[jnovek]

How do you “git gud” at this sort of troubleshooting?

I’ve built about a dozen synth modules over the last couple years and I’d say about half of them actually work 100% as intended. I’ve stopped building new modules because my success rate is so abysmal and I’m pretty stuck on the basics for troubleshooting.

I do flux and reflow any joints that look sus and I test with my meter to make sure that pins on chips don’t unexpectedly go to ground or +/-12v per the data sheet but that’s about the extent of my skills.

Any resources or tools to get better at this sort of thing?

[MisterTea]

You need a basic understanding of how components such as diodes, transistors and op-amps function. From there you can gain insight into the signal path and begin to work your way through the circuits.

Next up, tools. A must have tool is an oscilloscope to debug AC circuits as it enables you to visualize the signal. Buy one, and learn how to use it - you don't need anything fancy for audio work. A signal generator might be helpful to enable you to generate reference or test signals but is not necessary. You can buy scopes with signal generators built in. I'd also recommend an LCR meter to measure the values of inductors and capacitors though most multi-meters have a capacitor test function (though usually limited vs a real LCR.) Its also handy to have a little bench top amp/speaker combo or a set of computer speakers for testing. Do not use headphones for testing as erratic signals and/or connections can create LOUD transients that can cause hearing damage.

I'd start with the basics as you do - check supply voltages and look for bad solder joints. If the basics check out then you need to break out the scope and start tracing the signal path. Start by probing at the signal source to make sure you do have an incoming signal and then work your way through the circuit to the output. This is where it helps to understand the basics of electronics. A schematic of the circuit you are debugging is very helpful as attempting to follow traces on a board is not fun and error prone.

Some Resources: http://sound-au.com/ - can't recommend this DiY audio site enough. https://www.amazon.com/Troubleshooting-Analog-Circuits-Desig... good book I read years ago - Bob Pease was a fantastic engineer and writer.

[speed_spread]

I'm on that path too. I used to work mainly with a voltmeter but I recently bought a cheap oscilloscope. It does wonders by allowing me to actually see the signal at different points and confirm that it matches my expectations. That was a big unlock for me.

[abrugsch]

one starting point is to try and build one of the simplest synth components on the MFOS site (IIRC they are categorized by difficulty) as mentioned in my GP post, the MFOS explainers are incredible. But it does also help to have at least a little background in electronics. I started out by messing about with stuff as a kid (having a dad that was a ham helped) then eventually doing an EE degree at uni.

Fast forward 20 years and I didn't use the EE part of my degree in a career path except for the software eng part so I'd forgotten most of the theory. However the MFOS troubleshooting guide (for the WSG, and I assume most/all of their self build kits) is absolutely comprehensive, telling you exactly what points to probe and what you should see in a step-by-step logical progressive manner. You do definitely need a scope though.

For debugging analogue synths, you could probably even get away with a $25-40 STM32 based one from aliexpress as you're typically checking for DC voltage levels and human-audible waveforms ( < 20 KHz and probably realistically ~1-5 KHz) and the cheapest of cheap scopes can handle that.

so to recap:

tools:

reasonable to good multimeter - DC/AC voltage, current, resistance, continuity. (doesn't have to be a eevblog couple of hundred bucks one, but auto ranging would be useful, an audible continuity beep is a must though as is a diode check.)

Oscilloscope. for doing analogue synths, anything capable of displaying a waveform($). for anything else, the more you can budget the better. for debugging digital electronics you'll probably want multi-channel ($$) and/or MHz of range($$$). For doing RF, you'll want GHz ($$$$$)

Logic Analyzer - for digital electronics and reading the logic levels on multiple paths (bus lanes) at the same time.

Bus Pirate - for analyzing and debugging the content of various serial busses like I2C, SPI and CAN etc. not necessary but makes things easier than trying to do it with a scope or logic analyzer alone.

most importantly -

Curiosity! cannot overstate this enough, as most of the process would be very frustrating unless you have the innate curiosity to push you through.

Resources:

MFOS site in the OP. build guides with ELI5 level explainers and comprehensive troubleshooting guides

EEVblog youtube channel[0] (especially his older stuff) and forum [1]

Great Scott Youtube Channel[2]

BigClivedotcom youtube channel[3] for some hilarious teardowns and circuit explainers once you have a few videos of those in your watch history, you'll start getting plenty of relevant recommendations of other EE-tube creators

[0]https://www.youtube.com/playlist?list=PLvOlSehNtuHsCTtj-T_vk... (just one of many suitable playlists) [1] https://www.eevblog.com/forum/ [2] https://www.youtube.com/playlist?list=PLAROrg3NQn7cyu01HpOv5... [3] https://www.youtube.com/@bigclivedotcom

[jnovek]

Part of my problem is definitely understanding some aspects of these circuits on a deeper level.

I haven't done any MFOS builds but I have done a few "cores" to various synth modules (Moritz Klein's youtube is great for this) on the breadboard -- most successful was a square wave oscillator based on a schmitt trigger -- but often I'm overwhelmed by the supporting circuitry which is sometimes more complicated than the core.

For example: I cannot get my head around the design of an exponential converter.

I have trouble understainding the feedback loop between the two transistors. Ignoring that, even, as far as I can tell it creates an exponential relationship between voltage and current. However, my schmitt trigger is voltage-controlled, so don't I need a transimpedence amplifier of some sort to get back to voltage? But I can't see anything that looks transimpedence amplifier in the diagram...

I constantly feel like I have about 1/3 of the knowledge I need to proceed.

Perhaps I just need to accept certain things are outside of my understanding right now?

[abrugsch]

TBH I didn't understand any of the concepts you mentioned, and that's with background knowledge. So I'd say you're definitely talking about way beyond the basics concepts that understanding of will only come after learning the basics and then targeting your learning at synth specific designs, which will take a lot more time and discipline than a HN post can cover. I don't have that knowledge or resources to point to unfortunately (any any book covering the topic is likely 2" thick in dead tree format...)

[jnovek]

I feel like, as far as general electronics goes, I have the absolute basics. I understand basic components and circuits 101 stuff like voltage dividers and filters.

I have trouble with:

  - Putting all the concepts together to understand larger circuits

  - Concepts that are hard to visualize, like feedback

  - Recognizing circuits from the diagram on the PCB

Maybe I just need more practice?

Thanks for talking through this with me. :-)

[abrugsch]

this might help with the visualization aspect https://www.falstad.com/circuit/