Custom modules : pros and cons

[olkipukki]

I am wondering if somebody makes a custom modules for repeated/similar/modified/custom target designs rather than doing this from a scratch (or copy & paste) for each design.

Are you happy to share some details? Is there any major downside?

IMHO,

pros:
 - save time (target design schematics, routing, testing)
 - save money (modules can be 4,6+ layers while target less; parts and assembly cost)
 
cons:
 - packaging, storage, PnP
 - risky (not in use anymore)

[phil from seattle]

The term module is totally ambiguous in this case.  As a guess, I think you mean small single function boards - something akin to breakout boards. This is fine for some level of prototyping and experimentation. Lots of these are available out there right now. And there have been attempts to build up "libraries" of small boards with an interconnect system. These have been aimed at education and experimentation. Best I can tell they never got that much traction.

For a product (or one off project), it's better to have a single board with all the components on it.  More reliable, smaller, less expensive, better signal integrity, best shape for the application.  I don't think you save that much time in building and testing. Though, if you have to get a PCB made, that takes longer than cobbling something together from breakout boards.

[dmills]

Note that reusable bits of schematic and PCB are very much a thing, I have standard sheets that I just copy the file for used for such things as mic amps, audio IO, 3G SDI TX/RX, GPIO ports, motor drivers, jellybean PWM half and full bridges, half a dozen or so standard switching regulators, Ethernet PHYs, all that kind of thing.

It is a massive time saving to be able to just hook up the harness ports on a sheet and **KNOW** that whatever is on there will do what it says on the tin, it well worth developing these stock schematics.

Regards, Dan.

[olkipukki]

The term module is totally ambiguous in this case.  As a guess, I think you mean small single function boards - something akin to breakout boards.

Sorry for the confusion.
Yes, I mean small single function boards that that can be either presoldered and attached to a main target board, not really breakouts.

For example,

ESP-WROOM-32 or CC3220MODA modules (let's forget about the certification etc.) - would you do routing (very similar) on more than one target design rather than prefabricate & test your own module and use it when required?

FPGA module with hight-speed connectors - again, you can prebuild a generic module and reuse in more than one target...

[Mr. Scram]

The term module is totally ambiguous in this case.  As a guess, I think you mean small single function boards - something akin to breakout boards. This is fine for some level of prototyping and experimentation. Lots of these are available out there right now. And there have been attempts to build up "libraries" of small boards with an interconnect system. These have been aimed at education and experimentation. Best I can tell they never got that much traction.

For a product (or one off project), it's better to have a single board with all the components on it.  More reliable, smaller, less expensive, better signal integrity, best shape for the application.  I don't think you save that much time in building and testing. Though, if you have to get a PCB made, that takes longer than cobbling something together from breakout boards.

It's common for major manufacturers to have certain parts of devices in a board or modular form, so it can be reused in multiple designs. I wouldn't quite qualify those as breakout boards.

[Kean]

The most common reason is for pre-certification of the module, and use in small volume products.  But it also make sense if the module needs higher specification PCB than the rest - more layers, smaller track/gap sizes, smaller drills, blind vias, etc.
This can save a lot if the main PCB is significantly large can be two layer, and also assembled with lower end equipment, though if the module is large or heavy it could pose problems for automated placement.
If you are doing high volume products, then just consider the "module" a drop-in reusable piece of your final PCB designs.  The Gumstix Geppetto service is a perfect example of doing this in a largely automated fashion, but most ECAD tools should make it reasonably easy.

[Edison]

I save my time 
I make something - check if it can be bought - can not buy - I make    ...     ...      ...     ...            

Benefits - I know what I've figured out - I know what I put on parts - I do not know why it burned down   

disadvantages - visit of the shop - list of parts - 10% bought 90% not available - standard CZ