The 22ga DuPont ribbons are just too flimsy to solder and break, loose strands are hard to get into terminals and still have breakage issues.
I ruined a couple of expansion shields by trying to solder directly to them. The !@#$% wires break off. I was soldering them down and then using a drop of gel superglue but that was ... messy ... and slow.
Putting a DuPont into a terminal looks weird, takes a lot of space and the pins still shear.
The interconnects are the weak point. I can make the software more robust. I can beef up mountings and assemblies, but those silly thin wires are the source of much consternation.
I'm not in favor of the labor content of having to "roll your own" with 20ga (do DuPonts even come in 20 or 18 ga?), but what other solution is there? Anyone? ... Bueller? ... Bueller?
The labor cost of pigtailing EVERY WIRE is insane in a complex project. I can't sell it for a profit splicing without automated equipment. Tinning is bad enough for terminal insertion, splicing and then still having to tin is nightmare.
Extracting the pins from the housing and then crushing is one idea, but again it's not really a clean solve for production. Consistency is the key. The less handwork the better. I want to BUY a solution, not build one! I could pay 10x more for premade 18 ga DuPonts and still come out ahead of having to pigtail by hand.
Terminal strips are expensive when you need 100 stations and get big fast. DIN rails are no better. Mounting and additional costs go up in a hurry. Every $ increase in manufacturing cost translates to $x8 at retail. Labor costs are going up, up, up and if I can buy it instead of building (or even just modifying) it I'm $$$ ahead.
How many complex projects here look like explosions in spaghetti factories? The interconnects are complex, fragile, not color coded or easily labelled and just generally a PITA.
I love the hell out of Arduino EXCEPT the interconnects. I guess it's all a plot to make you design a PCB instead of trying to make a product out of piece parts. (sigh)
I'm a custom product designer. I'm targeting low volume, a few dozen to a few hundred. But even at the bottom end of that volume 150 connection points isn't unreasonable, and they all need to be solid to be sold as a product. A little shaking in transit cannot translate to high DOA calls. It's part of a larger question, not can I design a project, but can I design a RETAILABLE product with embedded Arduino parts at a low cost.
It's hard to set up even a small line if you're not sure the end result is going to be worth the effort.
The parts for a project may be $50 but if the labor is $500 it's not retailable, especially if 1/2 are DOA because of shipping.
dupont cables are fine for lashing up a quick test but otherwise very unreliable
even for small prototypes I use breakout boards with screw terminals , e.g. Arduino_Breakout_1_4
once prototype is proved design and build PCBs
I don’t understand your comments about “DuPont” and “soldering” at the same time. “DuPont wire” usually refers to wires with male and/or female connectors already crimped on.
(Also, few of them are 22g, though they may have ends designed to plug into Porto boards that like 22g)
If you’re looking for wires to solder from board to board, there are other possibilities.
Really if making any quantity make a PCB for a custom circuit or one where you plug the Arduino into.
Then you can have screw terminals on the edge of you board and so on .
Plugging wires into the uno itself is pretty unreliable .
You could use single row header pins , solder wires to them and insulate with heat shrink
Example : Custom “ experimenters” board for a Nano to plug in to ( or plug into a UNO)
I get it, it's all about tradeoffs. The leadtime to market stretches if you have a PCB outsourced, vs just barging ahead. But with PCB costs so trivially low today, it makes little sense to flywire something vs connectorizing and integrating on a PCB. Secondly, you're going to need power distribution and external regulation, so that becomes something easy-to-integrate and connectorize on the PCB at the same time. Fair warning - you'll lose money, and hair, the first time you do this. By about the third or fourth time, you'll wonder why you didn't do it five years ago. You'll develop a stable of subcircuits you'll re-use every time, plus an inventory of standard connectors (e.g. use this, this, and this for the power section, use that and that for the field inputs, use those and those for the outputs, etc. etc.). Finally, you'll learn what board footprint lends itself to your 'design style', and you'll have a pretty darn good idea of what you need to lay out right up front, which saves time and reduces rework. It's actually amazing how much a small board can contribute to speeding you on your way - but you have to jump over that first wall, and it will hurt.
BTDT
C
So I guess you cannot build a viable product without PCB fab. I was hoping that it would be a viable fab technique for just assembly of stock components. Having to design a PCB removes two of the main advantages of Arduino, low cost & modularity. Sometimes you want remote leads and not everything on one PCB.
I guess my feasibility study has to down check Arduino. It's a shame too, as the interconnect issue is the last stumbling block to approving the design. Damn!
Doesn't look like you can put TWO of those into a single terminal. And that only applies to signal leads. It still doesn’t solve multiple, reliable Vcc & Gnd issues without adding terminal blocks and associated hardware. Adds cost & bulk.
I suppose the cost argument depends on whether you create the PCB design yourself or pay someone to do it for you. The PCBs themselves are very cheap say something like $2 for a 100mm*100mm board in small quantities. There are free and good tools to do the design such as Kicad but there is a learning curve. Of course you do lose some flexibility but then you have reliable connections.
It is only really for crude experiments.
