Post by thread:[EE] Making Your Own Fine-Pitch PCBs

So I've done quite a bit of research on making your own PCBs. But I do have a few questions... Here's what I found out: The concept behind all this is to start out with a PCB completely covered in a thin sheet of copper (can be bought for low prices on ebay). The idea is to use an etching solution such as FeCl to eat away at the copper where you DON'T want the traces. Thus, one must "protect" the area that they want the traces with an etchant resistant material (such as permanent marker pen, printer ink/toner and other stuff). The method is as follows: 1. Materials: - Blank PCBs - Etching solution such as ferric chloride (FeCl) - Resistant transfer medium (glossy paper or specially made paper for this purpose) - Design software and a printer to print out your design to transfer onto the blank PCB. 2. Design your PCB and print it out on glossy photo paper using a laser printer. 3. Transfer the design from the glossy paper by pressing the glossy paper against the blank pcb with a hot clothes iron. 4. Prepare the FeCl etching solution. A warmer solution yields faster and more accurate results. 5. Soak the PCB that has your design on it into the FeCl solution. 6. Take it out after the copper has been eaten away and wash thoroughly in soap and water. Gently scrape off any toner ink still on the PCB. Complete. My questions: 1. What kind of paper works best for this? Glossy paper? What brand? Specially designed paper? What are your experiences? 2. What printer should be used? Laser or inkjet? What are your experiences? 3. When printing designs for fine pitch TQFP components such as the 100 pin PIC32 chip which has a pin pitch of 0.4 mm, where does the problem lie - during the ink transfer stage or during the actual etching stage? How can this be done more reliably so that no traces are broken and everything works out well? 4. When making double sided PCBs, does the same technique apply (other than pressing the glossy paper on both sides on a double sided blank PCB)? 5. Is there anything wrong/incomplete in my method above? Do you guys think it'll work? Does anyone have any advice? 6. Is it possible to make plated through-holes on your PCB practically? If so, how? I found some links on this but I don't really get it: http://www.thinktink.com/stack/volumes/VOLVI/copplate.htm http://www.diyaudio.com/forums/showthread/t-45235.html <= This guy got great results. I really don't want to waste $50 getting my PCB professionally printed (just for one PCB). Making your own seems so much more fun and cost effective :) -- ..::[ solarwind ]::..

On Dec 9, 2008, at 8:18 PM, solarwind wrote: > > I really don't want to waste $50 getting my PCB professionally printed > (just for one PCB). Making your own seems so much more fun and cost > effective :) Years ago, I made simple PCBs with rub-on resist transfers (are these still available?), masking tape, X-acto knife, and a Sharpie pen. For dips on 100mil centers and T0-5 transistors, this was fairly efficient for a one off board. The hardest part was hand drilling the board without busting up the #60 bits. Using this method when I was a teenager, a drill press wasn't in the budget. SOICs have a 65mil spacing. Much harder to do with a Sharpie. A trace that fine with only a permanent marker for resist is likely to be eaten by the etchant. Also, it is not very fun to work with SMD components unless your finished board is masked. Things tend to float around too much. Lastly, while silk screening is not mandatory, it sure makes assembly more pleasant. Honestly, for most one-off hobby projects I still try to use through- hole components. If through-hole parts are not available, for a one- off project, I find it easier to use Veroboard, Kapton tape, and superglue. The SOIC chips are superglued to the Kapton, which is cut slightly larger than the desired chip and placed in the desired location. #30 Kynar wire then can route from the chip pins to the appropriate Veroboard traces. 0805 components fit beautifully between the traces, and SOT-23 components can be staggered. Kapton is heat resistant to about 200c and has insulation qualities reaching several hundred volts. Wonderful stuff to prototype with. Veroboard aka Stripboard is fairly new to me. I've been using it for only about the past 5 years or so. For the life of me, I have no idea why this wonderful stuff never propogated across the Atlantic to the American shores. American hobbyists don't know what they're missing! A good friend of mine in the UK keeps me supplied, else I buy from a UK vendor on Ebay. For those of you in Europe, here in the States we otherwise use standard 0.1 perfboard (no copper) or 0.1 perfboard with copper donuts at each hole. Stripboard is essentially unknown, and to my knowledge not carried as a standard item by any American vendor. Joe

> > I really don't want to waste $50 getting my PCB professionally printed> (just for one PCB). Making your own seems so much more fun and cost> effective :)> -- > ..::[ solarwind ]::..> -- If you value your time at at least $10/hr (and even a hobbiest on their first project should be higher than that), $50 for a circuit board is nothing. The time spend screwing around with making a board yourself is better spend thinking through the design and learning to use a layout tool like Eagle (free if you can fit in the requirements, cheap otherwise) or 123PCB.com. And there are is a GPL released schematic capture and layout but the name escapes me at the moment, might be something like gEDA? Then think long and hard on how the circuit should work and have a board made. Othewise, stick to 0.1" (100mil spacing for those of you confused by the units) and chips available as DIP parts and 1/4W resistors, etc. Or Veroboard (spelling?) or 0.1" punched copper clad and a Dremmel tool. Or even "dead bug" style prototying. But etching a PCB when you don't have the ability to do through-hole plating and can't guarantee anything better than a 10/10 or 12/12 rule (and if you don't know what that is, Google is your friend) have it made professionally. AP Circuits and lots of others out there can just plain do a better job that you can fooling around with this stuff in the kitchen sink. Rob

On Dec 9, 2008, at 7:18 PM, solarwind wrote: > So I've done quite a bit of research on making your own PCBs. But I do > have a few questions... > > The method is as follows: [description of "toner transfer" method] > My questions: > > 1. What kind of paper works best for this? Glossy paper? What brand? > Specially designed paper? What are your experiences? It may depend on your printer. Probably specially designed paper works best, but it's expensive and harder to get. I've used glossy magazine paper (free!), which has worked ok for the NOT fine-pitch boards I've made using it. > 2. What printer should be used? Laser or inkjet? What are your > experiences? For this method, you MUST use a laser printer, because you're counting on the melting and adhesion properties of the laser toner. The people talking about inkjets are using a different (photographic) process that is better for finer pitches, but more expensive and somewhat more complex (see below.) > 3. When printing designs for fine pitch TQFP component ... where > does the > problem lie - during the ink transfer stage or during the actual > etching stage? printing and ink transfer mostly, I think. > How can this be done more reliably so that no traces > are broken and everything works out well? There are lots of theories... Most people give up at some level of complexity, because it has become increasingly easy to get a PCB easily made. (for instance, there are specialized heated lamination presses that replace the "clothes iron", and only cost, oh, 2x the cost of a single professionally made board...) > 4. When making double sided PCBs, does the same technique apply You add registration issues (making sure that the two sides line up exactly), and the lack of plated-through holes for getting signals from one side to the other that I mentioned back in my first message. > 5. Is there anything wrong/incomplete in my method above? Do you guys > think it'll work? Does anyone have any advice? I think the chances of someone without any experience being able to successfully create a toner transfer board with .4mm pitch packages (that's about .2mm lines and spaces, right?) the first time they try are almost zero. > 6. Is it possible to make plated through-holes on your PCB > practically? No. > http://www.thinktink.com/stack/volumes/VOLVI/copplate.htm Thinktink is "semi-pro", not amateur... > http://www.diyaudio.com/forums/showthread/t-45235.html <= This guy > got great results. He didn't plate his holes, did he? He made a double sided board where the top is mostly ground plane, and soldered some components on both sides. You can make changes to a PCB design so that the only places that carry signals from one side to another happen on holes where you CAN solder on both sides (and this works BETTER when most of your components are SMT, so most of these connections are vias), but it's generally a real pain. The photographic method for making PCBs is very similar to toner transfer, except that you start with a PCB blank treated with a photosensitive resist, which is then exposed to a light source (usually UV) and "developed", which removes part of the resist. After that the board is etched and drilled as normal; since the resist is applied professionally and "intimately" to the copper, it is more reliable that trying to transfer a resist pattern to the copper via amateur techniques. > I really don't want to waste $50 getting my PCB professionally [made] You'll spend that much on materials to make your home-made PCB too, though you'll end up with more than you need for ONE pcb... In the end, I suppose it depends on whether you want to make "PCB fabrication" part of your hobby or not... BillW

These guys have tons of information and mis-information. > http://groups.yahoo.com/group/Homebrew_PCBs Do it if you think it is fun to do and your time is valued at $2/hr*. Otherwise, look for a PCB house. MA > My questions: > > 1. What kind of paper works best for this? Glossy paper? What brand? > Specially designed paper? What are your experiences? I used gel coated paper ; the surface dissolved in water. The stuff I used is no longer sold. Someone sells "security paper" which dissolves in water. Hot water dissolves stuff faster than cold water. > > 2. What printer should be used? Laser or inkjet? What are your > experiences? Laser worked for me. Laser printers vary radically in quality. > > > 3. When printing designs for fine pitch TQFP components such as the > 100 pin PIC32 chip which has a pin pitch of 0.4 mm, where does the > problem lie - during the ink transfer stage or during the actual > etching stage? Both phases > How can this be done more reliably so that no traces > are broken and everything works out well? Cold solution etches slower than hot solution. Thin traces can disappear from being etched away from the side of the trace. > > > 4. When making double sided PCBs, does the same technique apply (other > than pressing the glossy paper on both sides on a double sided blank > PCB)? Use alignment holes > > > 5. Is there anything wrong/incomplete in my method above? Do you guys > think it'll work? Does anyone have any advice? Do it if you think it is fun to do and your time is valued at $2/hr*. Otherwise, look for a PCB house. {Quote hidden}> > > 6. Is it possible to make plated through-holes on your PCB > practically? If so, how? I found some links on this but I don't really > get it: > > http://www.thinktink.com/stack/volumes/VOLVI/copplate.htm > > http://www.diyaudio.com/forums/showthread/t-45235.html <= This guy > got great results. > > I really don't want to waste $50 getting my PCB professionally printed > (just for one PCB). Making your own seems so much more fun and cost > effective :) > -- > ..::[ solarwind ]::.. > -- * amortize the time spent getting it to work correctly. WFT Electronics Denver, CO 720 222 1309 " dent the UNIVERSE " All ideas, text, drawings and audio , that are originated by WFT Electronics ( and it's principals ), that are included with this signature text are to be deemed to be released to the public domain as of the date of this communication .

My real life experiences all have to do with through hole, and not really fine pitch. I've used two methods. The first is direct-resist transfers, and pens. In short, you draw your circuit directly on the circuit board using a combination of sharpies and direct-etch (rub-on) patterns. The direct-etch transfers seem to work better than the sharpies, but I had really good results using a combination, as long as the board is clean, the sharpie lines are dark, and the direct-etch transfers are actually rubbed down until flat. I might be tempted to try something similar if I was doing fine pitch surface mount. The second is positive photo-resist, using presensitized boards from circuit specialists (remember, I told you to bookmark them, you'd thank me later). Basically you print your design on a piece of overhead transparency film and then expose the board through the film. You can expose it using any fluorescent light source. Some experimentation will be required as to the right amount, but my recollection is that it isn't all that critical. Once it is exposed, you develop the resist using another chemical (also available from circuit specialists), so you now have hardened resist where the board wasn't exposed to light. The transfer method I've never actually tried, but I have heard that it is one of those methods which is quite hard to get right, but if you can make the process work (which means the right paper, toner, heat transfer method, and the like), it works well as well. Once you have a resist on the board you etch. I have used Ferric Chloride and Ammonium Persulphate. I am most partial to the dry Ammonium Persulfate since it is generally cheaper once mixed with water. Plus it is clear, and generally cleaner overall. Other people prefer the pre-mixed ferric chloride. I believe it does work a bit faster, but it is also nasty brown which likes to stain things. The one thing I can't recommend in good faith is dry Ferric Chloride. It is very Hydroscopic, so it can and will react violently when mixed with water. Others have had better luck, but the stuff scares me. I'd recommend trying both. In regards to a few specifics. solarwind wrote: > 1. What kind of paper works best for this? Glossy paper? What brand? > Specially designed paper? What are your experiences? > I have read all of the above. > 2. What printer should be used? Laser or inkjet? What are your experiences? > Always Laser, unless you are doing the photoresist thing which doesn't actually do a transfer. To do "toner transfer" you must, of course, have toner, which means a Laser, or a photocopied inkjet sheet. And I understand that the density of the toner is one of the biggest factor. > 3. When printing designs for fine pitch TQFP components such as the > 100 pin PIC32 chip which has a pin pitch of 0.4 mm, where does the > problem lie - during the ink transfer stage or during the actual > etching stage? How can this be done more reliably so that no traces > are broken and everything works out well? > It is always the resist, with the exception of over/under etching. That is, as long as you leave the board in the etchant just long enough that all of the unwanted copper is removed, and it doesn't have a chance to eat underneath the resist, then the limiting factor is the resist itself. Also, agitation or other fluid "circulation" is important so all the copper dissapears at nearly the same time, so that the board is consistent across it's surface That said, I'll be really interested to hear how you make out with a 100 pin PIC32. > 4. When making double sided PCBs, does the same technique apply (other > than pressing the glossy paper on both sides on a double sided blank > PCB)? > Yes, with the caveat that I have no direct-transfer experience. But for photo-resist, this is the case. Alignment is of course critical. Drilling with an appropriate backer board so you don't pop out the back of the board is also critical and lots of fun - since you are likely to lift a trace. > 5. Is there anything wrong/incomplete in my method above? Do you guys > think it'll work? Does anyone have any advice? > I'm skeptical about the 0.4mm pitch. 0.4mm is actually 15.7 mils. Assuming 50/50 pad/gap (the best case resolution wise - not necessarily footprint wise) you are talking at a minimum of 8mil resolution. Many professional board houses charge extra below 8mil, because of the extra care that is needed, so you are right at the edge of what professionals can do with "normal care". Whether you can do this with hobbiest grade materials will be interesting to see. > 6. Is it possible to make plated through-holes on your PCB > practically? I woudn't mess with it... just make sure any via's are either with a through-hole component that you can solder top/bottom to, or where you can use a wire instead. > I really don't want to waste $50 getting my PCB professionally printed > (just for one PCB). Making your own seems so much more fun and cost > effective :) > How big of one you are talking about... You can get fairly large ones for $33 from here http://www.4pcb.com/index.php?load=content&page_id=134 , assuming you qualify for the student discount. Plus if you have a smaller board, you can get them for $10+2.5 sq.in from batchpcb.com (a 4 sq inch board would only be $20 - which is big for SMD). -forrest

> I'm skeptical about the 0.4mm pitch. 0.4mm is actually 15.7 mils. > Assuming 50/50 pad/gap (the best case resolution wise - not necessarily > footprint wise) you are talking at a minimum of 8mil resolution. Many > professional board houses charge extra below 8mil, because of the extra > care that is needed, so you are right at the edge of what professionals > can do with "normal care". Whether you can do this with hobbiest grade > materials will be interesting to see. It's great that this should come up now, as tonight I completed a successful toner transfer etch (and subsequent reflow) of 6 boards each containing a 4x5 pad BGA with 0.4mm pitch. Some of the traces were 6 mil and the yield was 100%. Toner transfer is definitely doable for prototype work and there's nothing like same-day design turnaround. In the past I have used toner transfer to complete double sided 12x12 inch boards that would have been rather costly to have made due to the dimensions. So it's doable and once you get the hang of it, a little fun. Don't waste your time with dissolving paper transfer, use the slick backing of mailing labels and scuff/rub the waxy surface with a paper towel so the toner sticks. This peels right off instantly after you've pressd it to the board and you're ready to etch. I use one of the ubiquitous HP LaserJet 4M Plus printers with EconoMode and RET off. Can't get more basic/inexpensive than that! -Andrew

On Wed, Dec 10, 2008 at 1:02 AM, A K <.....00KILLspam@spam@case.edu> wrote: > It's great that this should come up now, as tonight I completed a > successful toner transfer etch (and subsequent reflow) of 6 boards each > containing a 4x5 pad BGA with 0.4mm pitch. Some of the traces were 6 > mil and the yield was 100%. Toner transfer is definitely doable for > prototype work and there's nothing like same-day design turnaround. In > the past I have used toner transfer to complete double sided 12x12 inch > boards that would have been rather costly to have made due to the > dimensions. So it's doable and once you get the hang of it, a little > fun. Don't waste your time with dissolving paper transfer, use the > slick backing of mailing labels and scuff/rub the waxy surface with a > paper towel so the toner sticks. This peels right off instantly after > you've pressd it to the board and you're ready to etch. I use one of > the ubiquitous HP LaserJet 4M Plus printers with EconoMode and RET off. > Can't get more basic/inexpensive than that! > -Andrew Can you please post a picture of what you made? I'm interested in the fine pitch part of it. -- ..::[ solarwind ]::..

"A K" wrote: > It's great that this should come up now, as tonight I completed a > successful toner transfer etch (and subsequent reflow) of 6 boards each > containing a 4x5 pad BGA with 0.4mm pitch. Some of the traces were 6 > mil and the yield was 100%. Toner transfer is definitely doable for > prototype work and there's nothing like same-day design turnaround. In > the past I have used toner transfer to complete double sided 12x12 inch > boards that would have been rather costly to have made due to the > dimensions. So it's doable and once you get the hang of it, a little > fun. Don't waste your time with dissolving paper transfer, use the > slick backing of mailing labels and scuff/rub the waxy surface with a > paper towel so the toner sticks. This peels right off instantly after > you've pressd it to the board and you're ready to etch. I use one of > the ubiquitous HP LaserJet 4M Plus printers with EconoMode and RET off. > Can't get more basic/inexpensive than that! Andrew, if you don't mind I'd like to clarify a few things because this is something I can definitely make use of (the instant gratification aspect is very appealing). - Which mailing labels are you talking about? Avery, FedEx? - I'm assuming you're printing on the side that the paper used to be on, after you peel the paper off? - How much do you have to rub the waxy surface in order for the toner to stick? - Do you have a favorite source of copper clad boards? - For double sided boards, how do you ensure the two sides line up? - Are there any tricks for the ironing part? Do you tin your boards after etching? In the past, I have used the photo method (2001), the plotter method (2002-2003, actually made hundreds of boards using this one), and relatively recently (2006) the direct toner transfer method. I wasn't very happy with the last one, as the toner would often rub off with the paper (plus it took a long time). By the way, what's considered the proper way to dispose of used etchant? Best regards, Vitaliy

>> It's great that this should come up now, as tonight I completed a >> successful toner transfer etch (and subsequent reflow) of 6 boards >> each >> containing a 4x5 pad BGA with 0.4mm pitch. It's not just a matter of pitch, but also of quantity/density. Being able to do 20 pins (?) worth on one of those annoying itty-bitty packages is a different issue than being to fan out 100 pins from a PIC32... As for batchpcb, I recall their "goodness" of deals going down quite a bit as you reach the larger sizes (near max sizes of other services.) For the tiny boards a lot of hobbyists produce, the $10 + $2.50/in^2 is a good deal, but some of the other PCB vendors have a max size of THEIR "proto deal" up to 30 in^2, at which point batchPCB is starting to look less attractive. (For instance, Olimex is $1.21/in^2 for their 3.9x6.4 inch board.) On the third hand, a lot of those "deals" are fraught with "issues"; don't include soldermasks or silkscreens, don't include small drill sizes, etc, etc. You have to be pretty careful (the whole obfuscation of real price thing is enough to drive one back to the kitchen...) (Your 0.4mm pin pitch may already have pushed you out of the realm of some "deals.") BillW

part 0 44 bytes his is a multi-part message in MIME format. part 1 3553 bytes content-type:text/plain; charset=ISO-8859-1; format=flowed (decoded 7bit) Vitaliy wrote: {Quote hidden}> Andrew, if you don't mind I'd like to clarify a few things because this is > something I can definitely make use of (the instant gratification aspect is > very appealing). > > - Which mailing labels are you talking about? Avery, FedEx? > - I'm assuming you're printing on the side that the paper used to be on, > after you peel the paper off? > - How much do you have to rub the waxy surface in order for the toner to > stick? > - Do you have a favorite source of copper clad boards? > - For double sided boards, how do you ensure the two sides line up? > - Are there any tricks for the ironing part? > > Do you tin your boards after etching? > > In the past, I have used the photo method (2001), the plotter method > (2002-2003, actually made hundreds of boards using this one), and relatively > recently (2006) the direct toner transfer method. I wasn't very happy with > the last one, as the toner would often rub off with the paper (plus it took > a long time). > > By the way, what's considered the proper way to dispose of used etchant? > > Best regards, > > Vitaliy > I'd be glad to expound.. This was the first time I had used the mailing label backing material so I just used what happened to be around. This was HP Q2593A address labels. The first time I didn't scuff it and the toner didn't stick well, so I scraped the toner off with a credit card, buffed it with a dry paper towel, and ran it through again (after a blank sheet of paper to remove excess toner particles from the fuser) and it worked perfectly. Even though this was the first time using this material, it worked better than any other I had tried. > - I'm assuming you're printing on the side that the paper used to be > on, after you peel the paper off? Yes. > - How much do you have to rub the waxy surface in order for the toner > to stick? In this case, just enough to remove the first failed run, or 'a few swipes with a credit card and a few rubs with a paper towel' > - Do you have a favorite source of copper clad boards? Not really, they all seem to work although the copper might lift off of the cheap stuff more easily if you do soldering rework. > - For double sided boards, how do you ensure the two sides line up? Registration marks, pre-drilled alignment holes in the corners, and a lot of patience. > - Are there any tricks for the ironing part? I use a hot laminating press at 125C for 3-4 seconds. I have heard ironing works too... > Do you tin your boards after etching? No, but have the stuff so should try it some time. One thing to remember, like most processes involving adhesion, is that you never want to touch the surfaces of the paper or board with your bare oily grubby little hands, no matter how clean you think they are. Touch the paper only at the edges or wear gloves. Also, as you can see in the photo attached, I scrub the copper board while wearing gloves with a brillo pad twice in orthogonal directions. This textures the copper, making the toner stick better. You can see a deeper scratch in the bottom left, it is not as deep as the lighting makes it look. After this, the board is rinsed with Ethanol and allowed to dry. Etched with preheated FeCl solution.. Once it is past its useful life I have heard you can neutralise the rest of the solution then it is safe to dispose of in a 'normal' way. -Andrew part 2 38368 bytes content-type:image/jpeg; (decode) part 3 35 bytes content-type:text/plain; charset="us-ascii" (decoded 7bit)

Those are really nice boards. I was not aware of them before. You have several choices for your connectivity. I will mention 2 possibilities. I am going to assume (*** warning **** ) that the adapter board is laid out on a 0.100 inch grid ( 2.54mm ) #1 You can lay your PCB (YPCB) out to match the holes of the adapter board and then connect YPCB to the adapter using the following connectors: http://www.gradconn.com/2-54mm/2-54mm.asp www.circuitspecialists.com/level.itml/icOid/993 The above represent an example of the many vendors who sell dual row connectors on 2.54 mm centers. #2 You can also connect using zebra strips www.gsweb.com.tw/silicone_products_no.php?id=13 MA > On Dec 10, 2008, at 8:01 PM, solarwind wrote: > > http://futurlec.com/Pictures/TQFP_100.jpg > -- WFT Electronics Denver, CO 720 222 1309 " dent the UNIVERSE " All ideas, text, drawings and audio , that are originated by WFT Electronics ( and it's principals ), that are included with this signature text are to be deemed to be released to the public domain as of the date of this communication .

Looking to the attached picture I see there is a very large insulation between routes. Component footprint is ugly. That's partially because of the etching process. Choose such dillution for the solution to keep the etching process slow. A 0.5 oz PCB is better for high pitch density than 1 oz. However the picture still shows a low quality homebrew process. If you really want to use 0.5mm pitch or lower and large number of pin package, my recommandation is to forget this methode and use CNC, which is much better for low quantity of PCB even if the CNC machine is homebrewed (of course using proffesional drill&mill bits) Vasile On 12/10/08, A K <.....00KILLspam.....case.edu> wrote: {Quote hidden}> > > Vitaliy wrote: > > Andrew, if you don't mind I'd like to clarify a few things because this is > something I can definitely make use of (the instant gratification aspect is > very appealing). > > > > - Which mailing labels are you talking about? Avery, FedEx? > > - I'm assuming you're printing on the side that the paper used to be on, > after you peel the paper off? > > - How much do you have to rub the waxy surface in order for the toner to > stick? > > - Do you have a favorite source of copper clad boards? > > - For double sided boards, how do you ensure the two sides line up? > > - Are there any tricks for the ironing part? > > > > Do you tin your boards after etching? > > > > In the past, I have used the photo method (2001), the plotter method > (2002-2003, actually made hundreds of boards using this one), and relatively > recently (2006) the direct toner transfer method. I wasn't very happy with > the last one, as the toner would often rub off with the paper (plus it took > a long time). > > > > By the way, what's considered the proper way to dispose of used etchant? > > > > Best regards, > > > > Vitaliy > > > > I'd be glad to expound.. > > This was the first time I had used the mailing label backing material so I > just used what happened to be around. This was HP Q2593A address labels. > The first time I didn't scuff it and the toner didn't stick well, so I > scraped the toner off with a credit card, buffed it with a dry paper towel, > and ran it through again (after a blank sheet of paper to remove excess > toner particles from the fuser) and it worked perfectly. Even though this > was the first time using this material, it worked better than any other I > had tried. > > > - I'm assuming you're printing on the side that the paper used to be > > on, after you peel the paper off? > > Yes. > > > - How much do you have to rub the waxy surface in order for the toner > > to stick? > > In this case, just enough to remove the first failed run, or 'a few swipes > with a credit card and a few rubs with a paper towel' > > > - Do you have a favorite source of copper clad boards? > > Not really, they all seem to work although the copper might lift off of the > cheap stuff more easily if you do soldering rework. > > > - For double sided boards, how do you ensure the two sides line up? > > Registration marks, pre-drilled alignment holes in the corners, and a lot of > patience. > > > - Are there any tricks for the ironing part? > > I use a hot laminating press at 125C for 3-4 seconds. I have heard ironing > works too... > > > Do you tin your boards after etching? > > No, but have the stuff so should try it some time. > One thing to remember, like most processes involving adhesion, is that you > never want to touch the surfaces of the paper or board with your bare oily > grubby little hands, no matter how clean you think they are. Touch the paper > only at the edges or wear gloves. Also, as you can see in the photo > attached, I scrub the copper board while wearing gloves with a brillo pad > twice in orthogonal directions. This textures the copper, making the toner > stick better. You can see a deeper scratch in the bottom left, it is not as > deep as the lighting makes it look. After this, the board is rinsed with > Ethanol and allowed to dry. > > Etched with preheated FeCl solution.. Once it is past its useful life I have > heard you can neutralise the rest of the solution then it is safe to dispose > of in a 'normal' way. > > -Andrew > > > -

On 12/10/08, Vasile Surducan <EraseMEpiclist9spam_OUTTakeThisOuTgmail.com> wrote: > Looking to the attached picture I see there is a very large insulation > between routes. > Component footprint is ugly. That's partially because of the etching > process. Choose such dillution for the solution to keep the etching > process slow. > A 0.5 oz PCB is better for high pitch density than 1 oz. More, if you send to a serious PCB house a design with 3mil traces in very high density ( 3 mil trace, 4 mil isolation) they will ask to start from a 0.25 oz copper and add copper to 0.5oz using electrochemical deposition. {Quote hidden}> However the picture still shows a low quality homebrew process. If you > really want to use 0.5mm pitch or lower and large number of pin > package, my recommandation is to forget this methode and use CNC, > which is much better for low quantity of PCB even if the CNC machine > is homebrewed (of course using proffesional drill&mill bits) > > Vasile > > On 12/10/08, A K <00spam_OUTcase.edu> wrote: > > > > > > Vitaliy wrote: > > > Andrew, if you don't mind I'd like to clarify a few things because this is > > something I can definitely make use of (the instant gratification aspect is > > very appealing). > > > > > > - Which mailing labels are you talking about? Avery, FedEx? > > > - I'm assuming you're printing on the side that the paper used to be on, > > after you peel the paper off? > > > - How much do you have to rub the waxy surface in order for the toner to > > stick? > > > - Do you have a favorite source of copper clad boards? > > > - For double sided boards, how do you ensure the two sides line up? > > > - Are there any tricks for the ironing part? > > > > > > Do you tin your boards after etching? > > > > > > In the past, I have used the photo method (2001), the plotter method > > (2002-2003, actually made hundreds of boards using this one), and relatively > > recently (2006) the direct toner transfer method. I wasn't very happy with > > the last one, as the toner would often rub off with the paper (plus it took > > a long time). > > > > > > By the way, what's considered the proper way to dispose of used etchant? > > > > > > Best regards, > > > > > > Vitaliy > > > > > > > I'd be glad to expound.. > > > > This was the first time I had used the mailing label backing material so I > > just used what happened to be around. This was HP Q2593A address labels. > > The first time I didn't scuff it and the toner didn't stick well, so I > > scraped the toner off with a credit card, buffed it with a dry paper towel, > > and ran it through again (after a blank sheet of paper to remove excess > > toner particles from the fuser) and it worked perfectly. Even though this > > was the first time using this material, it worked better than any other I > > had tried. > > > > > - I'm assuming you're printing on the side that the paper used to be > > > on, after you peel the paper off? > > > > Yes. > > > > > - How much do you have to rub the waxy surface in order for the toner > > > to stick? > > > > In this case, just enough to remove the first failed run, or 'a few swipes > > with a credit card and a few rubs with a paper towel' > > > > > - Do you have a favorite source of copper clad boards? > > > > Not really, they all seem to work although the copper might lift off of the > > cheap stuff more easily if you do soldering rework. > > > > > - For double sided boards, how do you ensure the two sides line up? > > > > Registration marks, pre-drilled alignment holes in the corners, and a lot of > > patience. > > > > > - Are there any tricks for the ironing part? > > > > I use a hot laminating press at 125C for 3-4 seconds. I have heard ironing > > works too... > > > > > Do you tin your boards after etching? > > > > No, but have the stuff so should try it some time. > > One thing to remember, like most processes involving adhesion, is that you > > never want to touch the surfaces of the paper or board with your bare oily > > grubby little hands, no matter how clean you think they are. Touch the paper > > only at the edges or wear gloves. Also, as you can see in the photo > > attached, I scrub the copper board while wearing gloves with a brillo pad > > twice in orthogonal directions. This textures the copper, making the toner > > stick better. You can see a deeper scratch in the bottom left, it is not as > > deep as the lighting makes it look. After this, the board is rinsed with > > Ethanol and allowed to dry. > > > > Etched with preheated FeCl solution.. Once it is past its useful life I have > > heard you can neutralise the rest of the solution then it is safe to dispose > > of in a 'normal' way. > > > > -Andrew > > > > > > --

I have a zebra strip connector on a Sharp brand microwave. It is a pain in the neck. No matter how hard I try to clean and protect the area, at least twice a year, there is an accumulation of corrosion that needs to be cleaned to make the start button function. :( Michael Algernon wrote: {Quote hidden}> Those are really nice boards. I was not aware of them before. > You have several choices for your connectivity. I will mention 2 > possibilities. > I am going to assume (*** warning **** ) that the adapter board is > laid out on a 0.100 inch grid ( 2.54mm ) > > #1 You can lay your PCB (YPCB) out to match the holes of the adapter > board and then connect YPCB to the adapter using the following > connectors: > http://www.gradconn.com/2-54mm/2-54mm.asp > www.circuitspecialists.com/level.itml/icOid/993 > The above represent an example of the many vendors who sell dual row > connectors on 2.54 mm centers. > > #2 You can also connect using zebra strips > www.gsweb.com.tw/silicone_products_no.php?id=13 > MA > > > >> On Dec 10, 2008, at 8:01 PM, solarwind wrote: >> >> http://futurlec.com/Pictures/TQFP_100.jpg >> -- >> > > > > WFT Electronics > Denver, CO 720 222 1309 > " dent the UNIVERSE " > > All ideas, text, drawings and audio , that are originated by WFT > Electronics ( and it's principals ), that are included with this > signature text are to be deemed to be released to the public domain as > of the date of this communication . > >

CNC routing is great. With a fine point engraver and a low runout spindle, doing a QFP with 11mil pads and 13 mil separation is no sweat as long as the board is small. Larger boards do run into flatness issues unless you have a floating head that presses the board down around the engraver. If the board is not flat against the table, the engraver goes deeper than intended. That not only routes out more of the FR4 core but if it's a fine-point engraver it's going to be V-shaped so it will cut wider than intended when it's going in deeper. Danny Vasile Surducan wrote: {Quote hidden}> Looking to the attached picture I see there is a very large insulation > between routes. > Component footprint is ugly. That's partially because of the etching > process. Choose such dillution for the solution to keep the etching > process slow. > A 0.5 oz PCB is better for high pitch density than 1 oz. > However the picture still shows a low quality homebrew process. If you > really want to use 0.5mm pitch or lower and large number of pin > package, my recommandation is to forget this methode and use CNC, > which is much better for low quantity of PCB even if the CNC machine > is homebrewed (of course using proffesional drill&mill bits) > > Vasile >

Naw there's a guy here in town who sells super-fine 20deg carbide engravers for like $4-ish. The point's real hard to measure. I'm not sure how much of the cut width comes from the tip size vs spindle runout or machine backlash at this point. I do know that with one pass there was less than the 13mil separation at the QFN pads. As part of the dialing in process with the PCB-gcode CAM tool I had to decrease the stated tip diameter enough so that it made more than one pass between the QFN pads to increase the separation. So I'm saying it's capable of more, and 8 mil *might* be possible. He goes by drillman1 on eBay. REALLY good supplier to have around. The cheap board makers often tack on costs for extremely fine spacing. Sometimes extra just for soldermask. Not that having a board mfg'ed isn't great and all, but there's a big place for home board creation too. Prototypes often don't work perfectly anyways and have to be redone over... and over. If you can do them at home this could save you weeks of waiting for several cycles of board revisions to come back. Besides, we were talking about home etching anyways- no soldermask or PTH in that process. With CNC you can make it drill all the leaded component/via holes for you too. BTW, you want fine lines in ferric chloride etch, SUSPEND THE BOARD UPSIDE DOWN and don't use a lot of agitation. As the etchant picks up copper it becomes heavier and falls to the bottom. This creates a vertical convection and a slightly more aggressive upward etching action, and the trace sides do end up more vertical with less undercutting under the resist. With double sided board I don't see how this'll work though, since only one side can be upside down. Danny William "Chops" Westfield wrote: {Quote hidden}> On Dec 11, 2008, at 9:47 AM, Danny Miller wrote: > > >> CNC routing is great. With a fine point engraver and a low runout >> spindle, doing a QFP with 11mil pads and 13 mil separation is no sweat >> > > Sounds about right. But that's not small enough for the 0.4mm pitch > IC we're talking about, which has 8 mil pads with 8mil spacing > (approximately.) > > And not having a soldermask or PTH later on is still a pain. > > (and of course those 'fine point engraver" bits can cost nearly as > much as having your board made professionally...) (I have an LPKF > circuit board router. It's nice, but it hasn't solved all my > problems. Sigh.) > > BillW > >

On Thu, 11 Dec 2008, Danny Miller wrote: {Quote hidden}> Naw there's a guy here in town who sells super-fine 20deg carbide > engravers for like $4-ish. The point's real hard to measure. I'm not > sure how much of the cut width comes from the tip size vs spindle runout > or machine backlash at this point. I do know that with one pass there > was less than the 13mil separation at the QFN pads. As part of the > dialing in process with the PCB-gcode CAM tool I had to decrease the > stated tip diameter enough so that it made more than one pass between > the QFN pads to increase the separation. So I'm saying it's capable of > more, and 8 mil *might* be possible. > > He goes by drillman1 on eBay. REALLY good supplier to have around. > > The cheap board makers often tack on costs for extremely fine spacing. > Sometimes extra just for soldermask. Not that having a board mfg'ed > isn't great and all, but there's a big place for home board creation > too. Prototypes often don't work perfectly anyways and have to be > redone over... and over. If you can do them at home this could save you > weeks of waiting for several cycles of board revisions to come back. > > Besides, we were talking about home etching anyways- no soldermask or > PTH in that process. With CNC you can make it drill all the leaded > component/via holes for you too. > > BTW, you want fine lines in ferric chloride etch, SUSPEND THE BOARD > UPSIDE DOWN and don't use a lot of agitation. As the etchant picks up > copper it becomes heavier and falls to the bottom. This creates a > vertical convection and a slightly more aggressive upward etching > action, and the trace sides do end up more vertical with less > undercutting under the resist. With double sided board I don't see how > this'll work though, since only one side can be upside down. > I haven't tried etching really fine tracks YET. But one thing that does come to mind though, if you know there is going to be undercutting, can't you just compensate for it? I mean, just make the tracks a tiny bit wider, perform a normal etch, remove all the remaining photoresist and etch again for a little while longer to remove the "overhang" on the copper. Yes it'll take a bit of experimenting to get right but the etching is a predictable repeatable process so I can't see why this shouldn't be possible. Regards Sergio Masci

When the spacing is only 8mil, compensating for undercutting by thickening the trace may not be possible. Of course if we grow the trace on either side by 4 mil then the cut simply disappears and no longer exists. I can't say how much you'd need to grow it to compensate. Note that the resist has its own thickness and a developed board is essentially a bunch of trenches in the resist. As the cut becomes extremely thin you will see problems with etchant failing to circulate into the trench as well as it does in open area, and fail to etch well. // Also fine line resolution can become a problem even with a perfect mask as light spills around the mask. This results in partially exposed resist both where it was supposed to be completely exposed and where it was not supposed to be exposed. Sort of a gradient. This area will be mechanically weak after developing and may result in inconsistency even with compensation. Danny sergio masci wrote: {Quote hidden}> I haven't tried etching really fine tracks YET. But one thing that does > come to mind though, if you know there is going to be undercutting, can't > you just compensate for it? I mean, just make the tracks a tiny bit wider, > perform a normal etch, remove all the remaining photoresist and etch again > for a little while longer to remove the "overhang" on the copper. > > Yes it'll take a bit of experimenting to get right but the etching is a > predictable repeatable process so I can't see why this shouldn't be > possible. > > Regards > Sergio Masci >