Aluminum Soldering

If the temperature coefficients of linear expansion of the brass and the aluminum don't match to within something like 1%, the joint will surely tear itself apart, no matter how well soldered the joint is.

Joe Gwinn

Hmm, OK thanks. I was thinking of trying a thin layer of epoxy too. But soldering would be better. The linear expansion coef. of brass and Al differ by ~ 4E-6/C

So if soldering happens at 300C (?) then the thin aluminum plate has to shrink by a part per thousand (or so). Maybe 2 mils for a 1.5 inch diameter cylinder. I have no idea if a thin Al plate can accomidate such strain. (I'll find out.)

Would a nice soft alloy like 1100 be best?

George H.

Nonsense.

If it's thin enough that it will yield.

I have the impression that the area of the plate is large, and 1.5" diameter is in the range. A few mils may not sound like much, but when metals fight, the stresses are large, and solder is weak.

The issue is that below the solidus temperature (where the solder becomes solid) as the two metals cool they shrink differentially, putting shear stress on the joint. If the solder gives enough (which depends on the max dimension between far points that are soldered), then no problem. Unless we have temperature cycling, which will fatigue the solder and break the joint.

Rigid epoxy will also fail. What can work is a silicon rubber adhesive, so long as the adhesive layer is thick enough to accommodate the shear strain without tearing.

War story: In the 1970s, I and a partner made lab instruments for measuring how much drugged rats run around (yes, this is useful to know). The instrument consisted of a big sheet steel base about 12" wide by about 24" long by maybe 4" high, with a milky plexiglass cover glued to the top. Light diffused through the plexiglass to light receptors within the base. The problem was that the glue joint between plexiglass cover and steel base was breaking, despite use of the best of adhesives (a silicon rubber adhesive caulk), the covers were coming loose, and customers were unhappy.

Plexiglass has something like five times the temperature coefficient of aluminum, never mind steel, so normal variation in temperature in the customer labs was enough to tear the joints apart, even in labs with only a few degrees of diurnal variation. We measured the strength of the joints - it was something like 50 or 100 pounds per inch in a lap joint, and yet the joints broke within a year.

The solution was simple enough: Put 1/16" thick spacers in the joints so there was always enough rubber that it didn't tear as temperature variation caused the cover to get larger and smaller compared to the steel base.

Anyway, is there any reason the aluminum sheet cannot be brass instead?

The other alternative is an interlayer having intermediate tempco, to spread the strain out. For instance, stainless steel has a tempco intermediate between aluminum and brass.

Joe Gwinn

Well I'll glue it at room temp and then use it at room temp, so maybe not as much of a problem.

Yeah I'm pretty much stuck with aluminum. (But I'm open to other materials if there are any suggestions.) I just started these acoustic experiments. I wanted to show impedance matching for acoustics. (Like a 1/4 wavlength AR coating in optics.) I've got a brass cylinder that's acting like an acoustic Fabry-Perot cavity. There are transducers on each side (one source and the other detector). The transducers have an acoustic impedance of about 3 (in some units.) The brass has an impedcance of about 40 in the same units. So to make an AR coating I need 1/4 wavelength of a material with a mean value between the two. (sqrt(3*40)) ~11. Aluminum has an impedance of ~17 which seems to be close enough. I should also try some glass which has about the right impedance. (10) ..Temp co wise it might be worse.

Grin.. yeah nice idea, but not for my application.

Here's a table of acoustic velocities and impedances.

George H.

But the acoustic loss may be high.

I think acoustic impedances may be expressed in ohms as well. Although it will depend on the system of units used.

Ahh. If the temperature range is small, epoxy will work, at least well enough for experiments. One can get slightly flexible epoxy from industrial suppliers like Master Bond and Loctite. They will know what's best. The strongest materials require very good cleaning of surfaces and cure at 250 F.

I have an old Bransonic ultrasonic cleaning unit with a stainless-steel tank. The ceramic transducer is epoxied to an aluminum interplate which is in turn epoxied directly the the stainless steel tank. This unit still works after decades of use.

Is there any reason not to just drill and tap the brass and mechanically attach the aluminum plate with a few steel screws, with some silicon grease between brass and aluminum masses?

Joe Gwinn

It IS possible, but you have to use something to scratch through the oxide coat first to bare aluminum. Usual practice is to puddle the solder on the aluminum and use a stainless brush through the molten solder until the surface is tinned, then sweat the other piece to it. My Alcoa book suggests that the high zinc-content solders work the best, another one recommends almost pure tin, go figure. Fluxes usually have some really active content to dissolve the aluminum oxide and will result in severe corrosion if not removed/neutralized.

If there's significant stress and it's a butt joint, you may have cracking and joint failure. In that case you might need some redesign, like going to solder tabs on your plates and maybe a thin aluminum shrink-fit tube over your brass to solder to.

Stan

[...]

I have not glued aluminum to brass but I have tried many adhesives for gluing brass to other materials including other metals. I find that most fail miserably.

The best for gluing steel to brass (which I do most often) is Household Goop.

Michael Koblic, Campbell River, BC

Hmm, Well in the deep dark past of grad school, we used stycast 1266 to glue Teflon 'washers' into TeCu plates. Brass can't be that much different.

And there may be better epoxies by now.

George H.

Interesting book and webpage. I think I'll read the book.

The issue is not that one cannot braze metal X to metal Y. For the most part, you can braze anything to anything. The issue is that if one does not design to handle dissimilar temperature coefficients, the joint will soon tear itself apart.

This is a big issue in high reliability engineering, especially in the engineering of packages for semiconductors. There is an accelerated-aging MIL-SPEC test for such things, where one alternately plunges a part being tested into hot oil then liquid nitrogen. If tempcos are not well matched, the part deconstructs itself. This is a very severe test, but is a good predictor of future failure rate. Especially in power electronics.

Joe Gwinn

I'm pretty sure it's a chemistry issue. For such things, one asks the glue manufacturer. If I recall, brass is the problem.

Joe Gwinn

I've a little bit of experience with HAST, but nothing as severe as what you describe. Military stuff is always pushing the limits.

The brass part could be changed. Bronze, TeCu, BeCu if any might be better?

An Al puck in an oversized hole is an interesting idea. I assume the solder will shrink less than the Al, Do I want the puck to be pulled from the sides as well as the face?

George H.

Greetings George, What Stan says about scratching the surface through the solder is right on. You can solder without any flux at all. I have used the aluminum solders available at the hardware store and the strength is remarkable. You can buy little stainless brushes that are about the size of a toothbrush at the hardware store too. Once the surface is tinned it is easy to solder to anything else. If it was me I would avoid the use of flux on the aluminum just for ease of cleaning. Once the solder is molten and you start scratching the aluminum through the solder you will see that the solder will follow the brush as the brush moves beyond the area initially covered with solder. Eric

Yeah, I figured I'd tin both sides and then mash 'em together.. (not sure exactly how to do the mashing, from center out at least.)

I never thought (or heard) of brushing it while under the solder, Thanks!

Yeah.. been there with some SS fluxes. Should be fine if I can tin it all nicely.

Well if solder doesn't work, then maybe some Al 'loaded' epoxy will, 'be the ticket'.

Thanks again Stan,

George H.

Since gluing is a possibility then there are a lot of other options. First, a press fit plug would probably be the best. If you want to solder, I wouldn't dick around with the zinc stuff. Get some La-Co Aluminum flux. This is for lead/tin solders. Scrape the aluminum clean and apply the flux (I usually scrape some more under cover of the flux) then tin the aluminum with regular plumbers solder. Tin the brass with whatever flux you prefer. Then clean both parts, apply regular flux and sweat them together. Practice with the La-Co on some scrap until you get your technique down.

If you want to glue, epoxy is a pain to bond with AL. The best stuff is an 2 part acrylic from Lord #406/16. This is the stuff they glue airplanes together with. You need a special dispenser gun and the first use is pretty spendy, but if you ever get a chance to try it you will be amazed at how well it works.

Paul K. Dickman

Nope.

Goop, pick your favorite flavor.

Personally I prefer Plumbers Goop for this job - thinner and easier to spread evenly.

Yes. The difference between theory and practice. I will have a look at Plumbers' Goop. Never tried it.

Michael Koblic, Campbell River, BC

[...]

The glue manufacturers I asked did not have a dicky. Indeed brass is the problem. It does not matter how you pre-treat it. Incantations and saying Lord' prayer backwards does not work either.

Just Goop.

Michael Koblic, Campbell River, BC

Another trick is to sand the aluminum clean and OIL it, then solder. Vegatable oil, kerosine, WD-40, or motor oil will all work. do the final cleaning/sandong/wire-brushing through the oil and it can't oxidize so the solder tins the aluminum very easily. After it is tinned you can solder anything to it.