[joeavaerage]

Hi,

According to my friend (Seaman Jolly back then) 'any adjustments are not permitted, the Navy never does things in such a shoddy manner, Seaman Jolly you will continue to scrape this mount until it is perfect'

Craig

[peteeng]

Hi Sus - I'm sure it will be good enough even with a long stick out. The machinist will be able to account for that. Once machined clock the surface against the machine bed and you will know.... Peter

scraping is done for two purposes 1) to rapidly knock down high spots on wavy surfaces 2) to texture the surface so it has oil passages or to allow it to come apart at some stage. if you make two perfect surfaces and bolt them together hard you'll have a very hard time getting them apart in the future. You will figure it out as you go. Peter

[joeavaerage]

Hi,
Seaman Jolly had another skill that impressed me mightily, he could be that drunk that if he was ashore you'd have to find a place to sit him down for fear that he would
hurt himself and yet......on a yacht, he was completely at home, drunk or otherwise, ready for a gybe'n'peal spinnaker drop in a gale....no trouble. The drunker he was the better
he seemed to get

Craig

[suspension]

Hi all

I got the quotation for CNC milling and its more than 200% above what I have allocated from the budget. So thought about some alternatives and wanted to get your views on it. Basically idea is to surface grind a large 10mm thick MS plate to the require precision and use that as the datum/reference for all steel plates which are fixed to the concrete cast. Here are the steps:

1. Surface grind the 10mm plate to required precision.
2. Prepare individual 50mm thick plates for Y axis rail plates, column fixing plates, motor bracket plate, floating bearing plate.
3. drill/tapper all holes for bolts used to fix them to the cast.
4. Drill/tapper holes required to fix linear rails, column, motor bracket, etc.
5. surface grind all these plates to the required tolerance. On the side facing up would be required.
6. Fiz all these plates to the reference plate. To fix them securely I can use small bolts. I can use surface ground spacer plates/block as required for positioning.
7. Fix the rest of the MDF mould
8. Do the casting with grout
9. Wait for 28 days
10. Check tolerances of each plate. If good, all good. If not, do the CNC as the last resort.

My guess is that above method can fail if casting/curing process exert large forces to plates making them deform/warp. But I will be using 50mm thick plates to reduce this effect.

Any thoughts?

Thanks
Sus

[peteeng]

Hi Sus - I'll have to think that through. How big will the 10mm reference plate be? Peter

[suspension]

Hi Sus - I'll have to think that through. How big will the 10mm reference plate be? Peter

Hi Pete,

It will be roughly 850 X 500 mm. Increasing the thickness of that plate is an options if needed.

Thanks
Sus

[badhabit]

I got the quotation for CNC milling and its more than 200% above what I have allocated from the budget.

...been there, done that :-) Have you asked the if there is anything particular driving the cost up? Or if there is anything you can do to lower the cost overall...

Once i got a quote for some parts, but because of a single feature they had to move the milling from a regular 3-axis machine to a huge 5-axis which more than doubled the cost of the part! The feature was purely cosmetic and could by no means be justified to double the cost of the part. So by adjusting the part it got a lot cheaper...lesson learned :-)

Tips for reducing your CNC cost by only having facing operations:
- Lose the fixed side-datum for your linear rails and "just" align them to eachother while you mount the rails. Lots of people do this with fine results.
- Drill/tap the holes yourself (rent/borrow a magnetic drill, thank me later!)


/Thomas

[suspension]

...been there, done that :-) Have you asked the if there is anything particular driving the cost up? Or if there is anything you can do to lower the cost overall...

Once i got a quote for some parts, but because of a single feature they had to move the milling from a regular 3-axis machine to a huge 5-axis which more than doubled the cost of the part! The feature was purely cosmetic and could by no means be justified to double the cost of the part. So by adjusting the part it got a lot cheaper...lesson learned :-)

Tips for reducing your CNC cost by only having facing operations:
- Lose the fixed side-datum for your linear rails and "just" align them to eachother while you mount the rails. Lots of people do this with fine results.
- Drill/tap the holes yourself (rent/borrow a magnetic drill, thank me later!)


/Thomas

Hi badhabit

I actually I did and one of the main costs was actually milling the side datum! They can reduce the price by around 18% without it. If I chose to go with CNC I will get rid of the datum. I can also get the drilling and tapering out also from CNC. I saw that you have used a magnetic drill for yours. Did you use it to tapper threads as well?

The rest of the cost was mainly for surface machining which is required anyway.

Thanks for thew tips.
Sus

[badhabit]

I actually I did and one of the main costs was actually milling the side datum! They can reduce the price by around 18% without it. If I chose to go with CNC I will get rid of the datum. I can also get the drilling and tapering out also from CNC. I saw that you have used a magnetic drill for yours. Did you use it to tapper threads as well?

With cost in mind i would just let them mill the flat surfaces, it is where you get the best "value" as i see it AND is not easy for you to do it yourself with decent results. The other stuff you can do yourself, if you take your time...

I use the magnetic drill for holes and threads. Here is a recent video of me doing M12 threads in the subtable for my machine:

.... remember that the drill has to be able to do reverse if you do threads. Most magnetic drills actually dont reverse, so just make sure it can do that...

I always use a torque-limiter when threading holes, especially blind holes!. Partly because i have the threadtap adapters for it so it is very easy to mount securely compared to a drillchuck, but mostly because the chance of snapping a tap is greatly reduced! And you are going to do a lot of blind holes, so i'd recommend the torquelimiter! When doing blind holes you can just continue down until the torquelimiter kicks in, then reverse out again.

I bought mine of alibaba some years back, it was a set similar to this one: https://www.alibaba.com/product-deta...439809663.html (not this set! Just similar!) Note that there is different types of holders for different taps, ISO, DIN, ANSI and so on, some are compatible and some are not! So a M12 tap is not just M12 tap, the square end and the shank has different sizes according to some standard. So make sure to order the correct ones matching the standard your taps use! Otherwise your M12 tap wont fit in the M12 adapter you bought! (ask me how i know! ;-) ).

/Thomas

[peteeng]

Hi Sus - A 850x500 at 10mm thick once ground may not be flat. You would have to get normalised or stress relieved plate. Normal structural plate will have lots of tension in it plus when they hold it down it will bend so when released it will have the inverse shape in it. You would have to work with a company that has a CMM or a suitable surface plate to contrast the surface to check its flat.

I've been researching the plastic phase growth of concrete with not much luck for true numbers. The concrete chemist I was chatting to some time ago said it was significant and that they use slips like plastic film on the base of slabs so that the top and bottom of the slab have a chance of growing together and shrinking together through the process. Thomas has been thru this so could comment better. But slabs are very long compared to a machine part so overall plastic growth/shrinkage should be small especially if the concerte is plastic and drying compensated.

The best procedure is to not be too precise with the placement of the fittings and correct after it has cured. But then you have to get the monolith to a suitable machinist. Do the best you can with what you've got I suppose. Peter

[badhabit]

[...] Thomas has been thru this so could comment better. [...]

I'm honored, but not sure i can live up to these expectations :-) My take on this is that "everything moves and everything is rubber". So, i wouldn't trust a steel-plate to hold pre-machined inserts rigid enoug to get the tolerances needed. Neither would i trust the concrete/epoxy-granite not to move the inserts while curing(remember that the 10mm steel-plate is made of "rubber" and can/will flex ).

I would defently machine after casting(and postcuring if needed) if at all possible/viable. If it isn't an option(e.g. due to cost) i would at least expect(and plan for!) manual processing of the inserts afterwards. So scraping/lapping/shimming/adjusting/whatever options must be considered before casting as it might require considerations in the design for this to be possible(like featues preventing scraping).

[suspension]

My machine design requires two plates of 25mm thick precision plates (as Y saddle). Its size is around half the size of the reference plate I require. So instead of using a 10 mm plate, I can use a 25mm plate as a reference plate and when casting is done, I can cut the 25mm to get the two saddle plates. This way there is less probability of warping/deformation and the cost of the plat + surface grinding will not be a waste. In any case I will be checking the tolerances after casting. I bought a surfaces plate, square, levels, etc required for this purpose. If this goes wrong, I can revert to CNC machining.

[peteeng]

Hi Sus and Thomas - My experience with moulded in features is in large vacuum cast fibreglass parts. The mass gets to around 90C during cure so the mould (fibreglass or aluminium) expands and the mass sets at that temp. Which is bigger than the desired dims. Then the mass and mould cure and cools which means the dims are off. So I have always post, machined the details. I'm about to cast some small parts using sika3350 which is dual compensated but even then this can't compensate for the thermal expansion contraction only the chemistry volume changes... Peter

[joeavaerage]

Hi,

It will be roughly 850 X 500 mm. Increasing the thickness of that plate is an options if needed.

Grinding flat and expecting it to stay flat......not in 10mm. I'd guess that the 10mm as-rolled plate will be better than once you've ground it!. Its machining, in this case
grinding, that causes the residual stresses to redistribute. So you'd grind it and once released it would warp or bow.

I just bought some 16mm med tensile steel plate, 650mm x 235mm, for the base of my trunnion table. Its flat enough as is but I dare not machine it otherwise I think it would be all over the place.
I do have to weld an upstand on one end....so I may yet have trouble......have to wait and see. I considered 20mm, as its more resistant to movement, but its cost and weight go up, while I'm not
getting a big enough improvement in properties. The next logical step is 32mm med tensile, but that would mean I'd be adding another 20kg and another $100NZD over and above the 16mm plate I've chosen.
The whole trunnion and fifth axis is already at about 70kg....I don't want to add any more than I have to. The base is bolted down to the cast iron bed, but I'm rather hoping its stays flat, even after I weld the
upstand. If it does not then I'll have to get it thermal stress relieved and bugger the cost.

Craig

[suspension]

Looks like I will have to go with the unexpected CNC cost as most of the other options are not straightforward and can fail. Before committing to it however I wanted to check one last option. Do you think it will be practical to just use hand srapping to get these steel plates flat? The holes/threading can be done without a CNC before casting.
Basically I will get all steel plates drilled/tapered first. Then will let them just rest on MDF sheet and make the rest of the mould and cast. here I am just dependent on the flatness of the MDF and steel plates. After casting I will just hand scrape until everything is flat. I have the tools already. Does this sound crazy?

[joeavaerage]

Hi,
scraping large areas is not going to work. If you were trying to remove 0.01mm, 0.02mm maybe, but tenths of a mm????

As I posted earlier I bought a piece of laser cut 16mm mild steel plate and it was flat to better than 0.1mm. Of course, I welded something to it so that 0.1mm has gone out the window now,
but it was OK to start with. Maybe not machinists flat but OK. If I wanted better, I would stress relieve then grind or mill.

I got a current price for thermal stress relief, works out at $6.80NZD/kg +GST (15% local tax) or $4.70USD/kg (tax inclusive)

As it turns out I was able to put it in the press and bend it back to about 0.2mm (over 650mm), and that will have to do for the time being.

Building CNC's is all about deciding on where to invest your resources. There is a perfect way to do everything, but nothing like enough money to do it all.
At best you find a compromise that means you get the best result for your money. I would say that final machining is essential, and who cares about the cost, it just has to be done.
Then getting the underlying flat plate 'machinists flat' is not that important. Its likely to be within a few tenths as is....don't sweat it. When I took my three axis beds (three by 115kg bits of cast
iron) in to be machined they quoted $6000NZD +GST, and I nearly died...it seemed extreme. I went ahead however, and the result has been perfect...the sort of thing you can build a machine around,
and that is exactly what I've done. Whatever imperfection or compromises I've had to make, I did not compromise on the axis beds.

This was in affect me deciding that this is where my money will be best rewarded. I did not frame it in my mind that way at the time....but is nonetheless what I did.

Craig

[peteeng]

Hi Sus - Doing everything by hand is possible. Thats how it was done 150+ years ago & is still done today. Just takes time and patience. To achieve whatever accuracy machine you want requires using a machine more accurate than what you want eg a commercial grade machine is made on a toolmaking grade machine, a toolmaking grade machine is made on a mother machine. A hobby machine is made on a commercial machine. Ultimately we do what we can with what we've got.... Its about now that you are finding out that maybe buying a mill is cheaper than making one? You just keep taking small steps towards your goal and manage expectations. You do need to be a little crazy to be a machine builder... Peter

[suspension]

Hi all

Like to get your opinions on how to restrict movement of rail blocks beyond the rail edge.

I have already put a proximity switch which will be controlled through HW/SW as the first limit and then a limit switch which triggers after the proximity switch. Limit switch trigger will cut the power to the whole machine. However none of this will prevent accident movement of saddle beyond the mechanical limit - which can get blocks out of the rail. The only mechanical blocker to this movement is the ball-nut which hits the bearing at the limit. I was wondering if I should have another mechanical blockers at each end of the axis (for each axis) instead of relying on ballnut/bearing combination. My worry is relying on the ballnut may damage the ballnut/screw?

I also created few mixes with the grout I selected. Mixed it with sand/chips/gravel and some sort of a powder at various rations. Waiting for 28 days to do a quick test such as following to get an idea about the stiffness. Plan is to select the mixture with best stiffness and use it for further lab testing to get an idea about flexural modulus.

Also received several main parts. The Spindle motor is much bigger than I thought!

[peteeng]

Hi Sus - Some machines have high density bumpers at the ends of the ballscrew. You could cut up your thongs/jandles/flip flops etc and make some bumpers. Petr

[joeavaerage]

Hi,
My preference would be to put a foam rubber ring on the ballscrew such the the ballnut will collide with the foam ring and the bearing carrier but BEFORE the car leaves the rail.
This not only stops the car from coming off the rail but softens the blow should your machine try to execute a move that would require the ballnut to 'depart for foreign climes'. The shock
of the ballnut hitting the bearing carrier will put dents in either the ballscrew shaft or the angular contact bearings. Ask me how I learnt this!!!

Craig