Creating my first footprint: questions as a beginner

[X4V1]

Hello,

I’m trying to do my first electronic project and I’m coming with not a lot of knowledge, so first of all, excuse me if I asking questions that could seems stupid to you.

My plan is to delegate the print and the assembly of the PCB to JLCPCB (seems to be low cost). I’m delegating the assembly to them because I don’t have the tools and experience to solder small components.

I would like to have a PCB on which I can plug some component. I’m not sure what are the best practice for that:

• Is it correct to create one footprint per sensor even if it needs 2 connector to be plugged in ?
• Is it ok to draw the outline of the sensor on the fabrication layer ? Or should I draw the outline of each connector separately ?
• If I can create one footprint for the whole sensor (containing the 2 female headers), won’t it be a problem to describe what must be assembled by the manufacturer ?
• I read that the anchor of the footprint must at the center, but some others says that the anchor must be at the pin 1 to make it clear for the manufacturer, what do you think about that ? In the case of a footprint that will need 2 connectors, what would you recommand ?
• I read that the fabrication layer is used to explain the manufacturer what is going where, but If I draw let’s say a big rectangle where the sensor will be, won’t it be an issue for them to assemble the 2 connectors ? (how will they know that I need 2 connectors for the same footprint).

The manufacturer is apparently needing a BOM and a CPL, I’m not sure how to make things clear for them.

Again, sorry if my questions seems obvious, I’m trying to learn but It’s not easy by doing that alone.

Thank you in advance

[PapaJ]

@X4V1

• Is it correct to create one footprint per sensor even if it needs 2 connector to be plugged in ?

There is no right way but if you deviate from the norm it probably will add cost, complexity and confusion (you can always make instructions), but typically there is one footprint per insertion. There are IPC* rules for what these should look like but I can’t say I’m an expert on IPC. *IPC is a standard for assembling PCB’s.

• Is it ok to draw the outline of the sensor on the fabrication layer ? Or should I draw the outline of each connector separately ?

See previous answer.

• If I can create one footprint for the whole sensor (containing the 2 female headers), won’t it be a problem to describe what must be assembled by the manufacturer ?

See previous answer.

• I read that the anchor of the footprint must at the center, but some others says that the anchor must be at the pin 1 to make it clear for the manufacturer, what do you think about that ? In the case of a footprint that will need 2 connectors, what would you recommand ?

KiCad will generate a pick location based on the 0,0 of the part. The CM can offset from this if it is not a good place to pick the part. They will be a lot happier with you if you make their life easy (put the 0,0 where the part will be picked from).

• I read that the fabrication layer is used to explain the manufacturer what is going where, but If I draw let’s say a big rectangle where the sensor will be, won’t it be an issue for them to assemble the 2 connectors ? (how will they know that I need 2 connectors for the same footprint).

See answer 1.

The manufacturer is apparently needing a BOM and a CPL, I’m not sure how to make things clear for them.

See answer 1.

Jacob

[straubm]

@X4V1

The manufacturer is apparently needing a BOM and a CPL

JLCPCB has sample files for both on their page IIRC and there is a KiCad plugin to simplify things further.

[retiredfeline]

@X4V1

The manufacturer is apparently needing a BOM and a CPL, I’m not sure how to make things clear for them.

The BOM tells the manufacturer what part they should provide for you (at your cost) or buy for you. The CPL tells they where and in what orientation it should be placed. There shoud be no need to put instructions in the Fab layer which can be misunderstood. I’ve never needed to submit Fab layers.

[straubm]

@retiredfeline

what part they should provide for you (at your cost)

They have an affiliate, LCSC. It is best to stick to their inventory. Components that are not offered by LCSC can be highly expensive and AFAIK it is not possible to send JLCPCB user supplied parts. Given the shipping cost this would not make any sense anyways for a hobbyist/maker.

[retiredfeline]

@straubm

They have an affiliate, LCSC. It is best to stick to their inventory.

Specific to JLCPCB, but you will learn about basic parts and extended parts. Basic parts are already in the factory, extended parts have to be ordered from LCSC and will cost more and have minimum order amounts. It’s all explained at their site so I won’t elaborate.

For THT parts like connectors consider doing them yourself. Othwerwise they will have to be done in a separate pass and not by machine. Anyway their quote will show you what you have to pay.

[straubm]

@retiredfeline

Othwerwise they will have to be done in a separate pass and not by machine

…and that will cost.

[Piotr]

@X4V1

I’m delegating the assembly to them because I don’t have the tools and experience to solder small components.

If in your plan this PCB is your first and last than it has a deep sense. But if you plan to play with electronic little more than in my opinion the better decision would be to accumulate tools needed and learn assembling PCBs yourself even on cost that this first one will be not success.
I have never ordered anything from JLCPCB but I suppose that if you order 10 your PCBs it will cost less than one with assembly.
There is no problem with assembling manually all THT elements. With SMD down to 0603 I see not a problem except that since about 15 years I need to do it with stronger and stronger glasses

@X4V1

Is it correct to create one footprint per sensor even if it needs 2 connector to be plugged in ?

If relative positions have to be fixed than I would do one footprint. If that two connectors are used to connect cables going from sensor I would do two footprints to be able to position each one as it will be better at PCB.

@X4V1

Is it ok to draw the outline of the sensor on the fabrication layer ? Or should I draw the outline of each connector separately ?

I have never send fabrication layer to our contract manufacturer. I understand fabrication layer as being something only for me. For example I used to place footprints with touching one to the next with their courtyard rectangles. But it happened that when I had a long line of SMB transils and placing them according courtyard made that line longer than connector pins for which these transils were destined I placed them little closer just looking at pictures at fabrication layer to verify how close they really can be.

@X4V1

If I can create one footprint for the whole sensor (containing the 2 female headers), won’t it be a problem to describe what must be assembled by the manufacturer ?

As I have written I have no experience with JLCPCB. For my contract manufacturer if that 2 holders have to be in fixed position relative (what I still am not sure how it is in your case) I would create one footprint and if he will be expected to assemble only two headers without sensor I will have in BOM that this part for example is 2x Female header 1x8. I would also make a picture with these two headers shown as separate rectangles.
I send to my manufacturer a picture with courtyard rectangles showing where each part is located and not the fabrication pictures (in courtyard rectangle I simply have more room for reference symbol). I have also pin 1 and polarization markings at courtyard (and I have courtyard checking switched off in DRC). So I suppose that for such sensor I would probably have 3 rectangles at courtyard in its footprint.
I know that I use courtyard something against KiCad intention and I am waiting when I will be able to stop doing it. When KiCad will allow to add at least one user layer pair (flipping top-bottom) I will move onto it with what I need for my pictures.

@X4V1

I read that the anchor of the footprint must at the center, but some others says that the anchor must be at the pin 1 to make it clear for the manufacturer, what do you think about that ?

I think that traditionally for THT anchor was at pin 1 and when SMD came for them anchor is in center.
Thanks to being at pin 1 all anchors were in grid when elements were placed on grid. May be it was important those days.

[X4V1]

Thank you all for your answers and advices, it helps me a lot !

@PapaJ

typically there is one footprint per insertion

I’m trying to stick to that but as Kicad is not allowing me to have 1 symbol mapped to 2 footprints, it would force me to split a correct symbol into 2 half. I’m not sure my statement is clear so I’ll give an exemple: an esp32 devkit has 2 row of pins, so 2 place 2 connectors I would need to split the symbol into 2 half esp32 symbol which I don’t really like.

@Piotr

I would create one footprint and if he will be expected to assemble only two headers without sensor I will have in BOM that this part for example is 2x Female header 1x8. I would also make a picture with these two headers shown as separate rectangles.

That could be a good option as it’s not making my symbols weird and solve the problem. Will have to chech with the manufacturer if they are ok with that.

@straubm

They have an affiliate, LCSC. It is best to stick to their inventory. Components that are not offered by LCSC can be highly expensive and AFAIK it is not possible to send JLCPCB user supplied parts.

Yesterday I requested JLCPCB to order some parts from Digikey for me, I’ll see what it cost (will post it here in case it could help someone else). I placed the order for parts, now waiting for them to get the parts. I did that for parts that I was not able to find in their catalog (or in lcsc catalog).

@straubm

JLCPCB has sample files for both on their page IIRC and there is a KiCad plugin to simplify things further.

I started yesterday night to check that more in detail, I think I was able to generate to BOM correctly, I’ll focus on the CPL asap ! Thank you for pointing that out.

@retiredfeline

For THT parts like connectors consider doing them yourself. Othwerwise they will have to be done in a separate pass and not by machine. Anyway their quote will show you what you have to pay.

Good remark ! I’ll solder all the one I can by myself ! But as I only have a basic soldering iron (weller w61), I don’t think it will be possible to solder SMD or small connector with pitch smaller than 2,54mm.

@Piotr

If in your plan this PCB is your first and last than it has a deep sense. But if you plan to play with electronic little more than in my opinion the better decision would be to accumulate tools needed and learn assembling PCBs yourself even on cost that this first one will be not success.

The plan is to start with something, if I really like that and continue playing with it, I’ll buy all the tools for sure. For now I’m not sure (and it seems to cost some pennies ^^) I checked quickly and apparently for SMD I need a microscope, a soldering station (heat gun + small tips), experience (that seems difficult to buy ), soldering paste, … I will probably buy those, but I first would like to be sure it’s not something I won’t use really often. I also need an oscilloscope, and many other things, it’s an expensive hobby

@Piotr

I have never ordered anything from JLCPCB but I suppose that if you order 10 your PCBs it will cost less than one with assembly.

I don’t know, I’ll tell you when it will be done ^^

@Piotr

There is no problem with assembling manually all THT elements. With SMD down to 0603 I see not a problem except that since about 15 years I need to do it with stronger and stronger glasses

Haha I know that feeliong, I’m looking to buy a microscope to see what I’m doing

@Piotr

If relative positions have to be fixed than I would do one footprint.

It is my use case.

@Piotr

I have never send fabrication layer to our contract manufacturer. I understand fabrication layer as being something only for me.

Good to know, I thought it was used by the manufacturer to delimit parts.

@Piotr

I send to my manufacturer a picture with courtyard rectangles showing where each part is located and not the fabrication pictures (in courtyard rectangle I simply have more room for reference symbol)

Do you take a screenshot of your pcb and add comments on it ? Like the name of the holder you want with an arrow the right location ? Do you have an example ? It can be good to avoid misunderstanding, good idea !

@Piotr

I have also pin 1 and polarization markings at courtyard

How do you indicate the first pin and polarization ? I’m currently creating a rectangular pad for the first pin thinking it was clear enough, but I may be wrong.

@Piotr

So I suppose that for such sensor I would probably have 3 rectangles at courtyard in its footprint.

Is it not an issue for the manufacturer to having one footprint with multiple courtyard rectangle ? I though the role of the courtyard rectangle was to indicate the limit of the component so Is it not confusing for them ?

@Piotr

I think that traditionally for THT anchor was at pin 1 and when SMD came for them anchor is in center.
Thanks to being at pin 1 all anchors were in grid when elements were placed on grid. May be it was important those days.

I’m not sure to understand what you mean.

Again a big thank you to all of you for your precious help

[retiredfeline]

@X4V1

Yesterday I requested JLCPCB to order some parts from Digikey for me,

I’ve read elsewhere that importing parts into China is an expensive can of worms.

[Piotr]

@X4V1

@PapaJ

typically there is one footprint per insertion

I’m trying to stick to that but as Kicad is not allowing me to have 1 symbol mapped to 2 footprints,

You don’t understand. If your element has two connectors but not at cables but at its PCB then they both are ‘one insertion’ so one footprint.

@X4V1

and apparently for SMD I need a microscope,

I started to assemble SMD myself around 1992 and till now I don’t have microscope. When I started to have to use glasses for reading I have bought a workshop lamp with magnifying glass.

@X4V1

a soldering station

I have no space at desk for station so I use:

GOOT PX-201 soldering iron

@X4V1

soldering paste,

Not needed. I think.
I use 0,5mm tin with rosin and metal container of 8mm film with rosin dissolved in it.

@X4V1

I also need an oscilloscope,

It was very expensive when I started to be interested in electronic so to have one I had to do one myself. Now enough good (as for amateur needs) digital oscilloscope is rather cheap. And you can also use USB oscilloscope adapter that is even cheaper, I think.

@X4V1

it’s an expensive hobby

No!
First silicon transistor I have ever had in my hands costed 3% of my father month salary. Now BC547 is about 0.004% of typical month salary (here in Poland).

@X4V1

I’m looking to buy a microscope

Try workshop lamp with magnifying glass first.

@X4V1

Do you have an example ?

Here is fragment of picture we put into our documentation. We have never send any KiCad files. Only gerbers (+ drill files) needed to manufacture PCB (fabrication, courtyard are not needed for it) and P&P files needed to program assembling machine. The rest is in form of pdf containing BOM + needed comments and remarks + pictures like this one (second is the same but with values, but I think they don’t need it). We have never send them any schematic of our devices.

@X4V1

How do you indicate the first pin and polarization ?

You see U5 first pin. In diode box I add one line as cathode marking.

@X4V1

I’m currently creating a rectangular pad for the first pin thinking it was clear enough, but I may be wrong.

Try to do it with QFN package
As you see I use light gray with copper at my pictures. It can be difficult to find rectangular pad at it, specially if it happens to have there some text.

@X4V1

Is it not an issue for the manufacturer to having one footprint with multiple courtyard rectangle ?

My manufacturer don’t get courtyard. He only get a picture. You can mark the whole module as M1 and one connector rectangle as S1, and second as S2 and write in text that M1-S1 and M1-S2 are the sockets listed in BOM pos.32.

@X4V1

I though the role of the courtyard rectangle was to indicate the limit of the component so Is it not confusing for them ?

And generally you are right. But I use courtyard a little different. I do it only because there are no free layer pair to be used by user.

[X4V1]

@Piotr

You don’t understand. If your element has two connectors but not at cables but at its PCB then they both are ‘one insertion’ so one footprint.

I’m not sure to understand what you mean.
Let’s take an example:

For this ESP32, it has 2 lines of pin headers.
So on the PCB, I’ll need 2 rows of female headers.
Do you consider those 2 female headers as 1 insertion ? It would need 2 female headers to be soldered to the pcb, therefore I see it as 2 insertions, is it correct ?
For now I created one footprint for the ESP32, containing the 2 female headers, I hope it won’t be an issue for JLCPCB for assembly. I’ll finish the PCB then try to generate BOM and CPL to see if they complain about that. I’m not sure they will like the 2 parts to place linked to the same footprint.

You convinced me, I bought a soldering iron for smd
Now I need to play with it and acquire some experience with smd soldering ^^

Also thank you for all the infos and example provided

[Piotr]

@X4V1

Do you consider those 2 female headers as 1 insertion ? It would need 2 female headers to be soldered to the pcb, therefore I see it as 2 insertions, is it correct ?

One insertion was not my therm. I have never used it before. I understood it (may be wrongly) that way that if ESP32 will be inserted as one element than all its pads have to be one footprint.
I understood that sentence (about one insertion) that way because for me it is obvious that they have to be one footprint to ensure that their relative position can’t be changed when you move footprint on PCB.
Such module can be soldered directly to PCB or you can use two female headers. In both cases footprint can be the same.
I didn’t got the idea that you can think of these headers as insertions. My fault.
I have never used modules like that inserted in several sockets at my PCBs so I never had a problem how to describe it.
Now I think - may be the good idea would be to use 3 footprints at PCB. At schematic female headers could be connected one to one to ESP32 pins. ESP footprint will show where place those headers.
But I have never done it.

@X4V1

You convinced me, I bought a soldering iron for smd

Good to hear.

@X4V1

Now I need to play with it and acquire some experience with smd soldering ^^

When the tip is too hot, the flux evaporates too quickly, when it is too cold, the tin does not want to dissolve. For amateur applications, classic tin with lead is much, much better than lead-free. When the solder point gets ugly add some solder with flux (I use 0.5mm) even if it makes too much solder.
In 2006, fearing that classic tin would disappear from the market, I stocked up.
For THT elements I use 100W transformer soldering iron. Something like:

ETP 3 100W | VARIOUS

In my opinion important parameter is that wire used as tip is horizontally installed (like in this one) and not vertically. I do my tips from 1.4mm diameter copper wire (they wear out pretty quickly). I also have many special shapes made of this wire - to heat at once all pins of 0805, 1206, SOT23, SO8, SO16 to desolder it.
Soldering with such iron requires more experience. You should never turn them on for more than a few seconds because the tip gets too hot and soldering becomes ugly.
At that link there is ETP3. In 1980 have bought ETP2. It had so strong switch that you had to squeeze so hard that your hand immediately began to tremble. The first thing I did was replace this switch made of strong spring with a lightly moving microswitch. The switch is so light that it is difficult to hold it in your hand not to turn it on.

[X4V1]

@Piotr

One insertion was not my therm. I have never used it before. I understood it (may be wrongly) that way that if ESP32 will be inserted as one element than all its pads have to be one footprint.
I understood that sentence (about one insertion) that way because for me it is obvious that they have to be one footprint to ensure that their relative position can’t be changed when you move footprint on PCB.
Such module can be soldered directly to PCB or you can use two female headers. In both cases footprint can be the same.
I didn’t got the idea that you can think of these headers as insertions. My fault.
I have never used modules like that inserted in several sockets at my PCBs so I never had a problem how to describe it.

I will try to keep my footprint (one footprint containing the 2 headers with the dimension of the esp on the courtyard layer) and describe what I want to jlcpcb. We’ll see if that works ^^

@Piotr

When the tip is too hot, the flux evaporates too quickly, when it is too cold, the tin does not want to dissolve. For amateur applications, classic tin with lead is much, much better than lead-free.

To be honest I bought 3 different brands of lead-free solder and it’s really annoying to me. Even with the solder at 480°C it does not dissolve correctly (I think the problem was the thermal transfert due to the too small tip). I also bought 2 lead solder from the BMJ brand, it was a world of difference, so much better and easier to use and clean result. I’ll continue to use lead solder for the moment. It’s so much easier for me

@Piotr

In my opinion important parameter is that wire used as tip is horizontally installed (like in this one) and not vertically. I do my tips from 1.4mm diameter copper wire (they wear out pretty quickly). I also have many special shapes made of this wire - to heat at once all pins of 0805, 1206, SOT23, SO8, SO16 to desolder it.

I bought this one: Velleman VTSS230, I’ll see what I can do with that. I first need to play a bit with it and acquire some experience before going with too smal packages.

[Piotr]

@X4V1

it was a world of difference, so much better and easier to use and clean result.

I have never tried Pbfree soldering.

See:
https://nepp.nasa.gov/whisker/
and specially some photos with whisker moss looks good:
https://nepp.nasa.gov/whisker/photos/index.html

In 2005 at IEEE: EMCPSTC list one man wrote that he collected about 4000 white papers about pbfree soldering (he had his page with all these articles) and he buys everything (electronic) he can to not need to buy anything pbfree and will wait till fatal accidents caused by pbfree will make people to get back to use Pb in soldering. And there were fatal accidents and … nothing.
Only 3% of Pb is enough to avoid all pure tin soldering problems.

[jmk]

A couple more soldering hints:
Keep away from active fluxes, many are conductive. Use rosin fluxes specifically for soldering electronics.

Most rosin fluxes, including the type in the solder wire, are easily cleaned from your PCB with a toothbrush and industrial alcohol.
Don’t be afraid to liberally paint the PCB with flux as an aid to soldering.

“Blu-tack” or “Tack-it” or whatever name the product is marketed in your country is a good way to hold a SMD part in place before soldering is started. Don’t leave the stuff on a board for longer than necessary and always solder any pads that have been exposed to Blu-tack within a few hours.
I’ve never tested, but there is probably some sort of corrosive effect if bare copper has been exposed to Blu-tack and then left for a few days.

The bigger the pad, the more heat required. Solder big areas with big tips. Don’t just turn up the heat.
Until you are proficient with soldering, with too much heat and spending too much time in one place, you can easily cause the track to separate from the board.

Good practice is to find an old PCB (with components attached) and re-solder all the solder connections. When the resoldered connections are all consistent in shape and size and shiny, you are ready to start on your new board.

[m852]

Can you clarify what is supposed to be under “Blu-tack” or “Tack-it”?

[jmk]

Nothing, except the PCB

Say, in the case of a 402 or 603 or whatever size, res. or cap. you place a blob of blu-tack on the PCB so it is halfway across the pad 2. You then place the component on the board and push that component up against the blu-tack, taking care not to allow the blu-tack to get under the component, and solder pad 1. Remove blu-tack then solder pad 2.
Don’t push component down onto blu-tack or there may be some blu-tack left underneath and this could cause problems.

With ICs, place IC correctly and place some blu-tac on the PCB while covering a corner (or end) of the IC also. Start soldering farthest away from blu-tack, When enough pins of the IC are soldered so the IC is not going to move, remove blu-tack and finish soldering IC.

Works well and saves having to grow a third arm and hand.

[Piotr]

@jmk

Works well and saves having to grow a third arm and hand.

I’m still not sure what Blu-tact is. Some kind of removable glue?
I have never needed 3 arms to solder SMD. When I solder 0603 I first tin one pad alone. At that moment I can’t put 0603 element at right position as a mound of tin disturbs, but there is some tin with enough of rosin at that pad. Then keeping 0603 with tweezers and iron in second hand I solder one 0603 pad. When tin is hot I can put element at right position. Then, if needed, I can correct its position several times. Then I solder the second pad and then once more the first with adding a little of tin with rosin to make the first one look good.

[paulvdh]

@Piotr

I’m still not sure what Blu-tact is. Some kind of removable glue?

Blu Tack

My prefered method for hand soldering SMT stuff is to first put a bit of solder on the leftside pad of each part, and do this as quickly as possible. The goal is not to evaporate the flux. Then with a soldering iron in my left hand and tweezers in my right hand, I positon an SMT part on it’s pad and solder the left side.
When all the left sides are soldered (provisionally), I swap the iron to my right hand and solder all the right side pads. And when necessary, I re-visit the left side pads to solder them more properly.

For QFP’s I also start with a quick dab of solder on a left side corner, and solder a corner (it may be two or three pads which get shorted, it does not matter). Then do an opposite corner and make sure the pins are aligned properly with the pads. Then I either solder the pads one by one, or do drag soldering over a row of pins. this also automatically solves any shorts I may have made in the first corner. Always use plenty of extra flux with the bigger IC’s. Resistors and capacitors do not need extra flux, but this also depends on the solder you use and your skill.