Hi,
I have the schematic with passive caps about some opamps. How do I configure KiCad to put the passive component beside the corresponding component as I have drawn in schematic? For example, not all 9v to ground caps are the same. I sometimes use the double flag technique to identify the net.
Is there a better way?
Thanks
Not right now but v6 will most likely come with build in support for net ties which will be able to do this.
Thanks, When do the powers to be expect a release date?
About two years from now.
1 Like
KiCad doesn’t really have auto placement. It is up to you to you as the designer to balance the compromise of the physical constraints of the PCB and component shapes with the electrical flow of signals.
That said, what I do is make sure on the schematic I place my bypass caps right next to the power pin of the IC that the cap is for. Too often I’ve seen all the bypass caps of a schematic all paralleled together near the power circuit w/o any indication of where they belong on the board. The way the multi-gate symbols are drawn in the v5 libraries is very similar to how I draw all my ICs, where the power pins are on their own separate gate. This allows me to put the IC power pins next to their bypass caps in a separate are of the schematic so I can focus on just that part when starting my layout and when double checking during my round(s) of final checks. See this simple schematic for what I mean, yeah it is probably overkill for such a simple schematic, but this was a demonstration schematic anyway.
KiCad5TestNew.pdf (17.0 KB)
I’m curious, what do you mean with the “double flag technique”? I don’t think I’m familiar with that. “Enquiring Minds Want to Know.” 
EDIT: corrected spelling of the quote. Sorry, I just realized that was probably a US centric joke. There is a tabloid called The National Enquirer that used that phrase as their advertising tagline back in the '80s. Here is some history: http://www.thisdayinquotes.com/2009/10/enquiring-minds-want-to-know-no-use-in.html
1 Like
I often annotate at least part of schematics by hand.
I draw all the decoupling caps next to each other in the power supply section, but use component numbering with the same numbers as the IC’s the caps should decouple.
So C42 is the decoupling capacitor for IC42.
For IC’s which have acompanying passives I tend to spread the IC numbers in multiples of 10, and then for the passives use the closest numbers.
For example, a dual rail opamp IC 20 will get decoupling caps C20 and C21 and resistors R20 through … R29.
It is a bit of work to do the annotation by hand, and I’m not sure it’s worth it, but it also seems to aid in locating components while designing the PCB, and in the end it may save a little time because of this.
It’s also just a convention, and no hard demand.
It makes sense. I also put my opamp formulas and other wordings on the schematic. Here is an example of double flag. In your schematic, you may have a run from microcontroller to test diode to resistor to ground. I may break the net between microcontroller and tag both ends Diode1. Now when I see the net on the PCB drawing, I know what it feeds. Because, I generally want the schematic readable, I remove just enough net so the two flags are back to back. That way when you look at the schematic, you can still see the flow.
You do not need to break the net to add a label to it.
When reading about EMC I sow information about special PCB techniques - VCC layer very close to GND and with special isolator to make capacitance big and blocking capacitors spreaded on PCB with no regard to other elements.
The use of planes is in part motivated by wanting a good cap between two potentials.
Spreading out caps without taking into account the components on the board sounds more like choosing one end of the expertise/effort vs quality tradoff. (Would be interesting to see your source as it should elaborate what their tradoffs where and how they determined that it is worth doing it that way.)
It might also be that in some usecases this might truly be an option (fast digital cmos circuits might not be one of these as you want the switching currents to stay as local as possible. Hence the use of decoupling caps very near to its supply pins.)
I was looking around in EMC subject in 2002…2006. That I remembered as curiosity with no practical meaning for me as I use 2 layer PCBs. It would took me lot of time to try to find where I sow it and with no gauarantee of success. I remembered that with normal FR4 the capacitance was too small so thay written about some ferrodielectric used in such PCBs. Some time ago I gived here (means at KiCad forum) a set of links to articles about PCB design for EMC. May be it was there. I don’t have that links at hand. I just searched them those time to give them in my post. I don’t know how to find that post - that links are not in my profile TOP LINKS (and under TOP LINKS there is no “More Links” as are under Replies or Topics).
I often put the mouser and/or digikey part numbers and prices on the schematic next to the part. That way, I only need to lookup the info once. Right now, I create the BOM by hand from that text.
I may also put the arcane lore next to the part as needed. For example, if the init needs to be done in a specific sequence, I note it. I will put he directory path to the schematic so I can find it again in the future.
This is what additional and hidden fields are for
2 Likes
As @davidsrsb already notes this is best done using symbol fields. These can already be set in your personal library.
If you use more than one distributor anyways then i would suggest to use a house part number instead and link that to your current favourite distributor in an external database (the same part can be sourced from different distributors depending on current stock, pricing, …)
Prices are something i would never include in a schematic.
Depending on usecase I might not even include it in the database that links my HPN to the real part. Prices change quite regularly and i would argue they are a property of a specific order at that distributor not a static value of a part.
For smaller organizations and hobbyists it might be enough to simply have the latest order price in the global database to have an idea of how much the stock is worth and how much a newly produced system would cost using that stock. After all most of the stock will have been bought at that price as one only orders something new if the stock is already quite depleted.
Note: The database can be something as simple as a spreadsheet file.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.