What temp do you set your solder pot for with 63/37 leaded solder? I just started using one and I don't know what the ideal temp is yet. Brian Kraut Engineering Alternatives, Inc. www.engalt.com -----Original Message----- From: piclist-bounces@mit.edu [mailto:piclist-bounces@mit.edu]On Behalf Of Forrest Christian Sent: Wednesday, March 14, 2007 5:45 PM To: Microcontroller discussion list - Public. Subject: Re: [EE] Small-Production PCB Manufacturer Experiences. Vitaliy wrote: > Yeah, these are relatively simple boards. Do you hand-solder them in-house? > Are they PTH? > We've been doing PTH in the past. We're switching to SMD. For PTH we've been doing solder-pot soldering, where you insert all of the components through the board into anti-static foam. Then you flux the entire board, preheat/dry it a bit and then dip it into a solder pot for 3 seconds. Fast, quick, efficient. Generally a board takes me no more than 30 seconds to solder after everything is in place. With the SMD, we're applying RoHS compliant solder flux using a manual stencil process. Small boards we're v-scoring so we're able to stencil a few at a time. After the stenciling, we manually place the components (generally 1206 or larger so it's easy for a human to place), and then bake. The switch to SMD is because it's actually faster to place SMD than PTH if you're good with the tweezers, and it's easier to deal with RoHS that way. PTH components (which we keep to a very low minimum) get soldered by hand. We generally try to keep the PTH pins under 10 or so per board. If it's more, we're leaving it in the Leaded environment and the entire board as PTH. Most of these devices are RoHS exempt anyways. >> Labor cost is roughly 0.003/second. ($10/hr, 16.666.../minute, >> 0.003/second). Even if you figure 5 seconds a placement, the cost is >> only about 1.5cents/placement. Heck, even if we could only place one >> every 30 seconds (even I can do this well), we'd still be in the same >> ballpark. >> > > Mmm... that doesn't really make sense to me. The relationship is linear, > unless you have a huge fixed overhead. > The cost to manually place a component is mainly time. Flux is dirt cheap per board, and our volumes are small enough that we don't really have to hire someone to do it. However, we do have to consider at what point is it cheaper to outsource. What I'm saying is that at the going rate of $0.10 per component at my quantities, I can hire lots of people to do the placement for me. If I could get assembly at $0.01/placement I'd outsource in a minute. > A lot depends on your market. If the potential market is big, and the demand > is elastic, it makes sense to produce larger quantities, but charge less per > unit so you can sell more. Keep in mind that your goal should be to maximize > your _total_ profit, so it doesn't matter how much profit you make on each > unit (many people forget this). > My market is relatively cost insensitive. My products are significantly lower priced than the competition (like 1/10th of the competition), but the market is extremely small. 500/year per product is my expectation. I also rev the products fairly quickly to respond to customer needs so I generally will buy 1-200 boards, go through those, make any design changes, and then proceed. My niche is specifically figuring out how to do those very low volume products which are simple to build (throw 40 components on the board, and 8 hours of programming), and that noone wants to build because they are interested in either selling 100 units at $5000 each or millions of units at $19.95 each. I'd rather sell 500 each of 20 different products at $150 each. I expect to hit a nerve at some point on some product and find out that 500/year is a decade or two too low of a quantity for that product. At which point I'll pick up the SMD design -forrest -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist