Gold thickness for wire bonding?

I have a friend who actually (still!) runs a hybrid operation. I'll ask him.

I assume you mean the standard ENIG finish, gold over nickel.

Ouch, I've been having some issues wire bonding gold. (I think mostly issues with the gold sticking to the under lying metal.

But for an upper bound on thickness, I ordered some blocks of Al2O3 with gold plated on four sides. I soldered those to pcb and there was no problem wire bonding to them. I can look up the company... It seems they were near to me. RF shorts?

George H.

I guess this is the palladium reference.

Funny this should come up, because I'm designing the electronics for an electro-optical gadget where we may wire-bond the active optical devices, little transparent slabs with sputtered contacts, to our PC board. That's the thing that needs the 1400 volt power supply and regulators and pulsers.

Watch out for drift due to the optical alignment walking around. If you can hang the PCB off the optical mount, rather than the reverse, it'll be a lot more stable.

Cheers

Phil "many tee-shirts" Hobbs

Fortunately I'm doing the electronic and thermal bits, not the optics or mechanics. We are making a few small PCB-based electro-optical gadgets that mount on a metal plate, a classic embedded optical bench. I don't think that wavelength-precision alignments matter; that would be nasty with FR4 controlling position stability. FR4 would shrink and swell and cold-flow at every mounting screw, which could be a real nightmare.

John "two pairs of jeans" Larkin

That would be very nice. Although I'll probably have to call in the order today. It's one of those fast-track experimental boards.

Yes.

How thick was the gold on those?

This isn't RF, all near DC.

Aha, thanks! Now we are finally getting somewhere. Quote "... with the Gold thickness being at least 80 microinches to allow for Gold wirebonding". For ENEPIG they say 1-2 uinches of gold is sufficient but

Then you are in the same boat except that mine doesn't need super high bond strength because it is just for R&D. Though I want to avoid frustration for the folks doing the bonding when they do their pull testing.

I'll be working with Mike at

Not many people do hybrids these days. This is terrifying:

Dang, I'm sorry Joerg, my google-fu is not working today. (I was web searching with no luck.) And the parts are over at UB with the company name on them.

Are you wire bonding to a pcb? I only know what I read, but some places said I needed soft gold for wire bonding. And when I went looking for board houses to do that it was expensive.

Ah I found it... (A thread by me on wire bonding.)

Anyway for prototyping these gold "bricks" were about $1 each (I bought ~200 or so) I stuck them onto the pcb and wire bonded to the gold pads... cheaper than getting the broad house to put down soft gold.

George H.

Can you let me know if that works?

George H.

Here's what I used.

Geo

It doesn't mention thickness but John's paper said 100 uinches. I can only get 50 max as the most expensive option but then I'll go for that.

I might have to do hundreds so don't want to have extra manual labor.

I will. And thanks, for the link to the hybrid producer, John. Running out of time here, got to get this board out.

Make sure you use aluminum / gold contacts... Guaranteed destruction via Purple Plague (and not the mimeo offset variety).

On a related note, I take that to mean you've got the layout going and all?

Don't know how many things are capable of doing that, at least in an integrated fashion. I recall reading Altium could do it, with what looked like a tedious setup (positioning the die lands and bond pads, and calling out each bond wire / pad pair), but that it apparently generates "PnP" data ready for wirebonding.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

Gentlemen,

Searched up and down the web but not much info. The PCB house (4PCB) offer three levels:

Anyhow, no NiPdAu so no palladium in there. We need to ultrasonically bond 1mil gold wires. Would this thin 10 uinch flash gold be ok? I heard that down to 2 uinches has been used but I don't want to mess this up.

-- Regards, Joerg

It's all done now. But by a layouter, I always outsource layouts.

Most of my layouts are done in PADS. The plating requirements go into the fab spec and fab drawing. We have spec'd 50 uinch which is the thickest the fab place can do.

People do sell bonding pads, little metal or ceramic blocks that you can pick-and-place and solder like components, but are optimized for wire bonding on top. That would work if you don't need a lot of bonds.

Hey, I wrote a couple of utilities for PADS, called PINK and BLUE. One compares the PCB and schematic netlists, and one compares a PADS netlist against our inventory list. Both check values and sizes and such. The Brat loves them.

That's what George suggested. But since I'll have to hand-solder 100 boards I really wanted to avoid that.

I always do a hand netlist check against the schematic, using ye olde crayons. Sometimes I find bugs and most are caused by yours truly.