Electronics, Flight, and Computing are my long time interests. So the videos found here mostly focus on lessons learned while attempting various DIY projects in these areas. Hopefully something here will help you complete your project, or better yet, inspire you to try something new.
Hi Pino Thanks for your interest & question. If you follow the process outlined in the 1st video ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-Hb9m60LtSJw.html ,@~the 5 minute mark, the directory references will automatically get corrected to match your local directory structure. The project was originally built on a Linux machine. But today I cloned it, on a Windows 10 machine, (with a totally different directory layout) Also, the process now is even more automatic, than when the referenced video was done. I found as the install process moved along, some error messages appeared, but I just followed the prompts, and they all resolved themselves. BTW, If you're really not that interested in source code, the much faster way to get going is to flash the .bin file to your ESP32. The GitHub site has links to the bin file, and the flash loader. GL & let me know how it works out for you. tnx, Jim(KW4KD) PS: The decoder section is still a work in progress. I at some point I hope to do another video explaining further whats going on under the hood. The general goal is to have a decoder that operates with as little intervention as possible. And final point, the decoder supposedly is optimized for "rag chew" QSO's
@@jmhrvy1947 Thanks Jim for your answer.... now I am going to listen to your video. I am a Linux user and I would like to modify your project to adapt it to my needs...i.e. to make it send decoded chars via Wi-Fi to my PC where I have a GAMBAS3 program that displays chars and also sends code to an IC-7300 via CAT. At present I have a simple Arduino program that sends chars via USB serial connection, but it is quite slow and not so precise. I lack experience with Code and PlatformIO....it seems to me too complex 😞 Best 73 Pino ZP4KFX
"....it seems to me too complex@@pinox61" QSL that. But what I like is how well it (VsCode/PlatformIO) works with a github project. Ideally you will have to do a minimal amount of configuration to get this project up and going (in its current form). Now the changes your proposing sound fairly complex. The unit/project, as it stands, is designed to just plug directly into your 7300's key input, and does wireless keying that way. To be honest, unless there's distance issue, it would seem that linking your blue-tooth keyboard directly your computer would be a quicker/simpler solution. But I probably don't appreciate what all your doing there. Bottom line, good luck, and again, thank you, for taking an interest in the work that's going on here. Jim(KW4KD)
Hi Jim, thanks for this project, iv been hoping to get in to CW for a while and i think this may just give me the push to do so. Please can you explain what i need to change in VScode to make this work with a 2.8 inch screen. I was unable to find any resolution setting for the ILI9341 in user_setup.h, however I've changed the setting in the ILI9341_defines.h and the display resolution has not changed and the text still goes off the screen. Any help would be greatly appreciated.
Steve, Appreciate your interest in the project. And I hope you'll see it through to completion. I know these display's look similar. But this project is built around a 480x320 display. Not a 320x240. The 480x320 allows for 16 lines, @40 characters/line, and 12 horizontal pixels/character. There a similar issue in the vertical direction, with the character height set at 20 pixels. In an earlier iteration of this project I did support both screen resolutions. And you can see evidence of that, in the TFTMsgBox.h file. There, if you comment out line 14, #define HiRes, this part of the code will try to accommodate the lower res screen. But today, that's only a fraction of code, and all the rest of the project makes no allowance, for change in the pixel count. You're welcomed to change the code any way you want, but if your main objective is learning CW, then time wise, I recommend getting the ILI9488 SPI display. That said, I congratulate you for getting as far into the project, as you apparently have. Other things to emphasize: 1. While most any generic Bluetooth keyboard will work, the project is really designed around the Logitech K380. So for the best user experience, that's the one to have. 2. The project is still in active development. So please check the github site regularly (last update was 2 days ago) Again, thanks for the comment, & congrats on your hard work. Jim(KW4KD)
Thanks for the comment. If you're interested in the decoder part. I hope you'll revisit the github site (next week). Should have an update in place that's an improvement over what's shown in this video Jim(KW4KD)
Thanks for the question. If I follow it, there are 5 memories, "F1" through "F5", on it now. Its possible to add more, by following that code. "F1" is intended (setup) to store the call of the station your talking to. And there's a 6th memory for your call (accessed through the "Settings" screen, & sent using the "enter" key). The others are intended for simple "canned" messages. Like "CQ". In my mind, the project is targeted for "ragchew" QSO's, over contesting. So it doesn't have an auto-indexing number memory. Hope that helps. Jim(KW4KD)
Thanks Jim. That looks quite useful, having encoding and decoding done by the same device. I'm not a fan of the ESP32, but I can see that having built-in Bluetooth for no extra cost is a good thing for projects like this. In any case, I will be at least studying your source code. Due to the inconsistencies in spacing, some of the code that your decoder is getting clean copy from, I wouldn't be able to copy even if it was slowed down to my current 9 WPM code skill level. And speaking of which, it seems a bit .. unfriendly the way I hear CW being used on the ham bands: a person will send a CQ at 10 WPM, and all of the responses will be between 15 and 20 WPM. Seems to me like the courteous thing to do would be to respond at roughly the speed the CQ was sent, which I guess would mean not responding to CQs if you're not willing to operate at their speed, so I don't know, maybe that justifies it for some. But as a very slow Morse reader, I can see the value of having a digital decoder to deal with that. I mean, copy what I can, and just rely on the decoder for the jackrabbits. And you never know - I might break down and use an ESP32 board. I already have at least one spare LCD module, and I'm curious about some of the dev boards made with the ESP32-C3, which is RISC-V based, making it more interesting to me than the ESP32-S3, which uses a Tensilica core. 73 de AF7VZ
QSL calling CQ, and having a speed demon return the call. That's not supposed to be how it works. But like to think, if the decoder can't copy the returning station, there's no embarrassment in saying, "TU for the call & hope to catch you next time, when conditions are better". The hobby needs more folks calling slow CQs. As for the ESP32, I too ignored it for a long time. (seemed like overkill for most of my projects) But being able to host bluetooth and running RTOS, seemed like a good fit for combining the two sides of Morse code. That (combining the two) proved to be a bigger challenge than expected. Took me several months to get the two play nice, on a single ESP. The problem is how to get usable crash reports (from code you didn't write). But that's another video. Thanks for the comment & hope to hear you calling CQ soon. Jim(KW4KD)
Well this was a bit of luck: I was trying to find your previous videos on this project, which inspired me to start on a Black Pill based radio as well, but life got in the way, and I've just now gotten back to it. Problem is, I couldn't remember your name or RU-vid channel. Then just now, RU-vid guided me to your latest update. Which on the surface isn't all that groundbreaking, but it points out the advantage of building something from scratch: you know every bit of it, and you can figure out to adapt it to things you hadn't anticipated. In industry, we would call that "feature creep", and it's everybody's biggest headache, but when you have a customer base of 1, you get to pick and choose the features you add! Thanks for the update and the inspiration.
Thanks for the comment. Especially appreciate the point, "the advantage of building something from scratch: you know every bit of it". I'll amplify that & say, "HB gear comes with an unlimited warranty & at home repair service." Hope things settle down long enough for you to see your project on the air. Jim(KW4KD)
Your software agility is truly impressive in extracting the good parts of the Arduino world and injecting them into a stubborn STM32 ide. This has become my favorite radio. Ak2😂b
Jim many thanks! This very clear video nicely covers the recent adds to the rig. Thank you! I am so grateful for this SDR project which has been one of the best and most value add projects I have ever built. The rig is fantastic to operate and brings me joy every time. I would not hesitate to recommend this project to all in our ham radio hobby. 73 Dick N4HAY.
Dick, Really appreciate your endorsement. As said earlier, w/o your input, the project wouldn't be what it is. What I fear is missed in the video is just how much bang for the buck (or fun) can be had with this lil unit. Jim(KW4KD)
I have been trying to do stand alone SDR based on STM32 since 2018 and some prototypes actually work. Included only basic things to be able to receive SSB, AM, and NFM. Now trying to continue work on ESP32-A1S. But your work presented here is really on another level, brilliant. It includes things that I have only dreamed of. I congratulate you on this comprehensive work and wish you to get further ideas to work on. Thank you for shearing this with us. Best Regards!
Thank you for your comments. I just watched your video "Stand alone SDR radio by S52UV". And especially enjoyed your description of the parts selected, and what the software was doing. My "take away" (and this true for for all computer based solutions) is there is more than one way to get the job done. I wish you success with the ESP32-A1S. Just subscribed to your channel. But please drop me a line when you post any progress related to that effort. GL, Jim(KW4KD)
Good Question. But No. Its CW only. Adapting it to do RTTY would be possible, if it had the program space. But the Blue Pill version does not. FT8 is a time synced encoding protocol. This setup has no geo-synced clock. So that's off the table too. Thank you for asking Jim(KW4KD)
Bonjour, j'ai réalisé votre décodeur cw. j'ai suivi tout votre tuto. Et cela fonctionne parfaitement. J'ai juste une demande, le décodeur ne reconnait pas le " ; " , je ne suis pas programmeur et je ne sais pas comment ajouter ce caractère dans les tableaux. Pourriez vous me décrire la marche à suivre..... Je vous en remercie par avance....Jean-Louis
Congratulations.. Nice Job and thank you for the comment. Given what you have done, this is what I would suggest you do. In the BluePill_CWdecoderW_ToneDetect.ino file, look at line 346. It should read, "#define ARSIZE 43". Change 43 to 44 Next, at line 390, which now reads "85" change it to "85," and then insert a new line after it, and enter the value "106" In the table/array "DicTbl1", at now line 438 and reads, " "." " . Change it to " ".", " And last change, insert a new line that reads " ";" " recompile and load your modified sketch to the bluepill. It should now decode "-.-.-." as ";" The binary value of decimal 106 is 1101010. The first "1" is a marker, or place holder, and the following 1's and 0's represent the dash dot pattern sent as morse code for the character. Now you have the power to add/decode any characters you like. Good Luck, Jim(KW4KD)
Merci pour votre réponse rapide. Je suis en train d'étudier pour passer le certificat de radioamateur. Je vous souhaite une bonne continuation. 73s de Jean-Louis-14FRS3818@@jmhrvy1947
Rebonjour, je viens d’effectuer la modification pour ajouter le caractère " ; " , et cela fonctionne correctement...Je vous remercie vivement pour votre aide et pour votre décodeur qui fonctionne à merveille...73s de Jean-Louis 14FRS3818
That is terrific. Was fearful, that the quotation marks would make the instructions too confusing to follow. Good luck with your studies for the exam. Its a great hobby. One which you can go in many directions. My personal favorite is building the hardware you use. Just worked Japan this morning with a 50 watt home brew station. At 76, that still puts a grin on my face. But so does reading that folks like yourself are looking at getting into the hobby 73, Jim(KW4KD)@@jean-louisrico-moulins3304
Good video. I had to print a sink cutout template which only came as a dxf. I tried converting dxf to pdf using online tools and nothing was ever to scale, and I tried using OnShape to make a pdf but it will only make a pdf of the lower corner of the drawing. So far this is the best way. I didn't need bullzip though, since Trueview came with its own virtual printers.
Appreciate your interest. The short answer is, no. There are no formal wiring diagrams; However, the "sketches" (found at the link in the intro description) are commented. And explain which pins are used, and what they do. As for powering up the processors, its your call. If you want to go deeper into CW, I have some other video projects that describe more advance decoder(s) and keyboard based senders. These do have wiring diagrams Tnx es GL Jim(KW4KD)
@jmhrvy1947 thank you. I guess I will need to Google the pinout for my bug connector, and go from there. I have several Uno and Nano boards I have used for various projects. This will be fun!
Hi Jim, thanks for another great video. I just now order boards from JLCPCB. I finished building and got working the receiver part of the Black Pill SDR. I see a slightly cleaner display at the beginning of your video. Are there updated files somewhere that I might download? Also if there are gerber files for the display and black pill board I would very much appreciate them. Thanks again. Tom, ak2b (rx works great)
I love your trigger mechanism. I appreciate your description of pulling back to get your line over just the one bough you want it on. I would like more details on how you lift and secure your antenna. Thank you.
Thanks for the comment. To your question, what I settled on, was attaching a pulley to one end of a steel cable (~ 18" to 2 ft in length. i.e., Long enough go around the limb I wanted to use as the attach point) and an "S" hook on the other end cable, To the pulley I looped parachute cord long enough to go from the ground to the limb. And attached the messenger line to the "S" hook. That way I could pull the cable over the limb. And by having tension on both ends of the cable could hook the "S" hook back on the pulley side of the cable. Once the "S' hook was in place, it basically made a "slip" knot/noose that the harder the pull on the pulley, the tighter the cable gripped the limb. The pulley allows me to raise an lower the antenna (an inverted"V") at will. Happily the setup is still working today.
Great project. It is a pity that you changed to the IDF. Most makers do not use it. They use the Arduino IDE. Which contains the IDF or is based on it. Also you might have a look at ready-made boards that already contain displays. It would further reduce complexity. Similar boards are now also used for LoRa APRS systems. So the ESP32 is used by many HAMS. Just with the Arduino IDE… 73 de Andreas, HB9BLA
Andreas, First, let me say thanks for all your videos. All informative, and all well done. Hope there're many more to come. Your correct, as a programmer, staying "main stream" is generally the path to take. There was a lot, in this project, that the video didn't cover. One, was the display library. Its in fact, an Arduino library. To make it work in the ESPIDF, a number of Arduino IDE core files were also ported. So "Arduino fans" will find commands like "delay()" will work here too. My belief is users will have the advantage of both worlds. As for implementing this project, using an "off the shelf" board, that's a great concept. On the flip side, the project repository includes the Gerber package needed to order the board from JLCPCB. Here in the US, I bought 10, for $5, plus postage. Beyond that, project concepts can readily be confirmed using just Dupont jumpers. Think 7 are needed to connect the display to the ESP32. So for those building just one, jumpers will work fine. But for a "group build" I'd recommend going the JLCPCB route. Thanks for the opportunity to cover some points missed in the video... 73, Jim(KW4KD)
@@jmhrvy1947 Indeed, JLCPCB offers a good service. We even use their assembly service for our projects. It is cheap and you can use smaller parts as well as ICs that can not be hand soldered. And it is a cool feeling when you get a professionally looking PCB of your own design…
No tuning knob, cool! And a particularly clear description of the key elements of your design. There's lots to learn and note in this video. Can you comment on which ADC board you're using?
What an interesting project, making a quality self contained SDR with common and relatively cheap components. It reminds me of a similar experimental SDR by Ian NA5Y. I'll watch the other videos with interest. Is your Tayloe detector schematic in your github repo?
Appreciate the comment. And yes, both the schematics & gerber files are out there (at least they should be; contact me if you don't find them). So the boards can be had through places like JLCPCB. IMO, in spite of its simplicity, its a remarkable CW transceiver, and fun operate. again tnx for checking it out. Jim(KW4KD)
Hi Jim I decided to build your project and unfortunately something is not working or something is wrong bluePill project from your files: BluePillTFTClock program works STM32_diagnose_TFT_support program works STM32_TouchScreen_Calibr_native program is not working STM32_CWdecoderW_ToneDetectR03 the program compiles and uploads but does not work, the screen is frozen and the led is not working... could i get a good design for STM32 main ID=0x1581 lcd best wishes Jim
Ryszard, I'm impressed you got as far as you did. For the Blue Pill version, you're correct the STM32_TouchScreen_Calibr_native was corrupt. I've updated the github repository with new calibrate & decoder files (for the blue pill). I don't own a display with an ID=0x1581, so can't confirm that 'out of the box' these new files wont have an issue with your display. But given that your display worked with the clock demo, there's every reason to believe they can be made to work with it. Please download them & check out the "readme" note found in the 'arduino' folder. Finally, if you have a connection with QRZ.com, & if needed, please Email me directly. Want to see you get your decoder up and going GL, Jim(KW4KD)
Thanks for taking to time to report that this method is still viable. It takes a number steps to get it done. But the software is free. This time of the year some might find it helpful for cutting Xmas figures that they have drawings for.
Hmm I suspect the compass should be moved away from the motor like up on a pole also I have seen some header pins that were magnetic. Choose pins that are non magnetic.
No, I'm sorry. That code is long gone. Doubt that it would compile today, since most all the libraries it used have had many updates of their own. FWIW, the code used here formed the core, of the balance bot videos, that followed. That code is out there. But again, wouldn't bet on it compiling, without 1st having to do a number of "tweaks" to get it to work with today's libraries. Thanks for your interest in the project, and taking the time to ask the question.
I also built one using small dc motor like you n20. By the way, I use nano 168p and your code looks enough to store it, but it doesn't work. 328p can work very well. Really, I still don't understand the minimum requirements of mid-project like that.
First & foremost, always appreciate folks taking the time to share their experiences. My guess as to why the code failed to work today, on a different MPU. 1. Parts of it were written specifically for the atmega32U4. and 2. The code , now 8 years old,. is trying to link to libraries that have, in all likely hood, been revised a number of times since then. All too often a revised library isn't backwards compatible. Rendering code that once compiled and ran fine, useless. Not complaining. Its just the way it is
As usual, your projects are interesting on several levels. Not just at hardware and software but in the subdivisions of both. I can't tell you how much I appreciate your efforts. Have you ever measured opposite sideband suppression on this radio? Tom, ak2b
Thanks Tom. Appreciate the encouragement. To your question about opposite sideband. I did.... Tried, this morning, to pull the plots back up but couldn't find them. Bottom line, it depends on which filter is selected. My favorite is the middle filter (with 97 taps, a theoretical BW of 2000Hz, and shifted 750Hz) . It will let some of the opposite side spill in. But for "normal" band conditions (and congestion) I like it. Its not a "brick wall", so it will "hear" really loud signals fairly distant from the "tuned" frequency. Don't know how that would be for true SSB work. But for CW, & for me, it puts the signal I'm listening to "front & center", while still providing "situational" awareness of what's going nearby. The Narrow filter, with 195 taps, is useful, when the QRN is high, or the signal is weak. But pushes the interrupt timing envelope, which can cause 'clicks' in the audio. For less than optimal conditions the "clicks" are often less annoying than the constant QRN. So each filter has its place. Sorry for not providing more of an analytical answer.
3:50 "...highest miles per watt..." A measure which completely ignores the inverse-square law. If I was awarded the "highest miles per watt", I'd send the award back, along with a copy of a Physics 101 text book. :-)
jmhrvy1947.....what would charge to build one for a man i had a stoke so i lost the ability to remember how to do CW or build such as this.Thank you ... would you mind texting me an answer PLEASE
Andy, Many thanks for your interest in the project. Its been through a couple iterations since the vid was posted. Today, I use a black pill (STM32F411). A faster processor, with more programming space. Build info for the F411 version is also available on Github. Several months back another viewer asked the same question. And I agreed to build one. That unit went to Canada. Shipping costs turned out to being more than twice the price of the unit. Bottom line, the experience was a big disappointment for both of us. On the flip side, the feedback from those who have built the unit themselves, has been favorable. I posted the project, thinking that it would/could be a Father/Son project. (Or some other pairing.) Often the younger members of our world know way more about Arduino like processors than we do, but never thought (or heard) of ham radio/CW. Maybe there's a young person near you, that you can team up with, and make this project a reality. If you all get into it, and need more info, please contact me (QRZ.com is a good way to go ) Tnx & GL, Jim(KW4KD)
Very interesting video, thank you. Great work. I notice the STM32F446 boards are cheap and offer 180MHz which I guess would offer an improvement in speed.
Agree. That higher clock speed would allow you to "see" a bigger picture of whats in the "Neighborhood" . The Sound board is capable of sample rates of >96Khz. But, with the F411 @ <100Mhz, had to run it at 48Khz. Thanks for the comment & interest in the project. Jim(KW4KD)
I doubt it, as the orientation does not impact the build dimensions. But as mentioned in the video, spacing was changed to allow for vertical mounting.
Amazing work. I'm also working on building a sdr transceiver with the blackpill. I'm using libopencm3 with cmsis dsp barebones adc input and pwm output. I saw in your code that you have some options to output coefficients for octave, very clever :) I'm trying to generate +/- 45 degree shifting fir filter coefficients with octave, but I'm still not too familiar with the theory and software. I know people have used the iowa hills software to generate the coefficients, but unfortunately their website is 'under construction' currently. My main problem is summing/mixing my two filtered channels, toggling the channels doesn't change the perceived audio so the unwanted sideband is not being cancelled. I'm also not sure if my current test setup using luaradio to shift around a iq.wav file and feeding that through an aux cable to the blackpill is the best idea. I should probably test with rf and use the tayloe mixer and si5351 topology as you have. I'm making steady progress though. I'm sampling usually at 48000 khz with a decent amount of coefficients, dma for both adc and pwm, half and complete transfer interrupts. I managed to overclock the f411 to 150MHz, it helps a lot with the processing, it was quite straightforward with libopencm3. Well this became retrospective documentation about my attempt :D I'm happy that I found your channel! I should publish my current code, when I do I can share it if you are interested.
Thanks for the report on the work you're doing. And yes, I'd be very interested in any documentation you care to share. So keep me posted. BTW, there's a 2nd video on this project. It reviews adding Si5351 control code to the project. Hope, if you have time, to check it out. Contact me directly through the QRZ e-mail address. Again, thanks for taking the time to post what & how you're using an F411 as an SDR/dsp processor. Jim(KW4KD)
Amazing explanation! Thanks for sharing. Did you face any issues due to the backlash of the N20 geared motors? Would it be effective if I increase the height of the robot and move the center of mass to the upper section to compensate for the motor backlash?
good questions... One reason I selected the N20 motors/gear set, is, for the money, they appeared to have a minimal amount of "blacklash". And to your 2nd question, the higher the center of mass the easier it is to balance. Try balancing a pencil on the palm of your hand, and then do the same with a broom. (sweep end up) Again, thanks for the interest in the project.
Hi, Jim Thank you for the project video on WSPR , I am kind of late of this WSPR, I am be able to get this VFO-AD9850 worked, So do you have any info for the BPF on how to? anywhere or link?
Hello Naiyim, Filter build instructions are found here: github.com/jmharvey1/100mw-WSPR-Transmitter/blob/master/EagleFiles/100mwAmp/CoilBuildInstructions.txt GL with the project, and let me know how it works out for you. Jim(KW4KD)
I was having a problem similar to what you mentioned, when I was using the MPU6050 in DMP mode with interrupt and at the same time driving motors with PWM. Turns out it was a problem with the libraries using the same timer (overlapping of timer usage), the interrupts etc and it was messing with the timings making the motors go beserk. Also the Gyro readings would not make sense sometimes. To be fair mine was a balancing robot though. If it helps, you may check if your problem is something similar perhaps.
