Podcasting since 2005! Listen to Latest SolderSmoke

Friday, February 17, 2017

Back from the Raspberry Pi SDR Brink

Earlier this week I shocked Pete Juliano by telling him that I was taking a break from my normal analog, discrete component, no-chips mode of construction so that I could put together a Raspberry Pi-based SDR receiver.  Even from 3000 miles away, his astonishment was clearly perceptible. He seemed briefly disoriented by it.  I'm sure some of you may have a similar reaction.  

I'd been lured in by that video of the Raspberry Pi RTL-SDR receiver with the very cool  touch screen display.  It has a waterfall!  And a touch screen! How could I resist?

I went to Amazon, but there I discovered that that attractive display is not exactly cheap. And maybe I'd need a new Raspberry Pi.  At this point, in search of economy and convenience, I began rummaging through my digital junk box.  There I found a Rasp Pi Model B.  And an old computer monitor.  This will be easy, I thought.  Just get some SDR code into that Pi, hook up the RTL-SDR dongle and Bob's my uncle, right? 

Not so fast.  I quickly began to run into daunting digital obstacles. Sure, the Raspberry Pi fired right up and filled the computer display with lines of code.  But it was all Linux.  Yuck.  Sorry Linux fans, but for some of us mere mortals,  Linux is a weird opaque world in which every little thing is somehow a lot harder.

I also began to suspect that my 2013 Model B might be sort of a Model T in the Rasp Pi world.  It might not be up to the computing task.

And finally, as I poked around the internet, I began to conclude that the Raspberry Pi software for SDR is not quite done yet.  All the sites seemed to have the word "experimental" in there.  And lots of "I'm pulling my hair out" comments Maybe I'm wrong, but maybe we just need to give this more time.

Let me ask the distinguished group some questions:

Is my Model B really useless for SDR purposes, even if I don't need all the bells and whistles?

Is there an SDR program that can be easily placed in a Raspberry Pi by someone who has NOT mastered the mysteries of Linux? 

For now, I have cleared the raspberries from the bench and am back to working on HDR stuff. 

New Posts to BITX HACKS

Don't miss the new posts on the BITX HACKS blog.  There are some great ideas from Don ND6T and some wonderful tribal knowledge from Pete N6QW.

http://bitxhacks.blogspot.com/

Sunday, February 12, 2017

Ken G4IIB's BITX Journey



The work of Ken G4IIB has been on this blog before -- he helped many of us make use of the amazing RTL-SDR Dongle SDR receivers.  He has recently turned his attention to the BITX40 Module and offers some great ideas for testing and for modification.  Ken's description of the smoothness of his audio adds a very evocative term to the SolderSmoke Enhanced SSB lexicon.     

Hi Bill, Pete

Many thanks for your respective responses to my plea for help in setting up SI5351 derived BFO to my BITX40 board. You were both on the money.


Pete suggested that I had too much gain in my HB amplifier from the SI5351 output to the modulator and indeed that proved to be true. Once sorted I also noted that I was getting extra hiss on switching to one of the sidebands as you pointed out Bill this proved to be due to incorrect placement of that particular BFO frequency.

These BITX40 boards that Ash Farhan has developed and released to the world wide community of Radio Amateurs are worth every penny. Because they are so hackable (not just the circuitry but now the Raduino code also) it means that you can tailor it to your specific specification and in the process you are likely to learn new stuff and make new friends. I describe my BITX40 incarnation and experiences below:

Upon first firing up the BITX I was getting quite a lot of mains hum from my PSU's (I thought that at least one of these PSU's was a quality item) but obviously not up to the job. I constructed a simple one transistor capacitor multiplier (this converted a humble 1000uF cap into a 1F cap) and the noise magically disappeared. By coincidence I note that Bill discussed this technique in a recent pod cast. Another advantage of this technique was that I got a 2V drop across the transistor so by running this on 13.8V I get 12V out so I run the PA section on un-smoothed 13.8V (this gives me 12 watts of RF out) and run the receiver section on the smoothed 12V output from the multiplier, happy days.

My thoughts were to turn my BITX into a multi band (several bands rather than all bands) rig and I figured that using high side mixing (running the VFO at 19Mhz (12Mhz + 7 Mhz) rather than the existing low side mixing (12Mhz - 7Mhz=5Mhz VFO)) would be a better option. For example running it on 17M would mean using high side VFO anyway. I also wanted the ability to be able to switch sidebands especially on the lower frequencies so that I could use the rig for Digital modes in my case this was to be achieved by coding the Arduino to run a BFO on one of the SI5351's clk ports.

I bought my BITX prior to the release of the Raduino so I had already commenced (with the aid of a new found radio friend and RF mentor) coding an Arduino VFO/BFO using a UNO and SI5351. Like I said at the beginning once you let folk know that you are starting on a new and interesting project you start to engage the more practical members of the ham community and they just want to get involved and help. Yet another good reason to buy a BITX . We used code originally developed by Jason Mildrum NT7S and Przemek Sadowski SQ9NJE and tailored it to suit the BITX40 and our requirements. This include high side VFO with frequency step adjustment and a BFO with long push BFO changeover. This meant that my BITX front panel should stay very minimalistic 2 knobs.
Getting the VFO to work was simple as the DDS socket was used and to better accommodate the high side VFO I modified the board by tombstoning caps C91 & C92.

Getting the BFO to work proved to be more problematic I was troubled with hiss and other noise. Words of wisdom from Pete Juliano when asked if I was doing something wrong were: " No –it is just that we tend to think our projects are like Lego type building blocks where everything mates and snaps together. Sometimes more is required". True Pete and that gives us the opportunity to learn new stuff!

To cut a long story short I found that the best place to connect the BFO was on the modulation transformer T4 thus bypassing the BITX BFO stage altogether. I was also getting hash noise believed to be emanating from the Uno. At this stage my after market Raduino arrived from India. I fired this up and noticed that I was not getting any hash noise from it. This pointed us to a coding problem and the LCD refresh was altered on our code and the problem disappeared. Below a picture of the module showing the BFO connections to T4 and the large heat-sink with the IRF510 insulated from it. Also shown is the capacitor multiplier and a glimpse of the Raduino in the foreground. Not the most elegant box but this is likely to change pending further refinements. It's still work in progress and this box gives me plenty of room.


The Raduino is a fantastic piece of kit for the money extremely neat and well thought out. The coding is comprehensive and innovative and works well. However, from an aesthetic and ergonomic point of view there were a few things that I personally did not like in terms of how it operates and performs. I could not get away with the potentiometer tuning, you can tune 50Khz of the band and then when you near the pot edge it increments/decrements and you can re-tune. I found this clunky to use and in addition the Raduino would hunt causing the last digit to increment then decrement causing an annoying warble on audio. In my opinion a Rotary Encoder would be better solution. On the plus side, although not mentioned on the Hfsigs web site the Raduino code does come with other functions such as changing sidebands by temporary high siding the mixer, a RIT, VFO B and CW tone. If you download and read the Raduino code from Github you will see this extra functionality which I believe you can make use of via extra switches (not supplied). The current Raduino code does not have any external BFO options as said it relies on the crystal BFO and temporally high siding the VFO to change from LSB to USB on 7Mhz.

The Raduino module itself is just too good and neat not to use. As I did not have the where for all to fully understand and amend Ash's code I decided to use the Raduino but to load it with the code that we have developed for he Uno and Addafruit SI5351 board. This would give me near conventional tuning via a rotary encoder, adjustable step sizes via quick push of the encoder switch and USB/ LSB switching via long push of the encoder switch by virtue of the SI5351 generating the BFO frequency. I have retained a copy of Ash's Raduino code just in case I wish to revert to it. I put a new header on the Raduino P3 connector so that I could connect a rotary encoder and use the 2nd clock output and then changed our code to run on a Nano. I had to add a correction factor in the code to cater for calibration differences in the SI5351's (in my case 1.21Khz).

As previously indicated I had a little trouble arriving at the correct BFO frequencies I found that 119940 and 119970 gave me LSB and USB respectively for my high side VFO (19Mhz) if you use low side VFO (5Mhz) then these would be reversed. We further refined these frequencies by injecting white noise into the mic amp and looked at each transmitted sideband on my RTL-SDR dongle via HDSDR (a useful piece of test equipment). By adjusting the carrier trimmer to show the carrier in the extended HDSDR spectrum display we could see how much to move the BFO frequency to best occupy the crystal filter pass band, see image below. This frequency adjustment being achieved by a coding change. The frequencies I consolidated on to cater for my particular crystal filter are 119941 LSB and 119969 USB. We then nulled the carrier back out. My audio is now as smooth as a maiden's inner thigh, trust me the image will follow!


So now I can get on and build an AGC and think about some sort of S meter. As for putting the BITX on other bands, whilst I now have a VFO capable of going anywhere, I would need to address band pass and low pass filter and switching arrangements. I may still experiment with this but, as pointed out by Ash in a recent pod-cast, the BITX single superhet design is not best suited to multi band operation but can be quite easily changed to operate on another single band. He also indicated that he was developing a dual superhet with consideration for multi band operation. Once released this might be a better option for multi-band use.

In the mean time folk should just get a BITX40, hack it to bits and share with us their customised versions.

Ken G4IIB


Saturday, February 11, 2017

VK4FFAB's FB Intro to LTSPICE


Rob VK4FFAB wrote a really nice series on how to get started with the LTSPICE circuit simulator.  I'm sure these articles will also have a lot to offer for those of us who've been using LTSPICE for a while now.  Thanks Rob!  

Rob's articles can be found here:

http://vk4ffab.info/lt-spice-for-radio-amateurs/


Friday, February 10, 2017

Falling for the SDR Waterfall...



This is going against everything I believe in, but I admit it -- I want one of these. 

Thursday, February 9, 2017

VK2EMU's Biscuit-Tin Direct Conversion Receiver


Bill and Pete:
First Sound

The enclosure came first, then the radio. The Homebrew challenge - build something in a Christmas Biscuit tin - only one rule, no mains power.

Step one - eat the contents.

VK2WI is an Australian version of W1AW and transmits weekly new bulletins on a number of bands from 160m to 23cm. On 80m the frequency is 3595kHz, so let's build a DC receiver for that.

I used a 7190kHz crystal, divided by 2 with a 74HC74 to get 3595kHz. The rest of the radio is pretty standard - double tuned front end, NE612 mixer, followed by a TL071 and a LM386. The reduction drive on the tuning cap gives a band spread of about +/- 300Hz.



The boards are all 2 inch x 2 inch and made on a PCB mill that I was given a few years ago. 

There should be enough room in the bottom of the tin a pack of 8 C cells to make the radio truly portable. Next step is to fit an audio low pass filter.

It is a pleasure to sit on the rear porch and listen to the Sunday night broadcast on a home brewed radio, while eating the contents of another identical tin - I wonder what I will do with that one?

73  Peter VK2EMU





Wednesday, February 8, 2017

VK3YE's GREAT "QRP by the Bay" Event

Last weekend Homebrew Hero Peter Parker VK3YE hosted another of his amazing twice-yearly QRP events.  It was at a park near the iconic Chelsea Pier in Melbourne.   Peter Marks VK2TPM sent a very nice write-up with pictures:

http://blog.marxy.org/2017/02/qrp-by-bay.html

And a nice audio report:

http://s3.marxy.org.s3.amazonaws.com/audio/QRP_By_the_Bay_2017.wav

Peter Marks reports that most of the on-the-air activity was on the 120 foot ham band (40 meters for you modernists).  Many BITX40's were on display.


Tuesday, February 7, 2017

ZL1UEM'S Tiny OLED Si5351 Arduino VFO



Hugh ZL1UEM has come up with a very creative way to take maximum advantage of the small size of the Si5351 board, the Arduino, and the OLED display.  He even has the rotary encoder in there.  Look carefully -- he uses both sides of the board.  Very nice. Thanks Hugh!

Hi Bill,

First let me say that I have been an avid follower of the SS blog and podcasts since the days of your podcasts involving Mike KL7R.

Like many others I was tempted to purchase the $49 surface mount module from HF Signals.  As a keen home brewer I felt guilty about employing a prebuilt board but excused my decision on the grounds that I would build a DDS and other accessories myself.

In addition to follow the SS blog I also check Pete's blog regularly and was excited by his OLED VFO for the Bitx40.

I constructed it on a small double sided matrix board with plated through holes.  A bit of noodling led to the layout shown in the photos.  There is only one board with components mounted on both sides.  The board came to life on the first power up but the text spilled off the bottom of the display.

I assumed that the sketch that I had downloaded from Pete's blog was for a different OLED module.  I knew that he had also used a yellow/blue OLED, the same as mine, previously so emailed him requesting a sketch for this OLED.

I was taken aback when he informed me that the sketch I had was the same for both the dual colour OLED and the black and white one too.  Pete suggested that I swot up on the use of OLEDs generally and that perhaps I should first experiment with the text size to begin with.  He also offered some advice about the mapping of the screen.

I soon discovered that the text size was not the cause of my grief and that I needed to look elsewhere.

I first tried running the ssd1306_128x64_I2c sketch from the sample sketch folder and was rewarded with the message "Height incorrect, please fix Adafruit _SSD1306.h".  A search of the Internet revealed that I needed to edit the .h file and find "#define SSD1306_128_64" which was commented out and uncomment it and make sure that the other two options, _32 and _16, were commented out.

My next problem was how to edit the specified .h file.  I tried notepad but the text all ran together.  Another internet search revealed that Notepad++ was a suitable choice and it did indeed cut the mustard.

A reload of the sketch completely restored the display to full functionality.

All this may be obvious to many but it was all new to me and if I had not been prompted by Pete l would not have had learnt so much and would not have had the same sense of achievement when it all came together.

Many thanks to you and Pete for providing a focus for my hobby.

73's
Hugh ZL1UEM

Schematic for the RTL SDR Dongle Front End

Here it is.  Very simple.   I used an obsolete 40673 dual gate MOSFET.  I didn't need both gates so I just soldered them together. You could probably substitute an easily obtained MPF-102  JFET.  The capacitors and the coil in in the gate circuit form a parallel LC filter circuit that resonates in the 40 meter ham band.   The 1 Megohm resistor maintains a very high impedance for the LC circuit, helping to maintain circuit Q (sharp response).  It also drains off any charge that builds up on the gates.  For the antenna input I just put 2 turns on the 4.5uH coil -- this too helps maintain LC filter response.  The 470 ohm resistor serves as the load and limits quiescent current through the MOSFET.  It should pull about 19 milliamps -- with a 9V battery that should give you around 26 hours of listening time between battery changes.  That's fine with me -- I don't use this thing that much. 

Realize that I'm using this with an RTL-SDR dongle that has been given the familiar modification that allows it to use direct sampling in the HF bands.   Some ideas here:
and here
and here



Monday, February 6, 2017

Adding a Homebrew Front End to an RTL-SDR HDSDR Receiver

I started playing with the RTL SDR dongles again.  I wanted to use them to confirm that my BITX signal is NOT 9 kc wide (it is not).  I also wanted to try to confirm the aircraft band frequencies in use in this area.

So I got the HDSDR software running and plugged in the dongle that I had modified for HF use.  It worked great, but I could see (literally SEE) that it needed some bandpass filtering and perhaps a bit of RF gain.  So today I threw together this device.  Dual gate MOSFET (gates tied together) with an LC filter on the gate.  Power from Malaysia via the 9V battery.

It works great. It was fun to add a bit of homebrew to an otherwise soul-less commercial receiver.   But beware: that waterfall is addictive, even for a hardcore Hardware Defined Radio guy like me.
Designer: Douglas Bowman | Dimodifikasi oleh Abdul Munir Original Posting Rounders 3 Column