Hi,
I've read in several posts about 'thinking of electrones as your car' so you should provide rounded corners or 45 degree bends. What's the need when component leads meet with the copper tracks at 90 degrees??
I've read in several posts about 'thinking of electrones as your car' so you should provide rounded corners or 45 degree bends. What's the need when component leads meet with the copper tracks at 90 degrees??
its a myth - audio frequency currents basically don't care about anything but the least resistance path
in PCB "design for manufacturability" some trace features are avoided because of etching limits, or rare failure mechanisms, some hypothetical, or out dated by improved materials, processing
at RF frequencies, modern digital, the rules change because the high frequency components are sensitive to the waveguide/transmission line properties of the trace/plane geometry - then you don't want the impedance discontinuity of sharp 90 degree bends
in PCB "design for manufacturability" some trace features are avoided because of etching limits, or rare failure mechanisms, some hypothetical, or out dated by improved materials, processing
at RF frequencies, modern digital, the rules change because the high frequency components are sensitive to the waveguide/transmission line properties of the trace/plane geometry - then you don't want the impedance discontinuity of sharp 90 degree bends
Last edited:
So in your case I think we can conclude this:
Short track as possible
Cut corners or rounded corners
= Looks good!
It looks very unprofessional to have 90 deg angles on a pcb.
Short track as possible
Cut corners or rounded corners
= Looks good!
It looks very unprofessional to have 90 deg angles on a pcb.
You need to think about what signals your tracks are carrying. If power, then wider tracks will give less voltage drop. If signal, then narrower tracks will give less capacitive coupling. In audio you may find that minimising loop area is more important than other things, so 'go' and 'return' paths should be near each other. Don't blindly follow rules (especially silly rules found on the internet); try understanding instead.
Rounded corners are used in R.F work to keep the track width uniform so there are no sudden changes in impedance, inductance, capacitance etc. They are not used to help the electrons turn corners. Electrons weigh next to nothing and can corner at light speed without a problem 🙂
At audio frequencies the inductance and capacitance is to small to have the same effect as it does at R.F and the lower impedance of a 90 degree corner (due to the wider track width) will only make a tiny difference to the overall series impeadence
Its evident different diyers have different reasons. Even if rounded corners are used in RF signals pcbs,it seems to work for the famous musical Aksa55.
Have you actually read post 6? Even where rounded corners are correctly used (RF) it has nothing to do with electrons' cornering speed.
yes i wondered about this too and acko explained it for me as above re impedance consistency/control. i would also think that perhaps even at some modern dac clock and DSP speeds that RF technique would not do any harm? certainly other RF technique is put to good use in dac layout.
...no actually I am being pedantic, let me explain the difference...
electrons do have mass, don't move too quick in room temp metals, strictly speaking it would be photons in their guise of EM fields that travel ~ 1/2 c in most transmission lines
for audio corners really don't matter except for how the size/orientation of loop areas change for mutual inducatance linkage
electrons do have mass, don't move too quick in room temp metals, strictly speaking it would be photons in their guise of EM fields that travel ~ 1/2 c in most transmission lines
for audio corners really don't matter except for how the size/orientation of loop areas change for mutual inducatance linkage
yes, but i was trying to get more info on the boundary of the increasingly complex and fast mixed signal designs of todays audio processors, DSPs, Clocks and even DACs. when clocks are running at over 100Mhz with opamps, comparators, receivers and transceivers to match, does this not blur the boundary somewhat between audio and RF design?
Yes, sharp pulses may benefit from rounded corners to reduce reflections. 'Sharp' means much faster rise time than any audio signal! Rough guess: frequencies above 100MHz, transitions measured in 10's of ns or less. Relevant to digital audio equipment, not analogue audio signals.
you really don't know which rules to apply just because it is a "audio product"...
normally one designing mixed signal circuits would make the distinction between the Audio Frequency signals and the Digital ones, often maintaining different gnd plane partitions, separate supplies, possibly shielding
even Class D typically isn't fast enough for full digital transmission line theory to be 1st order important in layout - just consider it as very dirty analog, pay attention to mutual inducatance, common gnd, power impedance coupling, part parasitics
normally one designing mixed signal circuits would make the distinction between the Audio Frequency signals and the Digital ones, often maintaining different gnd plane partitions, separate supplies, possibly shielding
even Class D typically isn't fast enough for full digital transmission line theory to be 1st order important in layout - just consider it as very dirty analog, pay attention to mutual inducatance, common gnd, power impedance coupling, part parasitics
At work I rarely get to see corners in pcb-tracks other than 45 degrees. But when frequencies get high, and I mean really high (e.g. on boards used in TV studios), will you see rounded corners. And tracks that do the slalom.
That might be true, but I can occasionally hear tires of very fast electrons squeaking in sharp corners! 😀
- Status
- Not open for further replies.