I'm a total design newbie. In fact, I am really crap at soldering too 😀 anyway, I designed my own guitar/bass "preamp" of sorts - in actual reality, a glorified overdrive stompbox. It's an original design, even though I did take inspiration from various circuits (Rod Elliott's website was instrumental).
Anyway, I really don't know how to go about designing PCBs. The circuit itself should work well enough (it's a bit low in headroom at 9V, but it should work well enough if placed in front of a guitar/bass amp), I simulated the thing countless times and it should be OK functionally speaking.
When designing the (dual-layer) PCB, I tried employing the best practices as instructed by various resources on the web - there isn't much dedicated to analog audio design, really, and everything assumes at least mixed signals - which I can summarise as follows:
Now, I can attach a picture of the (computer rendered) PCB, or maybe you need the schematic? Problem is, the silkscreen has yet to be fine-tuned (right now things are a jumble) so it is not really legible.
I also shied away from designing ground planes, since they can be more trouble than they're worth; I only, very timidly, tried making one around the switching circuitry, but I have my doubts as to whether it helps or breaks the design. As of now, it's all done with a star ground.
Anyway, I really don't know how to go about designing PCBs. The circuit itself should work well enough (it's a bit low in headroom at 9V, but it should work well enough if placed in front of a guitar/bass amp), I simulated the thing countless times and it should be OK functionally speaking.
When designing the (dual-layer) PCB, I tried employing the best practices as instructed by various resources on the web - there isn't much dedicated to analog audio design, really, and everything assumes at least mixed signals - which I can summarise as follows:
- keep power and signal grounds separate - with a caveat: the design features three dual opamps and two JFETs, so the JFETs have to share grounds with the opamps' power lines
- put any switching circuitry as far away from the signal processing as possible
- if power/ground and signal paths run close (on separate layers), then they should try and cross orthogonally rather than run parallel
Now, I can attach a picture of the (computer rendered) PCB, or maybe you need the schematic? Problem is, the silkscreen has yet to be fine-tuned (right now things are a jumble) so it is not really legible.
I also shied away from designing ground planes, since they can be more trouble than they're worth; I only, very timidly, tried making one around the switching circuitry, but I have my doubts as to whether it helps or breaks the design. As of now, it's all done with a star ground.
I'd not be too worried with a sigal-level only design - layout can match the natural flow of the signal path, normally power feeds in at the output end and can be C- or RC-filtered as it progresses towards the input to decouple any output stage currents from affecting the input.
Both schematic and layout are useful to see...
Both schematic and layout are useful to see...
One good way to learn about analog layout is to look at examples of other similar designs. Making note of how the power and grounds where routed is important.
Most stuff was done single sided so you are at an advantage of being able to do double sided, which allows for tighter layout to save pcb space. Lots of single sided designs needed a jumper here and there.
I try to make my protos <= 100x100mm if possible to get the lowest pcb pricing from China.
As said above get your design ready for review and we can take a look at it
Most stuff was done single sided so you are at an advantage of being able to do double sided, which allows for tighter layout to save pcb space. Lots of single sided designs needed a jumper here and there.
I try to make my protos <= 100x100mm if possible to get the lowest pcb pricing from China.
As said above get your design ready for review and we can take a look at it
Last edited:
I do not have Kicad,
1) generate a pdf of the schematic and
2) the pcb layers as pdf, jpg or better, the gerber/drill data is best.
1) generate a pdf of the schematic and
2) the pcb layers as pdf, jpg or better, the gerber/drill data is best.
Post a screenshot (or .pdf) of the schematic and a screenshot of the layout with all pours/planes shown. I prefer to see top and bottom copper along with the silk screen. Then create two more screen shots. One that shows just the top copper with the silk screen and one that shows just the bottom copper with the silk screen.
You don't have to disable all the other layers for this as long as they're not too distracting.
Tom
You don't have to disable all the other layers for this as long as they're not too distracting.
Tom
There you go. The silkscreen, as I said, is still a bit of a jumble, but I printed the courtyards as well so that you can sort of make sense of the components.
To summarise:
- power supply is on the far left, around with the 47R resistor (dimensioned so as to not burn out the diode even if applying 24V reverse voltage; the design can easily accomodate higher voltages than 9V)
- the clipping indicators' circuit is on the far bottom, as far from the rest of the circuit as possible. Keep in mind that this will go into a stompbox (a BBDD enclosure, to be precise) and potentiometers are placed over the left-hand side. The "hole" on the left is to possibly insert a battery for battery operation
- I tried making a copper fill around the clipping indicators' circuit, whereas the rest is simply a star ground.
- I wonder whether it would be more appropriate to make the clipping indicators' circuit share a layer with ground or with signal (as of now, signal is all on the front, while power and grounding is on the back).
It would help if you could produce screen plots as tomchr suggested.
One quick note, though. Your star ground would be better at C1 rather than at J1.
One quick note, though. Your star ground would be better at C1 rather than at J1.
I tweaked the silkscreen, it should be clearer now. (Although components are very close together, so I don't know what would the best way to arrange the silkscreen so that it is clearest).
Why should I move my star ground? Unfortunately C1 is a bit inconvenient as a star ground, unless I reshuffle things around.
Thinking about it some more, I would add a decoupling capacitor at J1 and leave the start ground there. That would deal with any noise picked up by the power lines.
Also, check out this article: https://hifisonix.com/wp-content/uploads/2019/02/Ground-Loops.pdf
Reducing loop areas in your power circuits and signal circuits will improve noise performance. For example, there is room to keep the V+ traces closer to the ground return traces. You can also use vias if you run into routing issues.
I would also increase the width of the traces for lower resistance.
Also, check out this article: https://hifisonix.com/wp-content/uploads/2019/02/Ground-Loops.pdf
Reducing loop areas in your power circuits and signal circuits will improve noise performance. For example, there is room to keep the V+ traces closer to the ground return traces. You can also use vias if you run into routing issues.
I would also increase the width of the traces for lower resistance.
Q1 is bootstrapped, so it is actually much higher resistance to ground. 3.3n is actually fine in simulations.
So something like this? (See attachment).
If I run VCC and/or GND closer together, they end up being (marginally) longer. The distance between them is, on average, half the width of an 8-DIP IC. That seems quite close together to me?
Signal traces are 10mil, 0.25mm (to reduce parasitic capacitance, currents are really small and impedances are kind of large) and power traces are 25mil, 0.65mm (actually slightly larger than 25mil). I took these numbers from a popular DIY stompbox designer. I also wanted to avoid vias as much as possible since they can act as antennas, or so I read.
You can reduce the big ground return loop around Q2, TP, and TP3 by moving their traces to the Red layer. Then you can straighten out the Green trace.
So, maybe, something like this?
I actually want to try and design a ground plane. I know how you do that in KiCAD, what I don't know is how I should design it.
Should I just fill the whole PCB with a ground plane (to GNDPWR)? Should I make one for GNDREF (4.5V)? Should I separate signal ground (the top part of the PCB), the power input ground (on the right), the switching circuit (at the bottom)? I placed the switching circuit entirely on the opposite side of the PCB with respect to the signal processing, did I make the right choice? So many questions 🤣
I actually want to try and design a ground plane. I know how you do that in KiCAD, what I don't know is how I should design it.
Should I just fill the whole PCB with a ground plane (to GNDPWR)? Should I make one for GNDREF (4.5V)? Should I separate signal ground (the top part of the PCB), the power input ground (on the right), the switching circuit (at the bottom)? I placed the switching circuit entirely on the opposite side of the PCB with respect to the signal processing, did I make the right choice? So many questions 🤣