The boards which I have designed have all been developmental/prototype or for my own personal use. As such I have hand assembled the vast majority.
Even where a prototype board has been machine assembled, there is the matter of tweaking prototype designs on the bench. Specifically related to compensating power converter feedback loops, this is a matter of adjusting a few resistor or capacitor values. While there are analysis methods which can calculate this for MANY designs, calculation and simulation do not always eliminate tweaking on the bench. And in many cases, I have seen that it is just a lot quicker to do that than it is to model and apply a trustworthy analysis.
With respect to some two-output power converters, I have proven to colleagues the advantage of taking feedback from both outputs. Analysis purists would argue not to do that, because (SFAIK) the available analysis techniques cannot handle a feedback loop which does not have a single point of loop closure. But closing the loop around both outputs eliminates a severe oscillation which can happen when only the secondary output is loaded. So the choice is whether to make the design one which can be analyzed on paper or to whether to make it tolerate all operating conditions. I would argue for the latter and this seems to imply tweaking the design on the bench. This requires some manual rework.
Beyond all of this, I have my own stash of components and I want to be able to use them. This does not reliably eliminate the need to buy some components for new designs, but I think of my own assembly as being an important part of what I do and I do not want to lose that capability.
These days, I model everything and try to build once. If there is a question leftover, I might use handsolder footprint. But since modeling, it is pretty rare.
For in house SMT assembly, and provided that parts that parts and pitch are not too small, I use a small reflow oven. It was about 250 on ebay. Fantastic. BH
This means that there is no such technology) just as there is no technology for tightening bolts by hand with a certain force without a special tool. Everything that I can do with my hands is suitable only for me but is completely unsuitable for production. I donât know how else to say it more simply) There is SMT technology and there is a SMD case type and there is no soldering iron there. If this technology is not followed during design, then your device will be impossible to produce and everything will have to be redone, this is usually how it happens. An electronics developer must take into account both the element base and these rules. A portable type is needed not only for a stand but also for practicing the technology, here a technologist is already connected and we all go home together with a soldering iron)
I think that is the key assumption where we differ. If I can build a board by hand and it works, then the method is valid and there is no problem. I have done this many times. KiCad includes âhandsolderâ footprints in the footprint library; this is a strong implication that some other people agree with me. My building a few boards by hand is not a statement that mass production is best done by hand.
If I were designing a board for mass production, I would either consult with a manufacturing engineer or I would hand off the design to another pcb designer. But my focus is power circuit design, and not pcb layout for production. So the main thing I want to do is demonstrate the functionality of my circuit design.
One other thingâŚIt is widely accepted that for power conversion circuits, the design engineer needs to either do or closely review the pcb layout. A pcb layout designer may design a board for mass production. But if it is not done with close supervision from the design engineer it is likely to not work well or not at all. The latter has happened to me onceâŚThe board was completely non-functional.
TLDNR
Every designer needs to confirm each footprint with every mfg data sheet. Donât ask me how I know.
There are millions of parts, new packages, variations, and you cannot trust any âstandardâ foot print. You have been warned.
Modify footprints as needed and save in your personal library.
Stock symbols are often too big, I change those too.
Standards are specifically to alleviate the inefficiencies and burden of ad-hoc. A few exceptions does not take away from the basic functionality of having a standard. When a manufacturer writes SOT-23-5, without caveat or qualification, we should reasonably expect to be able to drop in a sot-23-5 and have it be correct.
Unfortunately, everyone has a different concept of correct. If we look from the manufacturerâs point of view, then there are size recommendations for SMT technology. If we look as an amateur, then the size can be any without reference to the manufacturerâs recommendations. The manufacturer for SMD assumes that there are two technologies for applying solder paste to the board: a stencil or a dispenser. Based on this, we get a size that allows you to avoid defects during installation and make the device operational. This does not give a 100 guarantee since there are many other factors such as the quality of materials, equipment, design errors, etc. There are software systems for checking that compare landing sites with a library of standards, but thatâs another story and it is needed by those who need it)
Not really, for simple old SOT packages, usually no problem, but with high density pins, BGA, and RF packages you cannot take chances.
Then for some 0402 and 0201 packages you may want to have reduced pad sizes for impedance matching to traces, and push the production envelope for risk of rework, applied selectively.
Then every stock footprint use sharp angles on pads, no advantage for anything, including stencil cleaning, paste release or even high voltage.
Sharp corners are not really sharp) On a small step you may not see it, but there is always a rounding both on the pads and in the aperture. The angle of rounding can be different, but technically it is never 90 degrees, this is due to the technology of manufacturing printed circuit boards and stencils.Another point: rounding affects the area of ââthe contact pad and, accordingly, the amount of solder paste through the stencil.
The default âroundingâ is small enough to be considered non existent, it is not defined by a specific number.
Adding rounding to pad and stencil corners has insignificant reduction in solder volume, as sharp corners make paste release less effective there.
Do a microscopic paste inspection with and without rounded corners, and see for yourself.
Sharp corners offer exactly what advantage?
Minor rounding does not lead, but significant rounding does. In this case, the area is important for contact, and the shape affects the wave resistance in addition to mechanical properties. Rounding everywhere to an incomprehensible angle with an incomprehensible ratio is a thankless task.
I donât know if the above text is a Google Translate job from Russian. It is not understandable.
Some English words have like 20 synonyms, using the wrong one for the context and the meaning may get lost.
But thanks