exxos blog - random goings on

[exxos]

Have been looking more into soldering machines.. The one I was looking at in particular, and indeed the majority of others, are not designed to be run from a PC.. they all have their own programmer unit which would make sense as the machines can be programmed individually on site. But for my use, programming in 4 pages on a individual programmer to do every single solder joint would be a total nightmare If I was doing thousands of something and only one product, it would not be much of a issue, but because my items are very low quantity and many of different items, I think it would actually take longer to set up the machine than actually soldering the boards by hand in the first place

The only one I have found currently is a Windows-based one made in USA. I have emailed for a price but as my girlfriend pointed out last night, I did not take into account import costs as well which would cost $1,000+

The Windows-based one seem to operate in the same way and understandably, you have to set up each solder joint individually. Currently I have been unable to find anything which could simply take the Gerber files and at least use that for the hole locations. If the software used a text file I could probably write a script to convert the PCB hole locations to the solder program file to speed things up. I would still have to edit every single line to set up the soldering angle etc but it would be a order of magnitude faster than manually stepping the solder head to every single point one by one.

There are some other cheaper machines out there, but currently there is no English software for them. Overall it is looking like this might not happen, in part due to the huge cost it all, but also I do not want to create myself more work in having to set up the machine every time just to do a few different boards.

[JezC]

Our production line uses lots of auto-placement tools with IR ovens & selective solder machines (not to mention automated visual inspection etc.) & from what I was told they have the various programs for each of our many dozens (if not hundreds) of products which can be automatically loaded in & run (though human operators have to fit the correct feeders to the placement machines at present)...which sounds a lot like what you want.

The downsides? Well, apart from needing a massive factory to fit the 4 (or is it 5) surface mount assembly lines into there's also the cost of each machine. When we had the first 4-head auto-placement machine installed (many years ago) it was apparently the first in Europe & cost many tens if not hundreds of thousands of pounds...

Sounds mad, but it more than halved the time spent on that part of the production line & so almost doubled our throughput on that line.

Barring a big win on the lottery I suspect it won't be the approach that @exxos can follow though...

Sorry exxos

[exxos]

@JezC While cost is a factor, it is the cost of the machine versus how long it will take to get my money back on it which concerns me... Of course really I would need a solder machine, pasting machine, pick and place machine, larger reflow oven is... It will be cheaper to go out of business than buy all that stuff

What I was looking at is to increase production by just using one particular machine to do a series of jobs..Hence a toss up between soldering through-hole and SMT stuff.

For my current concern overall, is the amount of time it would take to program that machine in the first place.. If you took the remake motherboard for example, how long would it take to actually program in every single solder hole location individually... If I was selling a few hundred boards, it would be worth it...But for like 10 boards it would actually be creating more work.

It is like, getting all the components together, bagging and labelling them, boxing stuff up with instructions and posting it off... Waiting weeks or months for the job to be done etc versus not doing any of that and spending the time setting up a soldering machine myself instead...

Of course it is a bit more complicated than that but that is basically what things boil down to... Aside from what it costs in time and money in sending and getting things assembled third party, versus the time and the cost of assembling it myself... Which then brings us back to how many boards I could assemble on that machine, versus how long it would take to get my money back on the machine versus paying a third party company to assemble stuff...

I think I spent something like £2,000 - £3,000 easily over the past two years in getting stuff assembled by third party. I get really good prices so I cannot complain there. A lot of companies charge easily 10x that amount. Of course I cannot do entirely without the third party. I could potentially save I suppose £500 a year in assembly costs... but if a machine costs me £5,000, it will mean 10 years before I get my money back... By which time I would have to operate for another 10 years to actually turn a profit on it all... And I am really looking towards winding down in 10 years time anyway.

[JezC]

@exxos Yes, sounds like a very familiar story.

I keep having the same discussion with the management at work about adding in more automated /regression testing for our software. They all see the logic in doing it (and how expensive in time and development delays it is to have the s/w engineers doing the testing themselves) but always put doing another development for a customer ahead of making any progress with the automated testing.

So, I did a quick calculation that in the 16+ years I've been working here I've spent the equivalent of at least 18 months flipping a switch, watching an LED light up (or not) & checking that the reported status of the LED does change (or stay the same) in our PC s/w.

Of course, that sounds daft when put like that (and represents between 3 and 6 additional developments we haven't managed to do)...but then you think that we have in excess of a dozen embedded s/w engineers all spending time doing the equivalent manual testing for every development.

[exxos]

Rapidly losing the will to live and this is killing my fingers as well

IMG_5186.JPG (242.73 KiB) Viewed 3522 times

[exxos]

More progress.... Problem being I'm struggling for GAL's now shall be able to get five or six kits in my store soon, which will probably be enough for the time being anyway...

IMG_5187.JPG (127.89 KiB) Viewed 3494 times

[exxos]

FFS! Every bloody time I go on farnell its down

666.jpg (57.43 KiB) Viewed 3484 times

[stephen_usher]

They know when you're coming, obviously.

[exxos]

Yep! A few weeks ago farnell rang me up, think they noticed I don't order much from them anymore. The poor guy had half hour exxos rant on why I hate farnell. My mother was in the same room at the time, and when I came off phone she said, I bet he wishes he hadnt rang up now

[ube]

On it. (More or less). Could you add a 5V pin as well? Make it a four pin header with one blank (5-blank-gnd-sel) so people wont complain about soldering a physical switch the wrong way.

Lost me.. .Are you talking about the widget as was done on the earlier ROM boards ? I've moved the UBE ROM chip to the 555 reset circuit on the STE.. long thread about it all above somewhere...

Yeah, I read it all later. Hence the edit.