The Journey for a Master Square

[Brobdingnagian]

While pottering about the workshop today I started to think about making a set of two master squares,
The best design I can imagine for this is a box section or square section of cast iron, since then they could be self proving to each other to the accuracy of a surface plate. However I do not have any box section lying about and seeing as it is a little hard to come by around here, I'm putting the call out: Where to get hollow square cast iron?
-Josh

[dave4]

What about a cylinder square - they are pretty useful.

[Michael G]

I'm not sure that you would get square box section unless you had it especially cast. Thinking about how the metal would solidify, I suspect that those inside corners would tend to distort when cooling, running the risk of throwing the whole section out of square. No manufacturer would want to make it because the size specifications would be so broad to get a saleable product (and people would be unlikely to buy it because they would have to machine so much off to get the finished size they wanted - which would then distort in unpredictable ways because of the uneven amounts taken off...)

Could you use round hollow section and square it up on the outside? Round on the inside would at least have cooled evenly when cast.

Michael

[Pete F]

It's a shame this didn't come up in a few weeks or I could show you the finished product. I'm about to do this myself, and have been giving it a little thought, however just need to get a couple of jobs out the way first. I intend welding a "U" shaped section, however creative minds may well be able to come up with some steel sections. Assuming it's to be ground and not scraped, there is no particular reason squares should be made of cast iron, and many of the commercial cubes or grinding fixtures I've seen aren't. Instead of closing in the top of the cube with another plate I will weld in a brace. Once welded I will pop it in the oven, stress relieve it, then grind it. While I'll drill and tap mine for fixturing, I can't see any good reason that it couldn't be used as a master square. I've seen others in 4140 and hardened, however I will simply use mild steel. More technical information in this field can be found in "Design of Weldments" if I recall correctly.

Pete

Edit: A search for "Grinding cube" or "Grind cube" will possibly reveal more useful information.

[Brobdingnagian]

What about a cylinder square - they are pretty useful.

I really like cylindrical squares, I think they are the absolute best for quick inspection but they have two problems for what I wan't from a master square
1. they are not self proving (at least I don't think they are? anyone?)
2. you cant transfer squareness from them to another part.

The main reason I'm interested in making a set of master squares is so that I can easily square all my second hand angle plates on a surface grinder. They are all out of square by enough to make me want to true them up.

I'm scraping one of the larger plain ones ~200x200x200 for the same purpose at the moment but without 2 more I have to rely on indicators and parallels to prove square.

Ideally I would love to make a set of 3 scraped master surface plates as well, you know just for fun. :D

[Ueee]

Why not look into a small granite square to scrape everything to? I'm not sure how Christian is going with the mob i put him onto, but i can get a 00 400x250x40 out of China and shipped to my door for $220.....and if the Xinxing surface plates are anything to go by they should be pretty good. Can't you test then with the Laser anyhow?
I was planning to be the guinea pig on buying a square from them and reporting back on what it is like.

[Pete F]

A cylindrical square is self proving by first confirming it has no taper and then rotating it against an indicator.

Both have their advantages and disadvantages. It's typically not possible to fixture directly to a cylindrical square, and they need to be used in pairs for that. On the other hand a cube can be made with tapped holes for fixturing and is (in my opinion anyway) more easily proved using just an indicator.

Pete

[.RC.]

Probably not that hard to make, might be involved though.... I have been considering making a toolmakers cube myself...

My initial thoughts have been along these lines...

fabricate an open ended cube out of 16mm mild steel plate.. place in kiln and anneal the cube.. Rough machine the cube to size and squareness, all holes slots as required....

Case harden cube in kiln... Next up is the final grinding.... Cannot take too much off otherwise you will go beneath the hard skin....

To get square you could use the same technique I used in the post where I described making a rectangular block...

I doubt you will get it fantastically near perfect on a generic grinder.... Probably need lapping for the finish, depending of course on what perfection you desire...

[Stustoys]

Hi Josh,
What size are you thinking?
Whats the biggest Z you can get on the grinder(if you are going to grind them)

I wonder if Interlloy really have 310x310 cast square bar?
You could drill out the center and make a set of blocks like mateyoshka dolls?
Its only 727.8kgs a meter.

Stuart

[Ueee]

Stu, if the pricing i have seen and been charged for Flocast is right, it will set Josh back about $1250 just for a 310x310x310 cube......

Why not fab one from flat bar CI if you really want CI? Weld, stress relieve and then you can grind/scrape to finish. You really only need one cube and 2 angle plates to make them all square.

[Pete F]

Just for info, a cube can be proved using only an indicator, no other angle plates etc are needed.

Pete

[jhovel]

So you mean something like this, Josh?
Angle Plates,Box Angle Plates,Slotted Angle Plates (http://www.luthraprecision.com/angle-plates.html#box-angle-plates)

[Brobdingnagian]

Sorry I have been in the swing of scraping....

Yes Joe, that is what I have in my head.

The process I intend to make them is a little convoluted but it is the same idea as making a set of master plates by comparing 3 plates to each other until you get an even spotting patten, same goes for an angle plate you use 3 plates with the surface plate being you datum for flat. A cube box can be done with just 2 since you initially effectively have 4 faces, hence the pair. The other reason for the pair is so that they can spotted against each other for calibration latter on down the track with no need for a gauge.

I was thinking that I would take a few shortcuts with the surface grinder so get them pretty close.

and like pete said if the faces are parallel then I can also do a simple reversal test for square with an indicator.

Reason for cast iron over granite or mild/tool steel is just so that it can be easily trued by scraping. Also it would stick to the mag chuck.

And if push comes to shove I could always cast my own, but they would have to be a little smaller than what I had in mind which was about 200mm cube, and I would also have to get over the imperfections in the casting.

The Surface grinder will do about 305Zx300Y600X. so 200mm seems like a nice size for both it an the mill.

at $1250 for a solid 310mm cube kind of puts it out of contention. not to mention the weight.



So it looks like a special casting then. Nobody sells square hollow sections???

[Pete F]

Nobody sells square hollow sections???

You may consider importing a box parallel from the US, and regrinding it to tighter specs. If you're prepared to settle for a smaller size the freight should make it feasible. I don't know if they still sell them but Shars sold a 5 x 5 x 5 for US$58 weight is 17 lbs. Check their catalogue under 202-4361 as I doubt you could buy raw stock for that.

Pete

[Brobdingnagian]

Thats a great idea, I will have a look see.

[Stustoys]

I wonder if someone like this would ship you a couple "as cast" or roughed?
Cast Iron Hollow Box Parallel - Iron Hollow Box Parallel and Cast Iron Swivel Angle Plates Supplier & Manufacturer from Indore, India (http://www.jashprecision.com/cast-iron-hollow-box-parallel.html)

$1275! I guess thats to to bad "by the kg" ;)
And I hadnt thought the wieght issues.

Wouldnt they weight to much for you to cast? or did you mean making the pattern and getting it done else where?

Stuart

[Brobdingnagian]

the 10" version of these would be ideal. just need to find some crappy imports for a reasonable price

Taft-Peirce BOX PARALLELS by Suburban Tool, Inc. (http://www.subtool.com/tp/9160-64_taft-peirce_box_parallels.html)

[Brobdingnagian]

I wonder if someone like this would ship you a couple "as cast" or roughed?
Cast Iron Hollow Box Parallel - Iron Hollow Box Parallel and Cast Iron Swivel Angle Plates Supplier & Manufacturer from Indore, India (http://www.jashprecision.com/cast-iron-hollow-box-parallel.html)

$1275! I guess thats to to bad "by the kg" ;)
And I hadnt thought the wieght issues.

Wouldnt they weight to much for you to cast? or did you mean making the pattern and getting it done else where?

Stuart

yes and no. with enough iron, air and wood nothing is too big.. :P but our existing setup I could make a say a 100mm hollow cube. maybe 150mm, I would have to do the math. I will say the last piece of cast iron we (Ray and I) did was small (~150x100x30) and exciting, but the casting had quite a few defects, none that hurt the piece, but still... Now that I think about it we have bumped up the size of the R2D2 furnace since last time.

I might also consider getting it cast from my own pattern, If I knew a place that did cast iron, and the price was ok.

It would be much cooler to go from scrap to precision tool.

-Josh

[Stustoys]

It would be much cooler to go from scrap to precision tool.
Yes, very cool indeed.

[Dave J]

What about getting a little Douglas shaper and taking the table off it and use that. They can be had for around $300 and you could sell the rest as parts and easily make your money back or more. You would also have the benefit of having T slots for mounting it or gear to it.

It would probably be cheaper than buying something that size and the tables are not that heavy.

Dave

[Brobdingnagian]

What about getting a little Douglas shaper and taking the table off it and use that. They can be had for around $300 and you could sell the rest as parts and easily make your money back or more. You would also have the benefit of having T slots for mounting it or gear to it.

It would probably be cheaper than buying something that size and the tables are not that heavy.

Dave

Another great idea. I'm not sure I would have the heart to do it though. I would end up with two more shapers :)

[Dave J]

I think the guys name is Mal who Pipeclay knows that takes apart Hercus lathes and sells them as parts. You need the boxes and they are cheap, I know some times it's hard to do but if it's for a good course why not.

Dave

[jhovel]

What does the foundry (Furphy?) in Shepparton charge per kg if you provide the pattern? There is an iron foundry in Castlemaine with a good reputation as well.
I'm with Stu - get it cast the way you want it.

[Brobdingnagian]

Furphy don't cast anything any more they import it all, now days they are just a brand. :(

Castlemaine sounds promising anyone got stuff done by them?

[Brobdingnagian]

Just to get the idea of the weight. this design is 200x200x200mm and is 29kg and the raw casting ~35kg
255924

[Steamwhisperer]

Hi Josh,
we use Grigsbys foundry in Ballarat. I have found them pretty good.

Phil

[bollie7]

What about getting a little Douglas shaper and taking the table off it and use that. They can be had for around $300 and you could sell the rest as parts and easily make your money back or more. You would also have the benefit of having T slots for mounting it or gear to it.

It would probably be cheaper than buying something that size and the tables are not that heavy.

Dave

Or pick up an old, stuffed lathe and cut up the bed.
bollie7

[Machtool]

Castlemaine sounds promising anyone got stuff done by them?
Would Joe be thinking of the Old Thompson Kelly and Lewis site. Its now a division of the American company Flowserve.

I dont think they cast anything any more. Marko and I spent some time there over a year ago, moving a handfull of machines they wanted to keep into one bay. The third bay back from the main road. Every thing else seemed to be sold off. I didn't see any casting going on there. All the pump bodies seemed to be imported. The plan was to sell the front two building off. Very tricky because I think they are heritage listed.

Seemed to be only a few staff left, 30 - 40. I'm told they used to employ over 400.


Phil.

[.RC.]

Why not send giesser a PM, He might be able to do it http://www.woodworkforums.com/members/91734-giesser/

[Ueee]

There is also the wagga foundry Wagga Iron Foundry - Welcome (http://www.ironfoundry.com.au/)
For some reason the figure of $26 per kg is in my head, not sure where it came from though. That's 5 times the price of flocast.

[Neil317]

I have almost finished a pattern for a 200 x 300 x 50 (ish) CI square. It should be on its way to a foundry next week. Not yet sure of the price each, or which foundry the patternmaker uses.
Neil

[CGroves]

Why not look into a small granite square to scrape everything to? I'm not sure how Christian is going with the mob i put him onto, but i can get a 00 400x250x40 out of China and shipped to my door for $220.....and if the Xinxing surface plates are anything to go by they should be pretty good. Can't you test then with the Laser anyhow?
I was planning to be the guinea pig on buying a square from them and reporting back on what it is like.

I haven't chased this up further. I acquired a Wyler 0.02mm magnetic triangle square so the granite will have to wait.

Christian

[Machtool]

I acquired a Wyler 0.02mm magnetic triangle square
Christian

Where are you blokes finding those levels? I have a search set up for those, Tesa & Wyler. I get Ebay spammed daily, but never see them.

R.C, Greg Q and now you have found them.

Regards Phil.

[CGroves]

I found it on a UK site advertised as a "half box level".
USED METROLOGY EQUIPMENT SUPPLIERS USED ENGINEERING TOOLS TRIMOS TESA SUPPLIERS WESTERN TOOLING BRISTOL UK (http://www.radisol.com/western-tooling/index.htm)

I sent an email to ask if it was magnetic and triangular and luckily it was. It was from 2007 and I paid less than what "hotwheelinmum" was asking :-).

I'll post a picture in the euro tool thread.

Christian

[Brobdingnagian]

Neil,
I like your pattern, I would be interested to know where it ends up being cast.


Casting it myself seems the most appealing, second most appealing is getting someone else to do the fun part. Third is finding a crappy set of blocks and getting them up to spec, fourth is chopping up a machine beyond rescue.

I have sent off inquires to three different foundries, Grigsby, Wagga, and Billmans (Castlemain). I will see what comes back. If the price is too high it might be a trip down to the scrappy to find something I can make a cupola out of...

I did a quick FEA on the square and it will take a 100kg sideways force to push it out of square by 1.7µm, the box alone without the struts would move 0.04mm with the same loading.


PS Added a couple of handles.

255955

-Josh

[Brobdingnagian]

Grisbys got back to me, their prices are very reasonable, enough so to make it a feasible option, cheaper than buying an old worn out block and shipping it.

[Pete F]

Now just to wait a few years for the cast to stop moving :wink:

[Brobdingnagian]

Now just to wait a few years for the cast to stop moving :wink:

about 30-40 years should do it.. :P

Serious though, how long do you reckon they should be seasoned for in the heat treatment oven? a few days, a week, a month?

[Neil317]

Hi Josh, try this:

http://www.dtic.mil/cgi-bin/GetTRDoc?AD=AD620556&Location=U2&doc=GetTRDoc.pdf

Stress relief of grey cast iron.
An old US Navy document.
Neil

[Pete F]

about 30-40 years should do it.. :P

Serious though, how long do you reckon they should be seasoned for in the heat treatment oven? a few days, a week, a month?

It depends on what temperature you take it to. Basically the hotter the temperature, the shorter the cycle and the more complete the stress relief. However (oh but so often there's so often a "however" :p), IIRC there can be some be some metallurgical changes at the higher temperature that may not be desirable. I recall reading a paper on this some time ago, it's possibly on my machine, but I have no idea how to post a pdf. Around 600 C for 6 hours (at full material thickness) with a slow cool rings a bell, but don't for goodness sake quote me on that. I just remember the 600 for 6 rule.

I still have half a cuppa, and not much argon left, so I'll try to chase up the paper and then either see if I can either post it or find it on the net to post the link.

Pete

PS not trying to on any parade, nor doubt your casting ability, but if this is a master square you're building here, and based on some of the tolerances I've seen (admirably I might add) achieved and being called for, I wouldn't suggest casting your own. Anyone who has gone through the process of manufacturing or procuring precision cast references, such as the straight edges we bought a while back, knows that "iron ain't iron" when it comes to this field. My 2 cents for what it's worth, but hey, what would I know.

Edit: Oh Neil beat me to it, I'll need to type faster/drink slower. Yes I think that was the paper I had in mind.

[Brobdingnagian]

I think you are right, a professional foundry has a better control of the quality of the casting than I do.

Cool thanks for info guys, I'm also having a re-read of the cast irons sections of the Heat Treater's Guide - Practices and Procedures for Irons and Steels by The Materials Information Society.

-Josh

[Pete F]

'm also having a re-read of the cast irons sections of the Heat Treater's Guide - Practices and Procedures for Irons and Steels by The Materials Information Society.

Out of curiosity would you mind posting what is recommended there, as it's nice to file away. Although my oven is good for 1200 C I don't like the idea of taking it, or the material, and hotter or longer than I need to.

Pete

[Brobdingnagian]

Out of curiosity would you mind posting what is recommended there, as it's nice to file away. Although my oven is good for 1200 C I don't like the idea of taking it, or the material, and hotter or longer than I need to.

Pete

For low alloy Grey Iron with 2.5% to 4% carbon and 1 to 3% silicon.

Summary:
Ferritizing 700-760C cool at 110C-290C/hr

Full annealing: 790-900C furnace cooled from 790 to 675C

Graphitizing: 900-955C hold for several hr/inch
For maximum strengh air cooled to 540C
For maximum machineability furnace cooled to 540C
For both cool from 540-290 at rate no more than 110C/hr

Stress reliving with minimum carbides:
540-565C for stress relief > 85% use 595C soak minimum 1hr/inch
furnace cooled to 315C or lower to as much as 95C for intricate parts

Normalizing:
885-925C soak at 1hr/inch furnace cool

[Brobdingnagian]

Another Rendering.
255982

[jhovel]

Another Rendering.
255982

#### that's good! Thne rendering even has a workshop background.... How the devil do you do it?
Can your CAD program tell you what it weighs?

[Brobdingnagian]

Yeah the software can give me the weight and volume centre of mass, moments of inertia, thermal deformation etc. It started just a straight box section but the FEA showed it was to flimsy.

I also modelled up the pattern so that I could send it off the the foundries to get an accurate estimate on moulding and casting along with the weights and volume of the casting excluding the delivery system.

Attached is the 3D pdf of the split pattern, 2 degree draft, all edges filleted and a dimensions changed to allow for machining and shrinkage.

I think I said it earlier in the thread. The raw casting would be ~35kg and one finished square would be about 29kg

Some Impractical storage:
255989

[Brobdingnagian]

it needed a name plate
256013

actually it was just was just because I was reading the reading the etched bass name plates thread

[Stustoys]

Hi Josh,
You think it would be easy to control with handles like that?
Would some sort of T of D handle make things easier?
Oc course I dont have to design it :D

Stuart

[Brobdingnagian]

Thanks Stuart, Tee or Dee handles are a much better idea.

Ok so I decided last night that the prices to get it cast at a processional foundry are good enough to do it that way. Next step is making the pattern.

Design is not the problem making a pattern to match the design is. I always seem to have a problem getting the draft and fillets right at corners.

I'm gonna have a go at roughing out the pattern today in between other things. Any more fantastic ideas before I commit to physical form :)

My first step will be to make the 2 degree trapezoidal shape that will be the cross section of the external and internal "walls".

-Josh
'

[Ueee]

Can you make the pattern in machineable wax and cnc it?

I'd be interested to know what the pricing is like, if you don't want to post it I'll talk to you about it tomorrow.

[Brobdingnagian]

It is just a fraction to big for the little cnc. :(

But I also had the idea of making 2 more sets of patterns for a set of 150mm and a 100mm squares of similar design, and I could knock them out on the cnc out of machineable wax and then cast aluminium masters patterns, as the wax is not so good for multi time uses.

I don't think I should post the quotes, it does not feel ethical. I'm happy to tell you in private just not on a public space. I will say that it is much cheaper than buying new stock and machining the excess away.

-Josh

[RayG]

Hi Josh,
You think it would be easy to control with handles like that?
Would some sort of T of D handle make things easier?
Oc course I dont have to design it :D

Stuart


Hi Stuart,

Good point, I'm not entirely sure why handles are needed, Maybe temperature effects, but this is primarily for fixturing, not gauging... but I'm thinking they might just get in the way for clamping different jobs, on that basis I'm thinking removable handles might be better?

Regards
Ray

[Stustoys]

Hi Ray,
I assumed the handles would be removable:B. At 30kgs handles would be nice, I'd hate to drop one(maybe a threaded hole for an eyebolt and a hoist would be better for old guys like me). Though if a D or T handle is used, depending on how close those ribs come to the ends you might be able to get them below the surface? *

Stuart

*You could even cast them in for that extra pattern/core making challange

[Brobdingnagian]

Hi Stuart,

Good point, I'm not entirely sure why handles are needed, Maybe temperature effects, but this is primarily for fixturing, not gauging... but I'm thinking they might just get in the way for clamping different jobs, on that basis I'm thinking removable handles might be better?

Regards
Ray


Hi Ray,
I assumed the handles would be removable:B. At 30kgs handles would be nice, I'd hate to drop one(maybe a threaded hole for an eyebolt and a hoist would be better for old guys like me). Though if a D or T handle is used, depending on how close those ribs come to the ends you might be able to get them below the surface? *

Stuart

*You could even cast them in for that extra pattern/core making challange

Handles are there for all those reasons, as well as to keep my hands off the surfaces as much as possible. The way I see it the masters would not come out often other than to grind up other fixtures, like angle plates.

Handles will be removable on M12 thread to suit standard eyebolts or whatever you can put a M12 thread on, it also could be slung though the "cores".

30kg is a lot to lift up, it is about the same as our compound sine table and that is a real trial to move around carefully.

The other reason is that it will wring strongly to the mag chuck so a little extra leverage will be handy.

-Josh

[Michael G]

Just for info, a cube can be proved using only an indicator, no other angle plates etc are needed.

Pete

I've been trying to think how to do this and can't (parallel no problem but can't work out the right angle). Can you enlighten me please?

Michael

[.RC.]

How to Check if Your Solid Square is REALLY Square - YouTube (http://www.youtube.com/watch?v=SwHlyFo6FLo)

[Brobdingnagian]

How to Check if Your Solid Square is REALLY Square - YouTube (http://www.youtube.com/watch?v=SwHlyFo6FLo)

Currently re-reading (with more focus this time) Connelly Machine Tool Reconditioning chapter on automatic generation of gauges.
No section on squares though. I do remember reading it some where :S....

[Michael G]

The you-tube clip is interesting but relies on something else being square. I know that squares can be generated much like surface plates using a flat surface and three squares, but checking with a dial indicator has me flummoxed. The only way that I can come close to a way is with an indicator stand that will allow the indicator to travel vertically but once again that relies on the direction of travel being perpendicular to the surface.

Michael

[Pete F]

Currently re-reading (with more focus this time) Connelly Machine Tool Reconditioning chapter on automatic generation of gauges.
No section on squares though. I do remember reading it some where :S....

Put the cube on a plate and confirm top and bottom are parallel. If using a cylindrical square it's normally easier to do this with a micrometer (ie checked for taper). As they say in Thailand, Same Same ... but different. Turn the cube through 90 degrees so those faces are now perpendicular to the plate. Indicate one face, with the base of the stand up against the bottom of the cube and indicate the top. Spin it through 180 degrees and do the same for the opposite side. If the readings are the same the 3 faces are square with respect to each other (ie bottom on the plate, and the 2 faces just checked). You work your way around the cube that way. Squares are checked the same way, except a "bridge" needs to be built to butt up against. It doesn't take long to check and will be as square as your best measurement system.

Pete

[Brobdingnagian]

If you can check for parallel/ or create a parallel surface to indicate off, you can then use the reversal method with an indicator to determine squareness.

If your unclear as to what I mean I will take some photos/video

[.RC.]

The you-tube clip is interesting but relies on something else being square.

Michael

Sorry, I linked to the wrong video..... I did not check the video as at the moment it is raining like mad and my net connection goes to #### when it rains here so I just got the link from google instead..

Try this video Checking a Block For Squareness, Part II - YouTube (http://www.youtube.com/watch?v=oh3tY_WDdg8)

[Michael G]

Thanks, Pete, Richard
I understand now - I have an old cast iron "square" that should be but I've never been able to satisfactorily check. I should be able to adapt this method to check that for squareness. I could check it with a cylindrical square but I'm wary about checking a plane with a line contact. While in theory it will work it doesn't tell you how square and it doesn't take much of a gap to show some light.

Michael

[CGroves]

It is just a fraction to big for the little cnc. :(

But I also had the idea of making 2 more sets of patterns for a set of 150mm and a 100mm squares of similar design, and I could knock them out on the cnc out of machineable wax and then cast aluminium masters patterns, as the wax is not so good for multi time uses.

I don't think I should post the quotes, it does not feel ethical. I'm happy to tell you in private just not on a public space. I will say that it is much cheaper than buying new stock and machining the excess away.

-Josh


I might be able to help you out with that. Jelutong machines really nicely and my CNC cutting envelope is 450x, 250y, 350z.

Christian

[Stustoys]

While in theory it will work it doesn't tell you how square and it doesn't take much of a gap to show some light.

Hi Michael,
I'm led to believe you just need a couple of pieces of shim/feeler guage the same thickness. One at the top one at the bottom. If they "feel" the same it will be damn close.

Stuart

[Brobdingnagian]

Not often you need very square, But for a master....

Method 1.
Normally if I'm squaring up a block on the surface grinder I will be able to get to within 0.04mm using long wide brass shims like the picture bellow.
I grind the block parallel I then measure the block, determine how much it needs to be corrected find closest shim smaller than the error and then set the block up on the surface plate and move the shim further in or further out from the centre to get a finer degree of adjustment. I mark the corners of the block on the shim and transfer everything over to the surface grinder, setting the block aligned to the markings on the shim. Grind and test. Generally it works well on parts that are rigid enough not to be distorted, and takes not much more time than it does to measure squareness.
256216

The limitation with this method is I'm limited to how small the error can be corrected being a little less than the thickness of my finest shim. So if I was out by 0.02mm and that was not good enough I would have to correct another way.

Method 2. For smaller and non rigid parts that need a fine degree squareness instead of shims I use a 10" sine table with a specially ground gauge to cancel the error and with a little care on surface plate setup and measuring I can correct most squareness errors. The limitation with method 2 is that the sine table it self takes ~150mm of travel off my Z axis and adjusting a fine amount of error is a tedious endeavour. Plus the cleaning of the sine table after grinding is time very consuming.

Method 3 is one that Machtools told me about today. it involves making a fine step with the grinder parallel with the error, and the adjustment can be as fine as the down feed on the grinder in this case about 2µm. it is akin to having a set of 2µm shims. I thought it was a very neat solution and would not take long to do.

256217

Thanks for that one, I will be putting it to use for these master squares ;)

-Josh

Edit: Anyone got more techniques?

[Brobdingnagian]

I might be able to help you out with that. Jelutong machines really nicely and my CNC cutting envelope is 450x, 250y, 350z.

Christian

Seriously, That would be awesome.

[CGroves]

Seriously, That would be awesome.

Josh,

PM me with an IGES file of the square and i'll take a look to see that I can do it. Seeing the square is symmetrical its probably easiest to mill each half separately and then fix them together.

Christian

[Machtool]

Josh your pictures are far better than I'd do, scribled out on the back of a cigerrete pack.

http://www.woodworkforums.com/attachments/f65/256217d1362205757-journey-master-square-block-step.jpg

Just to expand. I meant to comment when R.C was knocking up blocks. If you can measure the squarness error, easy way to corect it is to step grind it out. And the accuracy is only dependant on the down feed accuracy of your grinder.

No offence Rich. Rather than hollow cut your blocks, them lap one land down, grind one off and then leave a land.

For the sake of an urgument your error was 1 thou out of square. You have to invision which way the square has to move to inprove it.

I'd take a dust cut to establish wheel to work height. Thermal stabilty and all that. Then take a thou cut from the side that needed reducing, until you leave a high land. Ie: dont complete the pass, but leave a high land. 3 -4 - 5 mm wide.

Then flip that over. Tipping it on the as Josh's renderings. Your high land will tip the part over correcting the out of squarness. Then flip flop & grind off the land and get back to parrallel.

I have trouble descibing it, it takes me seconds when I'm in front of a grinder.

[Stustoys]

If it's symmetrical do you even need to make 2 halves?

Stuart

[Brobdingnagian]

If it's symmetrical do you even need to make 2 halves?

Yes and no,

I was going to cast the single hand made half Pattern in Aluminium twice and then that would become the Master Pattern to be then used to make the molds for casting in iron.

And I have allowed for double shrinkage; once for the aluminium and once of the iron and there is no guarantee that the two halves would be mirror images. My thinking was that I would align the two aluminium halves as best as possible and bolt them together and then machine/clean up the Master Pattern together as one unit.

On the plus side the pattern would be extremely durable, and would be able to take very nice detail

But with the possibility of them being milled on a CNC a lot of the variability goes down between making 2 separate halves. If it was only ever just me getting it cast then a painted wooden pattern would be sufficient for the three casting I need. If other people wanted to have it cast? then I would make an aluminium master.

-Josh

[Brobdingnagian]

Josh your pictures are far better than I'd do, scribled out on the back of a cigerrete pack.

http://www.woodworkforums.com/attachments/f65/256217d1362205757-journey-master-square-block-step.jpg

Just to expand. I meant to comment when R.C was knocking up blocks. If you can measure the squarness error, easy way to corect it is to step grind it out. And the accuracy is only dependant on the down feed accuracy of your grinder.

No offence Rich. Rather than hollow cut your blocks, them lap one land down, grind one off and then leave a land.

For the sake of an urgument your error was 1 thou out of square. You have to invision which way the square has to move to inprove it.
256252
28 somthings in this case
256253

I'd take a dust cut to establish wheel to work height. Thermal stabilty and all that. Then take a thou cut from the side that needed reducing,
until you leave a high land. Ie: dont complete the pass, but leave a high land. 3 -4 - 5 mm wide.
256254

Then flip that over.
256255
Tipping it on the as Josh's renderings. Your high land will tip the part over correcting the out of squarness.

Then flip flop & grind off the land and get back to parrallel.
256256
I have trouble descibing it, it takes me seconds when I'm in front of a grinder.

Team work :)

[Machtool]

Team work :)
Do you think any one will believe us? Its such a good thing this isn't an international forum.

In full discloser. This method was spelt out with just the use of a beer coaster, and no available pen.

[.RC.]

Just to expand. I meant to comment when R.C was knocking up blocks. If you can measure the squarness error, easy way to corect it is to step grind it out. And the accuracy is only dependant on the down feed accuracy of your grinder.



I did not have a clue how to square up a block to get it right and relied on information on PM to work out a way.... Did not think of step grinding like that...

[Ueee]

Do you think any one will believe us? Its such a good thing this isn't an international forum.

In full discloser. This method was spelt out with just the use of a beer coaster, and no available pen.

Even i understood it with the beer coaster.....and i don't own a grinder and i was trying to control Mr too much energy and long car trip ahead.....

[Brobdingnagian]

In aid of my search for a better square, Ray earlier in the day extracted a Sony 0.5 micron indicator and display from the turbine test rig he picked up a while back. see picture

256454

So I need to fabricate a stand to for it to as it is too large for any of the existing stands. I'm hoping someone has a bright idea for a dedicated indicator stand to suit for testing square by reversal.

Thinking about it I think I have some 3" dia mild steel bar somewhere around here, it squared off and a V cut into the top might serve that purpose. what do you guys think?

[Brobdingnagian]

Do you think any one will believe us? Its such a good thing this isn't an international forum.

In full discloser. This method was spelt out with just the use of a beer coaster, and no available pen.

I don't think it the first time that a beer coaster has been employed in a educational capacity and I don't think it will be the last. :)

For all the illustrations and descriptions there is still a lot to be screwed up if your not paying careful attention to what you are doing and the orientation of the block or whatever in my case an it was angle plate. I parallel ground the sides of the angle to gauge the squareness by reversal, great worked no porblems +0.20mm +0.04mm so 0.16mm/2=0.08mm out of square and my depth of cut on the grinder. I worked out which way it had to go and put a dot on the corner where to put the the step, I put it on the grinder and ground the step where my mark said to, or so I thought. I ended up grinding the wrong square( rotated by 90 degrees, but it was not until I put it back on the table to measure and thought for 5 minutes before it clicked what I had done wrong. Now I'm marking where I need to leave the step with a line....

-Josh

[Michael G]

Josh, with respect to the Sony indicator, will you really be measuring to the 1/2 micron or will you be measuring thermal effects, vibration etc? You've probably seen the variations you get with the laser.
The other thing with trying to chase things to that degree of precision is Cosine error on the stand. I think I'd be looking at some masking tape for the last digit or two...

Michael

[Steamwhisperer]

Josh, with respect to the Sony indicator, will you really be measuring to the 1/2 micron or will you be measuring thermal effects, vibration etc? You've probably seen the variations you get with the laser.
The other thing with trying to chase things to that degree of precision is Cosine error on the stand. I think I'd be looking at some masking tape for the last digit or two...

Michael

Hi Michael,
I reckon if I was Josh I would be doing the same purely because I can, but I'm not, so, I would probably put tape over the last six digits. :D:D

Phil

[Stustoys]

Hi Josh,
If its a 0.0005mm indicator, why is the last number an 8?
How high is the stand going to be? 200mmish was it? Surely there is some sort of rule of thumb re base v height? Maybe find a cylindrical square the same height and see what diameter it is?


Hi Michael,

Given the distance over which Josh will be measuring, wont Cosine error be a couple more places the the right?
Even if its not, it should be repeatable so after the first few goes Josh should know what downfeed is needed for what error, even if the error isnt strictly an absolute measurement.


Stuart

[Brobdingnagian]

Hi Michael,
All those things are a real pain at this level but not as much as you might think, over the short time of the measurement thermal effects are minimal as it takes quite a long time for enough heat to be transferred to something with as much mass as these to change the geometry at micron level and the change that I have modelled seems like it will be reasonably uniform as long as the temperature gradient is not too steep. Cosine error is not an issues, there is more variability in the grinder downfeed than there is cosine error. From doing the 200mm angle plate the most error comes from dust on the part, surface plate, the mag chuck and dirty hands. But I can handle that just by being very careful and clean (super clean).

Stuart, the 8 is just an artefact from the photo, it should have been a zero. The stand would be for the 200mm blocks so I guess a height about 190mm would be ok. I wonder what that relation is, I will have to look it up.


The nice thing about the indicator for this purpose is its range of measurement 0-26mm and ease of reading, the display tracks the maximum and minimum values (no chance the kind of error from miss reading the small totaliser clock complication (sorry I can't think what they are called)) , its absolute accuracy wanders a little in the 5-10mm range to a maximum absolute error of 3µm. But, as Stuart says its more about the relative measurement. The other thing to note here is that all the indicators have a degree of repeatability and hysteresis at the finer end of this level. At the fine end of the scale a comparator would be better for the job because of its high degree of repeatability and low hysteresis but its a case of using what is available, and this was essentially a freebie since Ray only wanted the pelican case the test rig came in.

As far as motivation goes, I will probably only ever make one set of master squares from scratch so to speak, so I'm pushing my boundaries as far as they can go.

-Josh

[Pete F]

Yes and no,

I was going to cast the single hand made half Pattern in Aluminium twice and then that would become the Master Pattern to be then used to make the molds for casting in iron.

And I have allowed for double shrinkage; once for the aluminium and once of the iron and there is no guarantee that the two halves would be mirror images. My thinking was that I would align the two aluminium halves as best as possible and bolt them together and then machine/clean up the Master Pattern together as one unit.

On the plus side the pattern would be extremely durable, and would be able to take very nice detail

But with the possibility of them being milled on a CNC a lot of the variability goes down between making 2 separate halves. If it was only ever just me getting it cast then a painted wooden pattern would be sufficient for the three casting I need. If other people wanted to have it cast? then I would make an aluminium master.

-Josh

If it's just a few castings, all of which are going to be subsequently machined so don't need to be perfectly identical, personally I would just use sacrificial foam patterns. I don't see where you would need fine detail for this purpose, although there are foams that will allow some. They allow more flexibility in design as you don't have to consider how the pattern will draw. I'd expect the foundry would love you too; no cope and drag to contend with.

Pete

[Brobdingnagian]

If it's just a few castings, all of which are going to be subsequently machined so don't need to be perfectly identical, personally I would just use sacrificial foam patterns. I don't see where you would need fine detail for this purpose, although there are foams that will allow some. They allow more flexibility in design as you don't have to consider how the pattern will draw. I'd expect the foundry would love you too; no cope and drag to contend with.

Pete
How well does that work with iron?

[.RC.]

Is there going to be a group order for the castings?

Can I put my hand up for 1?

[Brobdingnagian]

I made a little progress this afternoon with the stand for the Sony indicator.

I still need to make a clamp to hold the indicator in place and mill the slot for the "curved blade". I was hoping to pick up some 100mm or 125mm BMS bar but no one around town had any. :( even though it is very rigid I might have to clamp it down to the surface plate to get good readings.

-Josh

256661

256662

[Stustoys]

Hi Josh,
Could you bolt the radius piece to the bottom?
I have a 12ishmm thick piece of 120mm cast bar in the shed if that would work for you.

Stuart

p.s. When I said one half pattern I was thinking they would just mount it to a flat plate. Cant remember what thats called. Two halves might be easier for them for all I know??

[Brobdingnagian]

Is there going to be a group order for the castings?

Can I put my hand up for 1?

No group purchase yet.

When I put in a request for quotes, I asked about any volume discounts, there was none. So there is no real incentive to get a whole bunch made.

I'm getting 3 cast for myself so that I can scrape the last few microns in and make a true and square set of masters. But that's just me.

If think you might want a casting let me know via a PM and we will sort something out. Maybe direct payment to the foundry or something like that.

-Josh

[Pete F]

How well does that work with iron?

Don't know as I've never had any done. But my mother's side of the family are 4+ generations of pattern makers and I've spoken with my uncle in the past about sacrificial foam patterns, how they're affecting the pattern making business etc and he's never said they don't work with iron. I personally can't see why it should make a difference what metal is being cast, I think the flasks and sand are the same right? If it were me I'd certainly be asking the foundry, as making a decent pattern is a job in itself.

Edit: Here's a bit more information on the process, I'm sure a bit of further research would provide more specifics on preparing the pattern
Lost-foam casting - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Lost-foam_casting)

BTW Josh, I know you are going to stress relieve this, but I keep meaning to mention that you may want to run that render/design past a pattern maker or foundry and specify precisely what its intended purpose is. To my eyes the design appears to be very highly stressed as cast. But what would I know.

[Brobdingnagian]

Got a couple of more things done on the sony indicator stand.

Just the indicator clamp to go, but that will have to wait till tomorrow now.

256679

The blade is a close fit and I have made it to suit 8.5mm thick parallels if i ever need to swap it out for some reason also added 3 cup grub screw to keep every thing snug.

-Josh

[Brobdingnagian]

Don't know as I've never had any done. But my mother's side of the family are 4+ generations of pattern makers and I've spoken with my uncle in the past about sacrificial foam patterns, how they're affecting the pattern making business etc and he's never said they don't work with iron. I personally can't see why it should make a difference what metal is being cast, I think the flasks and sand are the same right? If it were me I'd certainly be asking the foundry, as making a decent pattern is a job in itself.

Edit: Here's a bit more information on the process, I'm sure a bit of further research would provide more specifics on preparing the pattern
Lost-foam casting - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Lost-foam_casting)

BTW Josh, I know you are going to stress relieve this, but I keep meaning to mention that you may want to run that render/design past a pattern maker or foundry and specify precisely what its intended purpose is. To my eyes the design appears to be very highly stressed as cast. But what would I know.

It is an interesting process I'd admit and have been interested in trying it in my own foundry with bronze. But I still need to make the patterns and at the end of the day I am more familiar and articulate with making shapes out of wood and metal than polystyrene.

I have sent it to the pattern design to the foundries and what its intended purpose is and asked them to quote accordingly on the moulding and casting.

You will have to clarify as to what you think exactly would be the problem, as I can't see it.

-Josh

[Pete F]

What you have is essentially a square and wide wheel. In CI I believe the narrower "spokes" of the design will cool before the "hub" and "rim", the stresses thus produced can be very high, and is the reason many wheel castings fail at this point.

I'm not entirely certain what the purpose of the hub actually is? It seems to me to be redundant and would add considerably to the cost. I would personally be more inclined to look at a design that was basically a circle sitting inside a square (obviously extruded) if that makes sense? The hollow circle would make the design significantly lighter, yet form a web to resist racking forces. It would also largely avoid many of the casting stresses I alluded to above. Even after heat treatment, any unequal residual stresses in your design will be trying to distort the overall shape. Obviously bad. However in the design as mentioned, the nature of the design maintains the stress to a localised area, which would have little to no real effect on the overall shape ie squareness.

My 2 cents for what it's worth

[Michael G]

A foam pattern will work with CI. Essentially, the hot metal burns out the foam as it goes. The beauty of a foam pattern is that you don't have to worry about draft on the pattern as you would with a solid pattern. The other thing is that they can be rammed into the sand more easily rather than having to mess around with doing cope and drag sides sequentially. (This casting has some depth so may have to be done more carefully).

For stress in castings there are two things to remember.

Thick sections cool more slowly than thin sections
Internals cool more slowly than external features
Metal contracts after it cools


Quickly looking at Josh's renderings I would expect the spokes to solidify first, followed by the sides. As the sides are constrained by the sand, the spokes will be trying to pull the sides inwards. Last to cool will be the main box corners. The outsides will cool first followed by the insides of the corners. This will tend to pull the corners back and want to make them more acute. While not as experienced as a pattern maker in these things I suspect that the resulting casting will probably have 4 concave sides but by how much I would not want to guess. Depending on the ratio of spoke thickness to the main body, there may or may not be problems. Looking at the renderings I would think that they could be thicker and/or the main walls thinner but not having done much casting I could be wrong, especially if the casting shop is a good one. I would be concerned about gating as the sprues could well cool off before the rest of the casting, giving rise to voids. If thicker the spokes may be a better place to introduce metal to the (Most of my "casting design" has been for injection moulding - if this was a plastic part I would be slapping the designer around the head as I support designing parts that cause least problems in manufacture. However, Josh is bigger than me so I'll pass on that option)

Michael

[Ueee]

I was almost going to mention something similar earlier Pete. I don't know anything about casting stresses, but all the square frame levels i have looked at are made like that, a circle inside the square. There must be a reason....

[Stustoys]

but all the square frame levels i have looked at are made like that, a circle inside the square.
Hi Ewan,
Got a picture?
The ones I have seen are pretty much a square inside a square. Like this one,


Though I did like this one.


How about it Josh? ;)

I would have thought that regardless of shape.... IF it survived cooling, stress reliving would take care of any other issues. Maybe stress reliving before it cools would help? Looking at an engine block I wonder how they could cast then, of course they could have a couple of goes at it. Surly the foundary would say "well thats never going to work" if they thought they were going to have issues?? But then I wouldnt really have a clue and maybe they assume you know what you are doing?? Fun stuff this casting game.:p

Stuart

[.RC.]

May provide some ideas

Parallels on Busch Precision, Inc. (http://machining.buschprecision.com/category/parallels)

[Brobdingnagian]

The main considerations for the design for me in order of priority are: minimal weight for less 2 micron deviation of geometry with 1000N side load, 200x200x200mm size, able to not just measure but also transfer geometry ie: clamping space, simple to mould, ease of machining geometry and thermal stability.


I did few other designs design studies one of which was a circle in a square but the weight was higher than the + design for the same rigidity same went for the X and simple filleted box and a pitch circle kind of thing. Actually X was pretty close to the performance of the + design but the distortion was not uniform it introduced a wave in the faces under load which I did not like. The main problem I had with the "circle in the square"(the weight was not much more around a kilo from memory) was that it provided no easy way to clamp. And when using it as a spotting master you need to be able to clamp it to the surface plate or the wringing will drive you insane. Also need to be able to clamp to it with when used on a surface grinder to transfer it geometry onto the part. There were a few others that were a of various ribbed designs but that would have required cores that I ended up discarding because I thought they would be too complicated and hence too costly to mould.


The purpose of the hub is face to face rigidity and maintains parallel, it is also a handy place to have a handle to manurer it around being about the centre of rotation; as I would hate to drop one :S

Allowance has been made in the pattern for some distortion and rate of cooling problems (the "spokes" are a little thicker than I would have liked), but it is a balance between the processes and intended function. I think I have the balance just about right, but the only way to be sure is to get it cast. Failure is always an option but it is hard to see how it will fail so badly that it will end up being a useless artefact. The heat treatment oven is computer controlled and is capable of controlled ramps and I'm confident that I can remove enough of the residual stresses in the casting.

I'll also point out Taft Peirce honed in on the a similar design concept for their box parallels, maybe this is why they charge $2000 for each one.

-Josh

[RayG]

The ease of clamping with Josh's design is the thing I like, there are plenty of commercial examples of the same type of design.

http://2.imimg.com/data2/WA/TD/MY-2927515/22-250x250.jpghttps://encrypted-tbn3.gstatic.com/images?q=tbn:ANd9GcSGrI8fA2IPK_w-W5JcmpEAgTP5ecKCag49VgbUIs8QrewxN0X1
http://machining.buschprecision.com/ImgMedium/a1064.png

Series 5100, 5200, 5300,5400, 5500, 5600 , 5800 parallels : box , cube , giant cube , planer/boring machine , 3 dimension, steel , manufactured by Busch Precision (http://www.newmantools.com/busch/5000.htm#5400)


The FEA Josh did, showed surface distortion under load with the diagonal bracing version, but less distortion in parallelism with the orthogonal bracing.

After arguing around for a day or so, the final version is pretty much what you see,

Regards
Ray

[Brobdingnagian]

Hi Ewan,
Got a picture?
The ones I have seen are pretty much a square inside a square. Like this one,


Though I did like this one.


How about it Josh? ;)

I would have thought that regardless of shape.... IF it survived cooling, stress reliving would take care of any other issues. Maybe stress reliving before it cools would help? Looking at an engine block I wonder how they could cast then, of course they could have a couple of goes at it. Surly the foundary would say "well thats never going to work" if they thought they were going to have issues?? But then I wouldnt really have a clue and maybe they assume you know what you are doing?? Fun stuff this casting game.:p

Stuart

Oooo. I Like that second one, who makes that?

I would also assume that the foundry has a pretty good idea of what they are doing, and leave it in the mould long enough before knocking it out to allowing it to cool slower. I also assume that if it did not survive the initial cooling they would be able to cast another one with modifications to the moulding by adding chills and/or changing the ramming pressure.

-Josh

[Brobdingnagian]

May provide some ideas

Parallels on Busch Precision, Inc. (http://machining.buschprecision.com/category/parallels)


I really like their hugs and kisses parallel.
256715

:)

[Stustoys]

I really like their hugs and kisses parallel.

lol :D


So glad I took the links out. I had to go find it again lol. turns out its a company thats been linked before.


Machine Checking Square Frame - Machine Checking Frame and C.I. Testing Bed And Floor Plate Supplier & Manufacturer from Indore, India (http://www.jashprecision.com/machine-checking-square-frame.html)


Nervermind Ewan, while searching for the above link I came across a couple of circles. Thought the second one is a circle so you can move the level.

Stuart

[Brobdingnagian]

lol :D


So glad I took the links out. I had to go find it again lol. turns out its a company thats been linked before.


Machine Checking Square Frame - Machine Checking Frame and C.I. Testing Bed And Floor Plate Supplier & Manufacturer from Indore, India (http://www.jashprecision.com/machine-checking-square-frame.html)


Nervermind Ewan, while searching for the above link I came across a couple of circles. Thought the second one is a circle so you can move the level.

Stuart

Thanks Stuart.