[peteeng]

Hi Craig - Thanks for those, I have seen them before. My issue is getting the measuring equipment not so much the process. Justifying the cost for a once off project is the issue. I think I'll just have to say that for my purposes if the laps float they are good enough... unless I find a local certifier or some surplus equipment... Peter

[joeavaerage]

Hi peteeng,
yes I understand. I am very serious about several Aussie/Kiwi enthusiasts pooling resources for an autocollimator. With care and patience we could find a highly usable autocollimator
for under $2000USD, spread over four guys becomes doable. As you say you might use such a device for a period of a month or two and then not need it for three years......

Craig

[peteeng]

Hi Craig - although I think the idea of shared equipment idea is good, my experience with shared stuff is that it will go wrong eventually, even when people are in the same neighborhood let alone international distancers. I've had various things like vacuum pumps, motorcycles and even axes go pear shaped... when shared... Peter

I was looking at tech rental stuff and found two devices I had not seen before 1) a Holiday Detector and 2) a wobulator for vibrations!! maybe combine them into a wobbly holiday detector...

[joeavaerage]

Hi,
I understand the qualms about shared gear......but then if it came to it I cannot see myself ever having $2000USD to spare and am thus denied an Autocollimator.

It may well be the choice is to share (with attendant risks) or go without.
Pretty stark really. Ask yourself also, if an individual were prepared to stump up something in the order of $500AUD-$1000AUD for an autocollimator they can't be idiots or
fools, or at least any worse than you or I!

Craig

[peteeng]

Hi All - My thoughts so far on the plates are as follows:
1) The process requires the highest spots to be reduced to the lowest hollows. Initially removing high spots is quite quick but as I get to 80-90% contact removing large surface areas slows down progress. So use very coarse grit to start. I couldn't get P40 today, will try another shop tomorrow
2) Since I am working towards a "flat" surface for laminating (vs metrology) perhaps I can use the filling and cutting method for the proposed table. Once I get the coarse W&D onto it I'll see how its going
3) Patience and slowly slowly making dust will get the result.... Peter

been looking at comparators.... making a repeatometer is pretty basic

[ardenum2]

here's a short heller paper where topology optimization on a column failed.

[peteeng]

Hi Ard - The issue with topological "optimisation" using a static approach is that the result has many thin ligaments. These ligaments vibrate and contribute to poor dynamic performance (ie vibrations can be complex) The article does say that they improved on this approach. If the aim is to 1) decreases mass yet 2) increase rigidity, then thicker ligaments or extra geometry or webs or attention to lengths is needed. Using AI approach's all of these variables and performance targets can be included in the solution definition and can be accounted for. I'm sure they are working on this sort of thing already. Lastly the current solutions are typically not more economic than std solutions due to their complexity & many can only be made via 3D printing which is painfully slow. But that will get better... Research is always interesting stuff. Peter

[peteeng]

Hi All - Out for my morning workout. P80 is cutting well but I'll find some P40 today. Using water as a contrast I can see that the B plate (LHS Block) has a low spot at the top right and bottom left. So need to work it down until those spots level out. Contrasting the A plate to the B plate shows a small hollow at one corner in the A. Getting close as the blocks freeze sometimes due to suction... Just keep rubbing and the Genie will appear soon. Peter

[peteeng]

Hi All - Today I went wet. The wet cuts better and quicker. The B plate has a small hollow and when that's gone I'm going to call A&B a match and move to plate C. The takeaway is use P40 grit wet and you will get there fastest. P80 grit is still very smooth to the hand. I thought there would be some scratch to witness the cut but no. Seems P80 is a polish on this concrete. Even the slurry on the paper is getting to be very consistent. So I think I'm close to starting C... Peter

[ardenum2]

this looks awesome

https://www.youtube.com/watch?v=I_zS5RZ0LMc

[peteeng]

Hi all - I'm making a box for my neighbour. Its sort of a test build of a laminated aluminium part. It would be epoxied and screwed together at the lugs. Build is going well. Will need to spot drill the holes in the lugs next time, these weren't and the alignment is not spot on. These will be re-drilled and dowelled for the box. Peter

[pippin88]

Peter, I haven't read over it in a while, but I thought the 3 plate method required changing things up regularly, not matching A to B perfectly then moving on. If you match A to B first, you can end up with a perfect dish.

So it should be regular changing between
A - B
A - C
B - C
With changes in rotation as well (e.g. don't always work in the same direction, to truely guarantee a flat plane need to rotate A vs B by 90 degrees at times)

[peteeng]

Hi Pippin - I have thought about that as well. The usual description says match A-B then match A-C. Then B-C then back to A-B etc. I don't think it matters if you go as I'm going or you regularly change. You will eventually end up flat, its a "natural" flattening process.

When I was taught lapping of surfaces for microscopic samples and for preparation of ultrasonic device surfaces we always went in one direction until the scratch was even. But this was against a known lapped flat lapping plate. This is because you can observe the scratch and once its consistent at the working grit you then turn it 90degs and give it a wipe. If the scratch is then consistent at 90deg you then move to the next finer grit. You can also use contrast to determine the contacts. If you are working with a metal surface plate you identify the high spots and scrape these then lap, then contrast, then scrap high spots etc... Once you get the AB to match you move to the C... With granite and concrete I don't scrape just lap. I could move onto the C plate now using the A. If you end up with a dish and I think you would, then that's why you move to C. AB and AC will be dished and its the averaging of changing plates that works it to flat. Time and effort will end up with 3 plates flat. Prior to Whitworths 3 plate method they used 2 plates and this resulted in poor results.... I have been trying to get P30 or P40 paper with no luck, I think that would help move the process along faster.. If you don't work towards a match I think the process is slower as you have no reference or end point to work to or obtain. So yes you could say work AB for an hour then move to AC for an hour then BC for an hour etc and you would get to the same point eventually... I expect the time maybe the same as the amount of material to remove is the same... scraping would accelerate the initial high spot reduction as this is faster than lapping... I'm not sure what to scrape concrete with... any ideas?

Once I get it as best as possible with paper then I would move onto grit... the P90 paper I have leaves a very smooth finish to the hand & no visible scratch to guide me. Been interesting so far. Maybe for my laminating table I'll do it like a boat hull and fill the hollows with a filled epoxy and cut the hills to the average.... or just find some P30 and get into it!! Peter


moores book - peter Foundations Of Mechanical Accuracy : Free Download, Borrow, and Streaming : Internet Archive

[peteeng]

Hi Pippin et al - I started lapping A-C to guage what's happening. Then I decided to get out a straight edge and feeler gauges. AB&C have a hollow centre of about 0.0015". So some attention to the edges is needed. Moore discusses how square and round laps are better as rectangular plates or laps can cup or twist. But he concluded they were becoming more popular... I don't think I'll be able to get P30 or P40 paper. I looked at some catalogues on line and 80 maybe roughest in paper. So I'll order some grit. Grit cuts both sides at once so maybe that's faster than paper as paper cuts one side at a time.... I'm looking for a tray I can set up so the paper can be wet all the time.... Peter

[Mr-MaW]

This is my new product, share it with everyone.
The base of a gantry machining center.
The base has a length of 6.5 meters and a width of 3 meters.

[peteeng]

Hi All & Sundry - I have ordered some P16 grit so will update once I get that cutting. Peter

[peteeng]

Hi all you lot - I've been away for a week travelling to celebrate my birthday... Been to Uluru and back. Did a helicopter sunset ride which was spectacular. Always like to get close up to aircraft structures, so well built... Got home and the SiC is here. P16 looks a bit coarse but will give it a run and see how it goes.... I did a short stint as a cameraman and hung out of coppers for a couple of days some years ago. The signage is interesting. The rest of the placards describe how to take the door off, set up the external harness and emphases its not as safe as sitting in the inside.... Peter

[ardenum2]

whoa, you celebrate your birthdays for a week? I don't even do cakes anymore... well happy birthday Pete

[peteeng]

Hi Ard - It's a combined wife and me birthday, plus we haven't been anywhere in years... been looking after elderly parents, plus covid etc etc. Time for a break-out. May set a precedent!! Birthday week Yeh...Got to max these things out. Only here for a short time... Peter

I've had a bit of flight time and airport time to doodle a few things. Want to start on a new 5 axis machine and I keep coming back to formply as a good prototype material. Even a production material... Plus use a bit of aluminium laminate in the right areas and its a great material... If I look at it in terms of areal weight its the cheapest. Its thick enough (17mm) to edge screw and glue so has stiff connections. To build at same areal weight in alum and steel is Al 4mm and steel 1.5mm. I'd be keen to make a 1.5mm steel machine but don't look fwd to the hours spent soldering it together. I'd also be keen to braze 4mm Al but again it would be a mission... Looking at in-plane stiffness the steel and alum are about the same result... Then in bending the F17 ply is equivalent to 10mm thick aluminium or 7mm steel at far less areal mass. My recent machine called Frankie is the benchmark. Most parts are F17 ply or laminates but I want to half its weight with no stiffness change. This exercise tells me I can't do it by swapping materials. So I'll have to lightweight each part. I want it lighter so I can manage it by myself easier... and I expect a Maker would want the same. So its F17 ply with optimal cutouts and Al laminate. I use Al as I can cut that on Scoot my current router... This exercise also eliminated MDF. I have used it on two routers and its very successful. But I think the 4GPa stiffness is trumped by the formply E15 at same density.

Re 5 axis machine - I've been stumped with slew bearings. Found a Chinese company called ephant that makes lots of small to medium ones. They also make harmonic drive bearings which is a combined slew bearing and harmonic drive. Hadn't seen these before and they are perfect for this sort of thing. Cost is low as well. Early days on communicating with them, will see how they go... Peter

[peteeng]

sorry guys and gals thats Efant Bearings not Ephant... Peter