Thursday, March 10, 2022

It BREATHES!

With the antenna amplifier sorted and installed on the board.  I couldn't wait and I held my breath as I connected the antenna input to it.

The chain is as follows:  Antenna Amp - > SBL-1 -> Audio Amp> with CLK0 of the SI-5351 dumping into the LO port on the SBL-1.

Like clockwork, I couldn't hear anything.  Adjusting the volume knob doesn't help, at all.   Ok, maybe the signal is still too weak at the antenna amplifier.   I hooked up the signal generator, injected a 10-milliVolt signal and that sweet, sweet sound of a 1KHZ (roughly) tone popped up into my ears.  It wasn't very loud, but its there.  =)

I've got some sorting to do with the audio amplifier.  I may have to scrap & rebuild it, as, it's got a LOT of noise, the volume adjust doesn't work quite right, and its not quite loud enough.  It could be the speaker, but I'm going to start with the audio amp first.   Just in case, I ordered a pack of 10 pre-build audio amps from Bezos & Crew, and they should be here on Friday. 

Update:  I got it working using the previous LM380 Audio Amplifier.  I scrapped the LM386 board, and fiddled with the LM380 one until I got it working good.  =)

But I did fiddle with it, reconnected the antenna, and tuned down to the FT-8 portion of the band, and I could hear faint signals of the famed FT-8 "warble", so it does work.    But I had a broadcast station coming in on it as well.  I'm not sure if its from the length of cable that I have attached to the speaker, or if its something else with my configuration.    I slapped together a LPF from the uBitX schematic, threw it in the input before the antenna amplifier, and it didn't fix the problem.   Oh well, at least I have it built for when I do need it.  =)

So it looks like I've got a few issues to noodle around on before I can proceed, but for now, it does work as a direct-conversion receiver...just not very well.  =)

Update #2:  It works pretty well with a working audio amplifier.  =)

73s, 

-Jason

Saturday, March 5, 2022

The Flux Capacitor is missing...or maybe some clueless amateur hooked it up backwards

Well, not really.  Today was a bit of a success!  I built a half-stage of the W7ZOI/K3NHI Termination Insensitive Amplifier as an Antenna Amplifier.   Mainly because I was feeling too lazy to wind impedance matching coils.   Ha!  The CNC process went fine, except for when I didn't load the G-Code for drilling after doing the isolation routing.   You can see that mistake in the bottom left corner. Luckily, I caught it before it continued to route the board with a 2.5mm drill bit...and did not ruin the board.


Once that was done, I soldered all the components in place.  (Which reminds me, I need to order more 15-ohm and 180-ohm resistors!)  Hooked it all up, and of course, it didn't work.   But where would the fun be in that?   I started the usual troubleshooting, visual, power, signal, etc.   Power was fine, input signal was fine, but didn't seem right.   Also, I wound up replacing all the electrolytics with 100nF Tantalums, because...reasons.  Anyway, after about 20 minutes, I gave up and went to cook dinner for the family.  Then, watched a couple episodes of the SWMBO's current favorite sitcom with her, before heading up to the shack around midnight.  I'd had some time to noodle this over in my brain, and I think something was wrong with the layout.  Of course, it has to be me...


I finally figured it out after going back over the schematic.  It'd routed and placed the trace for the 10K-ohm resistor to the wrong pad.   Oh well, its homebrew, so I can fix it.  I left that trace alone, instead opting to route the 10-ohm power resistor to a different pad, as that was the easier fix at the time.   Bam!  it worked.


If you look closely, you'll see that a 50 milli-Volt peak-to-peak input signal  (purple trace) becomes a 1.62 Volt peak-to-peak output (yellow trace) signal.   I'm not a mathmagician, but sources on google tell me that's around 29dB of gain.  I'll take it for now.  =)

Tomorrow, I'm going to re-flash the Arduino and increase the drive of the SI5351 chip feeding the SBL-1 mixers. 

73s, 

-Jason

Friday, March 4, 2022

Everything Works...Till it Doesn't

I decided that the audio amplifier that I'd previously posted was probably a little too complex for my needs.   While noodling on the Idea, I stumbled on the JohnAudioTech YouTube channel, and rather liked his simple, less than a dozen component LM380 Audio Amplifier.  So, I did the layout and built one.  And like Mr. Murphy loves, of course, it didn't work.   

I started with the usual troubleshooting stuff.   Is it smoking?   No.   Does it have power?   Yes, at the input.   Does that power get to the chip?  No....first problem.   AHA!   Ok, now why does it do that?  If you look at the screenshot of the layout below, you'll notice why in the circle on the top left.   Yep...those two pads in the white circle on the top left?   They're supposed to be connected.  Easy fix...cut a jumper wire to bridge the gap and press on.   Good!   After cutting the jumper, soldering it in place, and testing with a voltmeter, I now had power to the Chip on Pin 14.  Good, I moved on.


I connected an audio source from a cheap E-Bay audio signal generator that I build a few months back.   I've never really tested it out, other than to know that it actually produces an audio output that can be heard on a speaker.  I turned the volume down on it and set it roughly to a 1000 Hz tone (I don't have perfect pitch, but it's close enough for amateur purposes), hooked it and a speaker to the board, powered it on, and ........nothing.   A close watch of the current draw on the power supply told me that it wasn't drawing enough power to be amplifying anything... 8-12 milliamps if my memory serves me correctly. It was low, that's for sure.

So, I spent about 20 minutes double checking voltages, and then I decided to put my audio signal generator on the oscilloscope to make sure it was actually producing a wave.    Well, I did, and what did I find?   The peak-to-peak voltage was around .5V, but the RMS was up around 6V.   Upon closer inspection, I realized that the sinewave was riding on top of DC.   I'll have to fix that later, so that test was junk.   I don't think it fried the board, as there's a capacitor on the audio input, but I pulled the LM380 out of its socket anyway and replaced it with a new one just to be safe.  Then, I dialed in a 0.2V, 1000 Hz tone from my Feeltech signal generator (which I know is good enough for testing) and connected it to the input of the audio amplifier circuit.   Still, no sound coming from the speaker.

After going on, I was struggling to figure out why my chip still wasn't amplifying. a known good input signal.  It had power, but there was no output at the speaker end.  I know my components were good before soldering to the board, as I tested all of them.   Plus, the likelihood of having two back-to-back dead LM380 chips was slim.   Something else had to be up.   I went to the datasheet to make sure that I had all my pins sorted out correctly.  If you were wondering why there were two white circles drawn on the layout pic, now you can begin to figure out why.  The other layout error is circled on the bottom right.  The datasheet says that pin 8, not pin 9, is the output pin of the LM380.   So, while still a simple fix, this one was a little more involved.   I cut the trace from pin 9 to the output cap and lifted the entire trace so that it wouldn't be a problem.  I then made up another jumper and soldered it from pin 8 to the cap.  I hooked everything back up again and Bob's your uncle.   

It was quite noisy though, so I did some noodling and reading.  I figured it had to be something with the coupling of the signal to ground.  So, I tested it by putting a 100nF cap across pin 6 and ground, and the problem went away.   I soldered that cap in place permanently and called it a night. 

With the audio amp sorted out, I'll spend the weekend researching on how I'm going to sort the antenna amplifiers out.  

73s,

Jason

Tuesday, March 1, 2022

Success! Time to move on..

 After 3 whole days burned on trying to get the CNC settings dialed in, I went into the shack yesterday with a newfound understanding of my past mistakes.   (Except that time with tequila in 2002...I don't much remember that one.)

I applied the new mathemagical properties that I'd discovered to my cutter diameter, generated new G-Code, and fired up the machine.   I probed the PCB to get a height-map, then nervously clicked the "send" button.   18 minutes later I was presented with as pristine of a board that could be made by a clueless amateur.  I still need to tweak some of the overlap settings, but overall, it turned out really good.  


Instead of immediately soldering the SBL-1 onto the breakout board like the last one, I took a hint from Chuck Adams and sprayed a thin coat of clear acrylic on it.  I'll solder it up in the next day or so, as it'll only take a few minutes.   If I don't forget, I'll post pics of the shiny-board before soldering it up.

Update:  Here's the main board with both SBL-1 mixers soldered to it.



Next step:  Audio Amplifier.

I spent a few minutes in the evening breadboarding an audio amp based on the LM380 from Charlie Morris. I was able to get it to make sounds with my cheap Ebay signal generator, but with an 8-ohm speaker attached, I couldn't get any real "volume" out of it no matter what I did with the input level, trim resistor adjustment, etc.  Plus, it seems like it distorts quite easily.  

Instead of that route, I think I'm going to build this audio amplifier based on an NE5532/LM386 from Vasily Ivanenko over at QRP Popcorn.    You can tell that Vasily spent a lot of time tweaking and testing, and there's no reason for me to doubt him (he's got Jazz Guitar ears).


I have a roll of 100 NE5532s, and haven't figured out what else to do with them at this point.   I'll build it on a breadboard and do some testing before I do the layout/milling/soldering. 

After the audio is finished, I'll probably tweak some settings in the Arduino code and determine what band I want this receiver to operator on for now.   Once I figure that out, I'll probably make up a band-pass filter to clean up the signals a little, which might require an antenna amplifier.  I'm not sure yet, I'll have to noodle on it for a little while.

73s,

-Jason

Monday, February 28, 2022

Well, there's your problem...



Have you ever watched someone do something stupid?   Then sometime in the future you wind up doing the exact same stupid thing, only to laugh at yourself just as hard as you laughed at the first-person being stupid?

Well.  That's me.   And this is why I'm stupid.  On another note, I may have found out why my cuts were so stinking crappy lately.

I was messing up so many circuit boards because I was using the wrong diameter cutter.   No, not making too deep of a cut, rather... I programmed the CNC wrong.    I was using a 30-degree Vee bit, with a cutting tip of 0.1mm.    So, my depth of cut was set at 0.1mm.   (Because I figured .1mm going down .1mm would leave a .1mm cut...)   Wrong!  I failed to add in the taper of the bit....30 degrees.

Luckily, FlatCam has a built-in calculator that I should have been using to calculate my "tool diameter".


So, instead of programming a .1mm diameter cutting tip, I should have been using a 0.1536mm diameter cutting tip.   The more I think about it, this would have a "broader" cutting edge when cutting through the copper of the PCB (the copper is higher up the bit) and the fine point only in the fiberglass.  I'll try a couple of boards tonight and see how it goes!

73s

-Jason

Sunday, February 27, 2022

If at first you don't succeed.....try again 21 times.


If you've read the last few posts, these should speak for themselves.  These are the failed CNC jobs.


But finally, after 21 tries, I succeeded.   It's soldered up, and I'll test it out tomorrow.  =)



73's

-Jason


Warning: Equipment should not be operated by Idiots


 Yep.   This one is all on Me. 

After getting a good night sleep and pondering my insufficiencies as a CNC operator, I awoke to a renewed sense of "I can do this."  After downing a cup of my favorite morning beverage (after all, us Vets are fueled by a love for coffee and a disdain for stupidity) I hit the computer and reprogrammed the CNC files to run.  I glued down a shiny-new piece of copper PCB material, probed the height-map, set the zeros, and started cutting.    All right!  The first set looks great!    Then it moves to the second cut, and I notice that the fiberglass isn't being cut, only the copper.   The bit isn't moving low enough to make it through the copper-clad material on top of the board.  

/insert Tazmanian Devil Sounds




(*@**!)(*$&&!)*(*)@#&&!(*&^^@#!@#*(^&!@#, !(@*&((!(*#^^!^!!!!!)

/end Tazmanian Devil Sounds

So, I junked another piece of copper, reprogrammed the CNC files to add another .025" of cutting depth, and tried again.  Maybe this is just me not cutting it deep enough.   IDK, but I was sure to find out.  

As luck would have it, I broke the tip off the bit.  So, once again, I fired up the G-Code software, and reprogrammed it to cut a single 30mm x 30mm board instead of a panel of 6 of them.   That way, it'll only take 28 minutes to cut one, instead of 2 hours.   Plus, I may waste a piece of copper, but at least I'll save the other spots on it for future cutting, resulting in less waste overall (I'd rather waste a 30mm x 30mm section than a 60mm x 90mm section).   I've prayed to the Radio Gods, bit I didn't have a chicken on-hand to sacrifice to Papa Legba.  Wish me luck.

73s,

-Jason

It BREATHES!

With the antenna amplifier sorted and installed on the board.  I couldn't wait and I held my breath as I connected the antenna input to ...