[suspension]

Hi all
Still no luck finding a CNC workshop. Watched many videos on scrapping but concerned if I ever be able to get the accuracy I need.

Checked the cost of 30mm ground plates and that too is high. One option I am considering now is to buy 30/40mm thickness, 40/50mm width ground plates and join them to get all surfaces required for rails/motors/etc. For example for a given axis I will use three plates, two for two rails and one in between those two for the ball screws bearings, motor, etc. To make sure that all three plates are parallel to each other (in each dimension), I keep two ground plates right angles to them with a distance of around say 300mm. This way, all three grounded plates would be resting on two joining ground plates. Since all are ground plates this should give a fully parallel set of plates. In order to attach plates together I can use some screws. Each plate will have bolts which will be attached to the EG casting.

Does this sound ok? Would the casting and curing process ruin the alignment/deform the plates?

[ardenum2]

Would the casting and curing process deform the plates?

yes it would, to mitigate this issue you cast the part with a hole for each anchor bolt and a land for the plates then glue/grout the plates onto the part after it has cured

[joeavaerage]

Hi peteeng,

Hi Craig if steel is $7/kg and Al is $14/kg in Kiwi then 100mm of al 1m2 is $3780 and the 70mm steel is $3828 same ? I'd go aluminium. I appreciate your married to the iron age, irons in our blood so our machines should be made from it!! some say. But I try to look at all the options. Peter

That's not quite accurate. I have taken delivery (Monday) of the steel for my trunnion table and including GST and plasma cutting it cost $245NZD for 43kg of steel, or $5.65NZD/kg.

The quote I just had from Altus for a 60kg piece of 6061 Al including shipping ($60NZD) and GST is $1219NZD or $20.30NZD/kg

So for a part of given stiffness steel is about 30% cheaper than a part made from Al. Having said that the price difference is small, and other considerations may influence your choice.
For example if you are dimension constrained then the Al part may just be to big, or I can weld steel but not Al, or the toolmaker will not put Al on his surface grinder, or in the case of
my steel plate supplier.......is 350m from the front door from where I work!

I think your approach has value, that is calculate the cost for a part of given stiffness. If a part has to be 100N/um who cares what you make it out of if one material costs $100
when another material cost $200 for the same part. Countervailing considerations aside the cheaper material is indicated.

Craig

[peteeng]

Hi Sus - I'm sure you would achieve the accuracy you need with lapping or scraping. They are slow but sure processes. But as Ard says, mould in pockets and bond the plates in place after casting. You still may have to lap at the end anyway. Some of the builds, even with machining had to be scraped at the end. Many decades ago I was having work done at a toolmakers. They had just bought a $2million AUD Mazak and it was being commissioned. The master builder arrived with a small bag of dial guages and some scrapers. He instructed on how the machine would be pulled down, did some scraping put it back together and it was tested good. That machine was built to micron stds but still had to be scraped, it moved in transit from Japan to OZ I suppose.

If you can't afford a Rolls Royce then think about a VW. Use the VW to gain experience and create the capability to make parts for the bigger machine. You never build "THE" machine at first attempt especially the way you are making cost down decisions that keep limiting your result. Or by a second hand manual mill to make your parts. Stepping stones to a result are sometimes better than trying to get there in one leap. All great machine builders build a series of machines never just one....Or build your imagined machine out of mass timber. Laminate good plywood into the shapes you imagine for the grout machine. Easy to change things, easy to flatten etc. Once built coat in epoxy or good paint and it will perform very well. Easily make aluminium parts for No2. Many builders in the forum build plywood machines then use that to make aluminium parts then swap those out to transform machine No1 into No2. Ply is 1/10th the weight of steel so is easy to manage and will give you insight into the mistakes youv'e made (there will be a few). My guess is that if you used aluminium ply laminates you could build 90% the machine you want in it anyway... There's many ways to skin this chicken. I think you need to put some time into CAD work and figure out what its going to look like. Peter

[suspension]

Hi Sus - I'm sure you would achieve the accuracy you need with lapping or scraping. They are slow but sure processes. But as Ard says, mould in pockets and bond the plates in place after casting. You still may have to lap at the end anyway. Some of the builds, even with machining had to be scraped at the end. Many decades ago I was having work done at a toolmakers. They had just bought a $2million AUD Mazak and it was being commissioned. The master builder arrived with a small bag of dial guages and some scrapers. He instructed on how the machine would be pulled down, did some scraping put it back together and it was tested good. That machine was built to micron stds but still had to be scraped, it moved in transit from Japan to OZ I suppose.

If you can't afford a Rolls Royce then think about a VW. Use the VW to gain experience and create the capability to make parts for the bigger machine. You never build "THE" machine at first attempt especially the way you are making cost down decisions that keep limiting your result. Or by a second hand manual mill to make your parts. Stepping stones to a result are sometimes better than trying to get there in one leap. All great machine builders build a series of machines never just one....Or build your imagined machine out of mass timber. Laminate good plywood into the shapes you imagine for the grout machine. Easy to change things, easy to flatten etc. Once built coat in epoxy or good paint and it will perform very well. Easily make aluminium parts for No2. Many builders in the forum build plywood machines then use that to make aluminium parts then swap those out to transform machine No1 into No2. Ply is 1/10th the weight of steel so is easy to manage and will give you insight into the mistakes youv'e made (there will be a few). My guess is that if you used aluminium ply laminates you could build 90% the machine you want in it anyway... There's many ways to skin this chicken. I think you need to put some time into CAD work and figure out what its going to look like. Peter

I considered the step by step approach, but to build a new CNC machine, you will need a bigger machine with better precision! However step by step process would surely give me insights on where mistakes are made, but it will not solve my main problem of getting precision surfaces of the required size.

If I were to scrape, I think I should just cast two steel plates without any milling (milling them is not practical as these plates will be 1000mm long) and then do the scrapping. With a straight edge I can understand how to get each surface flat along the length and maybe width. However how do I make sure both plates upper surfaces are parallel? Please see the pictures below.
Also I understand that for this process you need an engineers level. Can I rely on it for accuracy?


I am also re-considering the epoxy approach, but may be not with self leveling. But like this: https://www.youtube.com/watch?v=xtFzb6r6Np8

[peteeng]

Hi Sus - No that is not the case (regarding needing a bigger machine, yes its ideal but our forefathers got to where we are by building small machines) and historically it hasn't been the case either. Even though the machine you build is not as "precise" as you need it still is "repeatable" enough to get a precise part out of it. Accuracy, precision and repeatability are different.

I explained before but not good enough about the lapping of two surfaces, see attached. The lap has to span the two surfaces so they remain co-incident. If you use plain steel to start with you would not scrape. Scraping is too fine a process to begin with. You would rough lap it down. There is no need to scrape at all if you want to. Granite surfaces are not scraped for instance. Scraping creates a textured surface that is easy to pick up the hi/lo points with. Plus scraping is done on metal surfaces that are functional as bearings so to create oil cavities for the oil to flow otherwise the surfaces will clamp.

A good engineers level is very accurate. I use one occasionally and get impatient with it because it takes a long time for the bubble to settle. Its accuracy is specified. Keep at it. Peter

[joeavaerage]

Hi,
I think you are overestimating what scraping can do. Have a close look at the videos, these surfaces are already milled flat, possibly even ground and THEN they are scraped.
If you have to remove more than a few um of material I'd say its not going to happen.

(milling them is not practical as these plates will be 1000mm long)

I'm not sure where you get that idea....there are at least four machines in Christchurch a city of 250,000 on the backside of the earth they can do 4m plus. I'm not saying its easy or cheap but
it is possible and you'll have to find the equivalent capacity in your area or you are sunk.

Craig

[peteeng]

Hi - Re scraping pattern. There is a lot of discussion about scraping patterns. Functionally the pattern does not matter, if its pretty and consistent then thats Ok can be ugly and consistent as well. The Moore pattern is one that is ergonomically easy to do and looks good. But any pattern works, if it's for an oiled bearing surface the cuts need to overlap. If its purely for levelling they don't have to overlap... Peter

[suspension]

HI explained before but not good enough about the lapping of two surfaces, see attached. The lap has to span the two surfaces so they remain parallel.

How do I make sure they are parallel? Keeping a engineers level across them? But according to my drawing, even when the steel plates are twisted and not parallel, the level could say they are parallel as the highest edges/points are at the same level.
Can I use a grounded bar across the two plates to verify that top surfaces of two plates are parallel using dye method? Asking this because only small part of the surface plate would touch the steel plate when I do that and was concerned if this would give accurate information.

Thanks
Sus

[peteeng]

Hi Sus - You have to use dye and levels and observation, time and effort to achieve whatever shape you want or need. The lap and scratch tells you what the surface is doing. The lap has to be flat! You only lap in one direction so the scratch takes the dye off in one direction. Keep at it until the entire contrast/dye is gone and the scratch is consistent. Rotate the lap often,. Use a big lap. Check with level often in both directions. Once happy at coarse grit move to a finer grit and do the same. Observe the scratch and when the scratch is consistent do 90deg direction to check it has the same result. Then move to finer grit etc etc until you get to the flatness and finish you want... Stonemasons have been doing this for centuries....Peter

edit - Hi Sus - learn by doing. Get 3 pieces of steel say 100x300x20mm thick or 200x200x20 etc mark them A, B and C. Look up the Whitworth method and make them flat to whatever flatness you need. Get some SiC grit or wet and dry paper. Plus some dye or contrast. In a few days you will have great laps and great experience...

[suspension]

Hi Sus - You have to use dye and levels and observation, time and effort to achieve whatever shape you want or need. The lap and scratch tells you what the surface is doing. The lap has to be flat! You only lap in one direction so the scratch takes the dye off in one direction. Keep at it until the entire contrast/dye is gone and the scratch is consistent. Rotate the lap often,. Use a big lap. Check with level often in both directions. Once happy at coarse grit move to a finer grit and do the same. Observe the scratch and when the scratch is consistent do 90deg direction to check it has the same result. Then move to finer grit etc etc until you get to the flatness and finish you want... Stonemasons have been doing this for centuries....Peter

edit - Hi Sus - learn by doing. Get 3 pieces of steel say 100x300x20mm thick or 200x200x20 etc mark them A, B and C. Look up the Whitworth method and make them flat to whatever flatness you need. Get some SiC grit or wet and dry paper. Plus some dye or contrast. In a few days you will have great laps and great experience...

I believe you are referring to this process?
https://www.youtube.com/watch?v=dwTMZ2Qx0QY
@0:34

[peteeng]

Hi Sus - Yes - he has started with a granite flat as a lap. Or he has made it. I hope to see some reels of you doing it one day. He didn't use a contrast I suggest you do, then you know when you have done enough...Peter

I looked at other videos of this build. He has done a very good job. Is this the level at which you want to be? I'd not use such a dry epoxy mix, I'd use grout and self levelling grout. Easy, just pour and wait. If you made the lands in the mould you could lap the lands and not worry about metal inserts. Just as easy to lap concrete as it is steel... Then you don't have the issue of mixed materials and getting the right size steel etc etc Peter

[peteeng]

Hi Sus - No metal plates in these puppies! Peter
Ultra-high performance concrete (EPUDUR) | RAMPF Group (rampf-group.com)

[suspension]

Hi Sus - No metal plates in these puppies! Peter
Ultra-high performance concrete (EPUDUR) | RAMPF Group (rampf-group.com)

Yes, they look good.
But this also will require a mould with true parallel surfaces which will be hard!

[peteeng]

Hi Sus - The lands are made from one plate (so the surfaces are quite close to flat and colinear and then an insert makes the gutter. Lapping with coarse grit does not take much time. Nothing is "hard" usually some patience is needed and time. You either sell a kidney or two to pay for the machining or learn some patternmaking and lapping skills Didn't you read the disclaimer when you signed on here? Somewhere in the print its says something like "machine building is not for the fool hardy" be bold get a start....Peter

edit - If you go 100% concrete you can screed filled epoxy across the top of the land, this will save you lots of time vs lapping down to the lowest valley. So you start to lap and measure the low points, if they are way low screed, if close keep going. Once lapped I'd use thin epoxy to stabilise the surface then give it a final wipe.

[peteeng]

Hi Sus - Rampf publish these figures. They are good as far as I can see. I'm surprised by the really low damping figure for CI. But rest is good. Hard stone is solid granite.... Peter

[suspension]

I have hard time finding ball screws and linear guides from the ebay seller as much of his stock is either not preloaded or not precision grade.
Alternatively, I can buy HIWIN linear guides/rails from https://www.automation-overstock.com for very reasonable prices, but the model they have at these prices are old and discontinued. (LGH35 series). I am going to go ahead with those because life time calculation gives years when operated daily. This is a hobby machine, so by the time these needs to be replaced, I am probably dead anyways.

For ball screws also, automation-overstock has following from Steinmeyer.
https://www.automation-overstock.com...ts.asp?cat=211

Compared to THK/NSK/etc these are exceptionally cheap. Although they say these are Germán made, I cannot find much information about their quality and reliability in this forum nor on the internet.

What are your thoughts?

[ardenum2]

i don't like those carriages, single row bearings, maybe ask these guys for hiwins roller series instead, prices stated are correct, at least they gave me the same prices on the quote.

https://tnkbearing.en.alibaba.com/pr...RG_series.html

[joeavaerage]

Hi,
to be honest I think those ballscrews look the part.

I posted much earlier in this thread that linear components, and ballscrews in particular are the heart of the machine and good quality components are unaffordable new in most cases,
and that second hand or new old stock are really one of the few cost effective means that cash strapped hobbyists can get true quality parts.

Those German made screws are DIN P5, I would be pleased, even proud to have them in my machine. Those LGH linear rails and cars are OK too, they have four rows of bearings,
two on each side, an upper and lower. The sketch of the car might rather look like there is a single row on each side but that is deceptive.

In any event linear rails and cars are expensive new, but still very VERY much less than ground C5 and C3 ballscrews. If you can get good ballscrews, say new old stock, but bought
new linear rails/cars that would still be a good deal.

Downside is you take what you can get and design your machine around them......not design the machine and then try to find the parts. That would be fine if you bought
new parts, but now that you are acquainted with the cost of new parts you probably realise that they are prohibitively expensive.

Craig

[joeavaerage]

Hi,
I personally favor the 5mm pitch screws, you get greater thrust and therefore cutting authority, with a smaller servo. These 5mm pitch screws have max travel of 540mm, and probably 520mm
as a practical maximum.

The 8mm pitch screws have 580mm travel max and a practical max travel of say 560mm. Being 8mm pitch probably indicates that you should go up one size in servos, from say 750W to
900W or 1kW.

The 10mm pitch screws have 660mm and 589mm max travel respectively and a practical max travel of 640mm and 570mm. Being 10mm pitch will likely require 1.5kW servos to retain
the acceleration and thrust authority.

Its a classic case where if you want more travel then it costs. I can get a 750W Delta servo kit (servo, drive, cables) for $438USD + shipping whereas a 2kW Delta servo kit costs $895USD +shipping.

Pay your money and take your pick.

Craig