controllers; program in C

Oh, then where do I get these Inexpensive machines? I've looked on eBay, and total wrecks seem to go for $3000 and up.

Well, it would have to be a CNC machine, and since I make CNC systems, I could probably CNC one that wasn't. But, this is a pretty big investment.

Yes, I've looked at these, and I agree, they ought to work better than what I'm doing right now.

Thanks,

Jon

If you glue the connectors down, could you give the board a rap as you take it out of the pot?

Cheers

Phil Hobbs

the ~$10 STM boards include an ST-link, you just have to move a few jumpers if you want to use it with another board

afaics pretty much the only reason to use AVR is that will do 5V and it is in DIP, which doesn't matter if you buy a board and not a chip

-Lasse

I've even tried this without the glue or frame to hold them all in place, it didn't seem to help much. There's 30 connectors and other through-hole parts on this board, so a lot of stuff to mess with gluing.

Jon

Just curious, why do your own soldering? Assembly houses do this stuff pretty inexpensively if you have even a modest volume.

Jon - that does sound familiar. In the circumstance you are working in, getting another solder pot with tin/lead would be a reasonable step to see if your solder is causing trouble. A Walmart aluminum self heating skillet with the temp control jimied up 100 degf or so makes a good pot. A plain skillet with an eye on the temp works well, but it's not "good". With the solder you're using, and with it hanging to the board, try higher temperature. And, once again, be sure the flux is dry. If you have some good high quality temperature monitoring gear, put it in a drawer. Take some iron and constan wires insulated in teflon and wrap the bare ends together and stick the other end into a dvm at 200 mv. Temperature difference is 50 micro volts per degc. This plus ambient indicates temp at the other end. This provides max simplicity to verify other equipment.

Hul

J> > Jon - what are the make & model of the connectors causing trouble?

First, I guess I'm a control freak.

Second, I do not have what most would call "modest volume". I do

In these volumes, it seems most outfits want at least $20 per board, and this particular one has a LOT of connectors, so probably a lot more than that. $20 is pretty close to the profit I make on the board.

Jon

Almost certainly overkill for this particular project, but I just wanted to mention the Beagle Bone. I've used it on a couple projects, and it is just FANTASTIC. A complete Linux OS and development environment all on a board barely bigger than a credit card. If you need net connectivity, you can hardly beat it. Also has LOTS of I/O, plus the PRU (twin 200 MIPS microcontrollers that can do various bit-banging and other time-critical I/O tasks).

Well, you can tell I'm a convert!

Jon

I know about it because I almost designed it into a circuit for which the customer wanted all through-hole parts. Since it was a charity customer, and it became apparent that the guy he was going to arm-twist into manufacturing boards wasn't, I decided to use the 20-pin surface-mount counterpart (less $$, IIRC), and build the damned boards myself.

But I'm most certainly fully geared up to use the Cortex M0.

It does seem odd to be using 32-bit processors for a cruddy toy motor timer circuit (which is what this is). For the quantities that we're contemplating, it's going to cost me a lot less to use the slightly more per part Cortex M0 and not have to buy a $300 programmer that can do a "factory quality" job on a PIC, and I get to use the tool chain that I'm used to.

Well, yes, but using an 8-bit micro with thousands of gates to do something that could be done with a 7474, a few NANDs and a 555 seems wasteful, too -- until you realize that you're not talking about gates vs. gates, you're talking about pins vs. pins and square mm vs. square mm. Then it all starts to make a twisted sort of sense.

I think you just said you are making $400 - $600 per batch. Yes, an assembly house won't want to turn on the machine for much less than that. I also wouldn't write up a quote for that small a profit.

Too bad there aren't assembly houses that will work with the batchpcb model to assemble panels before being separated. I guess that would be a whole 'nother level of difficult coordination. With PCBs you just tell them what they get and everyone gets the same thing. With assembly they all will be using rather different parts with different packages and good luck on coming up with an XYRS file.

I regularly do runs of 25ish boards but they're generally pretty complicated. We have a couple of assembly houses that build for us in very small volumes. Don't know if they'd be happy doing simpler boards or be anywhere near affordable enough for you, though.

If you need RoHS,then drop suckie SAC and use the lower MP Sn96.3 Ag3.7 solder; it actually works easier than tin-lead on the bench.

OK, that's interesting. But, I guess with increased silver content, it is even more expensive than SAC305. I get NOS bar solder on eBay for about $20-25 a bar, which is a LOT less than list price. I haven't seen any of this (what do they call it, SA37?) on eBay.

This is for dip soldering in a big 5 x 8" solder pot.

Jon

Ah, indeed, going back up in the thread, it was! So, thanks, Sphero, then, for the preheat idea. I will have to try it on the next batch, probably will be a couple weeks before I need to do more.

Thanks all for your comments,

Jon

Good luck with it!

I recall seeing a process where the board was rolled into the bath and then rolled out rather than being lowered directly, but modern dip soldering systems seem to directly lower the PCB.

I have a small wave or fountain kicking around somewhere (maybe 100lbs when full of lead solder) - I'll have to fire it up some time.

What about wave soldering?

Yeah, that's the kicker. I got my self involved in an open source hardware project, and this board sells for $80 each. it has about $30 in parts on it, including the PC board. So, it has to be done cheap to make any money at all on them. Probably a big mistake to get involved in this at such low cost.

Jon

After doing some reading, it seems that preheat plus running the bath much cooler is the trick! So, I definitely am going to try this on the next batch. Possibly, also, a flux more optimized for dip/wave would also work.

Thanks,

Jon