Troubleshooting - Radio Design 101, Epilogue 2

This second follow-up to the Radio Design 101 series offers tips on how to troubleshoot electronic circuits. The problems found in Epilogue 1 are addressed and the radio performance is improved - but the main focus is on general techniques. A preview look at emerging Q-enhanced filter technology is also provided, and an example of applying the troubleshooting techniques to car repair is also shown.

[@alexgreysukh4875]

A very good course. Please continue!

[@MegawattKS]

Thanks! Glad it was helpful. There's now a website with links to all the videos - in this series as well as one on Antennas/Propagation and on Test Equipment. The website is still under construction, but if you haven't run across the other series yet, it might prove interesting. Thanks again for leaving the comment. ecefiles.org/

[@alexgreysukh4875]

@@MegawattKS Thanks, William. I will certainly follow the site. Given your academic position, it would be very useful if the website maintains periodic reviews of innovations in the field with a focus on what could be experimented with by the hams.

[@johnwettroth4060]

I just discovered this series- very enjoyable. I've done a lot of this kind of design in the past and its neat to see how someone approaches today especially with Tiny SA and VNA- amazing. One question I had is the biasing of the two cascode stages in the IF. It seems like they should have a divider from Vcc to the upper transistor base- its pegged at Vcc. I also thought it bypassing this base would be required to get the full effect of the cascode.

[@MegawattKS]

Thanks for the question. Normally at the university, we use an LM3189 or SA604 IC for the IF limiting amplifier and demod functions. But here I was trying to do everything discrete, and leverage prior discussions. As you noticed, this design departed from the traditional biasing. I always like to try new design ideas. The goal is to achieve a high-gain amplifier with an amplitude limiting behavior like in the ICs, but with single-ended rather than differential circuit topology. Here, the output swing at the collector can be as large as 4.3 to 5.7V (0.7Vpk or 0.5VRMS). That is because an NPN transistor collector with base at Vcc can actually swing lower than the base and still stay in the active region of operation - i.e. still be amplifying throughout the cycle. The rule is to just not let the base-collector junction become forward biased by more than 0.6 or 0.7V. It doesn't actually have to be reverse biased. (Recall VCEsat tends to be around 50 to 100 mV, and Q2 emitter is about 0.75 V below Vcc here.) At swings larger than 4.3 to 5.7V amplification falls as it goes into limiting - but that's the goal in an FM IF amplifier. Note that emitter-follower / buffer Q3 limits the swing on the high-side due to its base-collector becoming forward biased at 5.7V. I also agree that the base of the upper transistor in the cascode needs to be bypassed. In the actual build, that is done with the power supply bypass cap which is only a couple hundred mils away. In the simulation it works fine without it since the supply voltage source is an AC ground and simulators don't know about inductive parasitics unless we put them into the circuit explicitly. But in the build, a local bypass cap is needed at the base as you noted. Here it is the Vcc bypass cap. It doubles as the base bypass 🙂

[@johnwettroth4060]

@@MegawattKS - thanks for the thoughtful reply and I agree that the CB junctions on the cascodes don't have to be reverse biased- just not forward- I generally explicity reverse bias them but learned/relearned something. The benefit to having that upper transistor with a real bias R and bypass at base is PSRR and potential feedback issues that you've experienced. For the cost of an R&C , you can really keep junk on that rail out of the upper Q. It also would alleviate the feedback issue you had with the solved with 20 ohm R- an alternative solution really. Good puzzle, good presentation. Enjoyable.

[@jburchm1]

Excellent video. Might I suggest adding one general troubleshooting principle? Be cautious, or at least mindful, when making changes to test your theories, as your changes may mask the fault in the system under test without actually resolving the problem (bad), or interact with the fault in an unexpected way (good, if it gives you additional information to form more accurate theories). You demonstrated the good side of this quite well when the oscillation did not return after removing the probe/your finger. It provided additional information that helped to ultimately solve the problem.

[@MegawattKS]

Thanks. Good points. Troubleshooting is definitely an art.

[@katosembatya8353]

Thank you

[@MegawattKS]

You're welcome !

[@xDR1TeK]

Forget memory, I'm tattooing them. On my brain.

[@ftscotttinez33]

Is touching the circuit with your finger a trouble shooting technique? I guess I never would of thought to load the circuit by putting my finger on it. Is the behavior of the circuit a characteristic of an oscillating amplifier in general, and whether the amplifier is oscillating a question I should ask myself when wroking in RF? Without watching this video, I would of never thought about putting my finger on the circuit without knowing what behavior I should be looking for first haha.

[@MegawattKS]

Agreed. It looks silly, but it's absolutely a technique I've used many times. But ONLY when its safe (low voltage, low-power receive circuits, isolated from the mains). That warning given, its kind of like tapping/pressing on stuff to find bad solder joints, or using freeze spray to find temp-sensitive things. In this case, at VHF and above, the finger adds some capacitance, but also power losses - which can stop the oscillation and hopefully help suggest where the feedback path may be. Of course it's not very scientific beyond that qualitative description - and as you said, how do we map changes in the observed behavior to ideas on tracking down the oscillation? It could always be that we're getting false clues. But any clues are better than no clues, IMO 🙂

[@ftscotttinez33]

@@MegawattKS awesome! Thank you so much! Also, quick question. Are you going to do a series on the Doppler radar?

[@MegawattKS]

@@ftscotttinez33 I've thought about that, but for various reasons decided against it (at least for now). FWIW, that class (as well as the one that Radio Design 101 is patterned after) are still taught at K-State.

[@sullivanzheng9586]

Finally!

[@MegawattKS]

Yes - it did take a while. Kept messing with the slides (refining them)...

[@Homer19521]

Maybe I missed it, but why did the module go out simultaneous to the heater core mod? Bad luck?

[@MegawattKS]

That's a good question. It is not impossible that it was a coincidence. However, the probability of a dual failure at essentially the same time in unrelated systems is statistically very near zero. My best theory is that the head overheated. The ICM module is mounted to the head and uses it's temperature as part of the heat-sink for the electronics (its the case-temp for the module). While I did look at the temperature readings on the drive home from the bowling alley where the heater core failed and the temp did not get too high - there are actually two different temp sensors on the car. One feeds the analog gauge and the other feeds the digital readout on the dash. It could be that the one reading the cooling fluid for the head wasn't the one I was looking at and/or the head wasn't flush with fluid so it was higher temperature than what I was monitoring while driving home. All theories of course... :-)

[@katosembatya8353]

For those who need more reading , can you recommend some good books for this topic. Thank you once again

[@MegawattKS]

Sorry - I don't have any books on the topic. Doing some web searches, I notice there are a few, but they tend to be large and expensive - and like any books, one needs to be sure they fit what one is looking for. I'd like to pick one and suggest it, but it is unlikely I would pick a good fit without knowing a lot more about your background, goals and areas of interest. The points I mentioned in the video are just from my own life experience. I was lucky to learn some of the basics when I worked in a TV repair shop many decades ago. It is interesting how those skills applied in so many areas of life going forward.

[@xDR1TeK]

But, but what about designing the LO to be stable and tunable? PLL? What's a corvette? 😂

[@MegawattKS]

Sorry for the late reply(s). Agreed - a frequency synthesizer is definitely a good thing :-) We sometimes did some PLL designs in the class, but here I just used the TinySA. Synth/PLL chips can be purchased of course, but that at least requires software development to control them. I took the easy way out in this video series :-)

[@xDR1TeK]

@@MegawattKS your knowledge is valuable. I'm envious. I try hard but I'm afraid I'm limited. Currently building an AM transmitter, but I find myself stuck with what requirements to base the design by. Frequency, mixer, output impedance, etc... I have purchased many crystals, some overtone, some not. I wonder if the diode ring mixer works for transmission? I know it's not needed for modulation. Should I go with active filters? What works for basic operation and what works for more intricate designs?

[@MegawattKS]

@@xDR1TeK I think a lot is going to depend on frequency of operation. Are you targeting an HF-band (3 to 30 MHz range) design? Of course there are tons of options in a design. And I purposely stayed away from transmitter design in the series due to rules and regulations. But there are good books available (e.g. through the ARRL and other radio organizations).

[@xDR1TeK]

@@MegawattKS I understand. Part of engineering rules is to adhere to ethics. But having an engineering degree isn't fun if all you get to play with is LEDs. Thank you for everything.