Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


soldering and bonding on gold PCB's

mregalia

#18585

soldering and bonding on gold PCB's | 7 January, 2002

I work for a microwave company. On our lower frequency boards (<20GHz) we often use gold wire wedge bonding to tune SMT boards and to add bare die devices. Typically these boards are ENIG on various Teflon and Teflon/Glass substrates. In the past there has been no consistency on how thick of gold to use. My predecessor left the company just before I came on, and besides, he did not seem to be consistent on his recommendations. I am convinced that we need to keep the gold thickess below 10 microinches to assure long term reliability of the solder joints. But this thickness seems to preclude bonding on such soft board material. We have used selective build up of gold in the past in the areas where bonding is done. However, no one seems be sure what the best scheme is, or how much and what kind of gold is needed to assure good wedge bonding. My experience with bonding has always been on ceramic substrates. Has anyone had to deal with this?

reply »

#18591

soldering and bonding on gold PCB's | 7 January, 2002

For soldered areas, we like to keep gold thickness substantially below the maximum you propose. Search the fine SMTnet Archives for our discussions on the topic of acceptable levels of gold with solder connections.

Let�s fix this �no one seems be sure what the best scheme is, or how much and what kind of gold is needed to assure good wedge bonding� problem. Look to IPC-2221, �Generic Standard On Printed Board Design� [ http://www.ipc.org ] for guidance on your questions. In paragraph 4.4.4 Gold Plating, the standard uses numbers like gold in: * Edge connector areas: 0.8 um min * Soldered areas: 0.8 um max * Ultrasonic wire bonding: 0.05 um min * Thermosonic wire bonding: 0.3 um min

The standard goes on to discuss gold hardness and whatnot.

In this area, there is no difference between your experience in ceramic substrates and PTFE substrates.

reply »

mregalia

#18614

soldering and bonding on gold PCB's | 9 January, 2002

Thanks for cluing me into IPC-2221. Though the numbers they give seem to contradict what I have learned from people working with the materials. At least I now know where the guideline for a max of 30 microinches of gold for soldering comes from. But everyone who works with it tells me 8 max is a much better number. It seems that the board houses talk about 30 while the assembly houses insist upon 5 or 8 max. You also seem to agree with that. Specifying 12 microinches minimum for thermosonic bonding is also surprising. The people here claim that this is too thin.

Your comment about PTFE and hard ceramic boards bonding similarly contradicts my experience. Many years ago I was engineer on our project using duroid substrates. Bonding was much more sensitive than with ceramic. Though our process people eventually solved the problems. I was primarily doing electrical and mechanical at the time so I I didn't learn that much about what they did. Now I work for a company with almost no process engineering, so I have had to add a new hat to my collection Do you have experience with thermosonic and ultrasonic gold wire bonding to different substrates? I would be very interested.

reply »

Chris

#18615

soldering and bonding on gold PCB's | 9 January, 2002

Hi,

I have read many articles about theromsonically bonding to soft gold. Many of these articles state you can bond to 10uinches of gold. I have worked with both Hughes and Panasonic bonders at two different companies and I have never been able to get acceptable bond pull strength or a robust process with anything less that 25 u inches of soft gold over nickel. I assure you your process will be more reliable and produce higher quality produce if you are at 25 u inches or higher. Yes, your solder joints will be embrittled. I use selective gold on my product eliminate this problem.

Thanks,

Chris

reply »

mregalia

#18616

soldering and bonding on gold PCB's | 9 January, 2002

Thanks Chris. As I was writing my last message our mailperson dropped off the latest issue of Circuit Assembly. It has an article on mixing SMT and COB on the same board. The author recommends as a compromise 10-20 micro inches of gold over 70-100 of nickel for gold wire bonding. The articles actually says microns, but that is an obvious mistake. So we have quite a range of recommendations. I was hoping to find a middle ground that could preclude us having to do selective plating but I'm not convinced that there is one. Interestingly, I had our purchasing group get comparative quotes from three PCB manufacturers for the same board, done as ENIG, done with all thick gold, and done with thin gold over the entire circuit with selective plate up in areas. This was on a Rogers 4000 series board. The differences in price for two of the suppliers wasn't that great, about $10 per board if I remember right. The third supplier quoted the same price for all three processes. I talked to their process people about this. They felt that the extra step wasn't that big of an expense.

reply »

#18617

soldering and bonding on gold PCB's | 9 January, 2002

We follow IPC-2221 recommendations for Au and don't have a problem as long as the Ni is under 150 microinches.This ensures a consistent process for us.Also we do plasma clean after SMt and prior to wirebond.That makes a big difference. Rob

reply »

mregalia

#18618

soldering and bonding on gold PCB's | 9 January, 2002

Do you bond with Al or gold wire? What happens if the Ni is over 150 microinches? Thickness of nickel is one those things that we have not been consistent with. The industry seems to pretty much universally call for 100-200 microinches. Our old board expert, since gone, sometimes specified 50-100. PCB makers don't seem to like that at all. One supplier won't bid on them.

We don't have a plasma cleaner, though I have used them and I agree that they make a big difference. We solvent clean our modules after installing the board and before bonding.

reply »

Chris

#18620

soldering and bonding on gold PCB's | 9 January, 2002

I don't have a plasma cleaner either. It will help a lot. Actually I don't clean at all. Our wirebond pads are far enough away so the flux residue does not get on the wirebond pads. That's what we think anyway. I am sure we have some degree of contamination though. The Rogers 4000 I think will be much softer than FR4. Another potential problem for you. If your stitch pads are large enough, you could do a reinforcement bond. Basically put another ball bond ontop of the stitch bond.

reply »

#18622

soldering and bonding on gold PCB's | 9 January, 2002

When I said "In this area, there is no difference between your experience in ceramic substrates and PTFE substrates." I was referring to the area we were discussing ... gold plating. I am sorry if you thought I ment than the wire bonding process was the same.

reply »

Lead Free Wave Solder - 1 Click SMT

PCB X-Ray Inspection