On Saturday, January 14, 2023 at 12:08:03 PM UTC-4, John Larkin wrote:> On Fri, 13 Jan 2023 21:20:50 -0800 (PST), "gnuarm.del...@gmail.com" > <gnuarm.del...@gmail.com> wrote:=20 >=20 > >On Saturday, January 14, 2023 at 12:39:49 AM UTC-4, John Larkin wrote:==20> >> On Sat, 7 Jan 2023 09:49:24 -0800 (PST), "gnuarm.del...@gmail.com"=20 > >> <gnuarm.del...@gmail.com> wrote:=20 > >>=20 > >> >A small board with a 100QFP is being redesigned for a new FPGA due to=obsolescence. Gowin makes a 100QFP device that would be a good fit, but my= customer has said "no" to the 100% Chinese brand... US government customer= s, ya know!=20> >> >=20 > >> >So now I'm looking at a BGA. I don't want to get into fine PCB design=rules, so 1.0 mm ball pitch is my preference. The only devices I can find = that fit on the board have 196 or 256 pins. But the real problem is availab= ility.=20> >> >=20 > >> >Digikey has a few of the XC7S15-1FTGB196I and more a scheduled for de=livery in April. Add in the various speed and temperature flavors trickling= in (mostly in April) and I should be ok for the initial delivery in August= ... if I can get my hands on those. I don't know if Digikey factors in the = backlog orders in these counts.=20> >> >=20 > >> >Mouser shows great inventory of Efinix parts, particularly the T13 an=d T20 in a 0.8 mm 256 pin BGA, 10s of thousands in stock. But I'd rather wo= rk with a 1.0 mm BGA. Oddly enough, LCSC shows part numbers, but zero inven= tory.=20> >> >=20 > >> >Anyone work with 0.8 mm BGAs? What PWB feature dimensions did you use=? Did this impact the PWB cost?=20> >> The 0.8 mm 256-ball T20 isn't bad...=20 > >>=20 > >> https://www.dropbox.com/s/xjqgj2pz9mdhtma/P941_FPGA.jpg?raw=3D1=20 > >=20 > >I can't really see much detail. It looks like there are virtually no pad=s on the vias under the BGA. What size are they?> The BGAVIAs are 12.5 mil OD with 8 mil drills. The other vias on the=20 > board STANDARDVIA and POWERVIA are bigger.=202.25 mil (0.057 mm) is a pretty narrow via ring. Why not have a larger via= pad? You have just over 0.2 mm (8 mil) clearance between via pad and BGA = pads. Is this narrow annular ring buying you something? You could use a 0= .4 mm (16 mil) via pad for an annular ring of 0.1 mm (4 mil) and still have= 0.165 mm (6.5 mil) clearance between ball pads and via pads. Do you feel = that's not enough?=20> I have seen vias with no annullar ring, just a trace falling into a=20 > hole, but the PCB houses don't like that.=20No, and they don't like small annular rings, because that's a small target = to drill. I run into boards that the PCB fab house made badly at the vias = and they are disasters. The tiny cracks that develop are hard to find and = don't repair well.=20 Efinix recommends 0.46 mm (18.1 mil) ball pad, 0.5 mm (20 mil) via pad and = a 0.25 mm (10 mil) drill, with 0.1 mm (4 mil) trace/space and 0.085 mm (3.3= mil) clearance between ball pad and via pad. The trace/space seems fine, = but I'd like more clearance between via and ball pads. The question is, wh= ere to shave it from? Shaving 2 mil from the via pad leaves 4 mil annular = ring and 4.3 mil clearance. Shaving from the ball pad seems like a bad ide= a. But if it works...=20 This is something that should have a spreadsheet, with a diagram that adjus= ts the image to show the details. All the tradeoffs become apparent very q= uickly. lol=20> Filled via-in-pad would be cool but that's complex and expensive. As=20 > is buried vias. > >=20 > >=20 > >> The BGA pads are 16 mils. 8 mil drills on the BGA vias. 6 mil traces==20> >> mostly, except for the 50 ohm monsters. No big deal these days. Works==20> >> great.=20 > >=20 > >Yeah, 0.8 mm pad centers are doable, but I don't know where the line is =for higher pricing on the PWB. The via pads seem to be pushing the technolo= gy line at JLCPCB. Not that I'm using them, but if they can do it, pretty m= uch anyone should be able to do it. They build 0.45 mm via pads and 0.2 mm = drills (5 mil annular ring and 8 mil drill), but charge extra for a 0.4 mm = via pads (4 mil annular ring).> We use US suppliers for production boards, and they seem to think this=20 > 6-layer board is within the normal range. One advantage to using a big=20 > FPGA (256 balls in this case) is that you don't have to go deep to hit=20 > enough balls, so may save a PCB layer or two. The T20-256 is a nice=20 > part and Digikey has 29,000 in stock.=20I don't follow that exactly. If I could get a 100 ball BGA on 1.0 mm cente= rs, I would have zero trouble with routing and breakout. That could be rou= ted 100% on a double sided board. One side gets the two outer rings of pad= s leaving 6x6. The other layer routes the remainder. But FPGA companies d= on't like small packages. They have much more demand at the larger I/O cou= nts. Xilinx has a 196 ball, 1.0 mm package, but not much inventory and lea= d time is the standard 52 weeks.=20 The real problem is having to use the packages that are in stock. Everythi= ng other than Efinix is 52 week lead time, which is not a real forecast, ra= ther just the point where they stop counting. =20> Another project used a 484 ball Zynq and we used almost every ball.=20 > Lots of different power pours too. That took 10 layers. Another recent=20 > board has a 400-ball ZYNQ with a few unused PS pins and fits on 8=20 > layers.=20Yeah, when you have that many I/Os, it's tough to keep the layer count down= . =20> The ZYNQ has analog inputs but, crazily, they are all differential so=20 > they make you ground a perfectly good i/o pin for every analog input=20 > that you want. > >=20 > >=20 > >> We considered a T8 for a simpler application, but its 0.5 mm ball=20 > >> pitch looked nasty.=20 > >=20 > >I didn't price the T8, because they use the logic cells for routing in a=way they don't explain, so no way to factor it in. The T12 would be gravy = for my design I expect, but it's only $1 more for the T20, so why not? If i= t saves a day of work, it's a break even for 1,000 units. If it enables a f= uture expansion, it's worth much more than that! Both parts seem to have th= e same pin out, including I/O counts, so switching between them should only= be a recompile.=20> >=20 > >=20 > >> The efinix tool chain looks like it was developed in someone's garage,==20> >> which is actually praise. It's free and simple and just works without==20> >> 200 gbyte downloads and doing battle with FlexLM.=20 > >=20 > >The large downloads are from the support for the many, many products the=big three FPGA companies sell. Don't expect Efinix tools to continue to be= small... and they aren't really free. You have to buy a board. That's more= than I've paid for tools from FPGA vendors.> $150! That's in the noise, and an eval board is good anyhow. > >=20 > >I'd really like to use the Gowin parts (LQFP100). But the customer is hi=nky about parts from a Chinese company. They sell stuff to the US Governmen= t.> Yeah, we have a lot of aerospace customers and avoid Chinese parts.Where exactly are Efinix parts made? Their parts say China on them. =20 --=20 Rick C. +++ Get 1,000 miles of free Supercharging +++ Tesla referral code - https://ts.la/richard11209
PCB Layout for BGAs
Started by ●January 7, 2023
Reply by ●January 14, 20232023-01-14
Reply by ●January 14, 20232023-01-14
On Sat, 14 Jan 2023 10:05:33 -0800 (PST), "gnuarm.del...@gmail.com" <gnuarm.deletethisbit@gmail.com> wrote:>On Saturday, January 14, 2023 at 12:08:03 PM UTC-4, John Larkin wrote: >> On Fri, 13 Jan 2023 21:20:50 -0800 (PST), "gnuarm.del...@gmail.com" >> <gnuarm.del...@gmail.com> wrote: >> >> >On Saturday, January 14, 2023 at 12:39:49 AM UTC-4, John Larkin wrote: >> >> On Sat, 7 Jan 2023 09:49:24 -0800 (PST), "gnuarm.del...@gmail.com" >> >> <gnuarm.del...@gmail.com> wrote: >> >> >> >> >A small board with a 100QFP is being redesigned for a new FPGA due to obsolescence. Gowin makes a 100QFP device that would be a good fit, but my customer has said "no" to the 100% Chinese brand... US government customers, ya know! >> >> > >> >> >So now I'm looking at a BGA. I don't want to get into fine PCB design rules, so 1.0 mm ball pitch is my preference. The only devices I can find that fit on the board have 196 or 256 pins. But the real problem is availability. >> >> > >> >> >Digikey has a few of the XC7S15-1FTGB196I and more a scheduled for delivery in April. Add in the various speed and temperature flavors trickling in (mostly in April) and I should be ok for the initial delivery in August... if I can get my hands on those. I don't know if Digikey factors in the backlog orders in these counts. >> >> > >> >> >Mouser shows great inventory of Efinix parts, particularly the T13 and T20 in a 0.8 mm 256 pin BGA, 10s of thousands in stock. But I'd rather work with a 1.0 mm BGA. Oddly enough, LCSC shows part numbers, but zero inventory. >> >> > >> >> >Anyone work with 0.8 mm BGAs? What PWB feature dimensions did you use? Did this impact the PWB cost? >> >> The 0.8 mm 256-ball T20 isn't bad... >> >> >> >> https://www.dropbox.com/s/xjqgj2pz9mdhtma/P941_FPGA.jpg?raw=1 >> > >> >I can't really see much detail. It looks like there are virtually no pads on the vias under the BGA. What size are they? >> The BGAVIAs are 12.5 mil OD with 8 mil drills. The other vias on the >> board STANDARDVIA and POWERVIA are bigger. > >2.25 mil (0.057 mm) is a pretty narrow via ring. Why not have a larger via pad?I don't know. My PCB guy decides stuff like that. I'd guess that he wanted it to pass some design rule check, or maybe he started metric. The board houses haven't complained as far as I know. You should do your own thing and check with whoever will make your boards.
Reply by ●January 14, 20232023-01-14
On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote:> On Sat, 14 Jan 2023 10:05:33 -0800 (PST), "gnuarm.del...@gmail.com" > <gnuarm.del...@gmail.com> wrote: > > >On Saturday, January 14, 2023 at 12:08:03 PM UTC-4, John Larkin wrote: > >> On Fri, 13 Jan 2023 21:20:50 -0800 (PST), "gnuarm.del...@gmail.com" > >> <gnuarm.del...@gmail.com> wrote: > >> > >> >On Saturday, January 14, 2023 at 12:39:49 AM UTC-4, John Larkin wrote: > >> >> On Sat, 7 Jan 2023 09:49:24 -0800 (PST), "gnuarm.del...@gmail.com" > >> >> <gnuarm.del...@gmail.com> wrote: > >> >> > >> >> >A small board with a 100QFP is being redesigned for a new FPGA due to obsolescence. Gowin makes a 100QFP device that would be a good fit, but my customer has said "no" to the 100% Chinese brand... US government customers, ya know! > >> >> > > >> >> >So now I'm looking at a BGA. I don't want to get into fine PCB design rules, so 1.0 mm ball pitch is my preference. The only devices I can find that fit on the board have 196 or 256 pins. But the real problem is availability. > >> >> > > >> >> >Digikey has a few of the XC7S15-1FTGB196I and more a scheduled for delivery in April. Add in the various speed and temperature flavors trickling in (mostly in April) and I should be ok for the initial delivery in August... if I can get my hands on those. I don't know if Digikey factors in the backlog orders in these counts. > >> >> > > >> >> >Mouser shows great inventory of Efinix parts, particularly the T13 and T20 in a 0.8 mm 256 pin BGA, 10s of thousands in stock. But I'd rather work with a 1.0 mm BGA. Oddly enough, LCSC shows part numbers, but zero inventory. > >> >> > > >> >> >Anyone work with 0.8 mm BGAs? What PWB feature dimensions did you use? Did this impact the PWB cost? > >> >> The 0.8 mm 256-ball T20 isn't bad... > >> >> > >> >> https://www.dropbox.com/s/xjqgj2pz9mdhtma/P941_FPGA.jpg?raw=1 > >> > > >> >I can't really see much detail. It looks like there are virtually no pads on the vias under the BGA. What size are they? > >> The BGAVIAs are 12.5 mil OD with 8 mil drills. The other vias on the > >> board STANDARDVIA and POWERVIA are bigger. > > > >2.25 mil (0.057 mm) is a pretty narrow via ring. Why not have a larger via pad? > I don't know. My PCB guy decides stuff like that. I'd guess that he > wanted it to pass some design rule check, or maybe he started metric. > The board houses haven't complained as far as I know. > > You should do your own thing and check with whoever will make your > boards.Sounds good, but I've never been able to get a board house to even discuss these issues. They always take the approach that they will work with what I give them, which means, if it gives poor results, it's my problem. There are a number of board companies with published capabilities, but they all vary, enough that there seems to be no consensus. Just like your via pad size. I've never seen a board house that says that would be a standard board. I was just looking at one company who wants 10 mil annular ring on inner layers and 7 mil annular ring on surface layers. That's a huge difference from 2.25 mil. On the other hand, they will print 2.5 mil trace/space! I guess that's why we make prototypes. -- Rick C. ---- Get 1,000 miles of free Supercharging ---- Tesla referral code - https://ts.la/richard11209
Reply by ●January 14, 20232023-01-14
On 1/14/23 7:20 PM, gnuarm.del...@gmail.com wrote:> On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote:>> You should do your own thing and check with whoever will make your >> boards. > > Sounds good, but I've never been able to get a board house to even discuss these issues. They always take the approach that they will work with what I give them, which means, if it gives poor results, it's my problem. > > There are a number of board companies with published capabilities, but they all vary, enough that there seems to be no consensus. Just like your via pad size. I've never seen a board house that says that would be a standard board. I was just looking at one company who wants 10 mil annular ring on inner layers and 7 mil annular ring on surface layers. That's a huge difference from 2.25 mil. On the other hand, they will print 2.5 mil trace/space! > > I guess that's why we make prototypes. >IF you contact a "Good" board shop, they should be able to give you their specification to make the board. They may have several levels (of cost) with different requirements. If you board shop is NOT giving you a promise that the boards theya have built will be "successful", then I would not touch them. Yes, capabilities do vary a lot, so I always like to talk with my CMs about what shops they use for the sort of class board we are working on, and check with the shop on their requirements. We also keep a general idea of capabilities, so if one shop is a bit better on one spec, we might try not fully using that so other shops are likely able to handle it.
Reply by ●January 14, 20232023-01-14
On Sat, 14 Jan 2023 16:20:53 -0800 (PST), "gnuarm.del...@gmail.com" <gnuarm.deletethisbit@gmail.com> wrote:>On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote: >> On Sat, 14 Jan 2023 10:05:33 -0800 (PST), "gnuarm.del...@gmail.com" >> <gnuarm.del...@gmail.com> wrote: >> >> >On Saturday, January 14, 2023 at 12:08:03 PM UTC-4, John Larkin wrote: >> >> On Fri, 13 Jan 2023 21:20:50 -0800 (PST), "gnuarm.del...@gmail.com" >> >> <gnuarm.del...@gmail.com> wrote: >> >> >> >> >On Saturday, January 14, 2023 at 12:39:49 AM UTC-4, John Larkin wrote: >> >> >> On Sat, 7 Jan 2023 09:49:24 -0800 (PST), "gnuarm.del...@gmail.com" >> >> >> <gnuarm.del...@gmail.com> wrote: >> >> >> >> >> >> >A small board with a 100QFP is being redesigned for a new FPGA due to obsolescence. Gowin makes a 100QFP device that would be a good fit, but my customer has said "no" to the 100% Chinese brand... US government customers, ya know! >> >> >> > >> >> >> >So now I'm looking at a BGA. I don't want to get into fine PCB design rules, so 1.0 mm ball pitch is my preference. The only devices I can find that fit on the board have 196 or 256 pins. But the real problem is availability. >> >> >> > >> >> >> >Digikey has a few of the XC7S15-1FTGB196I and more a scheduled for delivery in April. Add in the various speed and temperature flavors trickling in (mostly in April) and I should be ok for the initial delivery in August... if I can get my hands on those. I don't know if Digikey factors in the backlog orders in these counts. >> >> >> > >> >> >> >Mouser shows great inventory of Efinix parts, particularly the T13 and T20 in a 0.8 mm 256 pin BGA, 10s of thousands in stock. But I'd rather work with a 1.0 mm BGA. Oddly enough, LCSC shows part numbers, but zero inventory. >> >> >> > >> >> >> >Anyone work with 0.8 mm BGAs? What PWB feature dimensions did you use? Did this impact the PWB cost? >> >> >> The 0.8 mm 256-ball T20 isn't bad... >> >> >> >> >> >> https://www.dropbox.com/s/xjqgj2pz9mdhtma/P941_FPGA.jpg?raw=1 >> >> > >> >> >I can't really see much detail. It looks like there are virtually no pads on the vias under the BGA. What size are they? >> >> The BGAVIAs are 12.5 mil OD with 8 mil drills. The other vias on the >> >> board STANDARDVIA and POWERVIA are bigger. >> > >> >2.25 mil (0.057 mm) is a pretty narrow via ring. Why not have a larger via pad? >> I don't know. My PCB guy decides stuff like that. I'd guess that he >> wanted it to pass some design rule check, or maybe he started metric. >> The board houses haven't complained as far as I know. >> >> You should do your own thing and check with whoever will make your >> boards. > >Sounds good, but I've never been able to get a board house to even discuss these issues. They always take the approach that they will work with what I give them, which means, if it gives poor results, it's my problem.We always specify bare-board testing and warpage and tolerances, so we don't get bad boards. What we can get is expensive boards.> >There are a number of board companies with published capabilities, but they all vary, enough that there seems to be no consensus. Just like your via pad size. I've never seen a board house that says that would be a standard board. I was just looking at one company who wants 10 mil annular ring on inner layers and 7 mil annular ring on surface layers. That's a huge difference from 2.25 mil. On the other hand, they will print 2.5 mil trace/space!Zero annular ring seems to be OK on inners. That reduces capacitance. 5 or even 4 mil traces are usually standard price. I don't know why my guy used 6 on the board that I posted. We do email our board houses and often they answer!> >I guess that's why we make prototypes.We don't prototype actual products; just go for it.
Reply by ●January 14, 20232023-01-14
On Saturday, January 14, 2023 at 8:39:59 PM UTC-4, Richard Damon wrote:> On 1/14/23 7:20 PM, gnuarm.del...@gmail.com wrote: > > On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote: > > >> You should do your own thing and check with whoever will make your > >> boards. > > > > Sounds good, but I've never been able to get a board house to even discuss these issues. They always take the approach that they will work with what I give them, which means, if it gives poor results, it's my problem. > > > > There are a number of board companies with published capabilities, but they all vary, enough that there seems to be no consensus. Just like your via pad size. I've never seen a board house that says that would be a standard board. I was just looking at one company who wants 10 mil annular ring on inner layers and 7 mil annular ring on surface layers. That's a huge difference from 2.25 mil. On the other hand, they will print 2.5 mil trace/space! > > > > I guess that's why we make prototypes. > > > IF you contact a "Good" board shop, they should be able to give you > their specification to make the board.Ok, which are the "good" ones? I assume you mean a place that makes the bare boards. A CM typically buys bare boards and assembles the parts. Like I said, they work with what you give them and will do their best. They expect the designers to do the design work.> They may have several levels (of cost) with different requirements. > > > If you board shop is NOT giving you a promise that the boards theya have > built will be "successful", then I would not touch them.They don't charge for boards that don't work, of course. But the parts that get thrown out are mine. I was a bit surprised to find out they actually expect to see 20% fall out because of parts not properly picked up. They get pulled off the tape, but if they are not aligned well enough, they get flung into space. One of the parts on my previous build had a $200 part on it. We lost some 40 or so, I don't recall the exact number. They recovered a few of them from various nooks and crannies. I've always been told about losses, but I thought that was mostly the tiny passives that don't matter. This was a pretty small, TSSOP-20.> Yes, capabilities do vary a lot, so I always like to talk with my CMs > about what shops they use for the sort of class board we are working on, > and check with the shop on their requirements. > > We also keep a general idea of capabilities, so if one shop is a bit > better on one spec, we might try not fully using that so other shops are > likely able to handle it.I got the "Award Letter" the other day and with the onerous conditions, I may not be able to accept the order. Two years ago, we went through this via a third party CM who did their integration. It took months to resolve the issues. They want prototypes in May. Ain't gonna happen. They want me to guarantee all manner of things that I can't guarantee, such as being able to manufacture the boards for 10 years. They want indemnifications for all manner of things. They even claim ownership of any "unpatented knowledge or information" that is disclosed to them is considered to be "part of the consideration for this Agreement". I believe this is what is called "trade secrets". This is far more onerous than what had been negotiated previously though their CM. Now, I have to start all over again. -- Rick C. ---+ Get 1,000 miles of free Supercharging ---+ Tesla referral code - https://ts.la/richard11209
Reply by ●January 14, 20232023-01-14
On Saturday, January 14, 2023 at 9:34:09 PM UTC-4, John Larkin wrote:> On Sat, 14 Jan 2023 16:20:53 -0800 (PST), "gnuarm.del...@gmail.com" > <gnuarm.del...@gmail.com> wrote: > > >On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote: > >> On Sat, 14 Jan 2023 10:05:33 -0800 (PST), "gnuarm.del...@gmail.com" > >> <gnuarm.del...@gmail.com> wrote: > >> > >> >On Saturday, January 14, 2023 at 12:08:03 PM UTC-4, John Larkin wrote: > >> >> On Fri, 13 Jan 2023 21:20:50 -0800 (PST), "gnuarm.del...@gmail.com" > >> >> <gnuarm.del...@gmail.com> wrote: > >> >> > >> >> >On Saturday, January 14, 2023 at 12:39:49 AM UTC-4, John Larkin wrote: > >> >> >> On Sat, 7 Jan 2023 09:49:24 -0800 (PST), "gnuarm.del...@gmail.com" > >> >> >> <gnuarm.del...@gmail.com> wrote: > >> >> >> > >> >> >> >A small board with a 100QFP is being redesigned for a new FPGA due to obsolescence. Gowin makes a 100QFP device that would be a good fit, but my customer has said "no" to the 100% Chinese brand... US government customers, ya know! > >> >> >> > > >> >> >> >So now I'm looking at a BGA. I don't want to get into fine PCB design rules, so 1.0 mm ball pitch is my preference. The only devices I can find that fit on the board have 196 or 256 pins. But the real problem is availability. > >> >> >> > > >> >> >> >Digikey has a few of the XC7S15-1FTGB196I and more a scheduled for delivery in April. Add in the various speed and temperature flavors trickling in (mostly in April) and I should be ok for the initial delivery in August... if I can get my hands on those. I don't know if Digikey factors in the backlog orders in these counts. > >> >> >> > > >> >> >> >Mouser shows great inventory of Efinix parts, particularly the T13 and T20 in a 0.8 mm 256 pin BGA, 10s of thousands in stock. But I'd rather work with a 1.0 mm BGA. Oddly enough, LCSC shows part numbers, but zero inventory. > >> >> >> > > >> >> >> >Anyone work with 0.8 mm BGAs? What PWB feature dimensions did you use? Did this impact the PWB cost? > >> >> >> The 0.8 mm 256-ball T20 isn't bad... > >> >> >> > >> >> >> https://www.dropbox.com/s/xjqgj2pz9mdhtma/P941_FPGA.jpg?raw=1 > >> >> > > >> >> >I can't really see much detail. It looks like there are virtually no pads on the vias under the BGA. What size are they? > >> >> The BGAVIAs are 12.5 mil OD with 8 mil drills. The other vias on the > >> >> board STANDARDVIA and POWERVIA are bigger. > >> > > >> >2.25 mil (0.057 mm) is a pretty narrow via ring. Why not have a larger via pad? > >> I don't know. My PCB guy decides stuff like that. I'd guess that he > >> wanted it to pass some design rule check, or maybe he started metric. > >> The board houses haven't complained as far as I know. > >> > >> You should do your own thing and check with whoever will make your > >> boards. > > > >Sounds good, but I've never been able to get a board house to even discuss these issues. They always take the approach that they will work with what I give them, which means, if it gives poor results, it's my problem. > We always specify bare-board testing and warpage and tolerances, so we > don't get bad boards. What we can get is expensive boards. > > > >There are a number of board companies with published capabilities, but they all vary, enough that there seems to be no consensus. Just like your via pad size. I've never seen a board house that says that would be a standard board. I was just looking at one company who wants 10 mil annular ring on inner layers and 7 mil annular ring on surface layers. That's a huge difference from 2.25 mil. On the other hand, they will print 2.5 mil trace/space! > Zero annular ring seems to be OK on inners. That reduces capacitance. > 5 or even 4 mil traces are usually standard price. I don't know why my > guy used 6 on the board that I posted. > > We do email our board houses and often they answer!You aren't paying attention. I don't use a PWB fabricator, I use a Contract Assembly house to assemble my boards. They are a middle man between me and the PWB fabricator. Once, no twice I've had to get on the phone with the actual PWB fabricator to convince him that he should not clip my silk screen. The use such a large clip radius that they made half the refdes illegible. That really makes the board hard to debug.> >I guess that's why we make prototypes. > We don't prototype actual products; just go for it.I don't have a choice, the customer wants 16 early protos, then 146 protos, then 100 pieces for pilot and 1100 FRS. So there is no "go for it". I'm sure I'll munge something up. Replacing the two main parts on the board and moving the connectors to the other side, means it's a complete redesign, at least for the artwork, if not the schematic. -- Rick C. --+- Get 1,000 miles of free Supercharging --+- Tesla referral code - https://ts.la/richard11209
Reply by ●January 14, 20232023-01-14
On 1/14/23 8:44 PM, gnuarm.del...@gmail.com wrote:> On Saturday, January 14, 2023 at 8:39:59 PM UTC-4, Richard Damon wrote: >> On 1/14/23 7:20 PM, gnuarm.del...@gmail.com wrote: >>> On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote: >> >>>> You should do your own thing and check with whoever will make your >>>> boards. >>> >>> Sounds good, but I've never been able to get a board house to even discuss these issues. They always take the approach that they will work with what I give them, which means, if it gives poor results, it's my problem. >>> >>> There are a number of board companies with published capabilities, but they all vary, enough that there seems to be no consensus. Just like your via pad size. I've never seen a board house that says that would be a standard board. I was just looking at one company who wants 10 mil annular ring on inner layers and 7 mil annular ring on surface layers. That's a huge difference from 2.25 mil. On the other hand, they will print 2.5 mil trace/space! >>> >>> I guess that's why we make prototypes. >>> >> IF you contact a "Good" board shop, they should be able to give you >> their specification to make the board. > > Ok, which are the "good" ones? I assume you mean a place that makes the bare boards. A CM typically buys bare boards and assembles the parts. Like I said, they work with what you give them and will do their best. They expect the designers to do the design work. > >There are a number of them, As I said, If you are using a CM, you will need to talk with them and find who they use and talk to them about their requirements.>> They may have several levels (of cost) with different requirements. >> >> >> If you board shop is NOT giving you a promise that the boards theya have >> built will be "successful", then I would not touch them. > > They don't charge for boards that don't work, of course. But the parts that get thrown out are mine. I was a bit surprised to find out they actually expect to see 20% fall out because of parts not properly picked up. They get pulled off the tape, but if they are not aligned well enough, they get flung into space. One of the parts on my previous build had a $200 part on it. We lost some 40 or so, I don't recall the exact number. They recovered a few of them from various nooks and crannies. I've always been told about losses, but I thought that was mostly the tiny passives that don't matter. This was a pretty small, TSSOP-20.So, you aren't using a good CM, as they aren't using a good board shop, or at least didn't give you their expected failure rates up front. Yes, there are loss factors for parts, but if they didn't give you those when you started to negotiate the contract when you indicated you will be supplying some of the parts, they aren't doing their job. Yes, it may be "cheaper" to use a shop like that, but you pay for it in those sorts of costs.> > >> Yes, capabilities do vary a lot, so I always like to talk with my CMs >> about what shops they use for the sort of class board we are working on, >> and check with the shop on their requirements. >> >> We also keep a general idea of capabilities, so if one shop is a bit >> better on one spec, we might try not fully using that so other shops are >> likely able to handle it. > > I got the "Award Letter" the other day and with the onerous conditions, I may not be able to accept the order. Two years ago, we went through this via a third party CM who did their integration. It took months to resolve the issues. They want prototypes in May. Ain't gonna happen. > > They want me to guarantee all manner of things that I can't guarantee, such as being able to manufacture the boards for 10 years. They want indemnifications for all manner of things. They even claim ownership of any "unpatented knowledge or information" that is disclosed to them is considered to be "part of the consideration for this Agreement". I believe this is what is called "trade secrets". > > This is far more onerous than what had been negotiated previously though their CM. Now, I have to start all over again. >Yes, some "customers" are not worth it.
Reply by ●January 15, 20232023-01-15
On Sat, 14 Jan 2023 17:52:45 -0800 (PST), "gnuarm.del...@gmail.com" <gnuarm.deletethisbit@gmail.com> wrote:>On Saturday, January 14, 2023 at 9:34:09 PM UTC-4, John Larkin wrote: >> On Sat, 14 Jan 2023 16:20:53 -0800 (PST), "gnuarm.del...@gmail.com" >> <gnuarm.del...@gmail.com> wrote: >> >> >On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote: >> >> On Sat, 14 Jan 2023 10:05:33 -0800 (PST), "gnuarm.del...@gmail.com" >> >> <gnuarm.del...@gmail.com> wrote: >> >> >> >> >On Saturday, January 14, 2023 at 12:08:03 PM UTC-4, John Larkin wrote: >> >> >> On Fri, 13 Jan 2023 21:20:50 -0800 (PST), "gnuarm.del...@gmail.com" >> >> >> <gnuarm.del...@gmail.com> wrote: >> >> >> >> >> >> >On Saturday, January 14, 2023 at 12:39:49 AM UTC-4, John Larkin wrote: >> >> >> >> On Sat, 7 Jan 2023 09:49:24 -0800 (PST), "gnuarm.del...@gmail.com" >> >> >> >> <gnuarm.del...@gmail.com> wrote: >> >> >> >> >> >> >> >> >A small board with a 100QFP is being redesigned for a new FPGA due to obsolescence. Gowin makes a 100QFP device that would be a good fit, but my customer has said "no" to the 100% Chinese brand... US government customers, ya know! >> >> >> >> > >> >> >> >> >So now I'm looking at a BGA. I don't want to get into fine PCB design rules, so 1.0 mm ball pitch is my preference. The only devices I can find that fit on the board have 196 or 256 pins. But the real problem is availability. >> >> >> >> > >> >> >> >> >Digikey has a few of the XC7S15-1FTGB196I and more a scheduled for delivery in April. Add in the various speed and temperature flavors trickling in (mostly in April) and I should be ok for the initial delivery in August... if I can get my hands on those. I don't know if Digikey factors in the backlog orders in these counts. >> >> >> >> > >> >> >> >> >Mouser shows great inventory of Efinix parts, particularly the T13 and T20 in a 0.8 mm 256 pin BGA, 10s of thousands in stock. But I'd rather work with a 1.0 mm BGA. Oddly enough, LCSC shows part numbers, but zero inventory. >> >> >> >> > >> >> >> >> >Anyone work with 0.8 mm BGAs? What PWB feature dimensions did you use? Did this impact the PWB cost? >> >> >> >> The 0.8 mm 256-ball T20 isn't bad... >> >> >> >> >> >> >> >> https://www.dropbox.com/s/xjqgj2pz9mdhtma/P941_FPGA.jpg?raw=1 >> >> >> > >> >> >> >I can't really see much detail. It looks like there are virtually no pads on the vias under the BGA. What size are they? >> >> >> The BGAVIAs are 12.5 mil OD with 8 mil drills. The other vias on the >> >> >> board STANDARDVIA and POWERVIA are bigger. >> >> > >> >> >2.25 mil (0.057 mm) is a pretty narrow via ring. Why not have a larger via pad? >> >> I don't know. My PCB guy decides stuff like that. I'd guess that he >> >> wanted it to pass some design rule check, or maybe he started metric. >> >> The board houses haven't complained as far as I know. >> >> >> >> You should do your own thing and check with whoever will make your >> >> boards. >> > >> >Sounds good, but I've never been able to get a board house to even discuss these issues. They always take the approach that they will work with what I give them, which means, if it gives poor results, it's my problem. >> We always specify bare-board testing and warpage and tolerances, so we >> don't get bad boards. What we can get is expensive boards. >> > >> >There are a number of board companies with published capabilities, but they all vary, enough that there seems to be no consensus. Just like your via pad size. I've never seen a board house that says that would be a standard board. I was just looking at one company who wants 10 mil annular ring on inner layers and 7 mil annular ring on surface layers. That's a huge difference from 2.25 mil. On the other hand, they will print 2.5 mil trace/space! >> Zero annular ring seems to be OK on inners. That reduces capacitance. >> 5 or even 4 mil traces are usually standard price. I don't know why my >> guy used 6 on the board that I posted. >> >> We do email our board houses and often they answer! > >You aren't paying attention.That's because you're obnoxious.
Reply by ●January 15, 20232023-01-15
On Sunday, January 15, 2023 at 12:10:54 AM UTC-4, John Larkin wrote:> On Sat, 14 Jan 2023 17:52:45 -0800 (PST), "gnuarm.del...@gmail.com" > <gnuarm.del...@gmail.com> wrote:=20 >=20 > >On Saturday, January 14, 2023 at 9:34:09 PM UTC-4, John Larkin wrote:=20 > >> On Sat, 14 Jan 2023 16:20:53 -0800 (PST), "gnuarm.del...@gmail.com"=20 > >> <gnuarm.del...@gmail.com> wrote:=20 > >>=20 > >> >On Saturday, January 14, 2023 at 3:17:24 PM UTC-4, John Larkin wrote:==20> >> >> On Sat, 14 Jan 2023 10:05:33 -0800 (PST), "gnuarm.del...@gmail.com"==20> >> >> <gnuarm.del...@gmail.com> wrote:=20 > >> >>=20 > >> >> >On Saturday, January 14, 2023 at 12:08:03 PM UTC-4, John Larkin wr=ote:=20> >> >> >> On Fri, 13 Jan 2023 21:20:50 -0800 (PST), "gnuarm.del...@gmail.c=om"=20> >> >> >> <gnuarm.del...@gmail.com> wrote:=20 > >> >> >>=20 > >> >> >> >On Saturday, January 14, 2023 at 12:39:49 AM UTC-4, John Larkin=wrote:=20> >> >> >> >> On Sat, 7 Jan 2023 09:49:24 -0800 (PST), "gnuarm.del...@gmail=.com"=20> >> >> >> >> <gnuarm.del...@gmail.com> wrote:=20 > >> >> >> >>=20 > >> >> >> >> >A small board with a 100QFP is being redesigned for a new FP=GA due to obsolescence. Gowin makes a 100QFP device that would be a good fi= t, but my customer has said "no" to the 100% Chinese brand... US government= customers, ya know!=20> >> >> >> >> >=20 > >> >> >> >> >So now I'm looking at a BGA. I don't want to get into fine P=CB design rules, so 1.0 mm ball pitch is my preference. The only devices I = can find that fit on the board have 196 or 256 pins. But the real problem i= s availability.=20> >> >> >> >> >=20 > >> >> >> >> >Digikey has a few of the XC7S15-1FTGB196I and more a schedul=ed for delivery in April. Add in the various speed and temperature flavors = trickling in (mostly in April) and I should be ok for the initial delivery = in August... if I can get my hands on those. I don't know if Digikey factor= s in the backlog orders in these counts.=20> >> >> >> >> >=20 > >> >> >> >> >Mouser shows great inventory of Efinix parts, particularly t=he T13 and T20 in a 0.8 mm 256 pin BGA, 10s of thousands in stock. But I'd = rather work with a 1.0 mm BGA. Oddly enough, LCSC shows part numbers, but z= ero inventory.=20> >> >> >> >> >=20 > >> >> >> >> >Anyone work with 0.8 mm BGAs? What PWB feature dimensions di=d you use? Did this impact the PWB cost?=20> >> >> >> >> The 0.8 mm 256-ball T20 isn't bad...=20 > >> >> >> >>=20 > >> >> >> >> https://www.dropbox.com/s/xjqgj2pz9mdhtma/P941_FPGA.jpg?raw==3D1=20> >> >> >> >=20 > >> >> >> >I can't really see much detail. It looks like there are virtual=ly no pads on the vias under the BGA. What size are they?=20> >> >> >> The BGAVIAs are 12.5 mil OD with 8 mil drills. The other vias on=the=20> >> >> >> board STANDARDVIA and POWERVIA are bigger.=20 > >> >> >=20 > >> >> >2.25 mil (0.057 mm) is a pretty narrow via ring. Why not have a la=rger via pad?=20> >> >> I don't know. My PCB guy decides stuff like that. I'd guess that he==20> >> >> wanted it to pass some design rule check, or maybe he started metri=c.=20> >> >> The board houses haven't complained as far as I know.=20 > >> >>=20 > >> >> You should do your own thing and check with whoever will make your==20> >> >> boards.=20 > >> >=20 > >> >Sounds good, but I've never been able to get a board house to even di=scuss these issues. They always take the approach that they will work with = what I give them, which means, if it gives poor results, it's my problem.= =20> >> We always specify bare-board testing and warpage and tolerances, so we==20> >> don't get bad boards. What we can get is expensive boards.=20 > >> >=20 > >> >There are a number of board companies with published capabilities, bu=t they all vary, enough that there seems to be no consensus. Just like your= via pad size. I've never seen a board house that says that would be a stan= dard board. I was just looking at one company who wants 10 mil annular ring= on inner layers and 7 mil annular ring on surface layers. That's a huge di= fference from 2.25 mil. On the other hand, they will print 2.5 mil trace/sp= ace!=20> >> Zero annular ring seems to be OK on inners. That reduces capacitance.==20> >> 5 or even 4 mil traces are usually standard price. I don't know why my==20> >> guy used 6 on the board that I posted.=20 > >>=20 > >> We do email our board houses and often they answer!=20 > >=20 > >You aren't paying attention. > That's because you're obnoxious.Wow! Talk about sensitive. What is going on with you??? --=20 Rick C. --++ Get 1,000 miles of free Supercharging --++ Tesla referral code - https://ts.la/richard11209
