0 Members and 1 Guest are viewing this topic.
If you want to learn more, please visit our website Circuit Board Cleaning Equipment.
NaxFM
Why gold plating PCBs?
«
on: July 28, , 06:07:32 pm »
I was wondering why high end equipment always have gold plated PCBs, and I can't give a precise answer, so I was wondering if you could list all the benefits of gold plating, especially for high precision circuits.
One obvious benefit is corrosion resistance, another one is thermal EMF reduction, since gold has a lower seebeck coefficient than copper.
Is there anything else?
bdunham7
Re: Why gold plating PCBs?
«
Reply #1 on: July 28, , 06:33:34 pm »
Gold (typically ENIG these days) is flatter than the usual alternative, HASL. This works better for placing very small or very fine pitch components. Also, if you have an edge connector or test pad that is exposed and not soldered in the finished product, gold makes better connections. For many things that you are prototyping or hand-making or even low-volume production with larger or TH components, HASL is actually the way to go.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
T3sl4co1l
Re: Why gold plating PCBs?
«
Reply #2 on: July 28, , 09:01:21 pm »
Longer shelf life as well.
OSP (organic solderability preservative, i.e. bare copper with a protective chemical applied) is also popular for mass production, where the relatively short shelf life (under a year) is managed by cycling inventory (say as a FIFO). Here's an example from a laptop of mine,
https://www.seventransistorlabs.com/Images/W_1.jpg
the orange rings are plated through holes; they're as dark now as the tantalums nearby! Obviously, they've aged a bit through soldering and about a decade of storage and use.
Bare copper without any kind of protection, depends very much on the atmosphere where it's stored, even with those anti-tarnish strips packed in with it; OSP helps stabilize that, giving a more predictable, if only moderately extended, shelf life.
Whereas platings do as well as, well, whatever the plating is. Gold is basically unlimited shelf life. Silver tarnishes about the same rate as copper (but mostly tarnishes due to sulfur, so the anti-tarnish strips help out more there?). Tin/solder oxidizes slowly, but boards some years old are typically usable without process changes.
HASL is better than tin flash, because tin flash is so thin it can oxidize fully through, whereas solder is dipped then blown off, leaving a fairly heavy layer. Similar shelf life to OSP I think?
And yeah, that's the catch with HASL, it leaves a heavy layer, but how much depends: it's uneven. Mainly a problem for large QFN/LGA/BGA type parts, which need very flat interfaces to solder reliably. Small QFN/SON/DFN I suppose could have issues with uneven solder dosing, which could lead to shorts, but shouldn't be a problem with opens.
Also mind copper balance in that case -- uneven copper distribution between opposite layers can cause board warpage, with similar consequences.
Leaded parts (SO, QFP, etc.) generally have enough slop that it doesn't matter.
Related: mind that ENIG is flash gold, no good for contact insertion. Edge connectors are always* done with 30µin hard gold, or something like that, which has reasonable longevity. A special setup is needed for plating (the fingers are shorted together on a short extended area, which is later cut off and chamfered to get the desired insertion blade edge), and obviously it's added cost (more steps, meaningful amount of precious metal).
*You know, proper like. ENIG is good for... maybe a couple insertion cycles? Let alone any kind of vibration.
Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life? Send me a message!
The following users thanked this post: lowimpedance
, daqq
langwadt
Re: Why gold plating PCBs?
«
Reply #3 on: July 28, , 09:20:20 pm »
Related: mind that ENIG is flash gold, no good for contact insertion. Edge connectors are always* done with 30µin hard gold, or something like that, which has reasonable longevity. A special setup is needed for plating (the fingers are shorted together on a short extended area, which is later cut off and chamfered to get the desired insertion blade edge), and obviously it's added cost (more steps, meaningful amount of precious metal).
*You know, proper like. ENIG is good for... maybe a couple insertion cycles? Let alone any kind of vibration.
Tim
yes, ENIG is on something like 0.1um think, and it has to be thin because when too much gold mixes in with the tin solder it becomes brittle
The following users thanked this post: lowimpedance
, T3sl4co1l
ejeffrey
Re: Why gold plating PCBs?
«
Reply #4 on: July 29, , 05:48:36 pm »
another one is thermal EMF reduction, since gold has a lower seebeck coefficient than copper.
Is there anything else?
The gold is an extremely thin layer that dissolves into the solder during assembly, so it doesn't have any impact on the electrical properties such as the seebeck coefficient. It's strictly there to extend shelf life before assembly, and also prevent (mostly cosmetic) corrosion of exposed copper such as vias and mounting holes.
If anything the nickel in the nickel layer in the ENIG would be a problem for thermoelectric effect but only if there is a temperature gradient across it. The seebeck coefficient of connector bodies and to some extent the solder itself is much more important than surface finish.
T3sl4co1l
Re: Why gold plating PCBs?
«
Reply #5 on: July 29, , 11:07:40 pm »
Incidentally, the nickel underplate could be a problem at RF, but as most of the current flows on the underside of the trace, it's a small effect in practice. When SMOBC isn't acceptable (due to SM dielectric losses), heavy gold or silver platings are preferable, I believe?
Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life? Send me a message!
Infraviolet
Re: Why gold plating PCBs?
«
Reply #6 on: July 30, , 12:04:13 am »
On the subject of ENIG usually being 0.1um thick, Is it true, or is it not, that when using leaded solder gold plating can actually be a bit bad as gold and lead experience the sort of "reaction" you mention which acts a bit similarly to the way gallium eats through aluminium? When using leaded solder one ideally sticks to HASL plating? Or isn't this effect all that significant really?
T3sl4co1l
Re: Why gold plating PCBs?
«
Reply #7 on: July 30, , 05:51:54 am »
Again, the gold dissolves, it's not a problem. It's intermetallics: too much gold and lead in one place.
Before gold flash was standard, processes might need to "wash" the gold off by repeatedly tinning and cleaning, thus dissolving the gold off. NASA has some documents on this.
The problem is embrittlement: intermetallics form at the interface layer, with lead being the primary issue. It's not like gallium diffusing extremely rapidly into aluminum; as far as I know, gold and solder do not move around at room temperature. (Well, gold by itself certainly does, though that might be on more of an atomic scale, I'm not sure if it's enough to matter in bulk. Lead and tin do diffuse notably at room temperature -- hence the spontaneous formation of tin whiskers, or the creepage rate of lead -- but I'm not aware that this is a problem for copper, silver or gold, or tin-soldered joints thereto.)
Incidentally, most solder joints are defined by an intermetallic layer, where the base metal is partially dissolved into the filler, which in turn diffuses some way into the base metal surface. Typically these are small amounts: copper, silver and gold are soluble in the low-percent range in tin/lead alloys not far above their melting points, and the diffusion layer is microns if that. (The amount and depth are greater in brazed joints, for obvious(?) reasons: the chemistry and mechanisms are essentially identical, just done at higher temperatures.) Intermetallics are generally brittle, but if the layers are thin, and reasonably strong (as CuSn's mostly are, for example), the joint is likely to fail by fracture of the filler or base metal first. (As it happens, you can usually peel the base metal off of a solder joint, so clearly those intermetallics aren't quite as strong as the solder. But a gray residue is left on the peeled material, indicating some solder came with. Hm, I don't know offhand if that's actually the intermetallic surface showing, or how much torn-off bits of solder are in that. Peeling stress is also just really severe!)
So it's not that it forms a layer, because it generally does; it's what's in that layer that matters.
Joints done with insufficient superheating and dwell time, or very low solder volume, are at greater risk of gold embrittlement. So, fast production lines doing LGAs say; compare with BGAs with a more relaxed cycle (more time at/above peak temp), tons of volume to dissolve platings into. The nickel might even dissolve! (Related concern: doing BGAs with lead-free balls on a leaded process, the balls won't fully melt, but rather the leaded paste sort of sloppily welds onto the lead-free part. I forget what the preferred resolution is for this; run a lead-free cycle whether you're using leaded or not, to ensure full melting??)
Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life? Send me a message!
Gyro
Re: Why gold plating PCBs?
«
Reply #8 on: July 30, , 09:51:24 am »
Paradoxically, HP used to Gold plate everything, this was in the THT pre silk screen era. This wasn't just in their test equipment, but straightforward logic stuff like the HPA editing terminals that we used to spend happy lunch times playing space invaders on (we used to have to swap RAM cards to get enough memory). I don't remember any increase in thickness on the edge fingers relative to the rest of the boards and the plating seemed quite heavy.
I have absolutely no idea why they did it, but it appears to be the exception that proves the rule as they never suffered joint failures.
« Last Edit: July 30, , 09:54:20 am by Gyro »
Best Regards, Chris
jonpaul
Re: Why gold plating PCBs?
«
Reply #9 on: July 30, , 10:06:08 am »
R&D by Bell Labvoruatoies, Western Electric in s..s for best plating on PCB and connectors prooved the value.
Gold is not subject to oxidation or corrosion like all other platings.
The cheaper plating like tin over nickel, etc do not stay corrosion free as long.
In high rel equipment, Avionics, Space, Medical, Telco CO underse cables the life can be 20..50 years.
All the Tektronix and Hewlett Packard instruments of PCBs (and many connetors, edge connectors) the epoc s...s were gold plated.
Yes it costs more but not much.
The hallmark of modern Chinese electronics is absolute minimum cost, sacrificing life, reliability, etc.
Jon
The Internet Dinosaur..
passionate about analog electronics since s
tszaboo
Re: Why gold plating PCBs?
«
Reply #10 on: July 30, , 11:19:18 am »
Paradoxically, HP used to Gold plate everything, this was in the THT pre silk screen era. This wasn't just in their test equipment, but straightforward logic stuff like the HPA editing terminals that we used to spend happy lunch times playing space invaders on (we used to have to swap RAM cards to get enough memory). I don't remember any increase in thickness on the edge fingers relative to the rest of the boards and the plating seemed quite heavy.
I have absolutely no idea why they did it, but it appears to be the exception that proves the rule as they never suffered joint failures.
The same reason they only used 0.1% resistors in all their desktop multimeters. Even the logic stuff had 0.1% defined as default.
mikeselectricstuff
Re: Why gold plating PCBs?
«
Reply #11 on: July 30, , 11:26:02 am »
Paradoxically, HP used to Gold plate everything, this was in the THT pre silk screen era. This wasn't just in their test equipment, but straightforward logic stuff like the HPA editing terminals that we used to spend happy lunch times playing space invaders on (we used to have to swap RAM cards to get enough memory). I don't remember any increase in thickness on the edge fingers relative to the rest of the boards and the plating seemed quite heavy.
I have absolutely no idea why they did it, but it appears to be the exception that proves the rule as they never suffered joint failures.
Gold was rather cheaper in those days. I suspect there may also have been some "quality feel" aspect to it.
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
jorgeh
Re: Why gold plating PCBs?
«
Reply #12 on: July 31, , 05:51:42 pm »
Im an Expert at nothing, so dont take too serious my opinion.
As a hobby I like to teardown old equipment. and usually military/avionics devices look pristine inside. also mil grade connectors. they look good as new. in contrast to much newer consumer it equipment.
I see also a big difference with old cards, gold contact vs silver contact. gold contact from CNC boards or ISA cards looks pristine. on the other side. silver contact tend to loose itself like the plating start pealing itself.
what i dont understood is why there were gold cap cpus. it is my understanding that silver has a bigger heat transfer coefficient.
a con i would say is the weight. I remember i got some scrap boards, probably prototypes of some sort. because some of them were gold plated and other nickel/tin. you could fell the difference immediatly.
tooki
Re: Why gold plating PCBs?
«
Reply #13 on: July 31, , 09:39:14 pm »
Does the plated gold layer also dissolve in low-temperature solders, e.g. Maker Paste (150 C melting point)?
Yes, because that’s how soldering works: the surface of the workpiece is slightly dissolved, creating the intermetallic layer.
T3sl4co1l
Re: Why gold plating PCBs?
«
Reply #14 on: August 01, , 12:23:38 am »
Does the plated gold layer also dissolve in low-temperature solders, e.g. Maker Paste (150 C melting point)?
Yes. I don't know offhand what gold-bismuth intermetallics may form (if dissolution doesn't occur, as on thick plating etc.) and to what consequence (are they actually OK? even worse?), but at a glance, it looks like the gold layer is just so thin it's hard not to dissolve it with basically anything.
Probably, the kinetics are slower at the lower temp, but that might also be balanced by different "solvent power" of Bi or the alloy, I don't know.
That is, consider trying to dissolve something in plain water, vs. plain ethanol, vs. a mix; obviously, it depends hugely on what kind of thing is being dissolved (many inorganic salts dissolve fine in water, not at all in ethanol; many organic (lipophiles) dissolve readily in ethanol, practically not at all in water; still others lie somewhere inbetween), but also appreciate that the mix has a much lower melting point, and may dissolve things easier, or at a similar rate but at a much lower temperature, than either pure substance. (Although, ethanol has quite a low melting point itself, so this isn't really a useful comparison, but I mean... condensed matter chemistry is hard, okay?)
Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life? Send me a message!
The following users thanked this post: tooki
, karpouzi9
Gyro
Re: Why gold plating PCBs?
«
Reply #15 on: August 01, , 08:58:36 am »
Yes, because that’s how soldering works: the surface of the workpiece is slightly dissolved, creating the intermetallic layer.
I guess my real question is "how deep is the intermetallic layer, and does its depth depend on solder chemistry and reaction temperature." If I know anything about chemistry, the answer to the latter is clearly a complicated but direct "yes."
You've missed one factor: duration - the time that the solder stays in its molten state. I don't know if any intermetalic layer chemistry continues at room temperature, but it must be insignificant compared to the dwell time in the molten state.
Best Regards, Chris
The following users thanked this post: tooki