Next Funbox module, Saturn! Saturn is a “spectral delay”, which is turning out to be one of my favorite effects. There are a handful of audio plugins that are advertised as spectral delays, but I have never seen one in pedal form. It works by using a STFT which lets you operate in the frequency domain, as opposed to time domain (same as the Venus module). The STFT breaks up the input signal frequencies into groups and delays each group by different amounts of time. Each group contains a frequency range of about 47Hz. The result is fragments of your guitar sound coming back at slightly different times, which to me kind of sounds like windchimes. There were a couple of techy blog posts that guided my code development on Saturn, so I’ll share those here for further reading if you’re interested:
I can’t speak for the FunBox, and am not half the EE tele and keith are, but I do recall from an issue with noise coming from a bluetooth receiver than a possible solution was to use the AGND for the audio signal. I’m not suggesting this would fix your issue, but it is perhaps not wise to tie AGND to DGND if you don’t have to.
It might also be worth mentioning that the other solution to reducing noise was to isolate the power going to the BT module and the amp using a B505 or somesuch (power isolator), though I seem to recall a note at the end saying that ended up reducing but not totally stopping noise.
That was of course the whole issue – the BT module was noisy, and it was passing this noise back up to the power supply which was shared by the amp.
Well, I didn’t have an answer, but since I just payed OpenAI $25 for extra access to it’s extra smart engine, I thought I’d make IT tell me the answer. (Note, I didn’t tell it which one was noisier). Full conversation is here https://chatgpt.com/share/66f01f1b-0900-800a-89e5-d43a1163de95 but I’ll see if I can post the highlights.
The differences in noise behavior between the TL072/TL074 and MCP6024 op-amps can be influenced by several factors, including gain bandwidth and slew rate, but also their architecture and how they handle different frequencies.
Key Parameters:
Gain Bandwidth (GBW): This specifies the frequency at which the op-amp’s open-loop gain drops to 1. A higher GBW (like 10 MHz for the TL072/TL074 vs. 5.5 MHz for the MCP6024) means the op-amp can amplify signals with higher frequencies before its gain drops off. While higher GBW improves the ability to amplify high-frequency signals, it may also make the op-amp more susceptible to noise or interference from higher-frequency components in your circuit, such as digital noise. The MCP6024’s lower GBW makes it less prone to picking up high-frequency noise but at the cost of bandwidth for amplifying high-frequency signals.
Slew Rate: This is the maximum rate at which the output of the op-amp can change in response to a step input (volts per microsecond). A higher slew rate (13 V/µs for the MCP6024 vs. 7 V/µs for the TL072/TL074) allows the op-amp to react faster to changes in input voltage, meaning it’s better at amplifying fast-changing signals. A higher slew rate can help avoid distortion in fast signals, but it can also make the op-amp more sensitive to noise, especially from digital switching noise.
Noise Architecture: The MCP6024 uses a CMOS input stage, which is known for having lower input bias current but can pick up higher noise in some situations, especially from digital circuits. The TL072/TL074 uses a JFET input stage, which tends to have better noise immunity but may have a slightly higher input noise voltage compared to CMOS op-amps.
Why One Might Be Noisier:
TL072/TL074: Due to their higher gain bandwidth and lower slew rate, they can amplify a wider range of frequencies. This can make them more susceptible to picking up noise from higher-frequency digital components. However, their JFET input stage generally handles low-frequency noise better.
MCP6024: Despite its faster slew rate, the MCP6024’s lower GBW makes it less sensitive to high-frequency noise. Its CMOS input stage might be more prone to picking up noise from digital circuits, depending on the specific layout of your circuit and grounding. The faster slew rate could also cause the op-amp to react more dramatically to fast digital signals, potentially amplifying switching noise.
Practical Consideration:
If the digital part of your circuit is causing noise, it might couple more into the MCP6024 if it is picking up fast transitions or spikes. The TL072/TL074, on the other hand, might pick up a broader range of frequencies but would be less sensitive to rapid changes in the signal.
Would you like further details on how circuit layout (such as grounding) might influence noise in your application?
Yes!
[spoiler]The layout of your circuit, especially concerning grounding, power supply decoupling, and component placement, can greatly influence how noise from digital parts couples into the analog components like op-amps. Here are some key factors:
1. Grounding (Star Grounding vs. Common Ground Plane)
Star Grounding: This method uses a central point (star point) where all grounds converge. Analog and digital ground lines are separated and only meet at the star point, minimizing the risk of digital noise coupling into the analog ground. This is effective at reducing ground loops and preventing digital switching noise from interfering with sensitive analog circuits.
Common Ground Plane: In a common ground plane, both analog and digital grounds share the same plane on the PCB. Although this is simple to implement, if not carefully designed, digital currents flowing through the ground can create voltage differences that induce noise in the analog circuit.
Impact on Noise: Poor grounding can cause digital switching noise to flow through the ground shared with analog components. This can be particularly problematic for op-amps with high gain bandwidth or fast slew rates (like MCP6024). TL072/TL074, while still susceptible, might handle this better because of its JFET architecture.
2. Power Supply Decoupling
Decoupling Capacitors: Using capacitors (typically 100nF ceramic close to the power pins and larger electrolytic caps further away) between the op-amp power supply pins and ground is essential to filter out high-frequency noise from the power supply. This prevents power supply noise from modulating the op-amp’s input/output signals.
Separate Power Supplies: If possible, having separate power supplies or power planes for digital and analog sections can help isolate noise. For instance, the digital logic and the analog op-amps would have separate 5V or 3.3V lines that only meet at a central point (like the star ground).
Impact on Noise: Without proper decoupling, noise from the digital section (such as clock signals or switching spikes) can couple into the op-amp’s power supply, leading to performance degradation, especially in high-gain applications.
3. Signal Routing and Trace Layout
Keep Digital and Analog Sections Separate: Physically separating digital and analog sections on the PCB minimizes the risk of noise coupling. For example, the digital section (microcontrollers, high-speed switching components) should be on one side of the PCB, and the analog section (op-amps, filters) should be on the opposite side.
Shielding Sensitive Signals: Use grounded traces or planes to shield sensitive analog traces from noisy digital traces. If possible, run sensitive analog signal traces away from digital signal traces, especially those carrying high-frequency signals (like clocks or data buses).
Short, Direct Traces for Analog Signals: Minimize the length of analog signal traces to reduce their susceptibility to picking up noise. Long traces act like antennas, amplifying noise, particularly in high-frequency environments.
4. Op-Amp Characteristics in Relation to Layout
MCP6024: Because of its faster slew rate and CMOS input stage, this op-amp can respond very quickly to voltage changes, including noise spikes from digital switching or fast changes in nearby signals. In a poor layout, this sensitivity can lead to more noise being amplified, especially if there are long, unshielded traces.
TL072/TL074: These op-amps, with their JFET input stage, are generally more robust in rejecting noise from slower, low-frequency interference. However, due to their higher GBW, they are more prone to picking up a broader range of noise frequencies if not well-isolated.
5. Capacitive and Inductive Coupling
Capacitive Coupling: This occurs when high-speed digital traces are near analog traces. The rapidly changing voltages in the digital traces can couple into analog signals via stray capacitance. Keeping digital signals away from sensitive analog signals and using ground planes between layers can help mitigate this.
Inductive Coupling: This occurs when large current loops in the digital section induce currents in nearby analog loops through magnetic fields. Minimizing loop areas, especially in power traces, can reduce inductive coupling. Placing analog and digital sections on separate layers with an intervening ground plane helps to minimize this issue.
Conclusion:
MCP6024 might be more sensitive to layout issues, particularly in how the ground planes and power supply decoupling are handled. Its faster slew rate and lower GBW can make it more likely to react to fast digital noise, especially if grounding is not optimized.
TL072/TL074 might handle noise better in some scenarios due to its JFET inputs and slower slew rate, but it still needs good layout practices, especially because of its higher GBW.
To reduce noise:
Use separate analog and digital grounds (star grounding if possible).
Properly decouple the power supplies with capacitors near each IC.
Keep digital and analog sections physically separated on the PCB.
Shorten and shield sensitive analog traces.
Would you like more details on any of these layout techniques?[/spoiler]
The disadvantage being that electro-mechanical devices are not small, so having them mounted on their own PCB like Joyo has done with the 3PDT means you don’t have to leave room on your main PCB (and you can re-use it in multiple design).
I’ll probably pull apart the “Analog Delay” (which is actually digital) next, and we’ll see how they designed it.
From OpenAI (take it or leave it), I’ll post an updated link and exhume the content from the message.
When using the Daisy Seed in a buffered guitar pedal, noise issues can arise from several factors, especially because the Daisy Seed is a digital processing platform that operates with relatively high-speed components (for audio purposes). The interaction between the digital section of the Daisy Seed and the analog circuitry of the buffer can introduce noise if not carefully managed. Here’s what to consider:
Fair point. I usually use OpenAI when programming, and I never cease to be amazed at how correct it can make code look while being ultimately disfunctional. I laugh to myself at the thought of a non-programmer trying to cobble something together with OpenAI.
So I really should know better than to trust it on a subject that I know far less about, despite how logical its arguments seem.
I will say that it mentioned three things that I already knew to be possible issues, and that I think I might be forced to agree with its suggestion that the Daisy Seed simply isn’t up to snuff for production designs. But I wouldn’t like to pass judgement until I actually build and test a FunBox myself.
After reading your posts on pedalpcb regarding adding an extra LPF, and reading this associated note from the TI datasheet on the ADC/DAC, it would seem that “out of band noise” gets “wrapped” back into the audible spectrum when output is coupled to some devices.
Solution: add a LPF. https://www.ti.com/lit/an/slaa313a/slaa313a.pdf
(Yes, there is already an LPF on the Daisy Seed, and it is an exact match for one of the suggestions by TI, but there is a second suggestion with more attenuation that I suspect is closer to what you ended up implementinig).
Great work as always Keith! I’m building this for sure with some modifications! I will update you…
This would sound really good with an acoustic guitar. Did you try with one? It would be nice to know if the buffer circuits designed for the funbox would work well with Piezo inputs…
Ilooked up KNA’s site. They specify that their pickups do well when the input impedance of our equipment is 1M ohms… What Funbox’s?
