12
\$\begingroup\$

I have a current version of my PCB, and I'll make a new one where I will delete some resistors here and add some more there.

What's the common rule for numbering? Should I re-use the numbers of deleted resistors for the new ones, should I re-number everything to avoid holes in the numbering sequences, or should I leave the holes in the sequence for the deleted resistors and add the new ones at the end of the sequence ?

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Is anybody ever going to service that board? \$\endgroup\$
    – PlasmaHH
    Jun 1, 2016 at 15:13
  • 9
    \$\begingroup\$ This is opinion only : I never re-used deleted component designator numbers. In the event someone didn't look at the board revision level, there could be no mistake about a component designator. \$\endgroup\$
    – Marla
    Jun 1, 2016 at 15:20
  • 2
    \$\begingroup\$ What @Marla said. Few more tips along those lines. (1) Make the designators for the newly added components distinct. For example, your rev 01 has capacitors from C1 to C42. Designate the capacitors added in rev 02 starting from C201. (2) In the BoM cross-out the lines for the removed parts and add a comment along the lines of "obsolete as of rev 02". But don't actually delete them. \$\endgroup\$ Jun 1, 2016 at 15:41
  • 1
    \$\begingroup\$ @NickAlexeev this is a good idea, I guess adding a note to the schema and BoM with the deleted designators could lift any doubt. \$\endgroup\$
    – nraynaud
    Jun 1, 2016 at 21:03

3 Answers 3

17
\$\begingroup\$

Leave the holes, it lets the schematics and the PCB indicate a mismatch if someone has to repair or diagnose such a device one day in the future.

Many consumer circuit producers number their resistors starting at round numbers to indicate the general circuit area involved. So radio parts starting in 100 may be the RF circuit, those in 200 may be the pre-amps, and the ones at 500 may be the audio amp components. Changing a part in one section will not affect the others even if you were to add another resistor it may increment the component number only in the local group and everything else is as before. It lets you reuse circuit sub sections easier. These days renumbering is much easier but some of the old reasons still have utility. Basically premeditated holes in the numbering system.

EDIT:
It also lets your Bill of Materials (BOM) maintain internal consistency over the versions. It will also generate more human readable file comparison diffs if resistor R203 is removed and R209 is added instead of resistor R203 removed and R203 added with a different value or watt rating (or what if the value and power stay the same the BOM diff will show no change after the circuit version was revised). If you have made a change to the value only then the diff will be descriptive too as it will indicate the change and you will know that no parts were removed or replaced, just the value or the particular circuit component R203 was changed in value.

Also if you have to provide replacement components to a service agent in future you will have to have each part number classified by revision as they may be different. If you were to add or remove just one component and renumbered ALL the parts following your admin will become very expensive.

Once the circuit is published (leaves your office/factory) it would be prudent to lock as much of the documentation down from spurious changes. If it is the Mark 2 with vector field nullifiers and no longer the proton precession coils you could have a totally new product and keep the documentation separate.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ Yep. I have also seen schematic styles where components start with 100 are on page 1, 200 on page 2, etc. But as far as holes and new numbers go, I agree with this 100%. \$\endgroup\$
    – user57037
    Jun 1, 2016 at 15:34
5
\$\begingroup\$

Where I work I generally leave the designators the same after a new rev if there is even a slight chance that someone might refer to old documentation while looking at the new board. It keeps continuity between the revs so that a technician can still follow the old schematic when troubleshooting the new board (test points, connector pins, etc will still stay the same) and positional references in documentation ("the pad immediately to the right of D14") will still be accurate. However, if the old documentation has never, and will never, be released, and will never be needed for reference (this is a very rare case) then you can probably re-annotate the entire board. Personally I recommend just leaving the gaps in annotation though.

\$\endgroup\$
2
  • \$\begingroup\$ Yes, my edit (simultaneously to your answer) also makes clear where the problem lies. It is when the documents are circulated wider than the design office, changes to them after that are the danger that needs to be mitigated. \$\endgroup\$
    – KalleMP
    Jun 1, 2016 at 15:39
  • 1
    \$\begingroup\$ @KalleMP Exactly right \$\endgroup\$
    – DerStrom8
    Jun 1, 2016 at 16:06
2
\$\begingroup\$

Generally I prefer to use numbering of components based on the functional block staring with 100, 200, aso. As an example, the power section will have components numbered 1XX (ie R101, C101, U101), the RF will have 3XX (ie R301, L301, U301, Q301), aso. Having it this way also helps me with the BOM and future service/debugging.

I never reuse the component identification (number) in a new version of the schematic. Helps everyone that has to look at the PCB/schematic in the future and keeps documentation consistent.

\$\endgroup\$
1
  • \$\begingroup\$ This is an excellent option for larger boards, especially. \$\endgroup\$
    – DerStrom8
    Jun 2, 2016 at 19:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.