Home |
Search |
Today's Posts |
#1
![]() |
|||
|
|||
![]()
Hi,
I would like to build a PCB board with fine pitch (LQFP32) and very small surface mount parts (1206 and smaller) parts. The board is a simple 24 bit A/D and D/A converter with a full speed (12 Mbs) USB2 interface; although I would prefer the USB2 high speed interface (480 Mbs). The temptation is too hard to resist using today's microprocessors such as SiLabs C8051F350 (basically an 8051 cpu). I know that a commercial 4 layer PCB with a solder mask would probably work for homebrew construction; but the board cost is high. So what is the chance of a double sided PCB without solder mask working with very small parts? The process I would use in either case would be: 1. Place solder paste on all pads. Typically, when heated the solder will vacate between pins and will suck up, like a magnet, under the SMD pin. 2. Place glue under parts that will possible move when the board is handled. 3. Heat the PCB in a toaster oven until soldered. Manually follow the heat up and cool down time curves for type of solder paste used. Comments/suggestions are appreciated. |
#2
![]() |
|||
|
|||
![]()
On Jun 30, 1:12*pm, "numeric" wrote:
Hi, I would like to build a PCB board with fine pitch (LQFP32) and very small surface mount parts (1206 and smaller) parts. The board is a simple 24 bit A/D and D/A converter with a full speed (12 Mbs) USB2 interface; although I would prefer the USB2 high speed interface (480 Mbs). The temptation is too hard to resist using today's microprocessors such as SiLabs C8051F350 (basically an 8051 cpu). I know that a commercial 4 layer PCB with a solder mask would probably work for homebrew construction; but the board cost is high. So what is the chance of a double sided PCB without solder mask working with very small parts? The process I would use in either case would be: 1. Place solder paste on all pads. Typically, when heated the solder will vacate between pins and will suck up, like a magnet, under the SMD pin. 2. Place glue under parts that will possible move when the board is handled. 3. Heat the PCB in a toaster oven until soldered. Manually follow the heat up and cool down time curves for type of solder paste used. Comments/suggestions are appreciated. LQFP32 is not ridiculously fine pitch (I do 0.65mm pitch parts by hand all the time), and 1206 is just enormous in the SMT world. I do both by hand all the time - just a Weller WTCPT with 0.015" solder. No solder mask for me (I use ExpressPCB cheapo boards) but pretinning does help. Sometimes flux comes in handy but is not really necessary. Just lay the solder down across the leads, melt on with the soldering iron, go back and clean up any blobs with solder-wick. Toaster oven is great for BGA's but if there's only a few leaded SMT parts it seems like way overkill. Tim N3QE |
#3
![]() |
|||
|
|||
![]()
On 30 Jun, 18:12, "numeric" wrote:
Hi, I would like to build a PCB board with fine pitch (LQFP32) and very small surface mount parts (1206 and smaller) parts. The board is a simple 24 bit A/D and D/A converter with a full speed (12 Mbs) USB2 interface; although I would prefer the USB2 high speed interface (480 Mbs). The temptation is too hard to resist using today's microprocessors such as SiLabs C8051F350 (basically an 8051 cpu). I know that a commercial 4 layer PCB with a solder mask would probably work for homebrew construction; but the board cost is high. So what is the chance of a double sided PCB without solder mask working with very small parts? The process I would use in either case would be: 1. Place solder paste on all pads. Typically, when heated the solder will vacate between pins and will suck up, like a magnet, under the SMD pin. 2. Place glue under parts that will possible move when the board is handled. 3. Heat the PCB in a toaster oven until soldered. Manually follow the heat up and cool down time curves for type of solder paste used. Comments/suggestions are appreciated. I don't have any problems making PCBs at home for 0.5 mm lead spacing QFPs, or soldering them. by hand (drag soldering). Leon |
#4
![]() |
|||
|
|||
![]() "Tim Shoppa" wrote in message ... LQFP32 is not ridiculously fine pitch (I do 0.65mm pitch parts by hand all the time), and 1206 is just enormous in the SMT world. I do both by hand all the time - just a Weller WTCPT with 0.015" solder. No solder mask for me (I use ExpressPCB cheapo boards) but pretinning does help. Sometimes flux comes in handy but is not really necessary. Just lay the solder down across the leads, melt on with the soldering iron, go back and clean up any blobs with solder-wick. Toaster oven is great for BGA's but if there's only a few leaded SMT parts it seems like way overkill. Tim N3QE Thanks for the reply. You are correct, a 1206 is a large SMD component and I will try to use them wherever possible. However, smaller (0402 and larger) may also be used. It is encouraging to hear about successful fine pitch results with hand solder techniques. In addition, it will certainly be useful for parts placed on both sides of the PCB. |
#5
![]() |
|||
|
|||
![]() "Leon" wrote in message ... I don't have any problems making PCBs at home for 0.5 mm lead spacing QFPs, or soldering them. by hand (drag soldering). Leon Thanks for the reply. I have though about using a laser printer to print a negative image on clear Mylar. Then expose a photo sensitive PCB to the negative image. The laser printer does print very sharp lines as visible using a 10X loop. The quality should be good enough for fine pitch. The etching process however may end up in disaster with fine pitch lines. How do you make the PCB? |
#6
![]() |
|||
|
|||
![]()
On Jun 30, 4:33*pm, "numeric" wrote:
You are correct, a 1206 is a large SMD component and I will try to use them wherever possible. However, smaller (0402 and larger) may also be used. It is encouraging to hear about successful fine pitch results with hand solder techniques. In addition, it will certainly be useful for parts placed on both sides of the PCB. For my purposes, 0603 is the "standard" size SMD part to use when I have a choice. Of course power parts are gonna be bigger. I mean, just so you don't scare yourself, look at the bottom of a TO-92 transistor that you've been soldering for 30 years or more. The pitch there 0.05 inches or 0.127 mm. And you do those all the time. Your LQFP is finer pitch, but not ridiculously so. So don't set yourself up with the mental attitude that it's gonna be hard, because it isn't going to be hard! I have occasionally used 0402 and even 0201 parts when they weren't available in 0603... I struggled with 0201, and 0402 required a lot of concentration. By comparison the 0.65mm pitch parts were a piece of cake. And finally, don't ignore the option of doing "dead bug" with SMT parts. Very fun for low-lead-count analog parts... and possible though questionable for digital stuff with lotsa leads. Tim. |
#7
![]() |
|||
|
|||
![]()
On 30 Jun, 21:45, "numeric" wrote:
"Leon" wrote in message ... I don't have any problems making PCBs at home for 0.5 mm lead spacing QFPs, or soldering them. by hand (drag soldering). Leon Thanks for the reply. I have though about using a laser printer to print a negative image on clear Mylar. Then expose a photo sensitive PCB to the negative image. The laser printer does print very sharp lines as visible using a 10X loop. The quality should be good enough for fine pitch. The etching process however may end up in disaster with fine pitch lines. How do you make the PCB? I use an inkjet printer and JetStar Premium film for the positive artwork. UV exposure on Fotoboard 2 with pre-coated resist and etching in ferric chloride. Leon |
#8
![]() |
|||
|
|||
![]()
On Jun 30, 10:12 am, "numeric" wrote:
Hi, I would like to build a PCB board with fine pitch (LQFP32) and very small surface mount parts (1206 and smaller) parts. The board is a simple 24 bit A/D and D/A converter with a full speed (12 Mbs) USB2 interface; although I would prefer the USB2 high speed interface (480 Mbs). The temptation is too hard to resist using today's microprocessors such as SiLabs C8051F350 (basically an 8051 cpu). I know that a commercial 4 layer PCB with a solder mask would probably work for homebrew construction; but the board cost is high. So what is the chance of a double sided PCB without solder mask working with very small parts? The process I would use in either case would be: 1. Place solder paste on all pads. Typically, when heated the solder will vacate between pins and will suck up, like a magnet, under the SMD pin. 2. Place glue under parts that will possible move when the board is handled. 3. Heat the PCB in a toaster oven until soldered. Manually follow the heat up and cool down time curves for type of solder paste used. Comments/suggestions are appreciated. I agree with Leon and Tim that making a board and mounting parts on it shouldn't be a problem, though I do greatly appreciate being able to do my soldering under a stereo microscope using a Metcal iron. I commonly use boards coated with positive resist (from Circuit Specialists in the Phoenix area), etched by floating them on top of ferric chloride. I find that (lacking a spray etcher) the float process maintains fine line resolution--I have a theory that it has to do with copper-laden etchant being heavier and falling away from the etch site, but whatever, it works well for me. Surface tension will hold fairly large boards up, but you can also just put support at the corners and let the board submerge. Resolution for me is always limited by the quality of the transparency. Thanks to Leon for posting a reference to another film material to try. I've had the best luck using an Epson inkjet printer: my HP printer doesn't do as well, and laser printers typically don't print a dense enough black. Pinholes WILL print on the board! I use a back-light film setting on the Epson for highest density. An alternative is to get a photoplotting shop to do the film for you: better quality but it lacks the immediacy of being able go from layout to board in minutes. Anyway, the best inkjet printed transparencies will allow me to do 5 mil lines with 5 mil spaces if I'm careful, and 8/8 pretty easily. Most of my work is with 0805 as the standard passive, since that's what I invested in an assortment of back a few years ago, but I also do 0603 with no trouble, and lead pitches down to 0.5mm. We have fancy equipment to do BGA rework here at work, and the trained techs have marginal success with it at best (for large parts), so I've avoided BGAs in my home projects. By the way, if you get boards from Circuit Specialists, be aware that what they call 1 ounce copper is really 0.5 ounce...seldom a problem and allows faster etching and it's easier to make fine lines, but be aware of it for high current traces. Cheers, Tom |
#9
![]() |
|||
|
|||
![]()
On 1 Jul, 00:30, K7ITM wrote:
On Jun 30, 10:12 am, "numeric" wrote: Hi, I would like to build a PCB board with fine pitch (LQFP32) and very small surface mount parts (1206 and smaller) parts. The board is a simple 24 bit A/D and D/A converter with a full speed (12 Mbs) USB2 interface; although I would prefer the USB2 high speed interface (480 Mbs). The temptation is too hard to resist using today's microprocessors such as SiLabs C8051F350 (basically an 8051 cpu). I know that a commercial 4 layer PCB with a solder mask would probably work for homebrew construction; but the board cost is high. So what is the chance of a double sided PCB without solder mask working with very small parts? The process I would use in either case would be: 1. Place solder paste on all pads. Typically, when heated the solder will vacate between pins and will suck up, like a magnet, under the SMD pin. 2. Place glue under parts that will possible move when the board is handled. 3. Heat the PCB in a toaster oven until soldered. Manually follow the heat up and cool down time curves for type of solder paste used. Comments/suggestions are appreciated. I agree with Leon and Tim that making a board and mounting parts on it shouldn't be a problem, though I do greatly appreciate being able to do my soldering under a stereo microscope using a Metcal iron. I commonly use boards coated with positive resist (from Circuit Specialists in the Phoenix area), etched by floating them on top of ferric chloride. *I find that (lacking a spray etcher) the float process maintains fine line resolution--I have a theory that it has to do with copper-laden etchant being heavier and falling away from the etch site, but whatever, it works well for me. *Surface tension will hold fairly large boards up, but you can also just put support at the corners and let the board submerge. *Resolution for me is always limited by the quality of the transparency. *Thanks to Leon for posting a reference to another film material to try. *I've had the best luck using an Epson inkjet printer: *my HP printer doesn't do as well, and laser printers typically don't print a dense enough black. Pinholes WILL print on the board! *I use a back-light film setting on the Epson for highest density. *An alternative is to get a photoplotting shop to do the film for you: *better quality but it lacks the immediacy of being able go from layout to board in minutes. Anyway, the best inkjet printed transparencies will allow me to do 5 mil lines with 5 mil spaces if I'm careful, and 8/8 pretty easily. Most of my work is with 0805 as the standard passive, since that's what I invested in an assortment of back a few years ago, but I also do 0603 with no trouble, and lead pitches down to 0.5mm. *We have fancy equipment to do BGA rework here at work, and the trained techs have marginal success with it at best (for large parts), so I've avoided BGAs in my home projects. By the way, if you get boards from Circuit Specialists, be aware that what they call 1 ounce copper is really 0.5 ounce...seldom a problem and allows faster etching and it's easier to make fine lines, but be aware of it for high current traces. Cheers, Tom I use a Metcal and microscope, as well. I use manual agitation in hot FeCl3. I don't have any problems getting down to 8 mil tracks or even less. Leon |
#10
![]() |
|||
|
|||
![]()
On 1 Jul, 00:30, K7ITM wrote:
On Jun 30, 10:12 am, "numeric" wrote: . *Thanks to Leon for posting a reference to another film material to try. *I've had the best luck using an Epson inkjet printer: *my HP printer doesn't do as well, The ordinary JetStar film is suitable for dye-based ink like that used in Epson printers. Premium Jetstar is intended for use with pigment- based inks used with HP printers. They are available from Mega Electronics and Farnell. Leon |
Reply |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
![]() |
||||
Thread | Forum | |||
home made Mobile Ant | Antenna | |||
home made pcb | Homebrew | |||
Home made antenna?? | Boatanchors | |||
Home Made Tower | Antenna | |||
Home made antennas | Scanner |