But I also mentioned that hand soldering/reflowing the affecting pin was easy and that means the surfaces are probably OK. I suspect how I applied the flux was not effective. I'll try dipping and also vertical spraying only. I've been spraying at an angle lately, cross patterns but not vertically up.
Here is a question I've been thinking: I need a pre-heater. Not sure to what temperature but at most 170DegC as most SMD reflow requires that initial slow heating to 170DegC. But at that temperature if I spray flux on the board, then the alcohol solvent will immediately evaporate in a sizzling sound (we're getting ready for outdoor grill season here). Probably a boiling temperature around 70degC for these solvents. So is that actually OK, since the flux may not evaporate or is that bad? I can also spray the board before the preheating but again alcohol will evaporate long before preheating is done. Just beginner question
I do get drops of boiling alcohol solvent rolling around on top of the solder bath at times when I sprayed too heavily. Such boiling action causes the solvent to sip through vias to the top of the board and makes it dirty looking. I've been ordering "tented" vias but I've not seen the fab actually paint over the vias at all
Sooner or later you will have to change the way you are manufacturing and go with organic acid fluxes and wash the boards after soldering. Or change the relay supplier and manufacturer. Concentrate on the relays, they seem to be the weak link in your process.
Is your quantity so large that you can't just hand solder the relays and be done with it? I notice frequently products in the marketplace with through hole components such as relays that are hand soldered.
I was wondering the same and am always curious to see the detail of projects on the fringes of the hobby/commercial market place like "what is it" (supported by schematic, board design and code) ? and who are the target consumers ? Of course, I also understand that the commercial aspects of the project may limit the scope for the disclosure of any such information.
Many moons ago I had a gap year working in an electronics company. They had a wave soldering machine and had the same issue with relays sticking. The quick fix back then seemed to be a sharp tap on the relay housing with the handle of a screwdriver!
I've been hand soldering and decided that it's time to upgrade to a faster method. I can solder pretty well and pretty fast but after doing thousands of junctions in one sitting, my hand and arm are stiff, not to mention my neck. It's very tiring to do thousands of junctions. My impending project has about 200 boards with 8 relays per board, 4 pins per relay, making almost 6,000 relay junctions and 4,000 terminal block junctions. It's a contractual project so I can't share designs etc. But I can share what I design and sell.
It's a 2" by 2" (5cm by 5cm) board on a 3-col*2row panel, with 29 thru hole pins per board. So the whole panel has 174 junctions. I hand solder fast, almost 13 junctions per minute. If I have 12 panels, that's 2,088 junctions or 164 minutes of hand soldering. If I dip solder, it takes about 15 minutes of spraying, dipping, and cooling. BUT, there is time for solder blobs and Kapton tape and one part I must hand solder. So altogether, I am twice as fast with dip solder, only if I manage to exceed 174 junctions per dipping, with at least a dozen or more panels, and only if I don't count parts prep time, just the non-part-prepping time. Parts prep time such as time to insert parts, use putty to hold some parts etc. is similar between dip soldering and hand soldering. Under these assumptions, I saved about an hour
The contract board has one board per panel (two actually but I must break them apart to fit my solder bath). This board has 57 junctions. So doing 57 junctions per dipping is probably not as advantageous as my own board with 174 junctions per dipping even if I need 200 boards vs my own 12+ panels, especially considering the possibility of failures.
I sympathize with your wanting to avoid the drudgery of hand soldering a bunch of boards. Since you are having random failures, I was just wondering if it might not be worth the extra care to insure success of your contract boards. Maybe you can talk to the relay manufacturer and get some pointers on preventing the problem should you need to manufacture more.
I believe that flux was activating and all soldered pins look good. The primary issue is stress on the relay pins tend to make a few of them not actuate reliably out of a thousand.
Thank you! I should read through the guide to see if I missed anything. I started with hand placing a single board at a time with a floppy Kapton stencil from a US stencil source and gradually upgraded to stainless steel stencils from the same source. Unfortunately they have rolled-up stainless steel sheets thus their stencils aren't very flat, essentially arching the wrong way (center propped up).
Hand placing and stenciling was limiting my throughput. I remember about 32 devices was my limit before I would get tired. So I eventually made the big leap to acquire a pick-n-place machine, a stencil printer to use framed stencils, and a reflow oven (was using toaster ovens and hand cranking a Variac). Later I even got a solder paste mixer. You pop a 500g solder paste can into the machine. It uses centrifugal force to spin the can in a way to make the paste creamy and easy to print.
Now I order panelized boards that fit on my pnp and not too big for my stencil printer, and fit inside my reflow oven without wasting space. This panel is the 3rd generation since I have started with this setup. The solder bath size becomes a 4th limiting factor of the panel size. I can make a couple hundred devices in one go now, still tiring but about 8-16 fold improvement in SMD parts throughput. Dip soldering helps this project because it has lots of pins. I just need to perfect it so I don't have to waste time removing solder blobs. Dipping the panel in flux will be what I have in mind next time vs vertically spraying upwards against the underside of the board.
I do SMD components first, flash firmware and test, hopefully in the summer when my garage is warm/hot instead of the freezing winter. After I have a stock of these panels, I can take time later in the season to do the thru hole parts in small batches when I receive orders.
I'm trying to be a one-stop shop for circuit design, firmware development, prototyping and small batch production. So far dipping my foot in production has been hard because I've got zero experience working in this field. Everything is learned from online and from mistakes.
Thanks for the comprehensive description. As I said it was curiosity and I now have a picture of what you are doing and it's scale. Very interesting to see what you are getting involved in and what is required to put a business like that together.
Was looking for some way to share my experience anyway
For me, going from stenciling one board at a time and hand placing components to pick-n-place plus stencil printing with framed stencils was definitely a nerve-wracking leap of faith. I wasn't comfortable with this move but everything pointed to the need to make the leap. Instead of going with an inexpensive community-supported type of pnp that need lots of assembling and tweaking, I went with an "expensive" commercial desktop machine that was ready to go out of the box. There's pros and cons. I can't recall a single day when I didn't have to fight the machine to get my work done but most of the time I ended up learning something out of my frustration or least didn't get shut down by the issues which I always eventually analyzed and found some solutions for. I give credit of my "winning against the machine" to my training as an experimental physicist. I can take the machine apart and analyze what it does and how it does it and put it back together, more or less (only did it once, partially).
Here is my machine. I'm not associated with the business or get any commission. I can tell you all of the frustrating things using this machine but in reality when it works well it purrs. I've made enough money to buy it several times over. Will I buy it again? NO! I might want to get a desktop machine that takes pneumatic feeders instead of this, rollers with motor etc. and I want more nozzles!
You need a lot of climate-controlled permanent space to set up a shop. The size of a one-bedroom apartment would be what you wish to have. I have about 400sqft basement space and could use more. I can't use my garage because it's freezing cold more than 6 months of a year.
I am wondering if several boards with flux could be pre-heated at the same time to takes out one at a time? It would effectively divide the time to preheat.
Using 3D print XYZ positioning, could the touch-reflow soldering be roboticized?
BTW; I have major respect for your work since 2011!
You may not like this, but....
ask yourself what your core competencies are and thus what constitutes your source of value towards your customer. Is this soldering boards, or translating functional customer requirements into a dependable electrical solution/product? In case of the latter, consider outsourcing the manufacturing work to a party that specializes in it.
I don't know.
If you see what has happened to Green technologies such as solar panels and electric vehicles or aluminium/steel etc. etc. with regard to tariffs to protect domestic industries, it may only be a matter of time before cheap PCBs., assembly services, components direct from China etc. are history. Then this guy will be well positioned. I'm certainly maintaining large stocks of TTH components and prototype boards for my hobby activities because I somehow don't believe this will all last.
There's a major difference between global trends & industrial evolution, and an individual decision in the short run at the level of one entrepreneur.
For one, outsourcing the work for one particular project doesn't mean his competence is lost.
Secondly, outsourcing is not the same as offshoring. I did not suggest to divest the work to East Asia; a suitable partner physically around the corner would be more sensible.
Thirdly, the situation he describes strongly suggest that his strengths and added value really are not necessarily in the domain of PCB manufacturing, but more in the engineering domain - so how 'well positioned' he really is, remains to be seen.
I could go on, but won't, because I don't feel it's all that relevant in the light of the practical question put on the table. Having someone else do the manufacturing job is a legitimate option that should be considered, especially if you run into (in all honesty) very mundane problems that are the bread & butter for any manufacturer. Sometimes there is good reason to re-invent the wheel - just not always.
Sorry about the long delay. Thanks for your honest opinion. I've decided to hand solder the batch of mechanical relay boards. It took me a lot of time and effort against an aggressive deadline but I managed to finish. I have a few reasons for making them myself:
The time line was extremely short. There's no way to get contract manufacturing done
I have to flash firmware, run function checks, fix boards anyway
I was able to find how how many boards I can build within what time frame. I think it's an important part of my business because a client either wants me to make boards or wants me to contract the boards to be made by others. I need to know which option has what kind of time line.
With future batches of comparable or larger quantities, I'm going for a regional contract manufacturer for quotes so I have some ideas of cost and time line.
It might help to maintain a bit of a relationship with one of these parties so that lead times do indeed work out the way you expect them when push comes to shove. Have them do work for you from time to time to keep the communication channels open. Just a thought.
