Starting SDR

[Ramiro Aceves]

Hi

I have just started building the SDR prototipe. You can watch the
picture at:

http://ea4nz.ure.es/sdr/SDR-1.jpg

On the left, AD9850 DDS signal source from previous experiments. In the
top, test virgin copper board with +-5V 7805 and 7905 power supply. At
the bottom, 74HC74 divide by four circuit. Close to it, BC547 transistor
just to rise RF voltage signal from the DDS. Without it the flip-flops
did not get excited enough.

74HC74 appears to work and generates 3.5 MHz clocks as I can hear on my
Elecraft K3 receiver tuned at 3.5 MHz. DDS is working at 3.5*4=14 MHz.
Going up beyond 17 MHz makes plenty of birdies on the receiver. DDS
internal clock is only40 MHz, Analog Devices recomends 1/3 of Fclock for
good quality.

So, 3.5 MHz band seems to be a good starting point to have some fun.

Now making trifilar winding for 74HC4052....

We keep in touch.

73, Ramiro. EA4NZ

[Earl Shaffer]

Hi Ramiro.

 

To make the 74HC74 more sensitive to drive, use bias resistors at the gate.

 

73, Jim Shaffer, WB9UWA.

--

Earl Shaffer,  WB9UWA

 

My EME array photos

http://www.gm4jjj.co.uk/WB9UWA/array_files/page33-1001-full.html

Detailed array photos

http://www.gm4jjj.co.uk/WB9UWA/

Facebook

http://www.facebook.com/earl.shaffer

Linkedin
http://www.linkedin.com/pub/earl-shaffer/12/881/735

[Ramiro Aceves]

El 07/04/11 21:15, Earl Shaffer escribi�:

> Hi Ramiro.
> To make the 74HC74 more sensitive to drive, use bias resistors at the gate.
> 73, Jim Shaffer, WB9UWA.

Thanks Jim, do you mean resitors at the clock input to raise level?

[Ramiro Aceves]

Hi all,

I am excited now. I have just switched on my simple SDR and it works! It
receives the 80m band very well with no filter at all. Also no OAmp at
all, only audio card preamp.

Some pictures of the ugly thing.

http://ea4nz.ure.es/sdr/SDR-2.jpg
http://ea4nz.ure.es/sdr/SDR-3.jpg

Now it is too late here, time to go bed, tomorrow I will post the
schematics and we can comment. In the middle of the 44.1 KHz passband
there are strong signals of 50 Hz hum and its harmonics. Also I can see
some 1 KHz spaced signals both sides. What are they?

Playing with Linrad 3.21 Suddenly it frecuently bombs out:

ramiro@debian-asus-64:~/linrad-03.21$ ./xlinrad64
xlinrad64: ../../src/xcb_io.c:249: process_responses: Assertion
`(((long) (dpy->last_request_read) - (long) (dpy->request)) <= 0)' failed.
Abortado

It makes that many times on SSB receive and makes it annoying. It seems
related to mouse movement.

See you

Ramiro, EA4NZ.

[Sid Boyce]

It looks like fun.
I'd like a copy of the circuit to try. Are you using a DDS?
73 ... Sid.

--
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
Senior Staff Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks

[Earl Shaffer]

Hi Ramiro.

 

Yes, on the clock input resistors.

Ground loops are very difficult to control on an SDR.

The ground plane is good for RF, but poor for ground loops.

The 50 hz is obvious. The 1 Khz might be a power supply, perhaps

the computer.

I found out the hard way the complexity of ground loops, but for starters

the antenna connector ground must be isolated from the audio ground.

Any ground potential at all will be heavily amplified in an SDR.

The feed to the audio amp is between the center winding and the switch output.

A shared ground at the center winding can be trouble.

Try using battery power to see if filtering needs to be improved. You may

need to repeat this test every time you improve ground loop control (less center hump).

Observe the level of the center hump with various parts of the circuit disconnected

and draw your own conclusions. When you have it right, the center bump is nearly gone.

 

If you wish to track down a source such as the 1 Khz noise, you can connect a pickup coil

in place of the SDR to the input of the PC. Place the coil near potential noise sources and

look for a harmonic pattern that matches the offending signal.

 

73, Jim Shaffer, WB9UWA.

[Edward R. Cole]

I discovered this working on my LP-Pan (simple SDRs) when trying to run two of them from a single 32.836 MHz xosc (to phase-lock both SDR),  I did not ground the shield of the coax lines to the pcb but only to the coax connector.  I tried splitting the output of the xosc but it appears to be a third overtone osc running 10.945 MHz with fund, 2nd, and 3rd harmonics all going to an op amp called a comparator.  Loading lowered drive too much to drive two op amps and using a mmic amp reduced the 3rd Harmonic level too much.

So I split the 32.836 MHz digital output of the op amp to clock the two IQ counters (provide 90-deg shifted LO for the quadrature mixer in each LP-Pan) using the mmic (MAR-4) to boost the level before splitting.  That is working nicely. 

But the point of not grounding RF to audio ground was noticed.  The RF input to the LP-Pan are floating above ground.  The audio output is either balanced or unbalanced for best noise performance.  LP-Pan provides several jumpers for choosing to ground (or not) for best performance.

Glad I am not designing SDR's!

73, Ed - KL7UW
32.836 = 4x 8.209 MHz,  8.209 is 6-hz offset of the K3 IF of 8.215 to lower LO feed thru or improve image response (not sure which).

73, Ed - KL7UW, WD2XSH/45
======================================
BP40IQ   500 KHz - 10-GHz   www.kl7uw.com
EME: 144-1.4kw, 432-100w, 1296-testing*, 3400-winter?
DUBUS Magazine USA Rep dubu...@hotmail.com
======================================

[Ramiro Aceves]

Hello again,

The schematics are on my WEB page at:

http://ea4nz.ure.es/sdr/SDR-EA4NZ-NACHROD.pdf

73, Ramiro. EA4NZ

El 08/04/11 08:29, Edward R. Cole escribi�

[Ramiro Aceves]

El 08/04/11 06:15, Earl Shaffer escribi�:
> Hi Ramiro.

Hi Jim,

> Yes, on the clock input resistors.

fine, will try next week, today I go on a trip.

> Ground loops are very difficult to control on an SDR.
> The ground plane is good for RF, but poor for ground loops.
> The 50 hz is obvious. The 1 Khz might be a power supply, perhaps
> the computer.

1 KHz signals are present even with the SDR off. I will investigate that!

> I found out the hard way the complexity of ground loops, but for starters
> the antenna connector ground must be isolated from the audio ground.

I can do it very easily removing the ground at toroid input winding.
Will try that. Also I should place the 9V power supply transformer in
other place, far from the receiver. I imagine that when I place de OAmps
the problems will get worse.

> Any ground potential at all will be heavily amplified in an SDR.
> The feed to the audio amp is between the center winding and the switch
> output.
> A shared ground at the center winding can be trouble.
> Try using battery power to see if filtering needs to be improved. You may
> need to repeat this test every time you improve ground loop control
> (less center hump).

Can not test batteries at the moment cause I need dual polarity power
supply. I will try with spare batteries from another QTH when I come
back on sunday.

> Observe the level of the center hump with various parts of the circuit
> disconnected
> and draw your own conclusions. When you have it right, the center bump
> is nearly gone.

Yeah, I will try and error.

> If you wish to track down a source such as the 1 Khz noise, you can
> connect a pickup coil
> in place of the SDR to the input of the PC. Place the coil near
> potential noise sources and

Good suggestion!

We keep in touch. Many thanks!

73, Ramiro. EA4NZ

[Ramiro Aceves]

El 08/04/11 01:44, Sid Boyce escribi�:

>
> It looks like fun.
> I'd like a copy of the circuit to try. Are you using a DDS?
> 73 ... Sid.

Hi Sid!

I am using a DDS because I do not have anything better at hand. But I am
going to use Si-570 chip in the near future. If you want schematics of
my DDS and even the PIC program I can publish it in the WEB page. There
is no PCB designed at the moment.

Ummm, I am using the DDS ___without____ band-pass filter. I should fix
that soon!

73!, Ramiro. EA4NZ

>

[Sid Boyce]

> El 08/04/11 08:29, Edward R. Cole escribi�:
>

Thanks very much indeed.
What are the transformer winding details, core type and trifilar turns.

[Sid Boyce]

That would be helpful. I could try a si570 and perhaps a spare si570
Controller I have lying around.

[Leif Asbrink]

Hi Ramiro,

> Playing with Linrad 3.21 Suddenly it frecuently bombs out:
>
> ramiro@debian-asus-64:~/linrad-03.21$ ./xlinrad64
> xlinrad64: ../../src/xcb_io.c:249: process_responses: Assertion
> `(((long) (dpy->last_request_read) - (long) (dpy->request)) <= 0)' failed.
> Abortado
>
> It makes that many times on SSB receive and makes it annoying. It seems
> related to mouse movement.

When searching the Internet I found:
I have an application that uses direct calls to X. This application
worked with a previous incarnation of X (before X.org) but crashes
with the following message on newer versions of X. I've tried 7.4+20 and
7.5+5, both Debian on a x86 machine.

xcb_io.c:249: process_responses: Assertion `(((long)
(dpy->last_request_read) - (long) (dpy->request))<= 0)' failed

------------------------ snip ---------------------------------

The main thread blocks with a the call XNextEvent(...). A secondary
thread sends a user event to X with XSentEvent(...). The application
immediately crashes on the this first call to XSendEvent with the above
error message. A normal event such as a mouse click generated by the
Xserver does not cause a crash.

XInitThreads() is called before any other X calls. All calls to X are
bracketed with XLockDisplay() and XUnlockDisplay() except for the
XNextEvent call in the main thread. As I mentioned, the app has worked
for several years.
-------------------------------- snip ------------------------

I can not interpret what this means for Linrad and I have never
seen this error myself.

What Linux distribution do you use?

73

Leif / SM5BSZ

[Ramiro Aceves]

El 08/04/11 16:36, Leif Asbrink escribi�:
> Hi Ramiro,

Hi Leif,

I use Debian Stable 64 bits.

Thanks

73, Ramiro. EA4NZ.

[Leif Asbrink]

Hi Ramiro,

> Playing with Linrad 3.21 Suddenly it frecuently bombs out:
>
> ramiro@debian-asus-64:~/linrad-03.21$ ./xlinrad64
> xlinrad64: ../../src/xcb_io.c:249: process_responses: Assertion
> `(((long) (dpy->last_request_read) - (long) (dpy->request)) <= 0)' failed.
> Abortado
>
> It makes that many times on SSB receive and makes it annoying. It seems
> related to mouse movement.

You say this is Debian Stable. I have been running linrad-03.22,
xlinrad64 for hours on my laptop with Centrino Duo.

I have not been able to provoke the error you report.
Neither with nor withoput MIT-SHM.

Please pack your par_* files and send them together with
a description of your hardware. Also a list of running
processes: ps aux > processes.txt and a list of selected
packages dpkg --get-selections > selections.txt

Also info on your motherboard. How many CPU cores etc.

I will try to make a similar setup with a hope to
be able to reproduce this error.

73

Leif / SM5BSZ

[Ramiro Aceves]

Hi Leif,

Many thanks for your help and efforts. I am now 200 km from Madrid and I
am not at my radio shack (I am in Valladolid to visit my parents). I do
not know if it makes sense to send you a processes list now that my
audio card is not with me. Anyway, I send you my packed files to your
personal mail (I do not know if files can be attached to this email
reflector)

Sound card is Maudio Fast Track Pro USB. Linux driver is snd-usb-audio.
I am using a Asus laptop with Intel(R) Core(TM)2 Duo CPU T5450 @
1.66GHz, 2GB RAM. Two processors.

I will send in a few minutes the information you asked to me.

73! Ramiro. EA4NZ

El 09/04/11 15:28, Leif Asbrink escribi�:

[Ramiro Aceves]

Hi

Preliminary tests of my earlier SDR receiver shows that it needs a RF
front end filter as a must. It receive not only in the 3.5 MHz band,
also multiples of the LO frequency as Leif stated.

1 KHz spurs come from the laptop power supply, it is a very dirty one
(swithed, I asume). It also adds plenty of noise into the receiver. 50
Hz and harmonics come from my shack main 30A 13.8 V power supply
(linnear, home made).

My audio card has got two preamps with two gain control knobs. These
controls are useful in removing the image signals in the receiver. Any
way, I think that I can not get more than 35 dB of image suppression
playing with that knobs.

I am going to desing a simple 3 order filter for 3.5 MHz. QUCS software
for Linux comes with a filter synthesizer tool. Very usefull for
designing band pass, low pass, and high pass filters of any order and
kind. It also has got microstrip tools for the MW fans.

Every simulation and desing of the receiver are intended to be donde
under the Linux operating system using free software tools. Kicad for
PCB and schematics, QUCS for analog simulation, TKGATE for digital
simulation and PIKLAB for microncontroller programming.

Will post the results on the WEB. Nothing new for you but perhaps useful
for the newbies.

73 Ramiro. EA4NZ.

[Sid Boyce]

Interesting, I have those tools installed but have yet to use other than
kicad and a quick look at qucs where I did a filter synth which said it
was sent to the clipboard ... I haven't found out yet how to view the
clipboard.
73 ... Sid.

--
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support

[Ramiro Aceves]

El 12/04/11 17:52, Sid Boyce escribi�:

> On 12/04/11 11:18, Ramiro Aceves wrote:

>
> Interesting, I have those tools installed but have yet to use other than
> kicad and a quick look at qucs where I did a filter synth which said it
> was sent to the clipboard ... I haven't found out yet how to view the
> clipboard.

Fine, the same happend to me the first time. DO not close the filter
design window. Go to the schematic window and "paste" the schematic. You
are done.

> 73 ... Sid.
>

[Sid Boyce]

Thanks, done. I shall be using it more as I've tried other tools that
aren't as useful

[Ramiro Aceves]

Hello

QUCS is not mature nor finished but it looks promising in the near
future. I use it cause it is the most easy to use Linux simulation
software at this moment. I used ngspice in the past but was very tedious.

I also recommend KICAD for schematics and PCB design.

Good luck.
73, Ramiro, EA4NZ.

El 12/04/11 23:05, Sid Boyce escribi�:

[Sid Boyce]

Thanks, I also use kicad and before that pcb.
73 ... Sid.

[Jeffrey Owen Katz]

Hi,

I use good old gschem (part of the geda suite) for schematic capture. It
works very well, has a lot of features, and includes a very nice component
library (even the old TTL 74XX logic family is there). For simulation I
tend to use Fortran or C++, but recently started learning gnucap which
looks like a nice package. I have been doing electronics on and off for
well over 50 years (remember the great RBC-3 with its 6-pole tuned front
end? Never tested it, but I bet its IP3 was terrific) and am just getting
started in SDR, so this thread is of great interest.

Jeff, AC2BQ

On Sat, 16 Apr 2011, Sid Boyce wrote:

> Thanks, I also use kicad and before that pcb.
> 73 ... Sid.
>
> On 15/04/11 10:37, Ramiro Aceves wrote:
>> Hello
>>
>> QUCS is not mature nor finished but it looks promising in the near
>> future. I use it cause it is the most easy to use Linux simulation
>> software at this moment. I used ngspice in the past but was very tedious.
>>
>> I also recommend KICAD for schematics and PCB design.
>>
>>
>> Good luck.
>> 73, Ramiro, EA4NZ.
>>
>>
>>

>> El 12/04/11 23:05, Sid Boyce escribió:

>>> On 12/04/11 17:08, Ramiro Aceves wrote:

>>>> El 12/04/11 17:52, Sid Boyce escribió:

[Sid Boyce]

Welcome aboard. I've been involved in electronics for over 53 years,
licensed going on 46 years. Until 1992 I always built my own gear
including synthesized transceivers and most of my test equipment. When I
saw the prices of ham gear in the USA, I couldn't help myself bring a
TS450S back in 1992 and a IC-737 in 1993.

One UK ham told how he had a fly-drive holiday in Orlando and brought
back a TS-850S, everything including the holiday cost him less than
buying the rig from the shop down the road from his home.

I slackened off building for some years when I did much heart searching
while thinking and re-thinking of buying a Flex-radio 1000 and while
googling happened on Softrock, ordered a SR v6.3. Since then the
commercial gear, as I've often said here, is there to keep the shelves
free of dust.

I always have gEDA installed, so there are many free tools to choose from.
73 ... Sid.

On 16/04/11 21:18, Jeffrey Owen Katz wrote:
> Hi,
>
> I use good old gschem (part of the geda suite) for schematic capture. It
> works very well, has a lot of features, and includes a very nice
> component library (even the old TTL 74XX logic family is there). For
> simulation I tend to use Fortran or C++, but recently started learning
> gnucap which looks like a nice package. I have been doing electronics on
> and off for well over 50 years (remember the great RBC-3 with its 6-pole
> tuned front end? Never tested it, but I bet its IP3 was terrific) and am
> just getting started in SDR, so this thread is of great interest.
>
> Jeff, AC2BQ
>
>
> On Sat, 16 Apr 2011, Sid Boyce wrote:
>
>> Thanks, I also use kicad and before that pcb.
>> 73 ... Sid.
>>
>> On 15/04/11 10:37, Ramiro Aceves wrote:
>>> Hello
>>>
>>> QUCS is not mature nor finished but it looks promising in the near
>>> future. I use it cause it is the most easy to use Linux simulation
>>> software at this moment. I used ngspice in the past but was very
>>> tedious.
>>>
>>> I also recommend KICAD for schematics and PCB design.
>>>
>>>
>>> Good luck.
>>> 73, Ramiro, EA4NZ.
>>>
>>>
>>>

>>> El 12/04/11 23:05, Sid Boyce escribi�:

>>>> On 12/04/11 17:08, Ramiro Aceves wrote:

>>>>> El 12/04/11 17:52, Sid Boyce escribi�:

[Ramiro Aceves]

Thanks for sharing your experiences.

I have uploaded current information of my "reinventing the wheel SDR" at:

http://ea4nz.ure.es/sdr/sdr.html

Enjoy.

Ramiro. EA4NZ.

El 17/04/11 04:07, Sid Boyce escribi�:

[Sid Boyce]

Nice work, keep us up to date with developments.
73 ... Sid.

[Ramiro Aceves]

Hi all

I am starting RF filter design. Not a trivial task for a begginer as I
tought at first glance.

Of course the best solution to get performance was Leif first proposal
of BC capacitors and tuned filters on the working receiving frequency. I
have discarded that solution because of the dificulty of tunning, even
using varicap diodes. Controlling the receiver remotely will be a
difficult task if we use that solution.

Doing a filter for each amateur band and frequencies inbetween is a hard
work. So, trying to keep things simple to start fun, I was thinking the
following. The problem with this receiver is that it receives in odd
hamonics of the LO, and some even harmonic too. The "idea" is to avoid
receiving frecuencies 2 times LO frecuency. So I am thinking in doing
four filters, each one having Fmax<Fin*2.

What do you thing about this arrange of filters:

1.8-3 MHz
3-5.5 MHz
5.5-9.7 MHz
9.7-17MHz
17-30 MHz

I am thinking about Butterworth 3rd order filters.

Please, let me know whether it makes sense and give me sugerences or
improvements before starting desinging with real parts.

Many thanks.

Ramiro.

[Earl Shaffer]

Hi Ramiro.

 

I suggest a test first.

Run the SDR as you normally would without filters.

Inject a signal at the harmonic frequencies to see how sensitive

it is to those harmonics. Run the signal at a level that matches

the largest signal you expect to reject. Attenuate the signal until

it is no longer a bother. The attenuation is the amount of attenuation

you would need from your RF filters. Something between 60 and 100 db

might be reasonable depending on expected performance.

Mine operates on a fixed frequency band, so I am using a 3 pole bandpass

filter with substantial selectivity.

After running this test, you may easily conclude the tunable bandpass filter is

a better option after all. You could use dual section capacitors to tune two

bandpass filters together. Because they may not track, you can add a single

tunable capacitor to fine tune the filter in use. If you use a series resonant

configuration with capacitors to ground between as I have, the bandpass filter

will have higher attenuation above resonance. This filter degrades into a low pass

filter well above resonance. The value of capacitors to ground determines the tradeoff

between insertion loss and selectivity. They also set the impedance of the filter.

Spice models this filter very well.

 

 Your upper frequency selection of 3 Mhz will place

stiff demands on your low pass filter considering 80 meter ham signals will be very strong.

 

73, Jim Shaffer, WB9UWA.

[Sid Boyce]

Are you thinking along the lines of the YU1LM tuneable HF BPF described
at http://ebookbrowse.com/bp-filter-hf-yu1lm-pdf-d37539060
73 ... Sid.

--
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support

Senior Staff Specialist, Cricket Coach

[Earl Shaffer]

Hi Sid.

 

This looks like a good start. The configuration is similar to what I suggest.

It might be ok if 30db of harmonic suppression proves to be sufficient.

That might be better than what a simple low pass filter would provide.

My filters also have capacitors to ground at both the input and output ports

and other than that it is about the same.

The extra capacitors from input and output to ground adds more harmonic suppression.

The Q of my coils I think are around 300. Insertion loss is below a db.

 

73, Jim Shaffer, WB9UWA.

[Sid Boyce]

I came across the sort of filters you described a while ago except they
were fixed for individual bands rather than tuned. I must go back and do
a search.
73 ... Sid.

[Ramiro Aceves]

El 18/04/11 16:06, Earl Shaffer escribi�:

> Hi Ramiro.
> I suggest a test first.

Fine suggestion Earl. The problem is that I do not have a signal
generator. The only thing I have is the DDS used to drive the LO in the
SDR and a simple dip meter, so it is difficult making useful
measurements. I am considering adquiring one in the near future cause
designing without test equipment is not possible, it is a nightmare itself.

I have done a quick and dirty test with the dip meter. Tuning the dip
meter at 3.5 MHz in the proximity of the receiver turns out into a 80 dB
signal on Linrad screen. A 10.5 MHz signal generates a close level value
on the screen (and plenty of spurs near de signal). 7 MHz signal results
a 30 dB weaker signal. So, it seems that the receiver is more sensitive
to the odd harmonic.

Another thing "discovered" is that if I adjust carefully input levels at
the sound card to get an image rejection, say 50 dB, this adjustment
gets destroyed when receiving the harmonics.

First thing I am going to do is a low pass filter for the DDS to avoid
spurs and then continue the tests.

73 and many thanks
Ramiro, EA4NZ.

[Ramiro Aceves]

El 18/04/11 16:59, Sid Boyce escribi�:

> Are you thinking along the lines of the YU1LM tuneable HF BPF described
> at http://ebookbrowse.com/bp-filter-hf-yu1lm-pdf-d37539060
> 73 ... Sid.

Thanks Sid, very interesting article!

[Ramiro Aceves]

Hello again,

More improvements were done on the RTW-SDR receiver:

-Low pass filter for DDS signal.
-Resistor voltage divider to improve FLIPFLOP sensitivity. Thanks to
Jim, WB9UWA for the suggestion.
-DDS clock improved 40 MHz--> 125 MHz.

WEB page was updated with new pictures and an index to see the evolution
of the receiver.
Enjoy, we will continue reinventing the wheel and having fun!

http://ea4nz.ure.es/sdr/sdr.html

Ramiro.
EA4NZ.

[Sid Boyce]

Looking at the pictures and the circuit diagram - U1/U2, are they 2
separate chips or just 1 74HC74?
73 ... Sid.

[Ramiro Aceves]

El 20/04/11 14:40, Sid Boyce escribi�:

>
> Looking at the pictures and the circuit diagram - U1/U2, are they 2
> separate chips or just 1 74HC74?

just one chip

> 73 ... Sid.
>

[Sid Boyce]

OK, if redoing the circuit you can relabel them as U1A and U1B.

[Earl Shaffer]

Hi Ramiro.

 

Credit goes to SM5BSZ for the voltage divider at the input of the flip-flop.

I learned a LOT from him as I was developing my own version of his WSE.

My biggest knowlege gain came in understanding ground loops better :)

They are very important to avoid in this project. My aproach is therefore modular.

 

73, Jim Shaffer, WB9UWA.

[Ramiro Aceves]

El 20/04/11 17:24, Earl Shaffer escribi�:
> Hi Ramiro.

Hi

> Credit goes to SM5BSZ for the voltage divider at the input of the flip-flop.

Thanks to Leif!!

> I learned a LOT from him as I was developing my own version of his WSE.
> My biggest knowlege gain came in understanding ground loops better :)
> They are very important to avoid in this project. My aproach is
> therefore modular.
> 73, Jim Shaffer, WB9UWA.

Thanks Jim

[Ramiro Aceves]

El 20/04/11 15:11, Sid Boyce escribi�:

>>
> OK, if redoing the circuit you can relabel them as U1A and U1B.

Hi Sid, I am very sorry. On the screen I can see U1A and U2B, it is an
error on my side. I have removed U2 and copied from U1 and changed to
unit B. Now in the screen I can see U1A and U1B, but when printing into
PS format the "A" and "B" letters disapear. I do not know what I am
doing wrong and why the letters diasppear when printing. I will
investigate. I have just changed the picture on the WEB but as you see,
it shows U1 and U1 instead of U1A and U1B.

73, Ramiro

>
> 73 ... Sid.
>
>

[Earl Shaffer]

Hi Ramiro.

 

Your current schematic does not show the voltage divider at the input of the flip-flop.

 

73, Jim.

On Thu, Apr 21, 2011 at 3:25 AM, Ramiro Aceves <ea1...@gmail.com> wrote:

El 20/04/11 17:24, Earl Shaffer escribió:

[Ramiro Aceves]

El 21/04/11 15:46, Earl Shaffer escribi�:

>
> Hi Ramiro.
> Your current schematic does not show the voltage divider at the input of
> the flip-flop.
> 73, Jim.

Oooooppppss!.

Thank you Jim! I have just updated the picture.

73, Ramiro. EA4NZ

[Sid Boyce]

Regarding the DDS, are you using the AD9850 or DDS-60 kit?
73 ... Sid.

[Ramiro Aceves]

El 21/04/11 18:32, Sid Boyce escribi�:

> On 21/04/11 17:16, Ramiro Aceves wrote:
>> El 21/04/11 15:46, Earl Shaffer escribi�:
>>>
>>> Hi Ramiro.
>>> Your current schematic does not show the voltage divider at the input of
>>> the flip-flop.
>>> 73, Jim.
>>
>> Oooooppppss!.
>>
>> Thank you Jim! I have just updated the picture.
>>
>> 73, Ramiro. EA4NZ
>>
>
> Regarding the DDS, are you using the AD9850 or DDS-60 kit?
> 73 ... Sid.

Hi Sid.

I am using an own home made version of AD9850 DDS with 16F84 microchip
microcontroller, dial encoder and 2x16 char LCD display. Firstable,
internal clock was 40 MHz, I have just changed it to 125 MHz. That
project was another "reinventing the wheel" one, hi. The DDS schematic
is here now:

http://ea4nz.ure.es/sdr/dds-schematic.jpeg

Assembler source code program and HEX file for programming is available
for anyone who request it.

In a future I will do a PCB board for it, but now my preference is the SDR.

73! Ramiro, EA4NZ

[Sid Boyce]

Thanks Ramiro,
I had to hunt around for a source of the 74HC chips yesterday, found 1
source for each and ordered a few of each.

Initially I'll go for the K5BCQ si570 Controller which I have lying
around. I just need to add a si570 to the board.

Many other projects on the go at present, but I think I can get it
together in the next couple of weeks.

[Ramiro Aceves]

Hello dear friends,

I have changed DDS low pass filter (60 MHz cuttof freq now) and now the
RTW-receiver works in 3.5, 7, 10 and 14 MHz with no RF filter at all.
It exhibit some spurs problems with strong broadcast stations and
unwanted signals due to the lack of filtering at all.

Last changes of my poor man SDR at http://ea4nz.ure.es/sdr/sdr.html

I have found some AM detection problems. You can see them at:

http://ea4nz.ure.es/sdr/AM.png

The AM signals does not depend on any DDS tuned frequency. My question
is: Is AM detected on the 74HC4052 switching mixer or it occurs inside
Maudio sound card? Any recomendations about it?

I am preocupied about AM detection because it will be a problem if I
want to listen to broadcast bands.

Another interesting subject for me to learn is the calculation of the
maximum frequency 74HC4052 can work. It appears to be working at 14 MHz
but not shure where the limit is and wether it is working not well (have
to compare with my HF rig to see if there is a big difference). I cant
check at higher frequencies cause I have reached my DDS frequency limit now.

I am doing the following calculations, not sure if I am doing it right
(too much time from university calculations and I have forgot near
everything about digital circuitry...).

74HC4052 datasheet is at http://ea4nz.ure.es/sdr/74HC4052.pdf

74HC4052 "typical" specs a VCC=5V, Cload=15 pF
Propagation delay IN to OUT, TPHL, TPLH= 4ns "typical"
Switching OFF times, TPHZ, TPLZ=21 ns "typical"
Switching ON times, TPZ, TPZL=27 ns "typical"

If we imagine the multiplexer at a fixed position, with a RF signal
passing through it, I believe that TPHL and TPLH are the important
values. So Fmax=1/(2*TPHL)=1/(2*4ns)=125 MHz. (we indeed use the higher
value, TPLH or TPHL. In this case they are the same). The datasheet show
a graph called "channel on bandwith" and we can see a curve that goes
down quickly in the proximitiy of 100 MHz. So it seems to match with the
calculations. Please, correct me if something is wrong. But that is not
the way the mixer works.

Now the switching specifications. I am not sure about how calculations
must be made, have not found anything in the web, but this is what I
think. We are controlling the multiplexer by feeding two square F
frequency signals 90 degree dephased at S0 and S1 inputs on the 74HC4052
chip. So the multiplexer changes its position 4 times during full cycle
at F frequency: position 1, 2, 3 and 4. For a choosen channel position,
for example, position 1, it will go through the following states: ON,
OFF, OFF, OFF.
So, maximum frequency should be Fmax=1/(4*Ton)=1/(4*27ns)=9.27 MHz.

In the datasheed another tables of "maximum times" can be found for
Cload=50 pF and various supply voltages. For example

74HC4052 typical specs a VCC=+5V, VEE=-4.5V, Cload=15 pF (not sure what
VEE voltage is used in the other CL=15 pF specs. �VEE=-5V?)
Propagation delay IN to OUT, TPHL, TPLH= 8ns "maximum"
Switching OFF times, TPHZ, TPLZ=38 ns "maximum"
Switching ON times, TPZ, TPZL=46 ns "maximum"

This will lead to the followinf maximum frequencies:

Fmax with channel on=1/(2*TPHL)=1/(2*8ns)= 62.5 MHz
Fmax for switching=1/(4*Ton)=1/(4*46ns)=5.4 MHz.

I will apreciate your replies to clear this subject. Thanks so much.

73, Ramiro. EA4NZ.

[Earl Shaffer]

Hi Ramiro.

 

I will answer what I know and leave the rest for others.

AM detection seen on the center carrier is a common issue when IMD is present.

RF filtering should take care of it. To test to see where the problem area is, install

an attenuator ahead of the suspect circuit. To test for audio amplifier overload,

place attenuator ahead of that. IMD should decrease more than the attenuator

value if it is placed ahead of the IMD limiting stage. DC is a valid  output frequency

for IMD.

 

You could double the frequency of your local oscillator by feeding it through a full

wave rectifier. You would then need to square it up and feed your divider chain.

 

You could try the FST3126 as used in the Softrock for the switch. It may not be

so bad and it should work at higher frequencies. Pushing the frequency of a switch

may lead to lower sensitivity. I don't know what the effect on IMD might be. You

can look at the output of the switch on a scope if you have a strong signal to

input to it.

 

Designing and redesigning by observing the results is a good way to go. The challenge

is then to make good measurements. I built a lot of test equipment in the process

such as crystal oscillators, attenuators and hybrid couplers.

 

73, Jim Shaffer, WB9UWA.

[Ramiro Aceves]

El 29/04/11 16:26, Earl Shaffer escribi�:
> Hi Ramiro.

Hi Jim,

Many thanks for your interesting replay.

> I will answer what I know and leave the rest for others.
> AM detection seen on the center carrier is a common issue when IMD is
> present.

Yes. The problem is worse at night when broadcast HF stations are
stronger. Also there are DDS birdies that are present without antenna,
of course.

I have a strong AM MW broadcast station on 585 kHz, 600 kW, and 264m
high antenna at 15 km from my QTH.

> RF filtering should take care of it. To test to see where the problem
> area is, install

My fear is: if broadcast strong stations produce IMD when listening
ham radio bands with no filters at all, will IMD be present when trying
to listen the broadcast stations with a broadcast station filter ahead
the mixer? I mean, if IMD is caused by the mixture of hundreds os
stations in the HF range, will IMD be reduced if I use a per band
broadcast filter?

> an attenuator ahead of the suspect circuit. To test for audio amplifier
> overload,
> place attenuator ahead of that. IMD should decrease more than the
> attenuator
> value if it is placed ahead of the IMD limiting stage. DC is a valid
> output frequency
> for IMD.

Fine, I like that method.

> You could double the frequency of your local oscillator by feeding it
> through a full
> wave rectifier. You would then need to square it up and feed your
> divider chain.

ok

> You could try the FST3126 as used in the Softrock for the switch. It may
> not be
> so bad and it should work at higher frequencies. Pushing the frequency
> of a switch
> may lead to lower sensitivity. I don't know what the effect on IMD might
> be. You
> can look at the output of the switch on a scope if you have a strong
> signal to
> input to it.

I seems that at this stage I need an oscilloscope and a generator to
continue with some confidence. I have understand that one does not need
complicated test gear if you duplicate others designs, but it is
impossible, or difficult to really design new things without that. I am
40 years old and I have been making my own transceivers and amplifiers
from HF to UHF since I was 16 years old, and I did survive with a hoome
made frequency counter, dip meter and a test probe. But I think it is
time to adquire a good generator and oscilloscope to have real fun.

Many thanks Jim for your valuable ideas, they are welcome.

Today I have found a problem in the receiver that I will describe in the
next post. Definitively I need equipment to see what it is happening.

Many thanks for all.

Ramiro. EA4NZ.

[Ramiro Aceves]

Hi again

As I said in the last post, I have found a curious problem that I want
to share with you. I am sure that someone with more experience can
discover what it is happening.

Today I was experimenting with the RTW-SDR receiver and I said: well,
let's listen to some MW AM stations for hearing local news. As I said
before, I have a strong 585 kHz AM station near my home. So I tuned the
DDS to 585*4=2340 kHz. No signal. Ummm, strange, I was really
preocupied. I switched my handheld radio on an yes, the station was
there at 585 kHz. Another local station is placed at 999 kHz. So I tuned
the DDS to 999*4=3996 kHz. Nothing at all. I was really puzzled, I am
sure that I have tuned that stations before, but that happened when I
used no lowpass filter on the DDS and a bipolar NPN transistor for
amplifying DDS signal before 74HC74 clock inputs.

Playing with DDS frecuency up and down searching for the stations, I
found them but wooooooow, DDS frequency should be tuned to F*3 instead
of F*4. The only test equipment I have is a frequency counter so I did
the following test. I tuned DDS at 4000 kHz and place the counter at one
of two 74HC74 outputs. Fine, frequency was 4000/4=1000 kHz. But when I
placed the counter at the other ouput, frecuency was 4000/3=1333.3 kHz.
It seems that there is a critical DDS frequency placed a bit higher
from 4 MHz. If DDS generated frequency is greater than that, by four
division is made correctly. But below that frequency, division by three
is done. After that quick test I had to leave home and I could not do
more tests.

Now I am out of home for the weekend and I can not make tests till
monday, but I suspect that 74HC74 is not working ok because of lack of
DDS signal strenght at lower frequencies. I have to rise the value of
150 pF capacitor ahead the low pass DDS filter to see what happens and
if that improves low freq response.

I am sure that owning an oscilloscope will tell me the real truth.

Thanls
73! Ramiro. EA4NZ.

[Earl Shaffer]

Hi Ramiro.

 

Let me answer with a possible example. The numbers are probably wrong, but the idea is there.

Lets say that two +20dbm signals can mix and cause IMD in your SDR. A more realistic condition

might be where you have many strong signals, such as 30 different signals all at -10dbm. These

could add together to cause IMD just as if two were at +20dbm. Again these numbers are probably

wrong, but the idea is there. If we use a band filter we might only pass 3 or 4 -10dbm signals.

By adding the bandpass filter we have rendered some in band signals harmless because we have

reduced out of band signals that lowered our threashold to cause in band IMD.

 

Test for DDS birdies with your attenuator ahead of the audio amp. It can be an interesting challenge to

suppress these, but it is possible. I have several crystals for my local oscillator. Each time I changed

crystal frequency I saw a new pattern of spurs. The high order spurs would move much more than the

low order spurs. This should give you an idea of what to look for. Part if not all of my spur problem was

that the local oscillator radiated signal into the audio amplifier. All harmonics of the local oscillator will

mix with each other to create new frequencies. If your local oscillator has some spurs of its own, then

the math becomes very complex indeed.

 

Don't overlook the idea of building a simple colpits crystal or LC oscillator.

Both of these can have far lower phase noise than a commercial oscillator.

A colpits should be good enough that all phase noise seen is due to your

local oscillator. I built two crystal controlled oscillators and combined them

with a resistive hybrid to do two tone IMD testing. Of course I can switch

one off and do one tone testing, so I can test IM3 and IM2 at little cost.

Leif has a favorite colpits crystal oscillator that is easy to duplicate and not

too bad in phase noise.

 

73, Jim.

On Fri, Apr 29, 2011 at 4:27 PM, Ramiro Aceves <ea1...@gmail.com> wrote:

El 29/04/11 16:26, Earl Shaffer escribió:

[Ramiro Aceves]

El 30/04/11 00:17, Earl Shaffer escribi�:

> Hi Ramiro.
> Let me answer with a possible example. The numbers are probably wrong,
> but the idea is there.
> Lets say that two +20dbm signals can mix and cause IMD in your SDR. A
> more realistic condition
> might be where you have many strong signals, such as 30 different
> signals all at -10dbm. These
> could add together to cause IMD just as if two were at +20dbm. Again
> these numbers are probably
> wrong, but the idea is there. If we use a band filter we might only pass
> 3 or 4 -10dbm signals.
> By adding the bandpass filter we have rendered some in band signals
> harmless because we have
> reduced out of band signals that lowered our threashold to cause in band
> IMD.

Thanks, I understand. I will aim my efforts now to the input filters and
see what happens.

> Test for DDS birdies with your attenuator ahead of the audio amp. It can
> be an interesting challenge to
> suppress these, but it is possible. I have several crystals for my local
> oscillator. Each time I changed
> crystal frequency I saw a new pattern of spurs. The high order spurs
> would move much more than the
> low order spurs. This should give you an idea of what to look for. Part
> if not all of my spur problem was
> that the local oscillator radiated signal into the audio amplifier. All
> harmonics of the local oscillator will
> mix with each other to create new frequencies. If your local oscillator
> has some spurs of its own, then
> the math becomes very complex indeed.

I understand..... If you have spurs with single frequency cristal
oscillator imagine the mixture of dirt a DDS can generate, hi... Hope
the Si-570 will be better than the AD9850.

> Don't overlook the idea of building a simple colpits crystal or LC
> oscillator.
> Both of these can have far lower phase noise than a commercial oscillator.
> A colpits should be good enough that all phase noise seen is due to your
> local oscillator. I built two crystal controlled oscillators and
> combined them
> with a resistive hybrid to do two tone IMD testing. Of course I can switch
> one off and do one tone testing, so I can test IM3 and IM2 at little cost.
> Leif has a favorite colpits crystal oscillator that is easy to duplicate
> and not

I will have a look into that. Leif WEB page is an excellent way to learn
electronics.
I have also seen some new chineese 100 MHz digital oscilloscopes on
ebay, but I think that it is better to wait for having more money and
buy a good oscilloscope of a well known brand.
Too cheap for beeing good, isnt it?:

http://cgi.ebay.com/New-UNI-T-UTD2102CEL-1G-Digital-Oscilloscope-100MHz-/320684975870?pt=BI_Oscilloscopes&hash=item4aaa5066fe

73! Ramiro. EA4NZ.

[Ramiro Aceves]

Hello.

Replacing 150 pF capacitor with a 2.2 nF one at the low paas filter
input solved the problem. It turned out that with 150 pF it was an
attenuation of 15dB at 4 MHz so that the 74HC74 did not worked properly.

Thanks to all.

Ramiro, EA4NZ.

El 29/04/11 23:53, Ramiro Aceves escribi�:

[Earl Shaffer]

Hi Ramiro.

 

Not sure if you need the scope for your purpose unless you wish to learn and possibly

improve the SDR substantially. I suggest looking for reviews of the scope. The seller

is already rated well. Of course low cost Chinese products can easily fail.

 

Linrad also has a built in scope function, but is limited bandwidth.

 

The effect of the low value capacitor is interesting. Perhaps it mainly passed

harmonics.

 

73, Jim Shaffer, WB9UWA.

On Sun, May 1, 2011 at 8:53 AM, Ramiro Aceves <ea1...@gmail.com> wrote:

El 30/04/11 00:17, Earl Shaffer escribió:

[Ramiro Aceves]

Hi Linraders

3-5 MHz and 5-8.5 MHz filters have been added to the RTW-SDR receiver.
This is a receiver now!
Last pictures, schematics and simulations at:

http://ea4nz.ure.es/sdr/sdr.html#RF%20filters%20%282011-05-05%29

I have removed antena isolation just for convenience at this stage. I
will return to ground loops issue later.

Many thanks to all, and specially to Jim, WB9UWA, Sid, G3VBV and
Leif, SM5BSZ for their technical assitance.

Ramiro.
EA4NZ.

[Earl Shaffer]

Hi Ramiro.

 

I have been thinking about your result when trying to use the receiver above 14 Mhz.

That could be a limit of the IC you are using. If that is not the case, there is something

else to look at. Is your DDS a sinewave or squarewave? The reason I ask is that

I found when coupling an RF squarewave using coax, there is severe ringing if the

coax is not terminated in its impedance value. This ringing could add new frequencies.

On the other hand, maybe the IC rings internally......

 

Softrock uses 3rd harmonic to receive the higher frequencies. I wonder if you have tried that?

For instance, set up your SDR to receive 7 Mhz. Replace the antenna 7 Mhz bandpass filter with a 21 Mhz

bandpass filter.

 

Nice job on the SDR. It's a simple and useful circuit.

 

73, Jim Shaffer, WB9UWA.

[Ramiro Aceves]

El 07/05/11 05:23, Earl Shaffer escribi�:

> Hi Ramiro.
> I have been thinking about your result when trying to use the receiver
> above 14 Mhz.
> That could be a limit of the IC you are using. If that is not the case,
> there is something

Hi Jim

My DDS clock is 125 MHz, so that 62.5 MHz is the maximum frequency.
When receiving 14 MHz DDS=56 MHz.

> else to look at. Is your DDS a sinewave or squarewave? The reason I ask
> is that

It is suposed to be sinewave, at least at low frecuencies. As we reach
Nyquist frequency I think it will we worse, who knows... I need an
oscilloscope.

> I found when coupling an RF squarewave using coax, there is severe
> ringing if the
> coax is not terminated in its impedance value. This ringing could add
> new frequencies.

Interesting. I can not see that without equiment.
Today I discovered I have some unstabilities in the by four division,
when DDS was tuned at 40.4 MHz to receive 10 MHz band. I have increased
DDS output by reducing the value of a resistor on the DDS and now
everything is ok. DDS output was only arround 0.5 V peak, now I have
doubled the value.

> On the other hand, maybe the IC rings internally......
> Softrock uses 3rd harmonic to receive the higher frequencies. I wonder
> if you have tried that?
> For instance, set up your SDR to receive 7 Mhz. Replace the antenna 7
> Mhz bandpass filter with a 21 Mhz
> bandpass filter.

I have just tried it. The problem is that 21 MHz filter has not got
enough attenuation to remove 7 MHz signals. Then, you have a misture of
both bands. But another annoying feature appears, 21MHz signals appear
the oposite way on the waterfall, I mean, If I tune DDS to 28 MHz, 7 MHz
signals start from the center of the screen to the right, and 21 MHz
signals start form the center to the left. Also image supression is
degraded.

> Nice job on the SDR. It's a simple and useful circuit.

At this point of experimentation I have understood that the key to get a
good direct conversion SDR is using very good RF filters with abrupt
response at the out of band frecuencies. You need higher order filters
in order to remove false 2*F and 3*F signals. High order filters need
too many toroids and winding work!

Anyway, for the price this receiver is a good toy.

73, Ramiro.