Homebrew 3 Channel Pure SIN Generator ( 0 ... 60MHz )

[Yucca]

Otto & GK, Cheers guys, your work in TPU seems deeper than average, it'd take me weeks of reading to catch up on what you guys have done. Runaways sound exciting otto!

While I wait for bits to arrive I've designed a touch sensitive circuit, I hope to have the one metal dial be also touch sensitive so scanning can be performed with one control in one hand and no strain from pushing switch.

The circuit sets up two of the NORs as a 100kHz RC oscillator. This oscillator then pumps up the capacitance (which is touchpad to ground) the capacitance is leaked through the variable pot.

When someone touches plate the capacitance to earth increases greatly and so the resistor can't pull it down fast enough. the fourth nand buffers the output and debounces with 22uF electrolytic.

The microcontroller interprets a momentary touch between 10ms and 200ms as a valid user press.

Maybe a working TPU will hash up the operation lol, but it does work fine a meter away from a 10W spark fired aircore LC at ~6MHz.

[BeeSee]

Hi Yucca!

Hope your project went on smoothly! Can you update us on its latest progress. Thanks!

Ernest

[dankie]

Give it time LOL , this is quite the engineering project .

When you *build stuff* you run into trouble and have to trouble shoot and solve that problem .

[Yucca]

As Dankie says, it does take time, I find most time on the project is thinking time away from the bench, and then programming time, probably  over 30Hrs now, if I don't like the code structure I often rip it back a little and then proceed.

But here's an update:

The software is done to initial beta stage and seems bugfree. Scanning with power sampling is now complete and runs at a reasonable 1200 samples per second.

Most of my parts have arrived, see photo.

On the left you can see the 3 x DDS boards with AD9851 synthesizers and AD8008 RF amps. I can't stress enough what a superb job Analog Devices have done with these ICs; I have a very good radio receiver (scanner) that is very stable. I programmed a module to output 60MHz. Tuned the radio in on sideband reception and it was spot on. When I listen to the heterodyne signal it is really solid and clean. I can set heterodyne to 500Hz and even at 60MHz the 500Hz tone is rock solid! When you try the same with analog oscillator the heterodyne will drift all over depending on temperature, proximity of hand, phase of the moon lol etc. And when you change the output load it will disappear from the radio.

Another good thing is that the 3 seperate modules are all VERY close in frequency. I can hear a small beat between two seperate heterodyne signals. But at the moment each module has its own xtal oscillator. I plan to feed them all from one clock in the end. But I'll build it up with three seperate clocks to start with.

Also pretty pleased with the jogwheel encoder. A nice bit of kit, made in japan for CNC machine control, ballraced and strong, nice clean output signal, works well with the micro. Also I will be able to isolate its case ground and my touchsense circuit works well on it.

The 3 bunches of black cables are chinese wifi extender cables, these were dirt cheap and superb quality. If you bought 6m of RG174 cable and the plugs and sockets from digikey etc it would cost quite a bit. These cost â,¬5 all up delivered to europe.

Before I expand the software to 3D sweeping I will upgrade to 32kb memory. I have plans to make the 3D scanning have 3D gradient ascent to quickly hunt into maximas. To do a brute force scan (raster scan) of 1000 x 1000 x 1000 samples = 1billion samples, it would take almost 10 days to scan that at 1.2 kS/s. With gradient ascent I should be able to shorten that to hours. I will also have raster scan as well. This will be nice for exploring areas of interest when the data is transfered via USB to the PC, I will in the end write a small PC app with openGL viewer so I can see the 3D data set and peel layers back to look inside.

I will also make small RF shields with cardboard and alu tape for each module.

I've also got to work out the best conditioning circuitry for the sampling input, I probably want an optoisolator and maybe a log opamp with small knob on box front to adjust gain.

I will order a small alu chassis case to put it in next paycheck.

It'll probably be done in a few weeks. Before then I'll probably make a vid showing performance of the  DDS modules.

[giantkiller]

A beautifully, extensive play of the mind and wonderful engineering!

Now when you want that device to run free at 60Mhz strap a hamster to the plate. That way you won't be around when it blows up.

Otherwise Saint Peter will be asking you "Well now son, What is that in your hand? You know we don't let that kind of technology in here. We have enough clouds as it is. And it also looks like your brought one of your own?"

--gk.