Non Electronics Question on Materials

[raptor1956]

I'm looking at the best methods for making a board or plate that's about 230mm x 150mm and is both flat and stiff.  I need it to be flat so that the peak-to-peak variation from flat is no more than 30um.  It does not need to be super smooth, ie polished, but needs to be flat and free from warpage.  It is also desirable that a magnet can be attracted to it.

As a first pass I'm looking to make it from 3/8 inch (9.525mm) steel as that would be stiff enough, would work with magnets, and can be fly/skim cut for flatness, but perhaps 1/8 inch (3.175mm) steel with bends on all four edges and with the pocket filled by an epoxy coated bit of wood or MDF.

What I'm looking to make is a grid board to track the motion of a spot from an optical assembly so flatness is critical.

The advantage of the 3/8 thick steel is that it's simple and should require few steps to manufacture, but the downside is will be pretty heavy at about 3kg and supported as it is at the corners the middle will tend to sink under it own weight even if the movement is small.  The advantage of the thinner steel bent for stiffness is the material costs will be less, it will weight less, and cutting 1/8 inch thick steel with laser or plasma is easier, quicker and cheaper than metal 3X as thick.  The disadvantage with the thinner bent steel is greater number of manufacture steps and the greater risk to warpage though if the pocket is filled with epoxy coated wood/MDF that problem should be reduced.

So, any ideas/pointers here?


Brian

[Rerouter]

Use a thicker piece of metal. Plane top and bottom. Then mill out most of the backside to form ribbing

[raptor1956]

Use a thicker piece of metal. Plane top and bottom. Then mill out most of the backside to form ribbing

Yeah, that will certainly work but the additional machining costs would be prohibitive.  A similar idea would be to mold/cast the piece then fly/skim cut the top surface but again, the cost would be higher -- I think.


Brian

[e100]

Instead of paying someone to mill ribs, just use an electric hand drill to makes lots of shallow holes to give a honeycomb effect.

If you go for a composite construction then you'll need to match the coefficient of expansion of the materials otherwise it'll want to turn itself into a bowl whenever the temperature deviates from the temperature at which the final finishing cut was done.

[raptor1956]

Instead of paying someone to mill ribs, just use an electric hand drill to makes lots of shallow holes to give a honeycomb effect.

If you go for a composite construction then you'll need to match the coefficient of expansion of the materials otherwise it'll want to turn itself into a bowl whenever the temperature deviates from the temperature at which the final finishing cut was done.

Yes, temperature coefficients are important no doubt.  Making it out of a homogeneous chuck of metal avoids that and may well be the only practical way to achieve and maintain the flatness I need as temp changes while permitting magnets to work.  That does kind of point to casting if this is made in quantity.  A well designed mold would limit machining needed and cut the weight by a factor of 2 or more.

Thanks for the feedback...


Brian

[Conrad Hoffman]

Can you use a piece of plate glass? I'd think 1/4" thick would do the job if properly supported. Otherwise, 3/8" steel, with a pass through the surface grinder should do the trick. 30 um is a mile! BTW, a profilometer I use has a big glass reference flat, with a PTFE pad that slides around on it, for a reference surface. That's pretty big though, about 10" diameter and about an inch thick.

[splin]

Cast iron is the traditional material used for surface plates and you get the ribbing/stiffners for free. I believe that properly aged (no idea how long though) cast iron can be more stable than steel.

Granite is also commonly used as it is very stable and easily machined, but it won't be light.

[Benta]

Cast iron is the favorite for form stability and strength, but will be unnecessarily heavy for this application.

My suggestion would be cast Aluminium with a high silicon content ("Silumin"). It is as form stable as cast iron, not as strong, but much lighter.

For both materials, you'll need to temper after casting and machine the surface.

https://en.wikipedia.org/wiki/Silumin

EDIT: dang, I missed the magnet part. Sorry.

[kosine]

OK, I've done a little experiment that might help...

Gauge plate (aka ground flat stock) should work. Cheap and readily available in a wide range of thickness and sizes, plus it's normally magnetic. A piece the size you're after is probably about $50.

Typical flatness is within 50um, usually less (1/1000 inch, 25um), so it might be good enough as is. If it's not, then any machine shop with a surface grinder should be able to dress the surface for you. Alternatively just check the supplier's stated tollerances, you can find some flat to within 5um.

Thermal expansion of steel is about 9um per metre per degree C, so it shouldn't affect the surface flatness. (The width and length will change slightly, but not by much.)

As for the warpage, I've just clocked out a 250mm length of gauge plate on my CNC machine. It's 9.5mm thick but only 50mm wide (sorry, I couldn't find any nearer to 130mm at the required length). Laid flat on the bed, it shows no more than 40um deviation in thickness, which is within the spec.

I then supported it at both ends to measure the sag, and it's also no more than 40um difference along the the whole length. So doesn't look like it's really sagging at all over a 250mm span. If I press down on the centre I can increase the deflection by about 30um, but that takes quite a bit of force.

Just to be sure, I also quickly tested some 3mm plate, and it showed a sagging deflection of around 200um in the middle. (Wow! Interesting experiment!) So thinner plate definitely sags under it's own weight, which implies you need 8-10mm thick plate.

I've also weighed the 9.5x50x250mm plate and it's a tad over 1kg, so you're spot on at around 3kg. If you can live with that, then you have a possible solution.

For good measure I also checked for any vibration issues, but couldn't really measure anything. However if the light source is not rigidly attached to the plate, then it's something you might want to consider.

[ajb]

Forum member robrenz would be an excellent person to talk to about this.  I don't know how much time he spends here these days, so you might try PMing him.  (For the unfamiliar, check out his youtube channel, he does some pretty crazy high precision machining work.)

[Conrad Hoffman]

useful info for support: https://www.engineeringnotes.org/metrology/airy-points-and-bessel-points/

Will this be horizontal, or in different orientations?

Just a crazy thought, they sell a magnetic paint; could you paint a piece of plate glass so magnets would stick? It's hard to lay down a perfectly even coating of paint, but it could be sanded flat.

[jmelson]

Use a thicker piece of metal. Plane top and bottom. Then mill out most of the backside to form ribbing

Wrought steel has a lot of stress built into it.  If you are worried about sag from gravity on a 3 kg piece, then you REALLY will have a worry about built-in stress relieving over time.  What you likely want is either white cast iron that has been stress-relieved by heat treatment, or possibly Blanchard-ground aluminum tooling plate.

Milling out part of one side of a cold-rolled steel plate will end up with a LOT of warping, as the stress is mostly in the skin, and if you remove the skin on only one side, it will then warp over time.  Make it perfectly flat one day, and next week it will be badly warped.

I've done a lot of hand scraping to make precision tools, and you have to do special processes to relieve the stress, or you won't get a stable result.

Jon

[Conrad Hoffman]

Piece of Durabar? AFAIK, that's pretty stable. I recently ground an 8" disc and it seemed to stay flat.

[CatalinaWOW]

For modest cost and easy acquisition you might try a thin gauge plate resting on a granite surface block (or the plate glass that Conrad suggests).  The surface block prevents sag, and helps resist warp while the gauge plate gives you the magnetic properties you want.  Since they are not bonded together the differences in CTE don't matter.  The only real drawback to this approach is weight, and difficulty in using in orientations other than horizontal.  (Not impossible, just requires some thought on support.)

[thermistor-guy]

Can you use a piece of plate glass? I'd think 1/4" thick would do the job if properly supported. Otherwise, 3/8" steel, with a pass through the surface grinder should do the trick. 30 um is a mile! BTW, a profilometer I use has a big glass reference flat, with a PTFE pad that slides around on it, for a reference surface. That's pretty big though, about 10" diameter and about an inch thick.

Or a hybrid - glass plate fastened to a steel plate. Amateur astronomers have techniques for making large flat glass pieces by hand. The steel plate would satisfy the magnetic requirement.

[rhb]

You might want to look into what is called "jig plate" .  It's cast aluminum plate and is made that way to ensure that it stays flat when machined.

[raptor1956]

Wow, been a couple days but a lot of great info -- thanks.

I hadn't thought of cast iron -- I'll give that a look to see how available and at what price.

I also need to look into gauge plate.

The material needs to be capable of attracting magnets as I plan to use them to mount/hold various things and do so when the plate isn't horizontal. 

Granite is, of course, used extensively in industry for flat and level surfaces and is an obvious choice save for the magnet thing.


Again, thanks for the feedback.


Brian

[rhb]

In searching for jig plate, I encountered steel pegboard which might be of interest.  I'm not sure how flat it is though.

[calexanian]

Use a thick steel plate and have it Blanchard ground. That is how large surfaces like that are made flat.

[CopperCone]

how do you ask a place to blanchard grind steel? I have a welding table that is stuck in limbo because I had ordered some blanchard ground plates for it, but they failed to deliver, and I realized I was paying out the ass.

What kind of prices would it cost to do a 3x3 foot 3/4 inch thick blanchard ground plate? Or two plates, like a 3x2 and a 3x1 so I can have a small gap for magnetic drilling and clamps.

I never delt with a machine shop directly before.

[Dubbie]

It completely depends on the shop. Not many machine shops have a Blanchard grinder, it’s not really “standard issue” for most toolmakers. You’ll probably have to find a shop that specialises in grinding. You’re probably much more likely to find a shop with a large surface grinder.

If you are on a strict budget, you’re probably best to look for a second hand cast iron surface plate. A new 18 x 18” Chinese one will be about $500 so that’s a place to start budget wise.

[Conrad Hoffman]

As an example of a grinding house and steel supplier, this place isn't far from me- http://www.niftybar.com/ They do all sorts of waterjet cutting and seem to have little problem dealing with artists and individuals on one-off projects.

The problem with steel is shipping it. One possible solution in the US is you can piggyback shipments on normal Fastenal shipments- https://www.fastenal.com/en/22/3pl-(third-party-logistics)

[CopperCone]

yea some place in california screwed me over, I got half my order but not the other half, then I called, so they said they reshipped (claimed I was not there to pick it up), then it turns out it never got mailed because I did not pay for some kind of special carrier, despite the fact that they said it was fine...

that shit was a bait and switch, they mail half then try to get you to pay more for the other half of the order. Super obnoxious.

But it was nice cold rolled blanchard ground steel. I made my table so I had two ground flat plates that were about 3 inches by 3 feet on two legs, so I can shim them to level out a large 3x2 and 1x3 plate with like a 2 inch gap between them, then drill and ream holes in those plates and basically fasten them to the other plates with some machine pins with a slight flare on them.. I welded together the other parts out of various bits of steel and some bolts so you can disassemble it.

Also made nice removable casters compartments so the nuts used to fasten the casters are each enclosed in a separate section of 4x4 inch square stock.. and I brazed nuts on the inside of the stock so there are little access panels that you unscrew... the idea was to make it easy to clean.

I noticed that tons of welding equipment is built super bootleg in regards to dust control or ease of cleaning.. stuff like tubes that were left open and bolts that are exposed just drives me insane.... if you do any grinding or brushing or even non gas welding you get all sorts of dust pileups around exposed bolts, then it just joins together because of oil residue and you get this perpetual crust that just seems to exist on those tools.................................

[calexanian]

A quick google search turns up lots of shops that do Blanchard grinding.