running chips in vacuum

Gentlemen,

We have a seies of chips that we have had made by MOSIS. In the past, we had them packaged in plastic packages. Some users run this chip in vacuum. We have the thermal issues well under control. We just revised the design, and MOSIS now seems unable to provide us with packaging in plastic. They are recommending a ceramic package. We are worried that with repeated cycling between air and vacuum, the lids may blow off. These systems may be cycled between air and vacuum several hundred times over their life. We are asking MOSIS about whether they have a process available to fill the package cavity. The candidate packages are 128-lead, 14x14 and 14x20 mm package outline.

Does anybody have experience with ceramic chips and cycling between air and vacuum? I'm guessing the space community has a lot of experience with this issue.

Thanks for any guidance anybody can give.

Jon

Reply to
Jon Elson
Loading thread data ...

Why can't MOSIS now provide the plastic packaging? Which foundry is your chip actually processed by? Maybe deal direct? ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

The foundry doesn't matter, fabrication and packaging are separate functions. Apparently, MOSIS is switching to a different vendor for plastic packages. The old vendor no longer has the lead frames available, and doesn't want to bother with lining up a new source.

It is a 128-lead 14x14mm body size, with 0.4mm lead pitch, and we need it made die down, with a heat spreader/sink inserted under the die attach pad. I didn't think this package was far outside the mainstream, but maybe now it is.

Jon

Reply to
Jon Elson

chip makers like Intel still attached dies flipped, with Silver filled epoxy, to the can lid, as you call it. That "lid" is actually the chip mount and heat sink interface for you. The grid array portion of the package that interfaces with the circuit board is where the cavity is. The epoxy is space approved IIRC. I know it is mil spec. It is the industry standard. I am pretty sure those cavities can take a vacuum/ pressure cycling. They are filled with nitrogen during a hot attach/bake process, and are very low pressure already inside at normal room temp. Kind of like food canning.

formatting link

Reply to
DecadentLinuxUserNumeroUno

Check my approximate math. Package is 1/4 square inch. Can you get more than 15psi pressure differential? Can the lid attach stand 4 pounds of force?

Reply to
mike

I just renewed my MOSIS account for an upcoming project that requires it... guess I'll be learning the hard way ;-) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

Is that a brazed-on lid? If so, you shouldn't have any worries. A lid that small won't flex much with a few pounds of air pressure. You can test it yourself with some good grade double-sticky tape and something to provide load and unload cycles. I'd be very surprised if you didn't get 100k cycles or more without issues.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

The sample packages they sent had no metallized ring around the cavity, so I think it must be an epoxy-attached lid. They said something about "TAPED ON" lid, which I immediately rejected for a production application. I still don't know if this will be a metal or ceramic lid. Metal will be thinner under the package (remember, we are going die-down) but would short out the vias on the board, so we'd have to deal with that, maybe with Kapton tape or such. Ceramic lids would be thicker, but no risk of shorts.

Another issue is if the package cavity was not hermetic, it could cause slow leaks which could contaminate the otherwise high vacuum in the experimental apparatus, and that would get people quite mad.

Thanks for the info!

Jon

Reply to
Jon Elson

Never more than 15 PSI differential. I have no idea how strong the lid is, or the attachment method. I've never faced this particular issue before.

Jon

Reply to
Jon Elson

I know nothing on the subject, but I like Phil's testing idea. Glue a wire on and "cycle" the lid with some mass.

George H.

Reply to
George Herold

Can you just poke holes in the lid and live with different pressures? (low voltage or at least no HV during the cycle.)

George H.

Reply to
George Herold

Glue a mass onto it, invert it, and place it on a vibe table.

The mass will break free before the lid does.

Unless you glue it on with some *really good* epoxy.

Reply to
DecadentLinuxUserNumeroUno

For how long time are these package subjected to vacuum at each cycle?

Materials to be sent to space are checked for out-gassing and plastic packages are bad in this respect. When some molecules out-gas leaving some kind of cavities in the material.

What happens, if such out-gassed material is returned to atmospheric conditions, are these cavities filled by atmospheric gases, including water vapor ?

Reply to
upsidedown

CERDIP has been around for ages, and flatpacks with brazed lids also have been around for ages. Many such parts were (are?) mil spec and used in satellites. There is no such thing as a vacuum; NOTHING THERE...

Reply to
Robert Baer

We'd really like to do something standard, not oddball. These chips and boards will need to be usable for 10 years or more, so anything that compromises the longevity of the chips should be avoided. Having them open to room air and possible vacuum pump oil sounds like a bad idea. Yes, the chips are passivated, but I think that is mostly to preserve them until they are packaged.

Jon

Reply to
Jon Elson

Generally, hours to weeks. Then, they might be stored in room air for a month or two before the next use.

We've run hundreds of the plastic-packaged parts in vacuum for years with no problems. This is under the same conditions as outlined above.

Jon

Reply to
Jon Elson

Right, I'm aware of that, and that's why I THINK it ought to be OK. Well, we run these systems at pretty good vacuum in the chamber, such as 1 x 10^-7 Torr. There could be 120 of these chips in boxes in a single vacuum chamber, and there could be circuitry just an inch away with up to 500 V on it. If several chips leaked gas and raised the pressure in the box, that could cause corona, another undesirable situation. So, there are a bunch of reasons we really don't want lidded packages with voids in them. But, so far, MOSIS doesn't have any other solution for us, such as some way to just go back to the plastic packages we have been using.

Thanks, all,

Jon

Reply to
Jon Elson

Cerdip is a fine package. Kind of heavy though. Side braze with solder on lid doesn't seem terribly secure to me. They are very common for prototypes in that the die can be probed or viewed with an emission scope.

Reply to
miso

There is vacuum and *HARD VACUUM* which one is it?

We never experienced any problems with stuff in pumped down to normal vacuum levels on turbopumps or diffs, but when you start baking vacuum systems hard for days in custom ovens to get the very last trace of water vapour out then things do tend to react eventually to the very aggressive thermal cycling. Mass specs being one obvious example where the Faraday amplifiers get subjected to harsh thermal treatment.

Noble gas machines were the worst for excessive and brutal thermal cycling. I can't recall how hot they baked but it was really nasty.

--
Regards, 
Martin Brown
Reply to
Martin Brown

This is the business end of a Fourier Transform Mass Spectrometer.

The pickup electrodes are buffered by a couple of unity-gain jfet opamps. They run uncooled in UHV, at about +120C or so. They need 1M resistors to ground to squash the opamp leakage, which costs 35 dB or so s/n ratio. The circuit was designed by chemists.

formatting link

formatting link

formatting link

They are out of business, of course.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.