Reply by Rick C June 22, 20212021-06-22
On Monday, June 21, 2021 at 10:00:07 PM UTC-4, Ed Lee wrote:
> On Monday, June 21, 2021 at 5:56:09 PM UTC-7, gnuarm.del...@gmail.com wrote: > > On Monday, June 21, 2021 at 7:43:38 PM UTC-4, Ed Lee wrote: > > > On Monday, June 21, 2021 at 4:34:38 PM UTC-7, gnuarm.del...@gmail.com wrote: > > > > Yeah, that's why I let assembly houses handle my boards. > > > Yes, for production, but i need to do it by hand myself for prototype. > > > > > > I am going to push some pins out for prototype, in order to expand the pads by 50 mils. > > I don't really understand what you are doing. Many people solder gull wing devices like SOPs, SSOPs, TSSOPs and QFPs by hand quite easily. I have no idea why you need to bend pins to use any of those. What are your concerns? > Sorry, i mean pushing out the pins on the circuit board, not the chip. > > You can look at it as Six-In-Line Package (the center goes into a ZIF 40), with columns at: > > _____550__450__350___350___450__550 (distance from center, in mils) > This allows around 10 mils for extended pads for the TQFP 100. > > For prototype, i am pushing out the second and fifth columns by 30 mils. > _____555__480__350___350___480__550 > This allows around 40 mils for extended pads. > > ______XL___LW____L_____R____RW___XR____|_______B/B____ > _1_____________Key____5v_______________|____B12____5v > _2_____________3v3___Gnd_______________|____B13___Gnd > _3_____________-*-____D2_______________|____B14___3v3 > _4______________B7____D1_______________|____B15___B10 > _5___Key________B8___C12____D0__Gnd____|_____A8____B2 > _6___-*-___B9___E0___C10___C11__-*-____|_____A9____B1 > _7___-*-___E1___A1____A6___-*-__-*-____|____A10____B0 > _8___-*-___E3_______________B5__-*-____|____A11____A7 > _9____E4__C13______________A10__-*-____|____A12____A6 > 10___C14___OI_______________C7___A9____|____A15____A5 > 11____OO___C2______________D12___C6____|_____B3____A4 > 12____C3___VA______________B15__D11____|_____B4____A3 > 13____A0__-*-______________B13__B14____|_____B5____A3 > 14___Gnd__3v3___A3____A2___-*-__Gnd____|_____B6____A1 > 15____A7___C4___A6___B11___E15__B10____|_____B5____A0 > 16____VB___E7___B1___E12___E10__48v____|_____B8_____R > 17_____________ B2___E14_______________|_____B9___C15 > 18_____________ B0___E13_______________|_____5v___C14 > 19_____________Gnd___3v3_______________|____Gnd___C13 > 20______________VB___48v_______________|____3v3____VB
I have no idea what any of the above means. It's incomprehensible in Google Groups.
> > The cost of setup is not so much that machine assembly of prototypes is really cost prohibitive. I think I am typically charged $200 or so. I suppose they might charge a bit more for the unit cost if you are only building a half dozen or so. One house all but refused to hand assemble three of a test fixture with SMD components. This was 13 years ago and I think they still had equipment to facilitate the hand insertion of DIPs in PCBs by shining a light where the chip should go. They said the labor was cost prohibitive. > > > > Sometimes it is hard for an engineer to understand the issues of an assembly house, particularly the difference between what a motivated engineer can do and what an hourly worker is willing to do. > I only need a handful to test it out first. Will ask how much they would charge. It should be cheaper if they can do it by hand without custom made stencil.
Of course you only need a few now, but you will need to pay the setup when you do go to production and if you have no changes it's already done. There are places that will make your boards and assemble it for very little, like JLCPCB. Are you using parts they can't supply? I know they won't assemble anything they don't sell. -- Rick C. ++ Get 1,000 miles of free Supercharging ++ Tesla referral code - https://ts.la/richard11209
Reply by Ed Lee June 21, 20212021-06-21
On Monday, June 21, 2021 at 5:56:09 PM UTC-7, gnuarm.del...@gmail.com wrote:
> On Monday, June 21, 2021 at 7:43:38 PM UTC-4, Ed Lee wrote: > > On Monday, June 21, 2021 at 4:34:38 PM UTC-7, gnuarm.del...@gmail.com wrote: > > > Yeah, that's why I let assembly houses handle my boards. > > Yes, for production, but i need to do it by hand myself for prototype. > > > > I am going to push some pins out for prototype, in order to expand the pads by 50 mils. > I don't really understand what you are doing. Many people solder gull wing devices like SOPs, SSOPs, TSSOPs and QFPs by hand quite easily. I have no idea why you need to bend pins to use any of those. What are your concerns?
Sorry, i mean pushing out the pins on the circuit board, not the chip. You can look at it as Six-In-Line Package (the center goes into a ZIF 40), with columns at: _____550__450__350___350___450__550 (distance from center, in mils) This allows around 10 mils for extended pads for the TQFP 100. For prototype, i am pushing out the second and fifth columns by 30 mils. _____555__480__350___350___480__550 This allows around 40 mils for extended pads. ______XL___LW____L_____R____RW___XR____|_______B/B____ _1_____________Key____5v_______________|____B12____5v _2_____________3v3___Gnd_______________|____B13___Gnd _3_____________-*-____D2_______________|____B14___3v3 _4______________B7____D1_______________|____B15___B10 _5___Key________B8___C12____D0__Gnd____|_____A8____B2 _6___-*-___B9___E0___C10___C11__-*-____|_____A9____B1 _7___-*-___E1___A1____A6___-*-__-*-____|____A10____B0 _8___-*-___E3_______________B5__-*-____|____A11____A7 _9____E4__C13______________A10__-*-____|____A12____A6 10___C14___OI_______________C7___A9____|____A15____A5 11____OO___C2______________D12___C6____|_____B3____A4 12____C3___VA______________B15__D11____|_____B4____A3 13____A0__-*-______________B13__B14____|_____B5____A3 14___Gnd__3v3___A3____A2___-*-__Gnd____|_____B6____A1 15____A7___C4___A6___B11___E15__B10____|_____B5____A0 16____VB___E7___B1___E12___E10__48v____|_____B8_____R 17_____________ B2___E14_______________|_____B9___C15 18_____________ B0___E13_______________|_____5v___C14 19_____________Gnd___3v3_______________|____Gnd___C13 20______________VB___48v_______________|____3v3____VB
> > The cost of setup is not so much that machine assembly of prototypes is really cost prohibitive. I think I am typically charged $200 or so. I suppose they might charge a bit more for the unit cost if you are only building a half dozen or so. One house all but refused to hand assemble three of a test fixture with SMD components. This was 13 years ago and I think they still had equipment to facilitate the hand insertion of DIPs in PCBs by shining a light where the chip should go. They said the labor was cost prohibitive. > > Sometimes it is hard for an engineer to understand the issues of an assembly house, particularly the difference between what a motivated engineer can do and what an hourly worker is willing to do.
I only need a handful to test it out first. Will ask how much they would charge. It should be cheaper if they can do it by hand without custom made stencil.
Reply by Rick C June 21, 20212021-06-21
On Monday, June 21, 2021 at 7:43:38 PM UTC-4, Ed Lee wrote:
> On Monday, June 21, 2021 at 4:34:38 PM UTC-7, gnuarm.del...@gmail.com wrote: > > Yeah, that's why I let assembly houses handle my boards. > Yes, for production, but i need to do it by hand myself for prototype. > > I am going to push some pins out for prototype, in order to expand the pads by 50 mils.
I don't really understand what you are doing. Many people solder gull wing devices like SOPs, SSOPs, TSSOPs and QFPs by hand quite easily. I have no idea why you need to bend pins to use any of those. What are your concerns? The cost of setup is not so much that machine assembly of prototypes is really cost prohibitive. I think I am typically charged $200 or so. I suppose they might charge a bit more for the unit cost if you are only building a half dozen or so. One house all but refused to hand assemble three of a test fixture with SMD components. This was 13 years ago and I think they still had equipment to facilitate the hand insertion of DIPs in PCBs by shining a light where the chip should go. They said the labor was cost prohibitive. Sometimes it is hard for an engineer to understand the issues of an assembly house, particularly the difference between what a motivated engineer can do and what an hourly worker is willing to do. -- Rick C. +- Get 1,000 miles of free Supercharging +- Tesla referral code - https://ts.la/richard11209
Reply by Ed Lee June 21, 20212021-06-21
On Monday, June 21, 2021 at 4:34:38 PM UTC-7, gnuarm.del...@gmail.com wrote:
> On Monday, June 21, 2021 at 1:41:43 PM UTC-4, boB wrote: > > On Sun, 20 Jun 2021 19:04:14 -0700 (PDT), Rick C > > <gnuarm.del...@gmail.com> wrote: > > > > >On Sunday, June 20, 2021 at 8:02:56 PM UTC-4, boB wrote: > > >> On Sun, 20 Jun 2021 05:30:08 -0700 (PDT), Rick C > > >> <gnuarm.del...@gmail.com> wrote: > > >> > > >> >On Sunday, June 20, 2021 at 4:19:25 AM UTC-4, Klaus Kragelund wrote: > > >> >> 19.06.21 18:29, Rick C wrote: > > >> >> >On Saturday, June 19, 2021 at 12:09:45 PM UTC-4, Ed Lee wrote: > > >> >> >> When i was using another PCB design tool, i built my own smd footprints . To do a quick prototype, i am using the linux "pcb" tool and the build-in footprints. The build-in footprints have pads with the same size as pins. > > >> >> >> > > >> >> >> When i build my own, i found it helpful to extend the pads beyond the pins. With enough flux, just let the solder suck in from the extended area into the pins. Also, it's easier to clear solder bridges on the extended area, rather than on the pins. > > >> >> >> > > >> >> >> I am adding 10 mils to the pads beyond the pins. Actually, easier to solder with longer extensions. > > >> >> >> > > >> >> >> What do you guys use? > > >> >> > > > >> >> >There are IPC guidelines that are generally pretty good. One for reflow that has minimal expansion beyond the pins, but enough and another I think is for wave soldering (yes, you can wave solder surface mount) with larger extensions. The 10 mil might be good at the toe where most provide a significant extension. The sides often are only a couple of mils if that and the heel can be as little as 5 mil, but don't hold me to any of these numbers. It's been a while since I laid out a footprint. > > >> >> > > > >> >> On the topic of wave solder, you need to glue the parts first, that's the normal way I guess > > >> >> > > >> >> Is it possible to use solder paste, pick and place, flip the board and wave solder? > > >> > > > >> >Not sure what you are describing. If you mean using glue, yes. One outfit wanted to do that with my board. I never found out why, but someone finally said it was not necessary. > > >> We do double sides SMT every day without using glue. I thought you > > >> had to use glue, too 20+ years ago. > > >> > > >> But we stuff the through hole parts on those double sided SMT boards. > > >> > > >> of course, no problem doing SMT on one side, through the oven and then > > >> stuff and wave solder the through hole parts. > > > > > >Double sided SMT reflow is one thing, but trying to run SMT through a wave solder bath without glue seems a bit risky. If you are describing single sided mixed reflow and then wave soldering the other side, that should work. > > No, we don't wave solder with SMT on the same side as the solder. That > > DOES sound risky but I guess that can be done somehow ? > Somehow? Yes, glue the parts on the board. > > However, we DO also have something called a "Selective solder > > machine" which automatically (programmed) solders individiual > > through-hole leads, usually larger bus-bars on the bottom which could > > have SMT parts on it. It isn't really any faster than soldering by > > hand but is a lot more repeatable. > Yeah, that's why I let assembly houses handle my boards.
Yes, for production, but i need to do it by hand myself for prototype. I am going to push some pins out for prototype, in order to expand the pads by 50 mils.
Reply by Rick C June 21, 20212021-06-21
On Monday, June 21, 2021 at 1:41:43 PM UTC-4, boB wrote:
> On Sun, 20 Jun 2021 19:04:14 -0700 (PDT), Rick C > <gnuarm.del...@gmail.com> wrote: > > >On Sunday, June 20, 2021 at 8:02:56 PM UTC-4, boB wrote: > >> On Sun, 20 Jun 2021 05:30:08 -0700 (PDT), Rick C > >> <gnuarm.del...@gmail.com> wrote: > >> > >> >On Sunday, June 20, 2021 at 4:19:25 AM UTC-4, Klaus Kragelund wrote: > >> >> 19.06.21 18:29, Rick C wrote: > >> >> >On Saturday, June 19, 2021 at 12:09:45 PM UTC-4, Ed Lee wrote: > >> >> >> When i was using another PCB design tool, i built my own smd footprints . To do a quick prototype, i am using the linux "pcb" tool and the build-in footprints. The build-in footprints have pads with the same size as pins. > >> >> >> > >> >> >> When i build my own, i found it helpful to extend the pads beyond the pins. With enough flux, just let the solder suck in from the extended area into the pins. Also, it's easier to clear solder bridges on the extended area, rather than on the pins. > >> >> >> > >> >> >> I am adding 10 mils to the pads beyond the pins. Actually, easier to solder with longer extensions. > >> >> >> > >> >> >> What do you guys use? > >> >> > > >> >> >There are IPC guidelines that are generally pretty good. One for reflow that has minimal expansion beyond the pins, but enough and another I think is for wave soldering (yes, you can wave solder surface mount) with larger extensions. The 10 mil might be good at the toe where most provide a significant extension. The sides often are only a couple of mils if that and the heel can be as little as 5 mil, but don't hold me to any of these numbers. It's been a while since I laid out a footprint. > >> >> > > >> >> On the topic of wave solder, you need to glue the parts first, that's the normal way I guess > >> >> > >> >> Is it possible to use solder paste, pick and place, flip the board and wave solder? > >> > > >> >Not sure what you are describing. If you mean using glue, yes. One outfit wanted to do that with my board. I never found out why, but someone finally said it was not necessary. > >> We do double sides SMT every day without using glue. I thought you > >> had to use glue, too 20+ years ago. > >> > >> But we stuff the through hole parts on those double sided SMT boards. > >> > >> of course, no problem doing SMT on one side, through the oven and then > >> stuff and wave solder the through hole parts. > > > >Double sided SMT reflow is one thing, but trying to run SMT through a wave solder bath without glue seems a bit risky. If you are describing single sided mixed reflow and then wave soldering the other side, that should work. > No, we don't wave solder with SMT on the same side as the solder. That > DOES sound risky but I guess that can be done somehow ?
Somehow? Yes, glue the parts on the board.
> However, we DO also have something called a "Selective solder > machine" which automatically (programmed) solders individiual > through-hole leads, usually larger bus-bars on the bottom which could > have SMT parts on it. It isn't really any faster than soldering by > hand but is a lot more repeatable.
Yeah, that's why I let assembly houses handle my boards. -- Rick C. -+ Get 1,000 miles of free Supercharging -+ Tesla referral code - https://ts.la/richard11209
Reply by boB June 21, 20212021-06-21
On Sun, 20 Jun 2021 19:04:14 -0700 (PDT), Rick C
<gnuarm.deletethisbit@gmail.com> wrote:

>On Sunday, June 20, 2021 at 8:02:56 PM UTC-4, boB wrote: >> On Sun, 20 Jun 2021 05:30:08 -0700 (PDT), Rick C >> <gnuarm.del...@gmail.com> wrote: >> >> >On Sunday, June 20, 2021 at 4:19:25 AM UTC-4, Klaus Kragelund wrote: >> >> 19.06.21 18:29, Rick C wrote: >> >> >On Saturday, June 19, 2021 at 12:09:45 PM UTC-4, Ed Lee wrote: >> >> >> When i was using another PCB design tool, i built my own smd footprints . To do a quick prototype, i am using the linux "pcb" tool and the build-in footprints. The build-in footprints have pads with the same size as pins. >> >> >> >> >> >> When i build my own, i found it helpful to extend the pads beyond the pins. With enough flux, just let the solder suck in from the extended area into the pins. Also, it's easier to clear solder bridges on the extended area, rather than on the pins. >> >> >> >> >> >> I am adding 10 mils to the pads beyond the pins. Actually, easier to solder with longer extensions. >> >> >> >> >> >> What do you guys use? >> >> > >> >> >There are IPC guidelines that are generally pretty good. One for reflow that has minimal expansion beyond the pins, but enough and another I think is for wave soldering (yes, you can wave solder surface mount) with larger extensions. The 10 mil might be good at the toe where most provide a significant extension. The sides often are only a couple of mils if that and the heel can be as little as 5 mil, but don't hold me to any of these numbers. It's been a while since I laid out a footprint. >> >> > >> >> On the topic of wave solder, you need to glue the parts first, that's the normal way I guess >> >> >> >> Is it possible to use solder paste, pick and place, flip the board and wave solder? >> > >> >Not sure what you are describing. If you mean using glue, yes. One outfit wanted to do that with my board. I never found out why, but someone finally said it was not necessary. >> We do double sides SMT every day without using glue. I thought you >> had to use glue, too 20+ years ago. >> >> But we stuff the through hole parts on those double sided SMT boards. >> >> of course, no problem doing SMT on one side, through the oven and then >> stuff and wave solder the through hole parts. > >Double sided SMT reflow is one thing, but trying to run SMT through a wave solder bath without glue seems a bit risky. If you are describing single sided mixed reflow and then wave soldering the other side, that should work.
No, we don't wave solder with SMT on the same side as the solder. That DOES sound risky but I guess that can be done somehow ? However, we DO also have something called a "Selective solder machine" which automatically (programmed) solders individiual through-hole leads, usually larger bus-bars on the bottom which could have SMT parts on it. It isn't really any faster than soldering by hand but is a lot more repeatable. boB
Reply by Ed Lee June 21, 20212021-06-21
On Sunday, June 20, 2021 at 9:15:32 PM UTC-7, Dave Platt wrote:
> In article <bc6b6907-cc47-4473...@googlegroups.com>, > Ed Lee <edward....@gmail.com> wrote: > >When i was using another PCB design tool, i built my own smd footprints . To do a quick prototype, i am using the linux "pcb" > >tool and the build-in footprints. The build-in footprints have pads with the same size as pins. > > > >When i build my own, i found it helpful to extend the pads beyond the pins. With enough flux, just let the solder suck in from > >the extended area into the pins. Also, it's easier to clear solder bridges on the extended area, rather than on the pins. > > > >I am adding 10 mils to the pads beyond the pins. Actually, easier to solder with longer extensions. > > > >What do you guys use? > You might want to take a look at the currently-available footprint > libraries for the KiCAD tool. Many of the SMT footprints come in both > standard, and "hand-solder" varieties, with the latter having extended > pads. > > I did something similar when I made some SMT-to-through-hole adapter > boards for the SMT versions of the Murata 10.7 MHz IF filters. I > ended up adding a couple of millimeters to the pads.
Changing footprint is easy enough, but i would need to relocate components for any more than 10 mils (1/4mm). Just wondering how much more i need for next batch of boards. For now, i am bending the pins inward to expose some pad area. It's lot of work and not very accuracy.
Reply by Dave Platt June 21, 20212021-06-21
In article <bc6b6907-cc47-4473-8e19-1e22344e796bn@googlegroups.com>,
Ed Lee  <edward.ming.lee@gmail.com> wrote:
>When i was using another PCB design tool, i built my own smd footprints . To do a quick prototype, i am using the linux "pcb" >tool and the build-in footprints. The build-in footprints have pads with the same size as pins. > >When i build my own, i found it helpful to extend the pads beyond the pins. With enough flux, just let the solder suck in from >the extended area into the pins. Also, it's easier to clear solder bridges on the extended area, rather than on the pins. > >I am adding 10 mils to the pads beyond the pins. Actually, easier to solder with longer extensions. > >What do you guys use?
You might want to take a look at the currently-available footprint libraries for the KiCAD tool. Many of the SMT footprints come in both standard, and "hand-solder" varieties, with the latter having extended pads. I did something similar when I made some SMT-to-through-hole adapter boards for the SMT versions of the Murata 10.7 MHz IF filters. I ended up adding a couple of millimeters to the pads.
Reply by Rick C June 20, 20212021-06-20
On Sunday, June 20, 2021 at 8:02:56 PM UTC-4, boB wrote:
> On Sun, 20 Jun 2021 05:30:08 -0700 (PDT), Rick C > <gnuarm.del...@gmail.com> wrote: > > >On Sunday, June 20, 2021 at 4:19:25 AM UTC-4, Klaus Kragelund wrote: > >> 19.06.21 18:29, Rick C wrote: > >> >On Saturday, June 19, 2021 at 12:09:45 PM UTC-4, Ed Lee wrote: > >> >> When i was using another PCB design tool, i built my own smd footprints . To do a quick prototype, i am using the linux "pcb" tool and the build-in footprints. The build-in footprints have pads with the same size as pins. > >> >> > >> >> When i build my own, i found it helpful to extend the pads beyond the pins. With enough flux, just let the solder suck in from the extended area into the pins. Also, it's easier to clear solder bridges on the extended area, rather than on the pins. > >> >> > >> >> I am adding 10 mils to the pads beyond the pins. Actually, easier to solder with longer extensions. > >> >> > >> >> What do you guys use? > >> > > >> >There are IPC guidelines that are generally pretty good. One for reflow that has minimal expansion beyond the pins, but enough and another I think is for wave soldering (yes, you can wave solder surface mount) with larger extensions. The 10 mil might be good at the toe where most provide a significant extension. The sides often are only a couple of mils if that and the heel can be as little as 5 mil, but don't hold me to any of these numbers. It's been a while since I laid out a footprint. > >> > > >> On the topic of wave solder, you need to glue the parts first, that's the normal way I guess > >> > >> Is it possible to use solder paste, pick and place, flip the board and wave solder? > > > >Not sure what you are describing. If you mean using glue, yes. One outfit wanted to do that with my board. I never found out why, but someone finally said it was not necessary. > We do double sides SMT every day without using glue. I thought you > had to use glue, too 20+ years ago. > > But we stuff the through hole parts on those double sided SMT boards. > > of course, no problem doing SMT on one side, through the oven and then > stuff and wave solder the through hole parts.
Double sided SMT reflow is one thing, but trying to run SMT through a wave solder bath without glue seems a bit risky. If you are describing single sided mixed reflow and then wave soldering the other side, that should work. -- Rick C. -- Get 1,000 miles of free Supercharging -- Tesla referral code - https://ts.la/richard11209
Reply by boB June 20, 20212021-06-20
On Sun, 20 Jun 2021 05:30:08 -0700 (PDT), Rick C
<gnuarm.deletethisbit@gmail.com> wrote:

>On Sunday, June 20, 2021 at 4:19:25 AM UTC-4, Klaus Kragelund wrote: >> 19.06.21 18:29, Rick C wrote: >> >On Saturday, June 19, 2021 at 12:09:45 PM UTC-4, Ed Lee wrote: >> >> When i was using another PCB design tool, i built my own smd footprints . To do a quick prototype, i am using the linux "pcb" tool and the build-in footprints. The build-in footprints have pads with the same size as pins. >> >> >> >> When i build my own, i found it helpful to extend the pads beyond the pins. With enough flux, just let the solder suck in from the extended area into the pins. Also, it's easier to clear solder bridges on the extended area, rather than on the pins. >> >> >> >> I am adding 10 mils to the pads beyond the pins. Actually, easier to solder with longer extensions. >> >> >> >> What do you guys use? >> > >> >There are IPC guidelines that are generally pretty good. One for reflow that has minimal expansion beyond the pins, but enough and another I think is for wave soldering (yes, you can wave solder surface mount) with larger extensions. The 10 mil might be good at the toe where most provide a significant extension. The sides often are only a couple of mils if that and the heel can be as little as 5 mil, but don't hold me to any of these numbers. It's been a while since I laid out a footprint. >> > >> On the topic of wave solder, you need to glue the parts first, that's the normal way I guess >> >> Is it possible to use solder paste, pick and place, flip the board and wave solder? > >Not sure what you are describing. If you mean using glue, yes. One outfit wanted to do that with my board. I never found out why, but someone finally said it was not necessary.
We do double sides SMT every day without using glue. I thought you had to use glue, too 20+ years ago. But we stuff the through hole parts on those double sided SMT boards. of course, no problem doing SMT on one side, through the oven and then stuff and wave solder the through hole parts.