Some interesting recent notes on PCB finishes, if you want the ultimate low resistance joint, then Immersion silver with Tin/lead/silver eutetic solder is the way to go
We use ENIG to IPC-4552, 3-6um nikel, 0.05um (min) gold, with proper solder (tin/lead) not allowed to use anything else.
Laying It Out: Off to the Races: Creating Winning Finishes
I will agree John that 20+ years ago FR4 was not what it is today, or was through hole plating, outgassing of joints was a pain if you didn't bake your boards.
These days a lot of packaging is apearing in leadless format (QFN's etc) with the die placed on a interposer forming the thermal pad, or BGA's where a PCB interposer is used. The size and pad pitch of a lot of todays SMD components require a CTE that is similar to the base PCB, that is another reason ceramic devices are not seen as often.
We use ENIG to IPC-4552, 3-6um nikel, 0.05um (min) gold, with proper solder (tin/lead) not allowed to use anything else.
Laying It Out: Off to the Races: Creating Winning Finishes
I will agree John that 20+ years ago FR4 was not what it is today, or was through hole plating, outgassing of joints was a pain if you didn't bake your boards.
These days a lot of packaging is apearing in leadless format (QFN's etc) with the die placed on a interposer forming the thermal pad, or BGA's where a PCB interposer is used. The size and pad pitch of a lot of todays SMD components require a CTE that is similar to the base PCB, that is another reason ceramic devices are not seen as often.
I agree with SY.
Polyimide is kapton, polyamid is nylon.(correct me if I'm wrong, you chem guys...) Had some guys here who didn't notice the spelling difference. Major radiation resistance difference, polyimide being the best.
My 67mm dia diodes were polyimide passivated. The junction edge was bevelled to spread the surface gradient and prevent reverse breakdown from occurring there, then a polyimide was spin coated to the edge. It was a first for powerex, but we needed it both for rad resistance as well as cryogenic survival. The standard passivation coat on the bevelled edged discs became brittle in LH2. The only remaining thing was I had to develop visual inspection criteria for the coat integrity. So far, of the 880 diodes in use, all have worked excellently through 13 run cycles with no passivation failures.
Some interesting recent notes on PCB finishes, if you want the ultimate low resistance joint, then Immersion silver with Tin/lead/silver eutetic solder is the way to go
We use ENIG to IPC-4552, 3-6um nikel, 0.05um (min) gold, with proper solder (tin/lead) not allowed to use anything else.
Aha...05 um... I wonder why the article I mentioned had the .8 to 1.3 um? I'm guessing the .05 you mention is the right one..
And, no question...tin/lead is the proper solder... I will admit however, I've not found either whiskers nor pest in the ton or two of tin/silver I used back in the mid 90's. But it still scares me, so no hi-rel app will have lead-free. Then again, I don't drive a toyota.
Quite accurate.
The leads of the old devices were so forgiving..I miss the good old days when I could actually see the parts...
Cheers, jn
So do I, I use two 24" wide screen monitor, and quite often have an area of PCb 20mm x10mm centred on one screen. The real thing is just a blur when I look at it.
Just double checked IPC-4552, 0.05 is the minimum, typical figures are quoted at 0.075 to 0.125 that match the figures you quoted. With todays gold price, and from talking to PCB manufacturers that they strive to minimise the amount of gold plated on the boards, I doubt I have ever had a board where it was near 0.125um. Once your signals get to >1GHz possible signal loss can occur due to skin effect and the Nickel barrier, there silver is the best finish, but storing the boards is a pain, as they can soon become unsolderable.
ENIPIG is become a popular finish and works well with lead free solder, used a lot for gold wire bonding, it is growing in popularity as a standard PCB finish, as they can get away with 0.05um Au and still get a good solder joint, again less gold minimises the chance of gold embrittlement.
Personaly, I have used and reccomended ENIG since early 2000's and have not had any major (or minor) issues with both real solder and with lead free, boards into harsh envoironments and high rel, with some boards topping 100,000+ per month, so I'm pretty shure of the finish
LOL, even with those magic light bending devices 0402 and 2001 chip components still look like a indistinc blob, and BGA's with 0.4mm pitch, well, thank goodness for AOI.
Just double checked IPC-4552, 0.05 is the minimum, typical figures are quoted at 0.075 to 0.125 that match the figures you quoted. With todays gold price, and from talking to PCB manufacturers that they strive to minimise the amount of gold plated on the boards, I doubt I have ever had a board where it was near 0.125um. Once your signals get to >1GHz possible signal loss can occur due to skin effect and the Nickel barrier, there silver is the best finish, but storing the boards is a pain, as they can soon become unsolderable.
ENIPIG is become a popular finish and works well with lead free solder, used a lot for gold wire bonding, it is growing in popularity as a standard PCB finish, as they can get away with 0.05um Au and still get a good solder joint, again less gold minimises the chance of gold embrittlement.
Personaly, I have used and reccomended ENIG since early 2000's and have not had any major (or minor) issues with both real solder and with lead free, boards into harsh envoironments and high rel, with some boards topping 100,000+ per month, so I'm pretty shure of the finish
LOL, even with those magic light bending devices 0402 and 2001 chip components still look like a indistinc blob, and BGA's with 0.4mm pitch, well, thank goodness for AOI.
Last edited:
There is a cool new product made out of specially shaped glass that may help!
If man were meant to see, he would have been born with them.
Me, I'm just tryin ta keep the rods and cones from suffering further loss of adhesion..
ps. I never knew "snow blindness" was color specific. The green laser toasted green reception for half an hour or so, the left side of my world was purple for a while..
jn
Aren't they a PITA when you drive???
Personaly, I have used and reccomended ENIG since early 2000's and have not had any major (or minor) issues with both real solder and with lead free, boards into harsh envoironments and high rel, with some boards topping 100,000+ per month, so I'm pretty shure of the finish
How do you handle tin whisker issues? You doing a conformal coat?
jn
Really? Lousy temperature range, lots of moisture regain, leaky as hell.
Yes, jn, Kapton is polyimide, Nylon is polyamide. To confuse things further, Torlon is polyamide-imide.
For the Blowtorch seekers here:
Over the decades we have tried to make the best circuits possible with some designs, and we have tried as well to use the very finest of circuit board materials and fabrication. Of course, we started with FR-4, everybody does, but with the Vendetta Research gold boards, we used Polyamide, and of course, it cost more. The basic idea is to minimize circuit 'hook' as defined in an old Tektronix article, that is caused by DA. IF we could, we would NOT use a circuit board at all, but just direct wire in 3D space, but it just is not practical for production.
A decade later, we moved to Teflon for the Blowtorch boards. They are quite different in construction, and almost impossible to trace, visually, because of the conformal coating, so I did not use the Blowtorch board, made by the same layout designer, as a reference, here. We spend time and money on board materials and construction, because we are out to win listening contests, no matter what the cost. It is quite possible that we could 'pull back' in some areas without anyone noticing, but it is a dangerous gamble if we change the wrong thing.
Over the decades we have tried to make the best circuits possible with some designs, and we have tried as well to use the very finest of circuit board materials and fabrication. Of course, we started with FR-4, everybody does, but with the Vendetta Research gold boards, we used Polyamide, and of course, it cost more. The basic idea is to minimize circuit 'hook' as defined in an old Tektronix article, that is caused by DA. IF we could, we would NOT use a circuit board at all, but just direct wire in 3D space, but it just is not practical for production.
A decade later, we moved to Teflon for the Blowtorch boards. They are quite different in construction, and almost impossible to trace, visually, because of the conformal coating, so I did not use the Blowtorch board, made by the same layout designer, as a reference, here. We spend time and money on board materials and construction, because we are out to win listening contests, no matter what the cost. It is quite possible that we could 'pull back' in some areas without anyone noticing, but it is a dangerous gamble if we change the wrong thing.
Never seen them yet with tin lead solder, and most designs I have done that use lead free solder are not valuable enough to investigate "fault not found" field failures. We have suspected possible tin shisker growth as possible field failures, but to prove it is another matter. Most I have seen from Nasa and NPL (National Physics Lab) research are that small that they will probably vapourise befor they can be found. For high rel stuff though we re-do the component terminations with Tin/lead solder, as most small components are now coming in with matt tin terminations (the worse for tin whisker growth) and all BGA's are reballed (2 pence per ball!!!). Even conformal coatings do not stop the damm things growing, there are photos of them growing through the coating. The best method of stopping thier growth is >3% lead weight for weight added to the tin (Thanks a lot RoHS). With the Toyota sudden acceleration and other problems I suspect tin whiskers is becoming more of a problem. but when it can take 2-3 hours to determine the fault, if at all (I have seen piles of no fault found boards), and you can populate maybe 5 or 100 boards in that time, the figures are not being collated. Also it is hard to prove it was tin whiskers.
Flexi PCBs are polyimide, not the best for ridgid PCB's, type of adhesive used and moisture content can cause them to delaminate under reflow/wave especially at lead free temps. Good quality high Tg FR4 with a high frequency weave of glass (finer glass even waft and weave, minimises Er discontinuityies as FR4s average of approx 4.2 is based on the mix of resin and glass) are just as good, if not better. Why make life harder than it needs to be.
Just for John
http://www.speedingedge.com/PDF-Files/tutorial.pdf
How relevant at audio frequencies one wonders...
Flexi PCBs are polyimide, not the best for ridgid PCB's, type of adhesive used and moisture content can cause them to delaminate under reflow/wave especially at lead free temps. Good quality high Tg FR4 with a high frequency weave of glass (finer glass even waft and weave, minimises Er discontinuityies as FR4s average of approx 4.2 is based on the mix of resin and glass) are just as good, if not better. Why make life harder than it needs to be.
Just for John
http://www.speedingedge.com/PDF-Files/tutorial.pdf
How relevant at audio frequencies one wonders...
Last edited:
You may not operate underwater, but in humid environments, polyamide soaks up moisture like a sponge. The dielectric properties do not improve.
The toyota/nasa paper did mention the minimum current to pop the whiskers, I believe.. obviously the connector assembly toyota had was not cleared by current.
I know. A throwaway society nowadays.
cheers, jn
- Status
- Not open for further replies.