| General > General Technical Chat |
| Is partial SMT board assembly a practical option? |
| << < (4/4) |
| KL27x:
--- Quote ---One day per 3 months of an engineer's time isn't too bad. --- End quote --- Realistically, I don't do real engineering 29 days a month. The EE stuff is the part of what I do, but the PCB sourcing/assembly is probably more of my time. I wouldn't enjoy doing hardcore EE, day in and out. I solve problems. When that requires coding or PCB design, I might spend several weeks doing only that. I also solve physical/mechanical/production problems. I also solder. The DIL chips, as I wrote, are not difficult to put on there, if you learn how. I did 300 in maybe 3 hours, earlier this week. That's the easiest 3 hours of physical labor I can imagine. Play some music. Or YouTube talk show, more likely, these days. Get in the zone. Next thing you know, the hour hand has moved and the stack of boards are done. No problems. No frustration. Clock in, clock out. C'ya. This is a nice change of pace from coding, where basically I lose all track of time for potentially days or weeks and sometimes work for 14 hours a day, cuz that's what works for me. Yeah pogo pin testers. I've made and used them (thousands of times). To me, this is a special kind of hell. Sometimes the board has to be wiggled to get a connection. Sometimes the connection starts out good but gets lost partway through. Sometimes there are certain errors that have to be categorized because of the cost of the board and the fact that some errors can be fixed. Boards of various sizes and shapes have to be set in bins, moved to other bins, and there is just a sea of bins requiring massive bench space. And labels. Fing labels. Clear the deck to do this crap. Doing this, the clock hand doesn't move. It's stuck in time. And here you are stuck in the hell that doesn't end. Having experienced this, I spare no expense to make the best jigs and testing protocols that I know how for my clients. But other than the design and testing of the test, I do not ever want to see or use any of these testing jigs ever again. I choose to not do that part (but emergencies do happen). I feel a real and deep bond and sympathy for anyone who has to do this kind of testing. Production soldering is like being a trained monkey. BUT a monkey that dictates his own pace and who has control over everything in this domain. And who can set this up so that in these hours his mind is free to his own thoughts. Board flashing/testing is like being a trained pigeon enslaved by and hopping to the beat of machines. In some cases, the flash time is not significant. But when the flash time is 20 seconds, and the test is 2 seconds, then you're extending this hell. To someone. Yes, you can have more testing stations. This is more work. More organization. More clutter to setup and to clean up everytime this production phase has to be done. And more time, even when things are running right, because the longer the chain, the more kinks you have to sort out. And ideally, this will require a short enough time and few enough testing setups that one can do this from a chair rather than having to walk around, constantly. The soldering is extra work. But the other benefits of partial assembly are the logistics and commitment. When the board is partially assembled, you aren't committing as much in the basket in case of a major unforeseen problem or obscolescence. You can still change the firmware or whatnot without adding or redoing work (for me, anyways, with my equipment and in my current work). It's like production on demand, rather than storing up lots of inventory at once. This is how I see it, and I could be missing some major things. But this is the little world I know. In China, it is probably a little different. In America, you can't find someone happy to be a good robot pidgeon for more than a short while, unless they are personally invested in profit and loss. And unless you have your own, you can't use children. When you do find people who are halfway productive at this, they are so for exactly 3 months. Then they qualify for unemployment. I see company owners and managers who bitch about this; they expect human beings to do this day in/out. If someone is good, you better find them other positions after they put some good time in at this job.*** Good employees are hard to find, and this will burn out any American. Americans have alternatives other than jumping out of a window. This is why we outsource this work to your country... when we're ok with the massive long-term commitment, upfront cost of paying in advance, long lead times, the paperwork, and the extra 21 emails setup and customs forms, and the whole trusting a Chinese manufacturer with the gerbers and the firmware. Of course it makes sense to you. You can find someone to flash/test boards cheaper than an IC vendor charges to load a machine and press a button, per unit cost. ***I work with a company who has had their own lessons to this problem. Imagine you have a worker doing assembly line, repetitive work to a device. They make, I dunno, sub $20 an hour. They do a procedure on a $100 part, once every 2 minutes or so. This worker willfully destroyed over a hundred of these parts by the time it was noticed. Been there a year, and has done this properly thousands of times, and all of a sudden he decides he would rather be on unemployment, and this is how he achieved his goal. Same guy, they talked highly of and were hoping to make him a manager, in the first several months of his employment. Repetitive work has a way of crushing the soul. |
| Navigation |
| Message Index |
| Previous page |