the fact that it is a square wave.
The Si5351 is their cheapest part at $1, there are other Silicon Labs parts
with far better jitter performance at around $10 to $15.
Would be interesting to compare results using one of those.
Make sure the other Si5351 outputs are shut down to avoid crosstalk.
I've heard it reported in this forum that the Si5351 has less phase noise (one way of
thinking about jitter) than most analog VFO's.? So I could be wrong.? Or they could
be wrong.? Or there might be good kinds of jitter and bad kinds of jitter.
This might be a case where you want to run the Si5351 with integer divide
on the output multsynth dividers, setting the frequency by adjusting the PLL multisynth.
That will give significantly less jitter than using fractional divides on the outputs.
It's is with all the stuff from QRP Labs does, but since the Si5351 has only two
PLL's/VCO's, this approach gives only two outputs.
On the uBitx, we use a fixed PLL VCO frequency of 875mhz, and divide it down
to each output using fractional output divides so we can get three independent outputs
from one chip.? Perhaps the uBitx should use?VCOa for the VFO at CLK1, VCOb for
the the mostly fixed LO into the second mixer at CLK2, and a fractional output divide
of VCOb for the relatively low frequency (and thus low jitter) CLK0 that is used for the BFO.? ?
That would burn more Nano flash for the firmware, but might be worth exploring.
Jerry, KE7ER
toggle quoted message
Show quoted text
On Sat, Nov 14, 2020 at 08:31 AM, Ashhar Farhan wrote:
Barry,
I too built a direct conversion transceiver a few months ago. I chose a fully analog design without any ICs.
A carefully designed free running vfo is used. It is so stable that I can continuously monitor the FT8 spot!
On the other hand, when I substitute it with an Si5351, the difference is immediately obvious. The CW is noiser and the SSB audio has less fidelity. I can no longer copy the signals that were barely above the noise floor.
Given a choice, I would use a free running VFO any day. Except that it is very expensive to build them with a good slow motion and implementing multiband design is always a?challenge.
- f
|
Re: Follow-up Antuino question
Ashar, I agree that a analog vfo would be superior but I have not had much luck building one that was stable enough. That was the reason I chose to use the 5351 or my PLL. Any chance you would share your vfo design with us? Barry
On Sat, Nov 14, 2020, 11:31 AM Ashhar Farhan < farhanbox@...> wrote: Barry, I too built a direct conversion transceiver a few months ago. I chose a fully analog design without any ICs. A carefully designed free running vfo is used. It is so stable that I can continuously monitor the FT8 spot! On the other hand, when I substitute it with an Si5351, the difference is immediately obvious. The CW is noiser and the SSB audio has less fidelity. I can no longer copy the signals that were barely above the noise floor. Given a choice, I would use a free running VFO any day. Except that it is very expensive to build them with a good slow motion and implementing multiband design is always a?challenge. - f
On Sat 14 Nov, 2020, 9:55 PM barry halterman, < kthreebo@...> wrote: Hi fellow DC rx fans. Back in 1992, Rick Campbell had a design in QST for a direct conversion receiver he called the R1. Later he upgraded this to the R2. I have the R1 that I have used the si5351 to drive the SBL-1 mixer with excellent results. I use a 6 db pad on the LO port, per a recommendation from Ashar. Later I changed the LO to a PLL oscillator with a sine wave output. I did not notice any difference at all between the two oscillators. The recovered audio is fantastic from this R1 board. Rick had a few different LC designs for filters, SSB narrow and wide and a cw filter.? For those who want a serious DC receiver, I highly recommend looking at his design. It is a little more complicated then just a mixer and LM386, but very well worth the effort. K3BO
Would be interesting to try sine wave vs square wave.
But I doubt you will notice much difference,
Assuming you have a low pass (or band pass) filter on the RF input,
the harmonics in a square wave should not cause any first order products in the resultant audio.
Also, your mixer is likely to have a non-linear voltage vs current relationship
at the local oscillator port anyway.
However, the clean audio one can get from a DC receiver make it?
an excellent testbed for trying this sort of thing.
And I have heard that a sine wave is somewhat preferred for driving
even a diode ring mixer, which has an extremely non-linear LO port.
Note that the Antuino has a very nice ADE-1 mixer.
Might be possible to include that in your DC receiver.
Jerry, KE7ER
On Fri, Nov 13, 2020 at 08:35 PM, Bob Lunsford wrote:
There was some comment in the webpage about the signal being a square wave. If this shows up in the signal, causing harmonics, then included in the amp should be a "smoothing" circuit or some way to make it show up as a sine wave. This is theoretically simple since an inductor's flywheel effect would actually convert it to a sine wave. How well it does this is another question and it would/could cause some refinement of the signal. A simple resistor-capacitor circuit followed by an amp may be better.
?
|
Re: Follow-up Antuino question
Barry, I too built a direct conversion transceiver a few months ago. I chose a fully analog design without any ICs. A carefully designed free running vfo is used. It is so stable that I can continuously monitor the FT8 spot! On the other hand, when I substitute it with an Si5351, the difference is immediately obvious. The CW is noiser and the SSB audio has less fidelity. I can no longer copy the signals that were barely above the noise floor. Given a choice, I would use a free running VFO any day. Except that it is very expensive to build them with a good slow motion and implementing multiband design is always a?challenge. - f
On Sat 14 Nov, 2020, 9:55 PM barry halterman, < kthreebo@...> wrote: Hi fellow DC rx fans. Back in 1992, Rick Campbell had a design in QST for a direct conversion receiver he called the R1. Later he upgraded this to the R2. I have the R1 that I have used the si5351 to drive the SBL-1 mixer with excellent results. I use a 6 db pad on the LO port, per a recommendation from Ashar. Later I changed the LO to a PLL oscillator with a sine wave output. I did not notice any difference at all between the two oscillators. The recovered audio is fantastic from this R1 board. Rick had a few different LC designs for filters, SSB narrow and wide and a cw filter.? For those who want a serious DC receiver, I highly recommend looking at his design. It is a little more complicated then just a mixer and LM386, but very well worth the effort. K3BO
Would be interesting to try sine wave vs square wave.
But I doubt you will notice much difference,
Assuming you have a low pass (or band pass) filter on the RF input,
the harmonics in a square wave should not cause any first order products in the resultant audio.
Also, your mixer is likely to have a non-linear voltage vs current relationship
at the local oscillator port anyway.
However, the clean audio one can get from a DC receiver make it?
an excellent testbed for trying this sort of thing.
And I have heard that a sine wave is somewhat preferred for driving
even a diode ring mixer, which has an extremely non-linear LO port.
Note that the Antuino has a very nice ADE-1 mixer.
Might be possible to include that in your DC receiver.
Jerry, KE7ER
On Fri, Nov 13, 2020 at 08:35 PM, Bob Lunsford wrote:
There was some comment in the webpage about the signal being a square wave. If this shows up in the signal, causing harmonics, then included in the amp should be a "smoothing" circuit or some way to make it show up as a sine wave. This is theoretically simple since an inductor's flywheel effect would actually convert it to a sine wave. How well it does this is another question and it would/could cause some refinement of the signal. A simple resistor-capacitor circuit followed by an amp may be better.
?
|
Re: Follow-up Antuino question
Hi fellow DC rx fans. Back in 1992, Rick Campbell had a design in QST for a direct conversion receiver he called the R1. Later he upgraded this to the R2. I have the R1 that I have used the si5351 to drive the SBL-1 mixer with excellent results. I use a 6 db pad on the LO port, per a recommendation from Ashar. Later I changed the LO to a PLL oscillator with a sine wave output. I did not notice any difference at all between the two oscillators. The recovered audio is fantastic from this R1 board. Rick had a few different LC designs for filters, SSB narrow and wide and a cw filter.? For those who want a serious DC receiver, I highly recommend looking at his design. It is a little more complicated then just a mixer and LM386, but very well worth the effort. K3BO
Would be interesting to try sine wave vs square wave.
But I doubt you will notice much difference,
Assuming you have a low pass (or band pass) filter on the RF input,
the harmonics in a square wave should not cause any first order products in the resultant audio.
Also, your mixer is likely to have a non-linear voltage vs current relationship
at the local oscillator port anyway.
However, the clean audio one can get from a DC receiver make it?
an excellent testbed for trying this sort of thing.
And I have heard that a sine wave is somewhat preferred for driving
even a diode ring mixer, which has an extremely non-linear LO port.
Note that the Antuino has a very nice ADE-1 mixer.
Might be possible to include that in your DC receiver.
Jerry, KE7ER
On Fri, Nov 13, 2020 at 08:35 PM, Bob Lunsford wrote:
There was some comment in the webpage about the signal being a square wave. If this shows up in the signal, causing harmonics, then included in the amp should be a "smoothing" circuit or some way to make it show up as a sine wave. This is theoretically simple since an inductor's flywheel effect would actually convert it to a sine wave. How well it does this is another question and it would/could cause some refinement of the signal. A simple resistor-capacitor circuit followed by an amp may be better.
?
|
Re: An English site of a French HAM, with interesting information about Ubitx.
Thanks, Evan, you¡¯re always on the cutting edge of information.
I note this aside for my next changes
Have a good weekend
|
Thanks Evan for the help.
I had been searching the CEC sketch instead.
My TXR already uses a 2X16 display and I was hoping to upgrade to the Nextion.
My TXR uses 2 swithced? 2nd IF oscillators for the sidebands.
I may upload some photos of my radio later for those interested.
I will keep playing!
many thanks
Richard? VA3NDO
|
Re: Follow-up Antuino question
Bob: My spare Raduino worked fine for that purpose feeding a 4SQRP ZZRX-40 DC receiver. Heard all states from Florida on an indoor wire.
James KE4MIQ
|
Re: An English site of a French HAM, with interesting information about Ubitx.
Gerard,
I went back and looked again.? I believe that F5NPV is also using the standalone signal analyzer option from KD8CEC.? That is a second Nano connected via the I2C lines to provide the "small scanner" as well as some other features.? The code in the Raduino Nano would be UBITX_CEC_V1.200_NX_S.hex.? There is then another program for the second Nano.? Here is the KD8CEC page that describes it:
73
Evan
AC9TU
|
Re: An English site of a French HAM, with interesting information about Ubitx.
Hello,
Thanks
An other link
cdt?
|
Re: An English site of a French HAM, with interesting information about Ubitx.
|
Re: Follow-up Antuino question
Would be interesting to try sine wave vs square wave.
But I doubt you will notice much difference,
Assuming you have a low pass (or band pass) filter on the RF input,
the harmonics in a square wave should not cause any first order products in the resultant audio.
Also, your mixer is likely to have a non-linear voltage vs current relationship
at the local oscillator port anyway.
However, the clean audio one can get from a DC receiver make it?
an excellent testbed for trying this sort of thing.
And I have heard that a sine wave is somewhat preferred for driving
even a diode ring mixer, which has an extremely non-linear LO port.
Note that the Antuino has a very nice ADE-1 mixer.
Might be possible to include that in your DC receiver.
Jerry, KE7ER
toggle quoted message
Show quoted text
On Fri, Nov 13, 2020 at 08:35 PM, Bob Lunsford wrote:
There was some comment in the webpage about the signal being a square wave. If this shows up in the signal, causing harmonics, then included in the amp should be a "smoothing" circuit or some way to make it show up as a sine wave. This is theoretically simple since an inductor's flywheel effect would actually convert it to a sine wave. How well it does this is another question and it would/could cause some refinement of the signal. A simple resistor-capacitor circuit followed by an amp may be better.
?
|
Re: Follow-up Antuino question
There was some comment in the webpage about the signal being a square wave. If this shows up in the signal, causing harmonics, then included in the amp should be a "smoothing" circuit or some way to make it show up as a sine wave. This is theoretically simple since an inductor's flywheel effect would actually convert it to a sine wave. How well it does this is another question and it would/could cause some refinement of the signal. A simple resistor-capacitor circuit followed by an amp may be better.
Still thinking about it.
Bob ¡ª KK5R
On Friday, November 13, 2020, 11:22:05 PM EST, Ashhar Farhan <farhanbox@...> wrote:
bob,? i have used it as a local oscillator. but there is a catch The primary purpose of the RF out was to be able to measure the frequency response by feeding?the RF out into the device/circuit under test and measuring the output through the RF-In jack. Hence, the RF output is low, about -20 dbm. This is insufficient?to drive a mixer. There are two ways to fix this 1. You can remove the three attenuator resistors from the RF out pad and directly get about?+10 dbm from the RF out jack. This can damage?the RF In if fed directly. You can build an outboard attenuator for similar to the original RF out attenuator, but remember to use it each time you are testing a device. 2. You can build a two-stage Feedback amplifier with about 16 db gain in each stage with a 6 db attenuator pad between the two stages to provide a useful block of 26 db gain. This circuit will be useful in many other places as well. Use this to boost the RF output to the level needed by the mixer.
- f?
On Sat, Nov 14, 2020 at 8:58 AM Bob Lunsford via <nocrud222= [email protected]> wrote: After rereading the specifications for the Antuino and its oscillator stability, I suddenly wondered if the
oscillator can be used for the oscillator in a Direct Conversion
receiver... Perhaps there is a way to set it up that the antenna jack can be the source for the oscillator signal.
The Antuino has a digital readout and this promises
(theoretically) to be a shortcut and result in an excellent signal
oscillator for a DC receiver. A mixer and an LM-386 audio amp and the
receiver would be simplified and made much superior to an LC
controlled oscillator plus make it usable for a full range of
frequencies. Has anyone considered doing this?
Bob ¡ª KK5R
|
Re: Follow-up Antuino question
bob,? i have used it as a local oscillator. but there is a catch The primary purpose of the RF out was to be able to measure the frequency response by feeding?the RF out into the device/circuit under test and measuring the output through the RF-In jack. Hence, the RF output is low, about -20 dbm. This is insufficient?to drive a mixer. There are two ways to fix this 1. You can remove the three attenuator resistors from the RF out pad and directly get about?+10 dbm from the RF out jack. This can damage?the RF In if fed directly. You can build an outboard attenuator for similar to the original RF out attenuator, but remember to use it each time you are testing a device. 2. You can build a two-stage Feedback amplifier with about 16 db gain in each stage with a 6 db attenuator pad between the two stages to provide a useful block of 26 db gain. This circuit will be useful in many other places as well. Use this to boost the RF output to the level needed by the mixer.
- f?
On Sat, Nov 14, 2020 at 8:58 AM Bob Lunsford via <nocrud222= [email protected]> wrote: After rereading the specifications for the Antuino and its oscillator stability, I suddenly wondered if the
oscillator can be used for the oscillator in a Direct Conversion
receiver... Perhaps there is a way to set it up that the antenna jack can be the source for the oscillator signal.
The Antuino has a digital readout and this promises
(theoretically) to be a shortcut and result in an excellent signal
oscillator for a DC receiver. A mixer and an LM-386 audio amp and the
receiver would be simplified and made much superior to an LC
controlled oscillator plus make it usable for a full range of
frequencies. Has anyone considered doing this?
Bob ¡ª KK5R
|
Follow-up Antuino question
After rereading the specifications for the Antuino and its oscillator stability, I suddenly wondered if the
oscillator can be used for the oscillator in a Direct Conversion
receiver... Perhaps there is a way to set it up that the antenna jack can be the source for the oscillator signal.
The Antuino has a digital readout and this promises
(theoretically) to be a shortcut and result in an excellent signal
oscillator for a DC receiver. A mixer and an LM-386 audio amp and the
receiver would be simplified and made much superior to an LC
controlled oscillator plus make it usable for a full range of
frequencies. Has anyone considered doing this?
Bob ¡ª KK5R
|
Good news. Thanks. That's a clincher for me.
Bob ¡ª KK5R
On Friday, November 13, 2020, 10:10:57 PM EST, Evan Hand <elhandjr@...> wrote:
My Antuino came fully assembled in the enclosure shown.? The screen may be different depending on the version of software that is now being supplied by HF Signals
73
Evan
AC9TU
|
My Antuino came fully assembled in the enclosure shown.? The screen may be different depending on the version of software that is now being supplied by HF Signals
73
Evan
AC9TU
|
I am preparing to order the Antuino. I wonder if it comes as shown in the image on the web page (in a box) or does it come as shown in the Calibration section...
I'm hoping it comes as shown below OR if a box is offered as an extra.
Bob ¡ª KK5R
|
Gerard
F5NPV
Bob Lunsford