BinaryRock

It was a regular crystal, one that I've had in stock for a while now and wanted to get used up.

Up until now, I had no idea TCXO's were even available in 2mm x 2.5mm packages, but, you made me look and now I have a choice to make (whether to add $3 more in parts per board or not). I might reconfigure the pads to accommodate both options and see how my crystal does first before ordering TCXOs.

Edit: Yea, I will do use TCXOs. .5ppm will have to be good enough because I am not spending more than a few bucks on the oscillator. And I am too poor to order the new micros and TCXOs in large quantity, I don't have thousands laying around to waste, so I will order small and only if they sell will I order more. Probably smarter that way anyhow.

TM86 said:

Sorry, thought you already knew about those. I was thinking about putting one of those, an Si5153C, and an Atmel micro on a board for a DSS with an FM option.

Well, once I finally sell the old house and get some capital to invest. Which is a whole different rant.

Click to expand...

Yea, for some reason I had it in my mind that these hadn't shrunk down like everything else. Since I now have a bunch of regular crystals, I wanted the option to use both. I added a separate set of pads for the TCXO, and when it is used instead of the normal crystal, a capacitor on the crystal pads will pass the TCXO output to XA (datasheet calls for a capacitor with external clock anyhow).

I think for AM a regular crystal is more than enough and I like the idea of being able to offer both options. The TCXO I chose is 0.5ppm and the original crystal I bought was 10ppm. I know 10ppm sounds like a lot, but considering most people will not be operating it at -20°C or at 70°C, I think its fine (and so does AI, but I welcome experienced opinions). If I make the assumption the ppm will only be 60 over a 20°C span, then I can calibrate it at room temp and AM users will never notice.


I am still going to wait for v1 to arrive to ensure there are no other issues hiding somewhere. I need to verify my mosfet buffer idea before ordering these black ones, but the file is in my cart ready and waiting.

I feel like I should add 6mm to the board and get the output away from the oscillators...

I decided to move my progress report from the PLL thread to here because I now have a finished board to share a pic of.

As I mentioned earlier, I did a noob mistake and assumed the internal pullup resistors could handle I2C and did not place externals on this board. I was under the impression that the internal pullups were 10k, but I now know they are about 35k. This will limit the I2C speed, but since my traces are 3.5mm long and the pins have low capacitance, I think I can still get 100k clock speeds to the 5351. I am also looking into bit-banging it so I can do it faster with push-pull outputs and ignoring the ACK the 5351 sends back. I may be able to get some good write speeds.

I have made mistake after mistake with this and its taking a while but it is slowly coming together. I would solder wires to this board and try programming it, but it's 2am and amazon says my edge connectors will be here tomorrow.

There will definitely be a clarifier, and if write speed permits, maybe an FM firmware update. As it sits, it has 4 extra solder pads and 7 I/O pins so there is plenty room to play with ideas. Im a little slow at code too so bear with me.



I was intentionally sloppy with the binary soldering to see how hard it was with a big iron tip and its really not that bad. And no wick is required, just swipe the iron down the board with some flux and they all separate. I should have specified mask between them and not a silkscreen line because that didn't print to the tinned GND pad like I thought it would. More updates after I finish the programmer.

Just had a thought. Since it has 7 IO pins, one can be the clarifier (which I will make one Hz for every 10-bit ADC step, or about 512Hz up and down), three will be for channel bank selection like the good old days, and the other three can be 5kHz, 10kHz steps or whatever. That way you should be able to set it to the main frequency using the binary and then bring the radio anywhere you want to afterwards, and still have fine clarifier control. IDK, too tired to think. That could be one version of the firmware anyhow. Bandwidth probably wouldn't justify more than 6 total bands anyhow. Three 440kHz jumps up and down from center.

nomadradio said:

Don't ask me to code it, but a velocity-sensitive clarifier would be cool. The slower you turn it the smaller the steps become. Might eliminate the need for a separate fine and coarse control.

73

Click to expand...

I agree, but I don't know how to make that intuitive.

The reason I am leaning towards the analog clarifier is because most radios already have a pot with a center detent. With a rotary encoder, especially a variable velocity one, how will people keep track of where they are without a display? They give it a spin or bump it without good conditions and their radio will be who knows where until they hear another station.

And if I have to do a screen, people will want it to show the radio's output frequency, and now I am doing math for every radio out there (or creating a configuration program for it, and then they would need a UPDI programmer). This is why I wanted to leave that to the DDS VFO guys that already have all that ironed out.

If you could give me some ideas on how we could do it without being able to see the frequency, I'm all ears. Maybe an encoder with a push-to-center feature?

I'm not against it, I'm just not sure how to do it in a way that doesn't get the operator lost. I am open to suggestions.

I should have added a RGB LED to the board and UV resined a single strand of fiber optic to it so an indicator light could have been brought through the smallest of hole in the radio's face or tucked in the corner of the meter. Too late for that.