From cast iron housings to aluminum

[geesamand]

I have a project where a customer is interested in using aluminum to build a differential housing instead of cast iron. All my past experience is in cast iron - as far as bearing fits, working stresses, machining, casting, etc.

The goal is to design a stronger/stiffer housing than the OEM cast iron unit. The aluminum option is intended to save weight.

What material should I consider for my aluminum casting? What is the most reasonable casting process for producing under 200 units? I intend to contact the bearing manufacturer for housing fit recommendations.

Any tips appreciated.

David

 

[MikeHalloran]

Aluminum founders may brag about their capabilities in using all sorts of strange and exotic alloys, but when you have passed the design stage and are asking them to actually pour something into a sand mold, what you will get is 356-T6. Don't design for any other alloy, because when push comes to shove, you won't get it.

About design for stiffness: The specific weight and the elastic modulus for steel vs. aluminum are almost exactly 3:1, for both numbers. Put another way, you will need three times as much aluminum, by weight, to get the same stiffness as a similar steel part. So aluminum has _zero_ advantage over steel, or iron, where stiffness is a design criterion.



Mike Halloran
Pembroke Pines, FL, USA

 

[swall]

Find out what grade of ductile iron you are working with on the existing housings, so you know what min yield strength you are dealing with. Then design your aluminum housing, taking in account the minimum yield strengths of the various casting alloys. All of the common aluminum casting alloys, except 206, will be below 50ksi yield strength. 356 is a good alloy, but you may be able to get by with something like 319, which is very common.

 

[btrueblood]

Adjust all press-fits (e.g. bearing races) to account for the lower stiffness of Al, as Mike points out. Plan on seeing some galling during installation of races, adjust tolerances accordingly, and learn some new swear words.

 

[TVP]

What MikeHalloran points out about stiffness and density is true when the load case is tension or compression. When bending is concerned, geometry will have an effect, which is why heavily ribbed, complex castings can see some improvement in stiffness when compared to sheetmetal designs with relatively little section modulus. Since aluminum castings can usually be more intricately detailed and have thinner sections, you may be able to use a greater number of thinner ribs than with iron castings and still see some improvement in specific stiffness.

 

[thundair]

I agree with Mike but as TVP said if you change the 'I' factor of Aluminum you can regain some of the lost stiffness.
In aircraft structure we are always fighting for room to stiffen something up...As far as casting Alum you can get 356 T7 same as Briggs & Stratton Go Cart motors.

There are tricks to casting bearing areas such as chill bars and post treatment to make a specified area stronger..

My choice (and I don't get to use it because I deal with aircraft) is CGI it is strong and lite..

http://www.moderncasting.com/archive/WebOnly/1002/CGI1002.asp

Cheers

I don't know anything but the people that do.

 

[MikeHalloran]

Actually, " three times as much aluminum, by weight, to get the same stiffness as a similar steel part" is incorrect. You need three times the volume (aluminum:iron), or the same weight. But you get the point.

The problem with diff housings is that the iron parts are already ribbed, and you can't put substantially deeper ribs on them without hitting things like floor pans. You probably could use _more_ ribs, but the stronger car housings, e.g. Ford 9" Nodular Iron, are already heavily ribbed in both directions, so you could double the number of ribs, but probably not triple it.

Maybe you could make an iron sub-housing to bear the gear separating forces with an aluminum spacer/ adapter between that and the axle housing, but you'd be adding two flanges, a gasket and a ring of bolts, so it would be hard to save weight.

Maybe you could make a somewhat skeletal iron sub-housing, again to bear the gear separating forces, and cast it into an aluminum housing. Have you got a budget to FEA something like that?



Mike Halloran
Pembroke Pines, FL, USA

 

[GregLocock]

And then you get into the exciting world of the different CTE.

Admittedly your axle oil shouldn't be running over 130 deg C, but that is enough to force a reconsideration of bearing fits and gear meshing.

I think you should be able to pull some weight out for a given stiffness, having been through a similar process on suspension arms, by exploiting better shape factors.

The resulting diff will be noisier, I suspect.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[MikeHalloran]

The ASME and SAE house rags regularly carry ads for some package that can optimize the structure of an A-arm, to look like a more skeletal A-arm, all automagically, they would have you believe.

I wonder what it would do with this problem? I.e., given interior and exterior envelope limits, bearing and bolt locations and loads, what would it generate?

Error messages, I presume.



Mike Halloran
Pembroke Pines, FL, USA

 

[GregLocock]

Optistruct is the one I use. It is fairly automatic, I've used it for a ladder frame chassis.

There's a lot of different parameters to fiddle with, which frankly don't make a huge difference. The important thing is to make sure the input data is good, ie that you have all the load cases covered, and that your desired repsonses (basically displacement limits at particualr points) are sensible.

Then it eats away at all the understressed material and leaves you with a Salvador Dali type shape, which with a good eye you may be able to turn into a realistic structure.

Even if the resulting shape is unmanufacturable, it gives you an ideal weight/stiffness target to aim for.

A diff housing would be an ideal project... but don't expect miracles.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[geesamand]

Wow, thanks for the interesting responses.

The budget/size of this project does not support multi-material structures, since the goal is to get strength at a reasonable cost over pure weight savings. So knowing that aluminum is generally the same net weight by stiffness, we're leaving it out right now. Definitely using a ductile iron to perform better under shock loads.

I am wondering about one issue if we use a ductile material - can the housing 'rack' under heavy load and not be immediately noticeable to the user?

Dave

 

[IceStationZebra]

The one thing that would worry me is the higher rate of thermal expansion - your hypoid mesh will vary quite a bit more than with iron. Then add in the possibility of reduced stiffness as discussed above. ISZ

 

[Tmoose]

For decades Lincoln Electric has been advocating replacing castings with steel weldments for improved stiffness and lighter weight.

Toyota has been producing welded axle housing for a while.

http://www.trailtough.com/images/toyhsng.jpg

These guys offer them for American muscle cars, but do not mention weight.

http://www.strangeengineering.net/catalog/pages/053.jpg

I'd think a one-off axle housing could be welded and properly stress relieved cheaper than a one-off casting. Machining could be a bit more costly on the steel part.

 

[geesamand]

Actually our project has changed a bit, and now we are taking another differential design that comes in aluminum housing and adapting the mounting to suit our vehicle.

So aluminum is back in, cast iron is out. Fabrications are definitely out. Qty expected to be around 20-50 units initially, so casting costs should remain superior.

Thanks again for the advice.

Dave

 

[GregLocock]

Just in case you need to, bear in mind you can weld aluminium castings, so long as you do the finish machining afterwards.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Tmoose]

"Fabrications are definitely out. Qty expected to be around 20-50 units initially, so casting costs should remain superior."

I suspect Toyota made more than 200 of their axle assemblies. It seems the rules have changed regarding economical production methods in the last few decades, at least for those with full CNC capabilities. Who would have thought that an economy car suspension woould look like this?

http://www.artsautomotive.com/CivicKnuckleBushings-004.jpg

 

[sbi]

# NOTE there is a small group of family owned no-ferius foundrys left in the USA (we are one) and we do pour 713, 35/27/5
74 br, It will air cure in 21 to 30 days from pour date.

for small runs to short production it works well and machine very well



SBI
Central Ne.,USA

 

[DKirkham]

SBI,

We cast about 100-150 aluminum differentials/year. We use 356-T6. I am very interested in other alloys and your company.

geesamand,

We have had NO trouble with our differentials. You can see what we use them for here on our website.

http://www.kirkhammotorsports.com

The original AC Cobra ran a cast iron differential and we changed it to aluminum. We did very little design changes, but we did thicken up the walls and ribs by about 20%. As you know thickness of a sample makes a tremendous difference on stiffness.

However, Aluminum looses strength at temperature quite quickly and differentials run much hotter than most people think--so you need to take that into consideration as you design your differential.

As for press fits on bearings, we tightened up our press fits by 0.0005-0.001 inches. We have never had a bearing race come loose and we use them under very demanding conditions.

David

 

[tbuelna]

geesamand,

Before accepting that a cast aluminum diff housing is more cost effective and lighter than ductile iron, you need to consider many things. Ductile iron is wonderful stuff for a gearbox housing. With modern foundry techniques, it can be cast very precisely and with very thin walls. And most importantly for gearbox applications, it is stiff, quiet and has a CTE close to that of steel bearings and gears. And if your product is cost sensitive, the price of ductile iron is low and stable. While the material cost of aluminum is relatively high and varies widely.

You did not provide much in the way of a description of the operating parameters of your gearbox, but if the bearings are highly loaded and/or operate at high DN, then you will need to provide bearing liners with a non-ferrous housing, to prevent fretting. This will add to production costs. Maintaining bearing fits/running clearances, and gear mesh geometry can be very challenging with a non-ferrous housing. Especially a housing that is required to operate over a wide range of temperatures.

 

[geesamand]

Thanks again all. This thread has been very helpful.

Now that the project has settled on a precise path (adapting the Ford 9" to a particular IRS platform) I am ready to push on. Clearly, our goal is cost-effective torque capacity. Light weight would be nice, but with the kind of capacities we're aiming for the extra weight of cast iron is not seen as a big obstacle. I suspect that commercially we'll be much better off if we develop the c/i option fully and see how it soaks into the market before re-engineering for aluminum.

Dave