DDS Buffer Amp

So, i got some focus back into my home brewing with the goal to complete the receiver. The receiver is direct conversion, the DDS VFO is built, but as it outputs 270mV RMS, it is not powerful enough to drive a ring mixer that needs 7dbm or 1V RMS input. So a buffer amp was needed.

The buffer is pretty standard, common emitter gain stage with a common collector buffer. The DDS output is 200ohm, so a 4:1 transformer was used to drop the impedance down on the input. C1 is needed to block DC flowing into the transformer, as the bias voltage is set by the feedback resistor R4 and the transformer was loading it down so there was no bias voltage to the base of Q1.

Initially the Q1 emitter was to ground, giving about 10x gain, i found that this put me in the 3V RMS / 20dbm power range, much too high for a 7 or 10dbm diode mixer, so R2 was added to reduce the gain to about 3x. This gave me a nice 700mV RMS into a 50ohm load, 9dbm, close enough for now.


When running at close to 10x gain the waveform was full of harmonics, dropping the gain back knocked those on the head and the waveform is now nice and clean.


Looking at the output of the spectrum analyzer shows that first and second harmonics are way down compared to the desired frequency. So the dropping of gain certainly cleaned up the output a lot.


And this is where we are at currently, DDS VFO, Arduino micro controller and buffer amp all assembled and confirmed working correctly. Next job is to add the mixer in, then start work on the front end, JFET LNA (Low Noise Amplifier) to follow the Low Pass Filter.



At anytime, there might be half a hundred dozen different things on my work bench slash radio desk that can take up my interest for a fleeting moment. It might be a new toy, a new component, or just something shiny.


I was reminded of this today when i picked up part of the Direct Conversion CW rig i have been building for what seems like forever. I had some boards build to house a micro controller and DDS unit and I started to fiddle about with it and noticed that the output was very low.

The DDS should be putting out about 250mv rms at 7mhz, but i was only seeing 50mv rms on the scope. Something was up and i had never noticed what or why before. Pulling things apart and checking things separately showed everything was working right, but when plugged into the board it all went wrong again.

Turns out, I was measuring things at different frequencies and amplitude out of the DDS depends on frequency. Measuring 40mhz in the board and 7mhz out of the board, but being distracted i did not notice my error til much later. So, i have now cleared off the work bench, the only things on it now are stages of the DC receiver and everything else i have been tinkering with has been shelved. Time to get some focus and get this things build already. So tonight, i will build the DDS buffer amp and add the mixer as well.

More to come on that soon. 🙂


PCB to CNC Tool Chain

So I thought i should write about the tools I use to take an idea for a circuit board and bring it to completion all nice and milled and sexy. I start in Sprint Layout 6. This is a windows program, but run perfectly under WINE on my Linux machine. Why Linux on the office computer, well, I dont like viruses and malware that hijacks my browser functions, and Linux fits that bill perfectly.

  • Sprint Layout 6.0
  • FlatCAM
  • CAMotics


Anyway, political rant aside, i layout the board in Sprint Layout then export the Gerber and Excellion drilling files and move onto step 2 FlatCAM. With the gerber and drill files in hand, we can then load them into FlatCAM and convert them to G-Code. FlatCAM is a multi platform program written in python. It will run on Windows and Linux equally as well.


The final step in the process is to check to see that the G-Code files we have created are any good. For this we can simulate their paths in CAMotics. For this you just load up the cutting and drilling G-Code files and hit the simulate button. Mostly I am just looking here to make sure the paths look like my circuit board.


Well there you have it, a simple work flow of how to take an idea for a board and get it to the point where you can plug the G-Code into your CNC software and start milling and drilling away to a brand new circuit board. The reasons why i chose these programs over others is that they all work on Linux, they are all rather simple and have very easy learning curves. You can pretty much learn how to use each program effectively in about 10 minutes each. And that just speeds up how quickly you can start turning out things you can use.


CNC Router Cont

So after a lot of DRAMA, i finally have the CNC machine working and ready to start routing boards for me. This has been quite the journey, with everything and then some going wrong along the way. In part because the instructions were not all that clear when it came to what individual aspects of the software and hardware will work together and partly due to the fact i really did not research things deep enough to ensure it would work out of the box.

No matter what i tried, i could not get this to work off the parallel port of any computer using any version of windows, i could not get linuxcnc to find and use my parallel port, and a whole lot of other crap i wish i knew earlier. Even the $200 motion controller (CN-100) i bought to bypass the parallel port issues did not like windows 10 and i have had to use YET ANOTHER COMPUTER with windows 7 on it to get it all working.

So it works, and that is a relief, as i was about to the point of putting an axe to it and rage quitting for good LOL.I have done a little testing just to make sure things are actually right. First test was to run some G-Code and just let the machine do its thing:

Next test was to tape a pen to the spindle and make sure that the output looks like the G-Code of the board i was routing.


And it looked just how it should, obviously the tracks are not to scale as the pen had a fat tip, but you can see the board looks like a circuit. All in all, i am pretty happy now that this is working and I can crack on into making some boards for my many projects.



A small update, i have done some routing, was going along just fine and PING the bit went, and I made a mess trying to correct it LOL.




This is looking much more like it, 3rd time is the charm. I still need to offset everything to the left about 0.5mm and up 2mm and it should be perfect, I know how to do all that now. Next problem is working out why my drill file is not right. Nothing is ever easy, but Its getting there.



Poor mans silk screen, SHARPIE LOL.

CNC Router